From 87339c425c49e4bc338f20fc6fc19e230f613b16 Mon Sep 17 00:00:00 2001 From: Kyle Mistele Date: Wed, 23 Sep 2026 10:20:18 -0700 Subject: [PATCH 1/2] feat: add GPT-6 and Opus 5.5 model support HumanLayer-Session: https://app.dev.codelayer.gg/sessions/01a0cf31-76ac-77af-a656-2e0e74a69323 --- agents/codelayer/src/agent.ts | 16 +- agents/codelayer/src/command.ts | 4 +- agents/codelayer/test/agent.test.ts | 119 + package.json | 2 +- packages/agentlayer-core/models.json | 339431 ++++++++++----- packages/agentlayer-core/src/models.ts | 4 +- .../test/compaction-engine.test.ts | 8 + packages/agentlayer-core/test/models.test.ts | 40 +- .../test/codex-ws-adapter.test.ts | 79 +- .../src/protocols/openai-responses.ts | 4 +- 10 files changed, 234505 insertions(+), 105202 deletions(-) diff --git a/agents/codelayer/src/agent.ts b/agents/codelayer/src/agent.ts index 2697270..4aaa4e7 100644 --- a/agents/codelayer/src/agent.ts +++ b/agents/codelayer/src/agent.ts @@ -135,6 +135,12 @@ function resolveAnthropicThinking(model: LanguageModel, effort?: string): Record effort: resolvedEffort, } } + if (modelId.includes('opus') && (modelId.includes('5-5') || modelId.includes('5.5'))) { + return { + thinking: { type: 'adaptive', display: 'summarized' }, + effort: resolvedEffort, + } + } if (modelId.includes('opus') && (modelId.includes('4-8') || modelId.includes('4.8'))) { return { thinking: { type: 'adaptive', display: 'summarized' }, @@ -200,8 +206,13 @@ export function buildProviderOptions( model: LanguageModel, overrides: CodelayerProviderOptionOverrides = {}, ): CodelayerProviderOptions { + const modelId = ((model as { modelId?: string }).modelId ?? '').toLowerCase() + const requiresAdaptiveThinking = + modelId.includes('opus') && (modelId.includes('5-5') || modelId.includes('5.5')) const anthropicThinking = - overrides.anthropic?.thinking === 'off' + requiresAdaptiveThinking + ? resolveAnthropicThinking(model, overrides.anthropic?.effort) + : overrides.anthropic?.thinking === 'off' ? {} : overrides.anthropic?.thinking === 'adaptive' ? { @@ -292,6 +303,7 @@ const EFFORT_RANK: Record = { medium: 1, high: 2, xhigh: 3, + max: 4, } /** @@ -302,7 +314,7 @@ const EFFORT_RANK: Record = { * - codex / firepass / copilot → `reasoningEffort = level` * - anthropic `4-5`/`4.5` (extended thinking, no adaptive support) → * `{ thinking: 'enabled', budgetTokens: LOW_ANTHROPIC_BUDGET }` - * - anthropic adaptive (`4.6`+/`4.7`/`4.8`, and any other model) → + * - anthropic adaptive (`4.6`+/`4.7`/`4.8`/`5.5`, and any other model) → * `effort = level` * * Guards (respect an explicitly-throttled parent — sub-agents never think diff --git a/agents/codelayer/src/command.ts b/agents/codelayer/src/command.ts index 751df72..8198aa4 100644 --- a/agents/codelayer/src/command.ts +++ b/agents/codelayer/src/command.ts @@ -74,7 +74,7 @@ function assertThinkingValue(args: { provider: ProviderType; modelId: string; th } } - if (args.provider === 'codex' && (modelId.includes('gpt-6-astra') || modelId.includes('gpt-5.6'))) { + if (args.provider === 'codex' && (modelId.includes('gpt-6-') || modelId.includes('gpt-5.6'))) { // These models also advertise ultra, but it assumes Codex CLI spawn-agent tools we do not expose yet. supported(['low', 'medium', 'high', 'xhigh', 'max']) return @@ -89,6 +89,8 @@ function assertThinkingValue(args: { provider: ProviderType; modelId: string; th if (modelId.includes('fable-5')) { supported(['low', 'medium', 'high', 'xhigh', 'max']) + } else if (modelId.includes('opus') && (modelId.includes('5-5') || modelId.includes('5.5'))) { + supported(['low', 'medium', 'high', 'xhigh', 'max']) } else if (modelId.includes('opus') && (modelId.includes('4-8') || modelId.includes('4.8'))) { supported(['low', 'medium', 'high', 'xhigh', 'max']) } else if (modelId.includes('opus') && (modelId.includes('4-7') || modelId.includes('4.7'))) { diff --git a/agents/codelayer/test/agent.test.ts b/agents/codelayer/test/agent.test.ts index e6d9354..a8511ed 100644 --- a/agents/codelayer/test/agent.test.ts +++ b/agents/codelayer/test/agent.test.ts @@ -525,6 +525,18 @@ describe('createCodelayerAgent', () => { expect(buildProviderOptions(model, overrides).openai.reasoningEffort).toBe('max') }) + test.each(['gpt-6-sol', 'gpt-6-luna'])('%s accepts every exposed CLI effort', (modelId) => { + const model = createMockModel(modelId) + + for (const thinking of ['low', 'medium', 'high', 'xhigh', 'max']) { + const overrides = applyCliThinkingOverride({ provider: 'codex', modelId, thinking, overrides: {} }) + expect(buildProviderOptions(model, overrides).openai.reasoningEffort).toBe(thinking) + } + expect(() => + applyCliThinkingOverride({ provider: 'codex', modelId, thinking: 'extreme', overrides: {} }), + ).toThrow('Unsupported --thinking value "extreme"') + }) + test('applies explicit CLI thinking for firepass kimi models', () => { const model = createMockModel(DEFAULT_MODELS.firepass) const overrides = applyCliThinkingOverride({ @@ -580,6 +592,43 @@ describe('createCodelayerAgent', () => { }) }) + test('uses summarized adaptive thinking and all exposed efforts for opus 5.5', () => { + const modelId = 'claude-opus-5-5' + const model = createMockModel(modelId) + + expect(buildProviderOptions(model).anthropic).toEqual({ + thinking: { type: 'adaptive', display: 'summarized' }, + effort: 'medium', + cacheControl: { type: 'ephemeral' }, + }) + for (const thinking of ['low', 'medium', 'high', 'xhigh', 'max']) { + const overrides = applyCliThinkingOverride({ provider: 'anthropic', modelId, thinking, overrides: {} }) + expect(buildProviderOptions(model, overrides).anthropic).toMatchObject({ + thinking: { type: 'adaptive', display: 'summarized' }, + effort: thinking, + }) + } + expect(() => + applyCliThinkingOverride({ provider: 'anthropic', modelId, thinking: 'extreme', overrides: {} }), + ).toThrow('Unsupported --thinking value "extreme"') + }) + + test('does not disable or downgrade opus 5.5 adaptive thinking through overrides', () => { + const model = createMockModel('claude-opus-5-5') + + for (const thinking of ['off', 'enabled'] as const) { + expect( + buildProviderOptions(model, { + anthropic: { thinking, budgetTokens: 10_000, effort: 'high' }, + }).anthropic, + ).toEqual({ + thinking: { type: 'adaptive', display: 'summarized' }, + effort: 'high', + cacheControl: { type: 'ephemeral' }, + }) + } + }) + test('uses summarized adaptive thinking with explicit xhigh CLI effort for opus 4.8', () => { const model = createMockModel('claude-opus-4-8') const overrides = applyCliThinkingOverride({ @@ -1132,6 +1181,47 @@ describe('subagentThinkingOverrides', () => { } }) + test.each(['gpt-6-sol', 'gpt-6-luna'])('%s preserves the root, child, research, and outline matrix', async (modelId) => { + const rootModel = createMockModel(modelId, 'codex') + const researchModel = createMockModel('gpt-5.6-terra', 'codex') + const agent = await createCodelayerAgent({ + model: rootModel, + researchModel, + cwd: '/tmp', + context7ApiKey: 'context7-test-key', + providerOptionOverrides: { codex: { reasoningEffort: 'max', fastMode: true } }, + }) + const rootConfig = getAgentConfig(agent) + const rootOptions = (rootConfig.providerOptions as unknown as (ctx: { runId: string }) => Record)({ + runId: 'root', + }) + const researchNames = new Set([ + 'rpi:codebase-locator', + 'rpi:codebase-analyzer', + 'rpi:codebase-pattern-finder', + 'web-search-researcher', + ]) + + expect(rootConfig.model).toBe(rootModel) + expect(rootOptions.openai).toMatchObject({ reasoningEffort: 'max', fastMode: true }) + for (const subagent of getSubagents(rootConfig.tools?.agent)) { + const config = getAgentConfig(subagent.agent) + const options = (config.providerOptions as unknown as (ctx: { runId: string }) => Record)({ + runId: subagent.name, + }) + if (researchNames.has(subagent.name)) { + expect(config.model, subagent.name).toBe(researchModel) + expect(options.openai.reasoningEffort, subagent.name).toBe('xhigh') + } else { + expect(config.model, subagent.name).toBe(rootModel) + expect(options.openai.reasoningEffort, subagent.name).toBe( + subagent.name === 'rpi:outline-implementer-agent' ? 'max' : 'low', + ) + } + expect(options.openai.fastMode, subagent.name).toBe(true) + } + }) + test('lets the outline implementer use parent anthropic effort while other sub-agents stay throttled', async () => { const agent = await createCodelayerAgent({ model: createMockModel('claude-opus-4-8'), @@ -1157,6 +1247,35 @@ describe('subagentThinkingOverrides', () => { expect(generalProviderOptions({ runId: 'general' }).anthropic.effort).toBe('low') expect(outlineProviderOptions({ runId: 'outline' }).anthropic.effort).toBe('max') }) + + test('keeps opus 5.5 adaptive across root and child roles while reducing ordinary children', async () => { + const model = createMockModel('claude-opus-5-5', 'anthropic') + const agent = await createCodelayerAgent({ + model, + cwd: '/tmp', + providerOptionOverrides: { anthropic: { effort: 'max' } }, + }) + const rootConfig = getAgentConfig(agent) + const rootOptions = (rootConfig.providerOptions as unknown as (ctx: { runId: string }) => Record)({ + runId: 'root', + }) + + expect(rootOptions.anthropic).toMatchObject({ + thinking: { type: 'adaptive', display: 'summarized' }, + effort: 'max', + }) + for (const subagent of getSubagents(rootConfig.tools?.agent)) { + const config = getAgentConfig(subagent.agent) + const options = (config.providerOptions as unknown as (ctx: { runId: string }) => Record)({ + runId: subagent.name, + }) + expect(config.model, subagent.name).toBe(model) + expect(options.anthropic.thinking, subagent.name).toEqual({ type: 'adaptive', display: 'summarized' }) + expect(options.anthropic.effort, subagent.name).toBe( + subagent.name === 'rpi:outline-implementer-agent' ? 'max' : 'low', + ) + } + }) }) describe('--subagent-thinking CLI knob', () => { diff --git a/package.json b/package.json index 2c7100e..94056a8 100644 --- a/package.json +++ b/package.json @@ -16,7 +16,7 @@ "hook:prepush": "env -i HOME=\"$HOME\" PATH=\"$PATH\" TMPDIR=\"$TMPDIR\" bun check", "models:refresh": "bun -e \"const data = await fetch('https://models.dev/api.json').then((response) => response.json()); await Bun.write('packages/agentlayer-core/models.json', JSON.stringify(data, null, 2) + '\\n')\"", "test": "bun test --concurrency 16 **/*", - "test:ci": "bun test --conditions=source ./packages/*/test/*.test.ts ./agents/*/test/*.test.ts", + "test:ci": "bun test --concurrency 8 --conditions=source ./packages/*/test/*.test.ts ./agents/*/test/*.test.ts", "typecheck": "bun --filter '*' --parallel typecheck", "release:prepare": "bun run scripts/release/prepare.ts", "release:publish": "bun run scripts/release/publish.ts" diff --git a/packages/agentlayer-core/models.json b/packages/agentlayer-core/models.json index 0d9132e..175878f 100644 --- a/packages/agentlayer-core/models.json +++ b/packages/agentlayer-core/models.json @@ -1,61 +1,57 @@ { - "requesty": { - "id": "requesty", + "deepinfra": { + "id": "deepinfra", "env": [ - "REQUESTY_API_KEY" + "DEEPINFRA_API_KEY" ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://router.requesty.ai/v1", - "name": "Requesty", - "doc": "https://requesty.ai/solution/llm-routing/models", + "npm": "@ai-sdk/deepinfra", + "name": "Deep Infra", + "doc": "https://deepinfra.com/models", "models": { - "xai/grok-4": { - "id": "xai/grok-4", - "name": "Grok 4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "tencent/Hy3": { + "id": "tencent/Hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-09-09", - "last_updated": "2025-09-09", + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 262144, + "input": 192000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75, - "cache_write": 3 + "input": 0.13, + "output": 0.53, + "cache_read": 0.033 } }, - "xai/grok-4-fast": { - "id": "xai/grok-4-fast", - "name": "Grok 4 Fast", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "meta-llama/Llama-3.3-70B-Instruct-Turbo": { + "id": "meta-llama/Llama-3.3-70B-Instruct-Turbo", + "name": "Llama 3.3 70B Turbo", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-09-19", - "last_updated": "2025-09-19", + "structured_output": true, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -64,411 +60,272 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 64000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05, - "cache_write": 0.2 + "input": 0.1, + "output": 0.32 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "meta-llama/Llama-4-Scout-17B-16E-Instruct": { + "id": "meta-llama/Llama-4-Scout-17B-16E-Instruct", + "name": "Llama 4 Scout 17B", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "structured_output": true, + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 327680, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.31, - "cache_write": 2.375, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.1, + "output": 0.3 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { + "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", + "name": "Llama 4 Maverick 17B FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "reasoning": false, + "tool_call": false, + "structured_output": true, + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 16384 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075, - "cache_write": 0.55 + "input": 0.2, + "output": 0.8 } }, - "google/gemini-3-pro-preview": { - "id": "google/gemini-3-pro-preview", - "name": "Gemini 3 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "XiaomiMiMo/MiMo-V2.5-Pro": { + "id": "XiaomiMiMo/MiMo-V2.5-Pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 16384 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 4.5 + "input": 1, + "output": 3, + "cache_read": 0.2 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "XiaomiMiMo/MiMo-V2.5": { + "id": "XiaomiMiMo/MiMo-V2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", "audio", - "video", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 1 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "openai/gpt-5.2-chat": { - "id": "openai/gpt-5.2-chat", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "thinkingmachines/Inkling-Small": { + "id": "thinkingmachines/Inkling-Small", + "name": "Inkling Small", + "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", + "family": "ling", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 524288, + "output": 1048576 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.45, + "output": 1.2, + "cache_read": 0.1 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "thinkingmachines/Inkling": { + "id": "thinkingmachines/Inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 524288, + "output": 1048576 }, "cost": { - "input": 21, - "output": 168 + "input": 0.95, + "output": 4.05, + "cache_read": 0.16 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma 3 27B IT", + "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text", - "audio", - "image", - "video" + "image" ], "output": [ - "text", - "audio", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.08, + "output": 0.16 } }, - "openai/gpt-5-chat": { - "id": "openai/gpt-5-chat", - "name": "GPT-5 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "google/gemma-3-12b-it": { + "id": "google/gemma-3-12b-it", + "name": "Gemma 3 12B IT", + "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": false, + "reasoning": false, + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text", @@ -478,44 +335,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.05, + "output": 0.15 } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", + "google/gemma-3-4b-it": { + "id": "google/gemma-3-4b-it", + "name": "Gemma 3 4B IT", + "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text", @@ -525,91 +367,70 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 272000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 15, - "output": 120 + "input": 0.05, + "output": 0.1 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4 Mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.28 + "input": 0.13, + "output": 0.38 } }, - "openai/gpt-5.1-chat": { - "id": "openai/gpt-5.1-chat", - "name": "GPT-5.1 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "google/gemma-4-26B-A4B-it": { + "id": "google/gemma-4-26B-A4B-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -619,93 +440,66 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.07, + "output": 0.34 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "google/gemma-4-E4B-it": { + "id": "google/gemma-4-E4B-it", + "name": "Gemma 4 E4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.02, + "output": 0.1 } }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT-5.1-Codex-Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "Qwen/Qwen3-VL-235B-A22B-Instruct": { + "id": "Qwen/Qwen3-VL-235B-A22B-Instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", @@ -715,146 +509,109 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.1, - "output": 9, + "input": 0.2, + "output": 0.88, "cache_read": 0.11 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "Qwen/Qwen3.8-27B": { + "id": "Qwen/Qwen3.8-27B", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.2, + "output": 2.5, + "cache_read": 0.05 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "Qwen/Qwen3-Next-80B-A3B-Instruct": { + "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.09, + "output": 1.1 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1-Codex-Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "Qwen/Qwen3.7-Max": { + "id": "Qwen/Qwen3.7-Max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -862,226 +619,184 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 100000 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "tiers": [ + { + "input": 5, + "output": 15, + "cache_read": 1, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 6.25, + "output": 18.5, + "cache_read": 1.25, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "openai/gpt-5-image": { - "id": "openai/gpt-5-image", - "name": "GPT-5 Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "Qwen/Qwen3.8-Max": { + "id": "Qwen/Qwen3.8-Max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-10-01", - "release_date": "2025-10-14", - "last_updated": "2025-10-14", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 256000, + "output": 131072 }, "cost": { - "input": 5, - "output": 10, - "cache_read": 1.25 + "input": 1.65, + "output": 4.951, + "cache_read": 0.206 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "Qwen/Qwen3.5-27B": { + "id": "Qwen/Qwen3.5-27B", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0.26, + "output": 2.6 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "output": 81920 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.1, + "output": 0.95 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo", + "name": "Qwen3 Coder 480B A35B Instruct Turbo", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 262144, + "output": 66536 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.03 + "input": 0.3, + "output": 1, + "cache_read": 0.1 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "Qwen/Qwen3-Max": { + "id": "Qwen/Qwen3-Max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -1089,42 +804,47 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 1.2, + "output": 6, + "cache_read": 0.24, + "tiers": [ + { + "input": 2.4, + "output": 12, + "cache_read": 0.48, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 3, + "output": 15, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B-A22B Instruct 2507", + "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": true, + "release_date": "2025-07-21", + "last_updated": "2025-07-21", "modalities": { "input": [ "text" @@ -1133,50 +853,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16000, - "output": 4000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01 + "input": 0.09, + "output": 0.55 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "Qwen/Qwen3.8-Flash": { + "id": "Qwen/Qwen3.8-Flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -1184,442 +886,411 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 30, - "output": 180, - "cache_read": 30 + "input": 0.113, + "output": 0.382, + "cache_read": 0.0141 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 40960, "output": 16384 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.08 + "input": 0.08, + "output": 0.28 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "GPT-5 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "Qwen/Qwen3.5-35B-A3B": { + "id": "Qwen/Qwen3.5-35B-A3B", + "name": "Qwen 3.5 35B A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-10-01", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "knowledge": "2025-01", + "release_date": "2026-02-01", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 81920 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.14, + "output": 1, + "cache_read": 0.05 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "Qwen/Qwen3.5-122B-A10B": { + "id": "Qwen/Qwen3.5-122B-A10B", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.29, + "output": 2.4 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "Qwen/Qwen3.5-9B": { + "id": "Qwen/Qwen3.5-9B", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.1, + "output": 0.15 } }, - "anthropic/claude-3-7-sonnet": { - "id": "anthropic/claude-3-7-sonnet", - "name": "Claude Sonnet 3.7", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen 3.5 397B A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-01", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "knowledge": "2025-01", + "release_date": "2026-02-01", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 81920 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.45, + "output": 3, + "cache_read": 0.22 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "Qwen/Qwen3.8-2.4T-A95B": { + "id": "Qwen/Qwen3.8-2.4T-A95B", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.2 + } + }, + "Qwen/Qwen3-30B-A3B": { + "id": "Qwen/Qwen3-30B-A3B", + "name": "Qwen3 30B A3B", + "description": "Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-04-28", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 40960, + "output": 16384 + }, + "cost": { + "input": 0.12, + "output": 0.5 + } + }, + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.32, + "output": 3.2 } }, - "anthropic/claude-opus-4-5": { - "id": "anthropic/claude-opus-4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek-V3.1", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 163840, + "output": 8192 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.25, + "output": 0.95, + "cache_read": 0.13 } }, - "anthropic/claude-sonnet-4-5": { - "id": "anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.06, + "output": 0.18, + "cache_read": 0.015 } }, - "anthropic/claude-opus-4-1": { - "id": "anthropic/claude-opus-4-1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Pro-0813": { + "id": "deepseek-ai/DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", "high", "max" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 1.3, + "output": 2.6, + "cache_read": 0.1 } }, - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "deepseek-ai/DeepSeek-V4.1-Flash": { + "id": "deepseek-ai/DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -1628,209 +1299,159 @@ "values": [ "none", "low", - "medium", "high", + "xhigh", "max" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-01", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 62000 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.2, + "output": 0.6, + "cache_read": 0.006 } }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", "high", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 16384 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 0.09, + "output": 0.18, + "cache_read": 0.018 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", "high", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 1.3, + "output": 2.6, + "cache_read": 0.1 } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "knowledge": "2024-12", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -1838,40 +1459,31 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 163840, + "output": 64000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.26, + "output": 0.38, + "cache_read": 0.13 } - } - } - }, - "qiniu-ai": { - "id": "qiniu-ai", - "env": [ - "QINIU_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.qnaigc.com/v1", - "name": "Qiniu", - "doc": "https://developer.qiniu.com/aitokenapi", - "models": { - "deepseek-r1-0528": { - "id": "deepseek-r1-0528", + }, + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", "name": "DeepSeek-R1-0528", "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-07", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -1882,22 +1494,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 163840, + "output": 64000 + }, + "cost": { + "input": 0.5, + "output": 2.15, + "cache_read": 0.35 } }, - "doubao-1.5-thinking-pro": { - "id": "doubao-1.5-thinking-pro", - "name": "Doubao 1.5 Thinking Pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "deepseek-ai/DeepSeek-V3": { + "id": "deepseek-ai/DeepSeek-V3", + "name": "DeepSeek-V3", + "description": "Open DeepSeek MoE chat model for coding, math, and general reasoning", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2024-12-26", + "last_updated": "2024-12-26", "modalities": { "input": [ "text" @@ -1906,50 +1523,70 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16000 + "context": 163840, + "output": 8192 + }, + "cost": { + "input": 0.32, + "output": 0.89 } }, - "qwen3-vl-30b-a3b-thinking": { - "id": "qwen3-vl-30b-a3b-thinking", - "name": "Qwen3-Vl 30b A3b Thinking", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "deepseek-ai/DeepSeek-V3-0324": { + "id": "deepseek-ai/DeepSeek-V3-0324", + "name": "DeepSeek V3 0324", + "description": "March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-09", - "last_updated": "2026-02-09", + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 163840, + "output": 163840 + }, + "cost": { + "input": 0.24, + "output": 0.9, + "cache_read": 0.135 } }, - "claude-3.5-haiku": { - "id": "claude-3.5-haiku", - "name": "Claude 3.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp": { + "id": "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", @@ -1959,73 +1596,100 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.44, + "output": 1.32, + "cache_read": 0.014 } }, - "deepseek-v3-0324": { - "id": "deepseek-v3-0324", - "name": "DeepSeek-V3-0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, - "reasoning": false, + "stepfun-ai/Step-3.7-Flash": { + "id": "stepfun-ai/Step-3.7-Flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16000 + "context": 262144, + "output": 256000 + }, + "cost": { + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 } }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235b A22B Instruct 2507", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, - "reasoning": false, + "MiniMaxAI/MiniMax-M3": { + "id": "MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-12", - "last_updated": "2025-08-12", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 64000 + "context": 524288, + "output": 512000 + }, + "cost": { + "input": 0.28, + "output": 1.1, + "cache_read": 0.056 } }, - "deepseek-v3": { - "id": "deepseek-v3", - "name": "DeepSeek-V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-08-13", - "last_updated": "2025-08-13", + "knowledge": "2025-06", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -2034,23 +1698,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16000 + "context": 196608, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0.15, + "output": 1.15, + "cache_read": 0.03 } }, - "kimi-k2": { - "id": "kimi-k2", - "name": "Kimi K2", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "MiniMaxAI/MiniMax-M2.7": { + "id": "MiniMaxAI/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -2059,17 +1730,24 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 196608, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0.25, + "output": 1, + "cache_read": 0.05 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -2077,188 +1755,295 @@ } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-04", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 40000, - "output": 4096 + "context": 262144, + "output": 16384 + }, + "cost": { + "input": 0.75, + "output": 3.5, + "cache_read": 0.15 } }, - "qwen3-max-preview": { - "id": "qwen3-max-preview", - "name": "Qwen3 Max Preview", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, - "reasoning": false, + "moonshotai/Kimi-K3": { + "id": "moonshotai/Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-06", - "last_updated": "2025-09-06", + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 2.85, + "output": 14.25, + "cache_read": 0.285 } }, - "claude-3.5-sonnet": { - "id": "claude-3.5-sonnet", - "name": "Claude 3.5 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-09", - "last_updated": "2025-09-09", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8200 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.68, + "output": 3.4, + "cache_read": 0.136 } }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, - "reasoning": false, + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-09-12", - "last_updated": "2025-09-12", + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, + "context": 262144, "output": 32768 + }, + "status": "deprecated", + "cost": { + "input": 0.45, + "output": 2.25, + "cache_read": 0.07 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "attachment": true, + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-04", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 202752, + "output": 16384 + }, + "cost": { + "input": 1.05, + "output": 3.5, + "cache_read": 0.205 } }, - "claude-4.5-haiku": { - "id": "claude-4.5-haiku", - "name": "Claude 4.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "attachment": true, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-10-16", - "last_updated": "2025-10-16", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 32768 + }, + "cost": { + "input": 0.75, + "output": 2.4, + "cache_read": 0.14 } }, - "kling-v2-6": { - "id": "kling-v2-6", - "name": "Kling-V2 6", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "zai-org/GLM-4.7": { + "id": "zai-org/GLM-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-01-13", - "last_updated": "2026-01-13", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 99999999, - "output": 99999999 + "context": 202752, + "output": 16384 + }, + "cost": { + "input": 0.4, + "output": 1.75, + "cache_read": 0.08 } }, - "glm-4.5": { - "id": "glm-4.5", - "name": "GLM 4.5", + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM-5", "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -2267,10 +2052,14 @@ } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-12", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -2279,79 +2068,119 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 202752, + "output": 16384 + }, + "status": "deprecated", + "cost": { + "input": 0.6, + "output": 2.08, + "cache_read": 0.12 } }, - "claude-4.1-opus": { - "id": "claude-4.1-opus", - "name": "Claude 4.1 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "zai-org/GLM-5.3-Flash": { + "id": "zai-org/GLM-5.3-Flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "attachment": true, + "zai-org/GLM-4.6": { + "id": "zai-org/GLM-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 64000 + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 0.5, + "output": 2, + "cache_read": 0.1 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "zai-org/GLM-4.7-Flash": { + "id": "zai-org/GLM-4.7-Flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -2360,49 +2189,65 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 202752, + "output": 16384 + }, + "status": "deprecated", + "cost": { + "input": 0.06, + "output": 0.4, + "cache_read": 0.01 } }, - "doubao-seed-2.0-pro": { - "id": "doubao-seed-2.0-pro", - "name": "Doubao Seed 2.0 Pro", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "zai-org/GLM-5.3": { + "id": "zai-org/GLM-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 128000 + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.2 } }, - "doubao-seed-1.6": { - "id": "doubao-seed-1.6", - "name": "Doubao-Seed 1.6", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "nvidia/Nemotron-3-Nano-30B-A3B": { + "id": "nvidia/Nemotron-3-Nano-30B-A3B", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -2410,96 +2255,122 @@ } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-15", - "last_updated": "2025-08-15", + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.05, + "output": 0.2, + "cache_read": 0.025 } }, - "doubao-seed-1.6-thinking": { - "id": "doubao-seed-1.6-thinking", - "name": "Doubao-Seed 1.6 Thinking", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "nvidia/Llama-3.3-Nemotron-Super-49B-v1.5": { + "id": "nvidia/Llama-3.3-Nemotron-Super-49B-v1.5", + "name": "Llama 3.3 Nemotron Super 49B v1.5", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-15", - "last_updated": "2025-08-15", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 131072, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 0.4 } }, - "gemini-2.0-flash": { - "id": "gemini-2.0-flash", - "name": "Gemini 2.0 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning": { + "id": "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning", + "name": "Nemotron 3 Nano Omni 30B A3B Reasoning", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ "text", "image", - "audio", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 8192 + "context": 262144, + "output": 65536 + }, + "status": "deprecated", + "cost": { + "input": 0.2, + "output": 0.8 } }, - "qwen-max-2025-01-25": { - "id": "qwen-max-2025-01-25", - "name": "Qwen2.5-Max-2025-01-25", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "attachment": false, - "reasoning": false, + "ByteDance/Seed-2.0-code": { + "id": "ByteDance/Seed-2.0-code", + "name": "Seed 2.0 Code", + "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -2507,22 +2378,39 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 256000, + "output": 131072 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.1, + "tiers": [ + { + "input": 1, + "output": 6, + "cache_read": 0.2, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "claude-4.0-sonnet": { - "id": "claude-4.0-sonnet", - "name": "Claude 4.0 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "ByteDance/Seed-2.0-pro": { + "id": "ByteDance/Seed-2.0-pro", + "name": "Seed 2.0 Pro", + "description": "Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", @@ -2534,24 +2422,43 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.1, + "tiers": [ + { + "input": 1, + "output": 6, + "cache_read": 0.2, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "doubao-1.5-pro-32k": { - "id": "doubao-1.5-pro-32k", - "name": "Doubao 1.5 Pro 32k", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, - "reasoning": false, + "ByteDance/Seed-2.0-mini": { + "id": "ByteDance/Seed-2.0-mini", + "name": "Seed 2.0 Mini", + "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -2559,21 +2466,48 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 12000 + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.02, + "tiers": [ + { + "input": 0.2, + "output": 0.8, + "cache_read": 0.2, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "qwen3-30b-a3b-instruct-2507": { - "id": "qwen3-30b-a3b-instruct-2507", - "name": "Qwen3 30b A3b Instruct 2507", - "description": "Tool-capable chat model for instruction following and agentic application workflows", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-04", - "last_updated": "2026-02-04", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -2582,28 +2516,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0.03, + "output": 0.14 } }, - "qwen3-next-80b-a3b-thinking": { - "id": "qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B Thinking", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-09-12", - "last_updated": "2025-09-12", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -2612,31 +2556,57 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, - "output": 32768 + "output": 16384 + }, + "cost": { + "input": 0.037, + "output": 0.17 } - }, - "qwen3-235b-a22b-thinking-2507": { - "id": "qwen3-235b-a22b-thinking-2507", - "name": "Qwen3 235B A22B Thinking 2507", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + } + } + }, + "perplexity-agent": { + "id": "perplexity-agent", + "env": [ + "PERPLEXITY_API_KEY" + ], + "npm": "@ai-sdk/openai", + "api": "https://api.perplexity.ai/v1", + "name": "Perplexity Agent", + "doc": "https://docs.perplexity.ai/docs/agent-api/models", + "models": { + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-12", - "last_updated": "2025-08-12", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -2644,25 +2614,44 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 4096 + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1 } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "attachment": false, + "anthropic/claude-opus-4-5": { + "id": "anthropic/claude-opus-4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -2670,26 +2659,44 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "doubao-1.5-vision-pro": { - "id": "doubao-1.5-vision-pro", - "name": "Doubao 1.5 Vision Pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "anthropic/claude-sonnet-4-5": { + "id": "anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -2697,80 +2704,134 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16000 + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "gemini-3.0-pro-image-preview": { - "id": "gemini-3.0-pro-image-preview", - "name": "Gemini 3.0 Pro Image Preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "gemini-2.5-flash-image": { - "id": "gemini-2.5-flash-image", - "name": "Gemini 2.5 Flash Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-10-22", - "last_updated": "2025-10-22", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -2778,53 +2839,38 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 64000 - } - }, - "claude-3.7-sonnet": { - "id": "claude-3.7-sonnet", - "name": "Claude 3.7 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, + "context": 1000000, "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "qwen3-30b-a3b-thinking-2507": { - "id": "qwen3-30b-a3b-thinking-2507", - "name": "Qwen3 30b A3b Thinking 2507", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "deepseek/deepseek-v4-flash-0731": { + "id": "deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-04", - "last_updated": "2026-02-04", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -2833,29 +2879,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 126000, - "output": 32000 + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.13, + "output": 0.26, + "cache_read": 0.028 } }, - "qwen2.5-vl-72b-instruct": { - "id": "qwen2.5-vl-72b-instruct", - "name": "Qwen 2.5 VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "perplexity/sonar": { + "id": "perplexity/sonar", + "name": "Sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -2865,23 +2914,44 @@ "limit": { "context": 128000, "output": 8192 + }, + "cost": { + "input": 0.25, + "output": 2.5, + "cache_read": 0.0625 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "attachment": false, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -2889,31 +2959,62 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "doubao-seed-1.6-flash": { - "id": "doubao-seed-1.6-flash", - "name": "Doubao-Seed 1.6 Flash", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-15", - "last_updated": "2025-08-15", + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-05", "modalities": { "input": [ "text", "image", - "video" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -2921,25 +3022,62 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "deepseek-v3.1": { - "id": "deepseek-v3.1", - "name": "DeepSeek-V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-19", - "last_updated": "2025-08-19", + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -2947,24 +3085,46 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "qwen3-235b-a22b": { - "id": "qwen3-235b-a22b", - "name": "Qwen 3 235B A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -2972,24 +3132,47 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "tiers": [ + { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 0.5, + "output": 3, + "cache_read": 0.05 + } } }, - "qwen3-coder-480b-a35b-instruct": { - "id": "qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, + "xai/grok-4-1-fast-non-reasoning": { + "id": "xai/grok-4-1-fast-non-reasoning", + "name": "Grok 4.1 Fast (Non-Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-14", - "last_updated": "2025-08-14", + "knowledge": "2025-07", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -2997,26 +3180,39 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 4096 + "context": 2000000, + "output": 30000 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "xai/grok-4.6": { + "id": "xai/grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-22", - "last_updated": "2026-02-22", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text", @@ -3028,23 +3224,55 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } }, - "mimo-v2-flash": { - "id": "mimo-v2-flash", - "name": "Mimo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "Nemotron 3 Super 120B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-12-01", - "release_date": "2025-12-16", - "last_updated": "2026-02-04", + "knowledge": "2026-02", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text" @@ -3055,90 +3283,129 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01 + "input": 0.25, + "output": 2.5 } }, - "qwen-vl-max-2025-01-25": { - "id": "qwen-vl-max-2025-01-25", - "name": "Qwen VL-MAX-2025-01-25", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "moonshot-ai/kimi-k3": { + "id": "moonshot-ai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "qwen2.5-vl-7b-instruct": { - "id": "qwen2.5-vl-7b-instruct", - "name": "Qwen 2.5 VL 7B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "moonshot-ai/kimi-k2.7-code": { + "id": "moonshot-ai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "GLM 4.5 Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "attachment": false, + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -3146,22 +3413,40 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 4096 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "claude-4.5-opus": { - "id": "claude-4.5-opus", - "name": "Claude 4.5 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -3173,29 +3458,44 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 200000 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "gemini-3.0-pro-preview": { - "id": "gemini-3.0-pro-preview", - "name": "Gemini 3.0 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "video", - "pdf", - "audio" + "image" ], "output": [ "text" @@ -3203,31 +3503,44 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "doubao-seed-2.0-mini": { - "id": "doubao-seed-2.0-mini", - "name": "Doubao Seed 2.0 Mini", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -3235,26 +3548,45 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "claude-4.0-opus": { - "id": "claude-4.0-opus", - "name": "Claude 4.0 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -3262,22 +3594,41 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 } - }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "gpt-oss-20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + } + } + }, + "bailing": { + "id": "bailing", + "env": [ + "BAILING_API_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.tbox.cn/api/llm/v1/chat/completions", + "name": "Bailing", + "doc": "https://alipaytbox.yuque.com/sxs0ba/ling/intro", + "models": { + "Ring-1T": { + "id": "Ring-1T", + "name": "Ring-1T", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "ring", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": false, + "tool_call": false, "temperature": true, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "knowledge": "2024-06", + "release_date": "2025-10", + "last_updated": "2025-10", "modalities": { "input": [ "text" @@ -3286,31 +3637,72 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 4096 + "output": 32000 + }, + "cost": { + "input": 0.57, + "output": 2.29 } }, - "gemini-3.0-flash-preview": { - "id": "gemini-3.0-flash-preview", - "name": "Gemini 3.0 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "Ling-1T": { + "id": "Ling-1T", + "name": "Ling-1T", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "ling", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-10", + "last_updated": "2025-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32000 + }, + "cost": { + "input": 0.57, + "output": 2.29 + } + } + } + }, + "poe": { + "id": "poe", + "env": [ + "POE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.poe.com/v1", + "name": "Poe", + "doc": "https://creator.poe.com/docs/external-applications/openai-compatible-api", + "models": { + "cerebras/qwen3-32b-cs": { + "id": "cerebras/qwen3-32b-cs", + "name": "qwen3-32b-cs", + "description": "Legacy model retained for compatibility with older integrations", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "temperature": false, + "release_date": "2025-05-15", + "last_updated": "2025-05-15", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -3318,26 +3710,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 - } + "context": 0, + "output": 0 + }, + "status": "deprecated" }, - "MiniMax-M1": { - "id": "MiniMax-M1", - "name": "MiniMax M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "cerebras/llama-3.1-8b-cs": { + "id": "cerebras/llama-3.1-8b-cs", + "name": "Llama-3.1-8B-CS", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "release_date": "2025-05-13", + "last_updated": "2025-05-13", "modalities": { "input": [ "text" @@ -3348,26 +3735,24 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 80000 + "context": 128000, + "output": 0 + }, + "cost": { + "input": 0.1, + "output": 0.1 } }, - "qwen-turbo": { - "id": "qwen-turbo", - "name": "Qwen-Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "cerebras/llama-3.3-70b-cs": { + "id": "cerebras/llama-3.3-70b-cs", + "name": "llama-3.3-70b-cs", + "description": "Legacy model retained for compatibility with older integrations", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-05-13", + "last_updated": "2025-05-13", "modalities": { "input": [ "text" @@ -3378,26 +3763,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 4096 - } + "context": 0, + "output": 0 + }, + "status": "deprecated" }, - "qwen3-30b-a3b": { - "id": "qwen3-30b-a3b", - "name": "Qwen3 30B A3B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "cerebras/gpt-oss-120b-cs": { + "id": "cerebras/gpt-oss-120b-cs", + "name": "GPT-OSS-120B-CS", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ "text" @@ -3408,26 +3798,28 @@ }, "open_weights": false, "limit": { - "context": 40000, - "output": 4096 + "context": 128000, + "output": 0 + }, + "cost": { + "input": 0.35, + "output": 0.75 } }, - "claude-4.5-sonnet": { - "id": "claude-4.5-sonnet", - "name": "Claude 4.5 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "cerebras/qwen3-235b-2507-cs": { + "id": "cerebras/qwen3-235b-2507-cs", + "name": "qwen3-235b-2507-cs", + "description": "Legacy model retained for compatibility with older integrations", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "temperature": false, + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -3435,92 +3827,85 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - } + "context": 0, + "output": 0 + }, + "status": "deprecated" }, - "doubao-seed-2.0-lite": { - "id": "doubao-seed-2.0-lite", - "name": "Doubao Seed 2.0 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "empiriolabs/deepseek-v4-pro-el": { + "id": "empiriolabs/deepseek-v4-pro-el", + "name": "DeepSeek-V4-Pro-EL", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-04-24", + "last_updated": "2026-05-02", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 1000000, + "input": 1000000, + "output": 384000 + }, + "cost": { + "input": 1.67, + "output": 3.33 } }, - "gemini-2.0-flash-lite": { - "id": "gemini-2.0-flash-lite", - "name": "Gemini 2.0 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "empiriolabs/deepseek-v4-flash-el": { + "id": "empiriolabs/deepseek-v4-flash-el", + "name": "DeepSeek-V4-Flash-EL", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-24", + "last_updated": "2026-05-02", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 8192 + "context": 1000000, + "input": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28 } }, - "doubao-seed-2.0-code": { - "id": "doubao-seed-2.0-code", - "name": "Doubao Seed 2.0 Code", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "anthropic/claude-haiku-3.5": { + "id": "anthropic/claude-haiku-3.5", + "name": "Claude-Haiku-3.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "temperature": false, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -3528,24 +3913,39 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 - } - }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "attachment": false, - "reasoning": false, + "context": 189096, + "output": 8192 + }, + "cost": { + "input": 0.68, + "output": 3.4, + "cache_read": 0.068, + "cache_write": 0.85 + } + }, + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude-Opus-4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 0, + "max": 31999 + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-11-07", - "last_updated": "2025-11-07", + "temperature": false, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -3553,26 +3953,47 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 100000 + "context": 196608, + "output": 32000 + }, + "cost": { + "input": 13, + "output": 64, + "cache_read": 1.3, + "cache_write": 16 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Moonshotai/Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude-Opus-4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 63999 + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-01-28", - "last_updated": "2026-01-28", + "temperature": false, + "release_date": "2025-11-21", + "last_updated": "2025-11-21", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -3580,24 +4001,43 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 196608, + "output": 64000 + }, + "cost": { + "input": 4.3, + "output": 21, + "cache_read": 0.43, + "cache_write": 5.3 } }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude-Opus-4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-08", - "last_updated": "2025-09-08", + "temperature": false, + "release_date": "2026-02-04", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -3605,25 +4045,32 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 100000 + "context": 983040, + "output": 128000 + }, + "cost": { + "input": 4.3, + "output": 21, + "cache_read": 0.43, + "cache_write": 5.3 } }, - "stepfun-ai/gelab-zero-4b-preview": { - "id": "stepfun-ai/gelab-zero-4b-preview", - "name": "Stepfun-Ai/Gelab Zero 4b Preview", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "anthropic/claude-sonnet-3.5-june": { + "id": "anthropic/claude-sonnet-3.5-june", + "name": "Claude-Sonnet-3.5-June", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-sonnet", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "temperature": false, + "release_date": "2024-11-18", + "last_updated": "2024-11-18", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -3631,25 +4078,45 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 189096, + "output": 8192 + }, + "status": "deprecated", + "cost": { + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "x-ai/grok-code-fast-1": { - "id": "x-ai/grok-code-fast-1", - "name": "x-AI/Grok-Code-Fast 1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": false, + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude-Opus-4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-02", - "last_updated": "2025-09-02", + "temperature": false, + "release_date": "2026-04-15", + "last_updated": "2026-04-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -3657,28 +4124,32 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 10000 + "context": 1048576, + "output": 128000 + }, + "cost": { + "input": 4.3, + "output": 21, + "cache_read": 0.43, + "cache_write": 5.4 } }, - "x-ai/grok-4.1-fast-reasoning": { - "id": "x-ai/grok-4.1-fast-reasoning", - "name": "X-Ai/Grok 4.1 Fast Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "anthropic/claude-haiku-3": { + "id": "anthropic/claude-haiku-3", + "name": "Claude-Haiku-3", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-19", - "last_updated": "2025-12-19", + "temperature": false, + "release_date": "2024-03-09", + "last_updated": "2024-03-09", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -3686,27 +4157,46 @@ }, "open_weights": false, "limit": { - "context": 20000000, - "output": 2000000 + "context": 189096, + "output": 8192 + }, + "cost": { + "input": 0.21, + "output": 1.1, + "cache_read": 0.021, + "cache_write": 0.26 } }, - "x-ai/grok-4.1-fast-non-reasoning": { - "id": "x-ai/grok-4.1-fast-non-reasoning", - "name": "X-Ai/Grok 4.1 Fast Non Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude-Opus-4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-19", - "last_updated": "2025-12-19", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -3714,28 +4204,31 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 1048576, + "output": 128000 + }, + "cost": { + "input": 4.2929, + "output": 21.4646 } }, - "x-ai/grok-4-fast-reasoning": { - "id": "x-ai/grok-4-fast-reasoning", - "name": "X-Ai/Grok-4-Fast-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "anthropic/claude-sonnet-3.7": { + "id": "anthropic/claude-sonnet-3.7", + "name": "Claude-Sonnet-3.7", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "temperature": false, + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -3743,28 +4236,39 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 196608, + "output": 128000 + }, + "cost": { + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "x-ai/grok-4-fast": { - "id": "x-ai/grok-4-fast", - "name": "x-AI/Grok-4-Fast", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude-Haiku-4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 0, + "max": 63999 + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-20", - "last_updated": "2025-09-20", + "temperature": false, + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -3772,27 +4276,39 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 192000, + "output": 64000 + }, + "cost": { + "input": 0.85, + "output": 4.3, + "cache_read": 0.085, + "cache_write": 1.1 } }, - "x-ai/grok-4-fast-non-reasoning": { - "id": "x-ai/grok-4-fast-non-reasoning", - "name": "X-Ai/Grok-4-Fast-Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude-Sonnet-4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 0, + "max": 31999 + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "temperature": false, + "release_date": "2025-09-26", + "last_updated": "2025-09-26", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -3800,25 +4316,33 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 983040, + "output": 32768 + }, + "cost": { + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "x-ai/grok-4.1-fast": { - "id": "x-ai/grok-4.1-fast", - "name": "x-AI/Grok-4.1-Fast", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": false, + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude-Sonnet-4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "temperature": false, + "release_date": "2025-05-21", + "last_updated": "2025-05-21", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -3826,25 +4350,32 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 983040, + "output": 64000 + }, + "cost": { + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "z-ai/autoglm-phone-9b": { - "id": "z-ai/autoglm-phone-9b", - "name": "Z-Ai/Autoglm Phone 9b", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "anthropic/claude-sonnet-3.5": { + "id": "anthropic/claude-sonnet-3.5", + "name": "Claude-Sonnet-3.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-sonnet", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "temperature": false, + "release_date": "2024-06-05", + "last_updated": "2024-06-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -3852,29 +4383,44 @@ }, "open_weights": false, "limit": { - "context": 12800, - "output": 4096 + "context": 189096, + "output": 8192 + }, + "status": "deprecated", + "cost": { + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "Z-Ai/GLM 4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude-Sonnet-4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "temperature": false, + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -3882,54 +4428,33 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 200000 - } - }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "Z-AI/GLM 4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-10-11", - "last_updated": "2025-10-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 983040, + "output": 128000 }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 200000 + "cost": { + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "Z-Ai/GLM 5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Claude-Opus-4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "temperature": false, + "release_date": "2025-05-21", + "last_updated": "2025-05-21", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -3937,105 +4462,106 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 192512, + "output": 28672 + }, + "cost": { + "input": 13, + "output": 64, + "cache_read": 1.3, + "cache_write": 16 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "OpenAI/GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": false, + "elevenlabs/elevenlabs-v2.5-turbo": { + "id": "elevenlabs/elevenlabs-v2.5-turbo", + "name": "ElevenLabs-v2.5-Turbo", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "elevenlabs", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-19", - "last_updated": "2025-09-19", + "temperature": false, + "release_date": "2024-10-28", + "last_updated": "2024-10-28", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 128000, + "output": 0 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "OpenAI/GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "elevenlabs/elevenlabs-v3": { + "id": "elevenlabs/elevenlabs-v3", + "name": "ElevenLabs-v3", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "elevenlabs", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": false, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 128000, + "output": 0 } }, - "xiaomi/mimo-v2-flash": { - "id": "xiaomi/mimo-v2-flash", - "name": "Xiaomi/Mimo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "elevenlabs/elevenlabs-music": { + "id": "elevenlabs/elevenlabs-music", + "name": "ElevenLabs-Music", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "elevenlabs", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-12-01", - "release_date": "2025-12-16", - "last_updated": "2026-02-04", + "temperature": false, + "release_date": "2025-08-29", + "last_updated": "2025-08-29", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01 + "context": 2000, + "output": 0 } }, - "stepfun/step-3.5-flash": { - "id": "stepfun/step-3.5-flash", - "name": "Stepfun/Step-3.5 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "novita/glm-4.6v": { + "id": "novita/glm-4.6v", + "name": "glm-4.6v", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-02", - "last_updated": "2026-02-02", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": false, + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ "text", @@ -4047,21 +4573,21 @@ }, "open_weights": false, "limit": { - "context": 64000, - "output": 4096 + "context": 131000, + "output": 32768 } }, - "meituan/longcat-flash-lite": { - "id": "meituan/longcat-flash-lite", - "name": "Meituan/Longcat-Flash-Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, + "novita/glm-4.6": { + "id": "novita/glm-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "temperature": false, + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -4072,21 +4598,25 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 320000 + "context": 0, + "output": 0 } }, - "meituan/longcat-flash-chat": { - "id": "meituan/longcat-flash-chat", - "name": "Meituan/Longcat-Flash-Chat", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "novita/glm-5": { + "id": "novita/glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-11-05", - "last_updated": "2025-11-05", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text" @@ -4097,47 +4627,69 @@ }, "open_weights": false, "limit": { - "context": 131072, + "context": 205000, "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "deepseek/deepseek-v3.1-terminus": { - "id": "deepseek/deepseek-v3.1-terminus", - "name": "DeepSeek/DeepSeek-V3.1-Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, - "reasoning": false, + "novita/kimi-k2.6": { + "id": "novita/kimi-k2.6", + "name": "Kimi-K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-05-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 262144, + "input": 262144, + "output": 262144 + }, + "cost": { + "input": 0.96, + "output": 4.04, + "cache_read": 0.16 } }, - "deepseek/deepseek-v3.1-terminus-thinking": { - "id": "deepseek/deepseek-v3.1-terminus-thinking", - "name": "DeepSeek/DeepSeek-V3.1-Terminus-Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "novita/kimi-k2-thinking": { + "id": "novita/kimi-k2-thinking", + "name": "kimi-k2-thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": false, + "release_date": "2025-11-07", + "last_updated": "2025-11-07", "modalities": { "input": [ "text" @@ -4148,25 +4700,30 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 256000, + "output": 0 } }, - "deepseek/deepseek-v3.2-exp-thinking": { - "id": "deepseek/deepseek-v3.2-exp-thinking", - "name": "DeepSeek/DeepSeek-V3.2-Exp-Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "novita/kimi-k2.5": { + "id": "novita/kimi-k2.5", + "name": "Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -4175,20 +4732,29 @@ "open_weights": false, "limit": { "context": 128000, - "output": 32000 + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "deepseek/deepseek-v3.2-exp": { - "id": "deepseek/deepseek-v3.2-exp", - "name": "DeepSeek/DeepSeek-V3.2-Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, - "reasoning": false, + "novita/glm-4.7-flash": { + "id": "novita/glm-4.7-flash", + "name": "glm-4.7-flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "temperature": false, + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -4199,22 +4765,21 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 200000, + "output": 65500 } }, - "deepseek/deepseek-v3.2-251201": { - "id": "deepseek/deepseek-v3.2-251201", - "name": "Deepseek/DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "novita/glm-4.7": { + "id": "novita/glm-4.7", + "name": "glm-4.7", + "description": "Legacy model retained for compatibility with older integrations", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -4225,22 +4790,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 - } + "context": 205000, + "output": 131072 + }, + "status": "deprecated" }, - "deepseek/deepseek-math-v2": { - "id": "deepseek/deepseek-math-v2", - "name": "Deepseek/Deepseek-Math-V2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "novita/glm-4.7-n": { + "id": "novita/glm-4.7-n", + "name": "glm-4.7-n", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-04", - "last_updated": "2025-12-04", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": false, + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -4251,15 +4820,15 @@ }, "open_weights": false, "limit": { - "context": 160000, - "output": 160000 + "context": 205000, + "output": 131072 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "Minimax/Minimax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": false, + "novita/deepseek-v3.2": { + "id": "novita/deepseek-v3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -4267,10 +4836,9 @@ } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -4279,17 +4847,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 128000 + "context": 128000, + "output": 0 + }, + "cost": { + "input": 0.27, + "output": 0.4, + "cache_read": 0.13 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "Minimax/Minimax-M2.1", + "novita/minimax-m2.1": { + "id": "novita/minimax-m2.1", + "name": "minimax-m2.1", "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -4297,10 +4870,9 @@ } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "temperature": false, + "release_date": "2025-12-26", + "last_updated": "2025-12-26", "modalities": { "input": [ "text" @@ -4311,56 +4883,47 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 128000 + "context": 205000, + "output": 131072 } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "Minimax/Minimax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "lumalabs/ray2": { + "id": "lumalabs/ray2", + "name": "Ray2", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ray", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", + "temperature": false, + "release_date": "2025-02-20", + "last_updated": "2025-02-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 5000, + "output": 0 } }, - "minimax/minimax-m2.5-highspeed": { - "id": "minimax/minimax-m2.5-highspeed", - "name": "Minimax/Minimax-M2.5 Highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "attachment": false, + "poetools/claude-code": { + "id": "poetools/claude-code", + "name": "claude-code", + "description": "Claude model for careful reasoning, writing, coding, and tool use", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "temperature": false, + "release_date": "2025-11-27", + "last_updated": "2025-11-27", "modalities": { "input": [ "text" @@ -4371,524 +4934,454 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 128000 + "context": 0, + "output": 0 } - } - } - }, - "alibaba-cn": { - "id": "alibaba-cn", - "env": [ - "DASHSCOPE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://dashscope.aliyuncs.com/compatible-mode/v1", - "name": "Alibaba (China)", - "doc": "https://www.alibabacloud.com/help/en/model-studio/models", - "models": { - "qwen2-5-math-72b-instruct": { - "id": "qwen2-5-math-72b-instruct", - "name": "Qwen2.5-Math 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + }, + "google/imagen-3-fast": { + "id": "google/imagen-3-fast", + "name": "Imagen-3-Fast", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "temperature": false, + "release_date": "2024-10-17", + "last_updated": "2024-10-17", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 4096, - "output": 3072 - }, - "cost": { - "input": 0.574, - "output": 1.721 + "context": 480, + "output": 0 } }, - "deepseek-r1-0528": { - "id": "deepseek-r1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "google/imagen-4-ultra": { + "id": "google/imagen-4-ultra", + "name": "Imagen-4-Ultra", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "temperature": false, + "release_date": "2025-05-24", + "last_updated": "2025-05-24", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.574, - "output": 2.294 + "context": 480, + "output": 0 } }, - "qwen3-omni-flash": { - "id": "qwen3-omni-flash", - "name": "Qwen3-Omni Flash", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, + "google/gemini-deep-research": { + "id": "google/gemini-deep-research", + "name": "gemini-deep-research", + "description": "Legacy model retained for compatibility with older integrations", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "temperature": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 1048576, + "output": 0 }, + "status": "deprecated", "cost": { - "input": 0.058, - "output": 0.23, - "input_audio": 3.584, - "output_audio": 7.168 + "input": 1.6, + "output": 9.6 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "google/imagen-4": { + "id": "google/imagen-4", + "name": "Imagen-4", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "context": 480, + "output": 0 } }, - "qwen-plus": { - "id": "qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], + "google/imagen-4-fast": { + "id": "google/imagen-4-fast", + "name": "Imagen-4-Fast", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", + "temperature": false, + "release_date": "2025-06-25", + "last_updated": "2025-06-25", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 0.115, - "output": 0.287, - "reasoning": 1.147 + "context": 480, + "output": 0 } }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "google/lyria": { + "id": "google/lyria", + "name": "Lyria", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "lyria", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2025-06-04", + "last_updated": "2025-06-04", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.216, - "output": 0.861 + "context": 0, + "output": 0 } }, - "qwen2-5-coder-7b-instruct": { - "id": "qwen2-5-coder-7b-instruct", - "name": "Qwen2.5-Coder 7B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "google/nano-banana": { + "id": "google/nano-banana", + "name": "Nano-Banana", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "nano-banana", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11", - "last_updated": "2024-11", + "temperature": false, + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 65536, + "output": 0 }, "cost": { - "input": 0.144, - "output": 0.287 + "input": 0.21, + "output": 1.8, + "cache_read": 0.021 } }, - "deepseek-v3": { - "id": "deepseek-v3", - "name": "DeepSeek V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "google/veo-3-fast": { + "id": "google/veo-3-fast", + "name": "Veo-3-Fast", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "temperature": false, + "release_date": "2025-10-13", + "last_updated": "2025-10-13", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 8192 - }, - "cost": { - "input": 0.287, - "output": 1.147 + "context": 480, + "output": 0 } }, - "qwen3-omni-flash-realtime": { - "id": "qwen3-omni-flash-realtime", - "name": "Qwen3-Omni Flash Realtime", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, + "google/gemini-2.0-flash": { + "id": "google/gemini-2.0-flash", + "name": "Gemini-2.0-Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "temperature": false, + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text", "image", + "video", "audio" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 990000, + "output": 8192 }, "cost": { - "input": 0.23, - "output": 0.918, - "input_audio": 3.584, - "output_audio": 7.168 + "input": 0.1, + "output": 0.42 } }, - "deepseek-r1-distill-llama-70b": { - "id": "deepseek-r1-distill-llama-70b", - "name": "DeepSeek R1 Distill Llama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "google/veo-3.1": { + "id": "google/veo-3.1", + "name": "Veo-3.1", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "temperature": false, + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 0.287, - "output": 0.861 + "context": 480, + "output": 0 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 38912 - } - ], + "google/nano-banana-pro": { + "id": "google/nano-banana-pro", + "name": "Nano-Banana-Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "nano-banana", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 65536, + "output": 0 }, "cost": { - "input": 0.287, - "output": 1.147, - "reasoning": 2.868 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "qwen-omni-turbo-realtime": { - "id": "qwen-omni-turbo-realtime", - "name": "Qwen-Omni Turbo Realtime", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, + "google/gemini-2.0-flash-lite": { + "id": "google/gemini-2.0-flash-lite", + "name": "Gemini-2.0-Flash-Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-05-08", - "last_updated": "2025-05-08", + "temperature": false, + "release_date": "2025-02-05", + "last_updated": "2025-02-05", "modalities": { "input": [ "text", "image", + "video", "audio" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 2048 + "context": 990000, + "output": 8192 }, "cost": { - "input": 0.23, - "output": 0.918, - "input_audio": 3.584, - "output_audio": 7.168 + "input": 0.052, + "output": 0.21 } }, - "qwen2-5-math-7b-instruct": { - "id": "qwen2-5-math-7b-instruct", - "name": "Qwen2.5-Math 7B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "google/gemini-3-flash": { + "id": "google/gemini-3-flash", + "name": "Gemini-3-Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "temperature": false, + "release_date": "2025-10-07", + "last_updated": "2025-10-07", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 4096, - "output": 3072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.144, - "output": 0.287 + "input": 0.4, + "output": 2.4, + "cache_read": 0.04 } }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini-3.5-Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.144, - "output": 0.574 + "input": 1.5152, + "output": 9.0909, + "cache_read": 0.1515 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": false, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini-2.5-Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "budget_tokens", - "max": 262144 + "min": 0, + "max": 32768 } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "temperature": false, + "release_date": "2025-02-05", + "last_updated": "2025-02-05", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" @@ -4896,51 +5389,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1065535, + "output": 65535 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.87, + "output": 7, + "cache_read": 0.087 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini-2.5-Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "budget_tokens", - "max": 262144 + "min": 0, + "max": 24576 } ], "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "temperature": false, + "release_date": "2025-04-26", + "last_updated": "2025-04-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -4948,61 +5429,58 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 1065535, + "output": 65535 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 0.21, + "output": 1.8, + "cache_read": 0.021 } }, - "qwen-long": { - "id": "qwen-long", - "name": "Qwen Long", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "google/imagen-3": { + "id": "google/imagen-3", + "name": "Imagen-3", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01-25", - "last_updated": "2025-01-25", + "temperature": false, + "release_date": "2024-10-15", + "last_updated": "2024-10-15", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 10000000, - "output": 8192 - }, - "cost": { - "input": 0.072, - "output": 0.287 + "context": 480, + "output": 0 } }, - "qwen-math-turbo": { - "id": "qwen-math-turbo", - "name": "Qwen Math Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, - "reasoning": false, + "google/gemini-3-pro": { + "id": "google/gemini-3-pro", + "name": "Gemini-3-Pro", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "temperature": false, + "release_date": "2025-10-22", + "last_updated": "2025-10-22", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -5010,291 +5488,256 @@ }, "open_weights": false, "limit": { - "context": 4096, - "output": 3072 + "context": 1048576, + "output": 65536 }, + "status": "deprecated", "cost": { - "input": 0.287, - "output": 0.861 + "input": 1.6, + "output": 9.6, + "cache_read": 0.16 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "google/veo-2": { + "id": "google/veo-2", + "name": "Veo-2", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "temperature": false, + "release_date": "2024-12-02", + "last_updated": "2024-12-02", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.861, - "output": 3.441 + "context": 480, + "output": 0 } }, - "qwen2-5-omni-7b": { - "id": "qwen2-5-omni-7b", - "name": "Qwen2.5-Omni 7B", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, + "google/veo-3.1-fast": { + "id": "google/veo-3.1-fast", + "name": "Veo-3.1-Fast", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-12", - "last_updated": "2024-12", + "temperature": false, + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ - "text", - "audio" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 2048 - }, - "cost": { - "input": 0.087, - "output": 0.345, - "input_audio": 5.448 + "context": 480, + "output": 0 } }, - "qwen3-8b": { - "id": "qwen3-8b", - "name": "Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 38912 - } - ], + "google/gemma-4-31b": { + "id": "google/gemma-4-31b", + "name": "Gemma-4-31B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 262144, "output": 8192 }, "cost": { - "input": 0.072, - "output": 0.287, - "reasoning": 0.717 + "input": 0, + "output": 0 } }, - "qwen2-5-14b-instruct": { - "id": "qwen2-5-14b-instruct", - "name": "Qwen2.5 14B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "google/veo-3": { + "id": "google/veo-3", + "name": "Veo-3", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "temperature": false, + "release_date": "2025-05-21", + "last_updated": "2025-05-21", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.144, - "output": 0.431 + "context": 480, + "output": 0 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini-2.5-Flash-Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "budget_tokens", - "max": 131072 + "min": 0, + "max": 24576 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-14", - "last_updated": "2026-04-14", + "temperature": false, + "release_date": "2025-06-19", + "last_updated": "2025-06-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 128000 + "context": 1024000, + "output": 64000 }, "cost": { - "input": 0.87, - "output": 3.48, - "cache_read": 0.17 + "input": 0.07, + "output": 0.28 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "google/gemini-3.1-pro": { + "id": "google/gemini-3.1-pro", + "name": "Gemini-3.1-Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "max" + "low", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "qwen3-next-80b-a3b-thinking": { - "id": "qwen3-next-80b-a3b-thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini-3.1-Flash-Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "effort", + "values": [ + "minimal", + "low", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "temperature": false, + "release_date": "2026-02-18", + "last_updated": "2026-02-18", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.144, - "output": 1.434 + "input": 0.25, + "output": 1.5 } }, - "qvq-max": { - "id": "qvq-max", - "name": "QVQ Max", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qvq", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "xai/grok-3": { + "id": "xai/grok-3", + "name": "Grok 3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "temperature": false, + "release_date": "2025-04-11", + "last_updated": "2025-04-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -5306,25 +5749,25 @@ "output": 8192 }, "cost": { - "input": 1.147, - "output": 4.588 + "input": 3, + "output": 15, + "cache_read": 0.75 } }, - "qwen-plus-character": { - "id": "qwen-plus-character", - "name": "Qwen Plus Character", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "xai/grok-4.20-multi-agent": { + "id": "xai/grok-4.20-multi-agent", + "name": "Grok-4.20-Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01", - "last_updated": "2024-01", + "temperature": false, + "release_date": "2026-03-13", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -5332,67 +5775,63 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 4096 + "context": 128000, + "output": 0 }, "cost": { - "input": 0.115, - "output": 0.287 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Moonshot Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, + "xai/grok-4": { + "id": "xai/grok-4", + "name": "Grok-4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "temperature": false, + "release_date": "2025-07-10", + "last_updated": "2025-07-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.574, - "output": 2.294 + "input": 3, + "output": 15, + "cache_read": 0.75 } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "xai/grok-4-fast-reasoning": { + "id": "xai/grok-4-fast-reasoning", + "name": "Grok-4-Fast-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "temperature": false, + "release_date": "2025-09-16", + "last_updated": "2025-09-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -5400,41 +5839,31 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 2000000, + "output": 128000 }, "cost": { - "input": 0.574, - "output": 2.294 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "qwen3.5-flash": { - "id": "qwen3.5-flash", - "name": "Qwen3.5 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "xai/grok-4.1-fast-reasoning": { + "id": "xai/grok-4.1-fast-reasoning", + "name": "Grok-4.1-Fast-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -5442,149 +5871,96 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.172, - "output": 1.72, - "reasoning": 1.72 + "context": 2000000, + "output": 30000 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "xai/grok-4-fast-non-reasoning": { + "id": "xai/grok-4-fast-non-reasoning", + "name": "Grok-4-Fast-Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "temperature": false, + "release_date": "2025-09-16", + "last_updated": "2025-09-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, + "context": 2000000, "output": 128000 }, "cost": { - "input": 1.1, - "output": 3.851, - "cache_read": 0.275, - "cache_write": 0 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "xai/grok-3-mini": { + "id": "xai/grok-3-mini", + "name": "Grok 3 Mini", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 131072 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "temperature": false, + "release_date": "2025-04-11", + "last_updated": "2025-04-11", "modalities": { "input": [ - "text", - "image", - "video" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.1875, - "output": 1.125, - "cache_write": 0.234375 - } - }, - "qwen2-5-vl-72b-instruct": { - "id": "qwen2-5-vl-72b-instruct", - "name": "Qwen2.5-VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 131072, "output": 8192 }, "cost": { - "input": 2.294, - "output": 6.881 + "input": 0.3, + "output": 0.5, + "cache_read": 0.075 } }, - "qwen3-vl-plus": { - "id": "qwen3-vl-plus", - "name": "Qwen3-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "xai/grok-code-fast-1": { + "id": "xai/grok-code-fast-1", + "name": "Grok Code Fast 1", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "temperature": false, + "release_date": "2025-08-22", + "last_updated": "2025-08-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -5592,27 +5968,26 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.143353, - "output": 1.433525, - "reasoning": 4.300576 + "input": 0.2, + "output": 1.5, + "cache_read": 0.02 } }, - "qwen-vl-ocr": { - "id": "qwen-vl-ocr", - "name": "Qwen-VL OCR", - "description": "OCR model for extracting structured text from documents and screenshots", - "family": "qwen", - "attachment": false, + "xai/grok-4.1-fast-non-reasoning": { + "id": "xai/grok-4.1-fast-non-reasoning", + "name": "Grok-4.1-Fast-Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-10-28", - "last_updated": "2025-04-13", + "tool_call": true, + "temperature": false, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ "text", @@ -5624,253 +5999,205 @@ }, "open_weights": false, "limit": { - "context": 34096, - "output": 4096 - }, - "cost": { - "input": 0.717, - "output": 0.717 + "context": 2000000, + "output": 30000 } }, - "qwen-mt-turbo": { - "id": "qwen-mt-turbo", - "name": "Qwen-MT Turbo", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "family": "qwen", - "attachment": false, + "ideogramai/ideogram-v2": { + "id": "ideogramai/ideogram-v2", + "name": "Ideogram-v2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ideogram", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01", - "last_updated": "2025-01", + "tool_call": true, + "temperature": false, + "release_date": "2024-08-21", + "last_updated": "2024-08-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 16384, - "output": 8192 - }, - "cost": { - "input": 0.101, - "output": 0.28 + "context": 150, + "output": 0 } }, - "qwen-math-plus": { - "id": "qwen-math-plus", - "name": "Qwen Math Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "ideogramai/ideogram-v2a-turbo": { + "id": "ideogramai/ideogram-v2a-turbo", + "name": "Ideogram-v2a-Turbo", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ideogram", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-08-16", - "last_updated": "2024-09-19", + "temperature": false, + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 4096, - "output": 3072 - }, - "cost": { - "input": 0.574, - "output": 1.721 + "context": 150, + "output": 0 } }, - "qwen-mt-plus": { - "id": "qwen-mt-plus", - "name": "Qwen-MT Plus", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "family": "qwen", - "attachment": false, + "ideogramai/ideogram-v2a": { + "id": "ideogramai/ideogram-v2a", + "name": "Ideogram-v2a", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ideogram", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01", - "last_updated": "2025-01", + "tool_call": true, + "temperature": false, + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 16384, - "output": 8192 - }, - "cost": { - "input": 0.259, - "output": 0.775 + "context": 150, + "output": 0 } }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], + "ideogramai/ideogram": { + "id": "ideogramai/ideogram", + "name": "Ideogram", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ideogram", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "temperature": false, + "release_date": "2024-04-03", + "last_updated": "2024-04-03", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.573, - "output": 3.44, - "reasoning": 3.44 + "context": 150, + "output": 0 } }, - "qwen-omni-turbo": { - "id": "qwen-omni-turbo", - "name": "Qwen-Omni Turbo", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, + "fireworks-ai/kimi-k2.5-fw": { + "id": "fireworks-ai/kimi-k2.5-fw", + "name": "Kimi-K2.5-FW", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01-19", - "last_updated": "2025-03-26", + "temperature": false, + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 2048 + "context": 262144, + "input": 245760, + "output": 16384 }, "cost": { - "input": 0.058, - "output": 0.23, - "input_audio": 3.584, - "output_audio": 7.168 + "input": 0, + "output": 0 } }, - "qwen2-5-72b-instruct": { - "id": "qwen2-5-72b-instruct", - "name": "Qwen2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "runwayml/runway": { + "id": "runwayml/runway", + "name": "Runway", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "runway", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "temperature": false, + "release_date": "2024-10-11", + "last_updated": "2024-10-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.574, - "output": 1.721 + "context": 256, + "output": 0 } }, - "deepseek-r1-distill-qwen-7b": { - "id": "deepseek-r1-distill-qwen-7b", - "name": "DeepSeek R1 Distill Qwen 7B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "runwayml/runway-gen-4-turbo": { + "id": "runwayml/runway-gen-4-turbo", + "name": "Runway-Gen-4-Turbo", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "runway", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "temperature": false, + "release_date": "2025-05-09", + "last_updated": "2025-05-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 0.072, - "output": 0.144 + "context": 256, + "output": 0 } }, - "deepseek-v3-1": { - "id": "deepseek-v3-1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "trytako/tako": { + "id": "trytako/tako", + "name": "Tako", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "tako", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "temperature": false, + "release_date": "2024-08-15", + "last_updated": "2024-08-15", "modalities": { "input": [ "text" @@ -5881,185 +6208,182 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 65536 - }, - "cost": { - "input": 0.574, - "output": 1.721 + "context": 2048, + "output": 0 } }, - "qwen2-5-coder-32b-instruct": { - "id": "qwen2-5-coder-32b-instruct", - "name": "Qwen2.5-Coder 32B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "stabilityai/stablediffusionxl": { + "id": "stabilityai/stablediffusionxl", + "name": "StableDiffusionXL", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "stable-diffusion", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11", - "last_updated": "2024-11", + "temperature": false, + "release_date": "2023-07-09", + "last_updated": "2023-07-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.287, - "output": 0.861 + "context": 200, + "output": 0 } }, - "qwen-flash": { - "id": "qwen-flash", - "name": "Qwen Flash", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "temperature": false, + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.022, - "output": 0.216 + "input": 2.2, + "output": 14, + "cache_read": 0.22 } }, - "deepseek-v3-2-exp": { - "id": "deepseek-v3-2-exp", - "name": "DeepSeek V3.2 Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "openai/dall-e-3": { + "id": "openai/dall-e-3", + "name": "DALL-E-3", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "dall-e", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "temperature": false, + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 65536 - }, - "cost": { - "input": 0.287, - "output": 0.431 + "context": 800, + "output": 0 } }, - "moonshot-kimi-k2-instruct": { - "id": "moonshot-kimi-k2-instruct", - "name": "Moonshot Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, + "openai/sora-2": { + "id": "openai/sora-2", + "name": "Sora-2", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "sora", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "temperature": false, + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.574, - "output": 2.294 + "context": 0, + "output": 0 } }, - "deepseek-r1-distill-qwen-14b": { - "id": "deepseek-r1-distill-qwen-14b", - "name": "DeepSeek R1 Distill Qwen 14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4-Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "temperature": false, + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.144, - "output": 0.431 + "input": 27, + "output": 160 } }, - "qwen3-vl-235b-a22b": { - "id": "qwen3-vl-235b-a22b", - "name": "Qwen3-VL 235B-A22B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5-Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2023-09-13", + "last_updated": "2023-09-13", "modalities": { "input": [ "text", @@ -6069,105 +6393,123 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 16384, + "output": 2048 }, "cost": { - "input": 0.286705, - "output": 1.14682, - "reasoning": 2.867051 + "input": 0.45, + "output": 1.4 } }, - "qwen-deep-research": { - "id": "qwen-deep-research", - "name": "Qwen Deep Research", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5-Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01", - "last_updated": "2024-01", + "structured_output": false, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 7.742, - "output": 23.367 + "input": 27.2727, + "output": 163.6364 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Moonshot Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": false, + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4-Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "temperature": false, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.574, - "output": 2.411 + "input": 0.18, + "output": 1.1, + "cache_read": 0.018 } }, - "qwen3-vl-30b-a3b": { - "id": "qwen3-vl-30b-a3b", - "name": "Qwen3-VL 30B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ "text", @@ -6177,29 +6519,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.108, - "output": 0.431, - "reasoning": 1.076 + "input": 1.6, + "output": 13, + "cache_read": 0.16 } }, - "qwen-vl-max": { - "id": "qwen-vl-max", - "name": "Qwen-VL Max", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-08", - "last_updated": "2025-08-13", + "temperature": false, + "release_date": "2025-11-12", + "last_updated": "2025-11-12", "modalities": { "input": [ "text", @@ -6211,29 +6562,30 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.23, - "output": 0.574 + "input": 1.1, + "output": 9, + "cache_read": 0.11 } }, - "deepseek-r1-distill-qwen-1-5b": { - "id": "deepseek-r1-distill-qwen-1-5b", - "name": "DeepSeek R1 Distill Qwen 1.5B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "temperature": false, + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -6241,26 +6593,21 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 + "context": 128000, + "output": 8192 } }, - "qwen-max": { - "id": "qwen-max", - "name": "Qwen Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "openai/gpt-4o-search": { + "id": "openai/gpt-4o-search", + "name": "GPT-4o-Search", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", + "temperature": false, + "release_date": "2025-03-11", + "last_updated": "2025-03-11", "modalities": { "input": [ "text" @@ -6271,102 +6618,106 @@ }, "open_weights": false, "limit": { - "context": 131072, + "context": 128000, "output": 8192 }, "cost": { - "input": 0.345, - "output": 1.377 + "input": 2.2, + "output": 9 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/sora-2-pro": { + "id": "openai/sora-2-pro", + "name": "Sora-2-Pro", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "sora", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "temperature": false, + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2 + "context": 0, + "output": 0 } }, - "qwen3-235b-a22b": { - "id": "qwen3-235b-a22b", - "name": "Qwen3 235B-A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 38912 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.287, - "output": 1.147, - "reasoning": 2.868 + "input": 1.1, + "output": 9 } }, - "deepseek-r1-distill-llama-8b": { - "id": "deepseek-r1-distill-llama-8b", - "name": "DeepSeek R1 Distill Llama 8B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "temperature": false, + "release_date": "2025-06-25", + "last_updated": "2025-06-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -6374,105 +6725,109 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.22, + "output": 1.8, + "cache_read": 0.022 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Moonshot Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "openai/gpt-image-2": { + "id": "openai/gpt-image-2", + "name": "GPT-Image-2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "knowledge": "2025-01", + "reasoning": false, + "tool_call": false, + "temperature": false, "release_date": "2026-04-21", "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 0, + "output": 0 }, "cost": { - "input": 0.929, - "output": 3.858 + "input": 5.0505, + "output": 32.3232, + "cache_read": 1.2626 } }, - "qwen3-coder-480b-a35b-instruct": { - "id": "qwen3-coder-480b-a35b-instruct", - "name": "Qwen3-Coder 480B-A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2-Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.861, - "output": 3.441 + "input": 19, + "output": 150 } }, - "qwq-plus": { - "id": "qwq-plus", - "name": "QwQ Plus", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-03-05", - "last_updated": "2025-03-05", + "temperature": false, + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -6480,101 +6835,111 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.23, - "output": 0.574 + "input": 0.99, + "output": 4, + "cache_read": 0.25 } }, - "qwen2-5-32b-instruct": { - "id": "qwen2-5-32b-instruct", - "name": "Qwen2.5 32B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "temperature": false, + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.287, - "output": 0.861 + "input": 0.99, + "output": 4 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "temperature": false, + "release_date": "2026-02-10", + "last_updated": "2026-02-10", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.43, - "output": 2.58, - "reasoning": 2.58 + "input": 1.6, + "output": 13, + "cache_read": 0.16 } }, - "tongyi-intent-detect-v3": { - "id": "tongyi-intent-detect-v3", - "name": "Tongyi Intent Detect V3", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "yi", - "attachment": false, + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01", - "last_updated": "2024-01", + "tool_call": true, + "temperature": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -6582,26 +6947,27 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1024 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.058, - "output": 0.144 + "input": 0.09, + "output": 0.36, + "cache_read": 0.022 } }, - "qwen3-coder-flash": { - "id": "qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/o3-deep-research": { + "id": "openai/o3-deep-research", + "name": "o3-deep-research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "temperature": false, + "release_date": "2025-06-27", + "last_updated": "2025-06-27", "modalities": { "input": [ "text" @@ -6612,29 +6978,41 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.144, - "output": 0.574 + "input": 9, + "output": 36, + "cache_read": 2.2 } }, - "deepseek-r1-distill-qwen-32b": { - "id": "deepseek-r1-distill-qwen-32b", - "name": "DeepSeek R1 Distill Qwen 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "temperature": false, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -6642,100 +7020,132 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.287, - "output": 0.861 + "input": 0.045, + "output": 0.36, + "cache_read": 0.0045 } }, - "qwq-32b": { - "id": "qwq-32b", - "name": "QwQ 32B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-5.2-instant": { + "id": "openai/gpt-5.2-instant", + "name": "GPT-5.2-Instant", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-12", - "last_updated": "2024-12", + "temperature": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.287, - "output": 0.861 + "input": 1.6, + "output": 13, + "cache_read": 0.16 } }, - "qwen3-14b": { - "id": "qwen3-14b", - "name": "Qwen3 14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "openai/o1": { + "id": "openai/o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 38912 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2024-12-18", + "last_updated": "2024-12-18", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 14, + "output": 54 + } + }, + "openai/chatgpt-4o-latest": { + "id": "openai/chatgpt-4o-latest", + "name": "ChatGPT-4o-Latest", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": false, + "release_date": "2024-08-14", + "last_updated": "2024-08-14", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 128000, "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0.144, - "output": 0.574, - "reasoning": 1.434 + "input": 4.5, + "output": 14 } }, - "qwen3-asr-flash": { - "id": "qwen3-asr-flash", - "name": "Qwen3-ASR Flash", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5-Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": false, - "knowledge": "2024-04", - "release_date": "2025-09-08", - "last_updated": "2025-09-08", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ - "audio" + "text", + "image" ], "output": [ "text" @@ -6743,29 +7153,29 @@ }, "open_weights": false, "limit": { - "context": 53248, - "output": 4096 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.032, - "output": 0.032 + "input": 14, + "output": 110 } }, - "qwen-doc-turbo": { - "id": "qwen-doc-turbo", - "name": "Qwen Doc Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, + "openai/gpt-3.5-turbo-instruct": { + "id": "openai/gpt-3.5-turbo-instruct", + "name": "GPT-3.5-Turbo-Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01", - "last_updated": "2024-01", + "temperature": false, + "release_date": "2023-09-20", + "last_updated": "2023-09-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -6773,37 +7183,25 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 3500, + "output": 1024 }, "cost": { - "input": 0.087, - "output": 0.144 + "input": 1.4, + "output": 1.8 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "openai/gpt-5.3-codex-spark": { + "id": "openai/gpt-5.3-codex-spark", + "name": "GPT-5.3-Codex-Spark", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "temperature": false, + "release_date": "2026-03-04", + "last_updated": "2026-03-04", "modalities": { "input": [ "text" @@ -6814,38 +7212,40 @@ }, "open_weights": false, "limit": { - "context": 202752, + "context": 128000, "output": 16384 }, "cost": { - "input": 0.86, - "output": 3.15 + "input": 0, + "output": 0 } }, - "qwen-turbo": { - "id": "qwen-turbo", - "name": "Qwen Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 38912 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11-01", - "last_updated": "2025-07-15", + "temperature": false, + "release_date": "2025-11-12", + "last_updated": "2025-11-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -6853,57 +7253,57 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 16384 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.044, - "output": 0.087, - "reasoning": 0.431 + "input": 1.1, + "output": 9, + "cache_read": 0.11 } }, - "qwen2-5-7b-instruct": { - "id": "qwen2-5-7b-instruct", - "name": "Qwen2.5 7B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "openai/gpt-5-chat": { + "id": "openai/gpt-5-chat", + "name": "GPT-5-Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "temperature": false, + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.072, - "output": 0.144 + "input": 1.1, + "output": 9, + "cache_read": 0.11 } }, - "qwen2-5-vl-7b-instruct": { - "id": "qwen2-5-vl-7b-instruct", - "name": "Qwen2.5-VL 7B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "temperature": false, + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", @@ -6913,40 +7313,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 124096, + "output": 4096 }, "cost": { - "input": 0.287, - "output": 0.717 + "input": 0.14, + "output": 0.54, + "cache_read": 0.068 } }, - "qwen3.6-max-preview": { - "id": "qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "openai/o3-pro": { + "id": "openai/o3-pro", + "name": "o3-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 131072 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-21", + "temperature": false, + "release_date": "2025-06-10", + "last_updated": "2025-06-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -6954,27 +7356,24 @@ }, "open_weights": false, "limit": { - "context": 245800, - "output": 65536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.32, - "output": 7.9, - "cache_read": 0.132 + "input": 18, + "output": 72 } }, - "qwen-vl-plus": { - "id": "qwen-vl-plus", - "name": "Qwen-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "openai/gpt-5.3-instant": { + "id": "openai/gpt-5.3-instant", + "name": "GPT-5.3-Instant", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-08-15", + "temperature": false, + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", @@ -6986,40 +7385,31 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 128000, + "input": 111616, + "output": 16384 }, "cost": { - "input": 0.115, - "output": 0.287 + "input": 1.6, + "output": 13, + "cache_read": 0.16 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], + "openai/gpt-4o-aug": { + "id": "openai/gpt-4o-aug", + "name": "GPT-4o-Aug", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "release_date": "2024-11-21", + "last_updated": "2024-11-21", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -7027,159 +7417,130 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 2.2, + "output": 9, + "cache_read": 1.1 } }, - "MiniMax/MiniMax-M2.7": { - "id": "MiniMax/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "openai/o3-mini-high": { + "id": "openai/o3-mini-high", + "name": "o3-mini-high", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": false, + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.99, + "output": 4 } }, - "kimi/kimi-k2.5": { - "id": "kimi/kimi-k2.5", - "name": "kimi/kimi-k2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": false, + "openai/gpt-5.1-instant": { + "id": "openai/gpt-5.1-instant", + "name": "GPT-5.1-Instant", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2025-11-12", + "last_updated": "2025-11-12", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 1.1, + "output": 9, + "cache_read": 0.11 } }, - "siliconflow/deepseek-r1-0528": { - "id": "siliconflow/deepseek-r1-0528", - "name": "siliconflow/deepseek-r1-0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-image-1": { + "id": "openai/gpt-image-1", + "name": "GPT-Image-1", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-11-25", + "temperature": false, + "release_date": "2025-03-31", + "last_updated": "2025-03-31", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 163840, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 2.18 + "context": 128000, + "output": 0 } }, - "siliconflow/deepseek-v3.1-terminus": { - "id": "siliconflow/deepseek-v3.1-terminus", - "name": "siliconflow/deepseek-v3.1-terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4-Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-11-25", + "temperature": false, + "release_date": "2026-03-12", + "last_updated": "2026-03-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -7187,29 +7548,83 @@ }, "open_weights": false, "limit": { - "context": 163840, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 1 + "input": 0.68, + "output": 4, + "cache_read": 0.068 } }, - "siliconflow/deepseek-v3-0324": { - "id": "siliconflow/deepseek-v3-0324", - "name": "siliconflow/deepseek-v3-0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2025-11-25", - "modalities": { + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT-5.1-Codex-Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "temperature": false, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", + "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.1, + "output": 9, + "cache_read": 0.11 + } + }, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "temperature": false, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" @@ -7217,61 +7632,73 @@ }, "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1 + "input": 1.6, + "output": 13, + "cache_read": 0.16 } }, - "siliconflow/deepseek-v3.2": { - "id": "siliconflow/deepseek-v3.2", - "name": "siliconflow/deepseek-v3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 163840, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 0.42 + "input": 4.5455, + "output": 27.2727, + "cache_read": 0.4545 } }, - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/o4-mini-deep-research": { + "id": "openai/o4-mini-deep-research", + "name": "o4-mini-deep-research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "temperature": false, + "release_date": "2025-06-27", + "last_updated": "2025-06-27", "modalities": { "input": [ "text" @@ -7280,42 +7707,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1, - "output": 5 + "input": 1.8, + "output": 7.2, + "cache_read": 0.45 } - } - } - }, - "regolo-ai": { - "id": "regolo-ai", - "env": [ - "REGOLO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.regolo.ai/v1", - "name": "Regolo AI", - "doc": "https://docs.regolo.ai/", - "models": { - "llama-3.1-8b-instruct": { - "id": "llama-3.1-8b-instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + }, + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-04-07", - "last_updated": "2025-04-07", + "temperature": false, + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -7323,29 +7740,30 @@ }, "open_weights": false, "limit": { - "context": 120000, - "output": 120000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.05, - "output": 0.25 + "input": 1.8, + "output": 7.2, + "cache_read": 0.45 } }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax 2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4-classic-0314": { + "id": "openai/gpt-4-classic-0314", + "name": "GPT-4-Classic-0314", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-10", + "temperature": false, + "release_date": "2024-08-26", + "last_updated": "2024-08-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -7353,29 +7771,30 @@ }, "open_weights": false, "limit": { - "context": 190000, - "output": 64000 + "context": 8192, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 0.8, - "output": 3.5 + "input": 27, + "output": 54 } }, - "mistral-small3.2": { - "id": "mistral-small3.2", - "name": "Mistral Small 3.2", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4-classic": { + "id": "openai/gpt-4-classic", + "name": "GPT-4-Classic", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "temperature": false, + "release_date": "2024-03-25", + "last_updated": "2024-03-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -7383,25 +7802,26 @@ }, "open_weights": false, "limit": { - "context": 120000, - "output": 120000 + "context": 8192, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 0.5, - "output": 2.2 + "input": 27, + "output": 54 } }, - "qwen3-reranker-4b": { - "id": "qwen3-reranker-4b", - "name": "Qwen3-Reranker-4B", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "qwen", - "attachment": false, + "openai/gpt-4o-mini-search": { + "id": "openai/gpt-4o-mini-search", + "name": "GPT-4o-mini-Search", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": false, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "release_date": "2025-03-11", + "last_updated": "2025-03-11", "modalities": { "input": [ "text" @@ -7410,59 +7830,61 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, + "context": 128000, "output": 8192 }, "cost": { - "input": 0.12, - "output": 0.12 + "input": 0.14, + "output": 0.54 } }, - "qwen3-embedding-8b": { - "id": "qwen3-embedding-8b", - "name": "Qwen3-Embedding-8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", - "attachment": false, + "openai/gpt-3.5-turbo-raw": { + "id": "openai/gpt-3.5-turbo-raw", + "name": "GPT-3.5-Turbo-Raw", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": false, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "release_date": "2023-09-27", + "last_updated": "2023-09-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 4524, + "output": 2048 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 0.45, + "output": 1.4 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "temperature": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -7470,28 +7892,30 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.7 + "input": 0.36, + "output": 1.4, + "cache_read": 0.09 } }, - "qwen-image": { - "id": "qwen-image", - "name": "Qwen-Image", + "openai/gpt-image-1-mini": { + "id": "openai/gpt-image-1-mini", + "name": "GPT-Image-1-Mini", "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", - "attachment": false, + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "tool_call": true, + "temperature": false, + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "image" @@ -7499,26 +7923,46 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 0, + "output": 0 + } + }, + "openai/gpt-image-1.5": { + "id": "openai/gpt-image-1.5", + "name": "gpt-image-1.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] }, - "cost": { - "input": 0.5, - "output": 2 + "open_weights": false, + "limit": { + "context": 128000, + "output": 0 } }, - "qwen3.5-122b": { - "id": "qwen3.5-122b", - "name": "Qwen3.5-122B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4-Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "temperature": false, + "release_date": "2023-09-13", + "last_updated": "2023-09-13", "modalities": { "input": [ "text", @@ -7528,22 +7972,22 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.9, - "output": 3.6 + "input": 9, + "output": 27 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT-OSS-120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1-Codex-Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -7556,9 +8000,9 @@ } ], "tool_call": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "release_date": "2025-11-12", + "last_updated": "2025-11-12", "modalities": { "input": [ "text" @@ -7569,56 +8013,119 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 400000, + "output": 128000 }, "cost": { - "input": 1, - "output": 4.2 + "input": 0.22, + "output": 1.8, + "cache_read": 0.022 } }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "Qwen3-Coder-Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "temperature": false, + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.8, + "output": 7.2, + "cache_read": 0.45 + } + }, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 1.1, + "output": 9, + "cache_read": 0.11 } }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5-9B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/o1-pro": { + "id": "openai/o1-pro", + "name": "o1-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "temperature": false, + "release_date": "2025-03-19", + "last_updated": "2025-03-19", "modalities": { "input": [ "text", @@ -7628,28 +8135,75 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 8192 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 140, + "output": 540 } }, - "mistral-small-4-119b": { - "id": "mistral-small-4-119b", - "name": "Mistral Small 4 119B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "topazlabs-co/topazlabs": { + "id": "topazlabs-co/topazlabs", + "name": "TopazLabs", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "topazlabs", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": false, + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 204, + "output": 0 + } + } + } + }, + "cerebras": { + "id": "cerebras", + "env": [ + "CEREBRAS_API_KEY" + ], + "npm": "@ai-sdk/cerebras", + "name": "Cerebras", + "doc": "https://inference-docs.cerebras.ai/models/overview", + "models": { + "qwen-3.8-27b": { + "id": "qwen-3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-03-15", + "release_date": "2026-08-14", + "last_updated": "2026-09-03", "modalities": { "input": [ "text", @@ -7659,20 +8213,20 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 65536, + "output": 32768 }, "cost": { - "input": 0.75, - "output": 3 + "input": 0.99, + "output": 1.49 } }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT-OSS-20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", "family": "gpt-oss", "attachment": false, "reasoning": true, @@ -7687,9 +8241,10 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "release_date": "2025-08-05", + "last_updated": "2026-06-10", "modalities": { "input": [ "text" @@ -7700,41 +8255,39 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 40960 }, "cost": { - "input": 0.4, - "output": 1.8 + "input": 0.35, + "output": 0.75 } } } }, - "stackit": { - "id": "stackit", + "groq": { + "id": "groq", "env": [ - "STACKIT_API_KEY" + "GROQ_API_KEY" ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.openai-compat.model-serving.eu01.onstackit.cloud/v1", - "name": "STACKIT", - "doc": "https://docs.stackit.cloud/products/data-and-ai/ai-model-serving/basics/available-shared-models", + "npm": "@ai-sdk/groq", + "name": "Groq", + "doc": "https://console.groq.com/docs/models", "models": { - "cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic": { - "id": "cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic", - "name": "Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "whisper-large-v3": { + "id": "whisper-large-v3", + "name": "Whisper", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": false, + "tool_call": false, "temperature": true, - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "release_date": "2023-09-01", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" @@ -7742,30 +8295,25 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.53, - "output": 0.76 + "context": 0, + "output": 0 } }, - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma 3 27B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "llama-3.3-70b-versatile": { + "id": "llama-3.3-70b-versatile", + "name": "Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2025-05-17", - "last_updated": "2025-05-17", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7773,26 +8321,24 @@ }, "open_weights": true, "limit": { - "context": 37000, - "output": 4096 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.53, - "output": 0.76 + "input": 0.59, + "output": 0.79 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "allam-2-7b": { + "id": "allam-2-7b", + "name": "ALLaM-2-7b", + "description": "ALLaM-2-7b instruction tuned model by SDAIA", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2025-01-23", + "last_updated": "2025-01-23", "modalities": { "input": [ "text" @@ -7803,30 +8349,28 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 16384 + "context": 4096, + "output": 4096 }, "cost": { - "input": 0.53, - "output": 0.76 + "input": 0, + "output": 0 } }, - "Qwen/Qwen3-VL-Embedding-8B": { - "id": "Qwen/Qwen3-VL-Embedding-8B", - "name": "Qwen3-VL Embedding 8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", - "attachment": true, + "whisper-large-v3-turbo": { + "id": "whisper-large-v3-turbo", + "name": "Whisper Large V3 Turbo", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "temperature": false, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", "modalities": { "input": [ - "text", - "image" + "audio" ], "output": [ "text" @@ -7834,30 +8378,25 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 4096 - }, - "cost": { - "input": 0.09, - "output": 0.09 + "context": 0, + "output": 0 } }, - "Qwen/Qwen3-VL-235B-A22B-Instruct-FP8": { - "id": "Qwen/Qwen3-VL-235B-A22B-Instruct-FP8", - "name": "Qwen3-VL 235B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "llama-3.1-8b-instant": { + "id": "llama-3.1-8b-instant", + "name": "Llama 3.1 8B", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2024-11-01", - "last_updated": "2024-11-01", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7865,36 +8404,25 @@ }, "open_weights": true, "limit": { - "context": 218000, - "output": 16384 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.76, - "output": 2.05 + "input": 0.05, + "output": 0.08 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "groq/compound-mini": { + "id": "groq/compound-mini", + "name": "Compound Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "groq", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-09-04", + "last_updated": "2025-09-04", "modalities": { "input": [ "text" @@ -7903,29 +8431,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131000, + "context": 131072, "output": 8192 - }, - "cost": { - "input": 0.53, - "output": 0.76 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "groq/compound": { + "id": "groq/compound", + "name": "Compound", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "groq", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-09-04", + "last_updated": "2025-09-04", "modalities": { "input": [ "text" @@ -7934,28 +8456,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 131072, "output": 8192 - }, - "cost": { - "input": 0.18, - "output": 0.29 } }, - "intfloat/e5-mistral-7b-instruct": { - "id": "intfloat/e5-mistral-7b-instruct", - "name": "E5 Mistral 7B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "mistral", + "meta-llama/llama-prompt-guard-2-86m": { + "id": "meta-llama/llama-prompt-guard-2-86m", + "name": "Prompt Guard 2 86M", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": false, - "release_date": "2023-12-11", - "last_updated": "2023-12-11", + "release_date": "2025-05-29", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" @@ -7966,121 +8483,102 @@ }, "open_weights": true, "limit": { - "context": 4096, - "output": 4096 + "context": 512, + "output": 512 }, + "status": "beta", "cost": { - "input": 0.02, - "output": 0.02 + "input": 0.04, + "output": 0.04 } - } - } - }, - "vercel": { - "id": "vercel", - "env": [ - "AI_GATEWAY_API_KEY" - ], - "npm": "@ai-sdk/gateway", - "name": "Vercel AI Gateway", - "doc": "https://github.com/vercel/ai/tree/5eb85cc45a259553501f535b8ac79a77d0e79223/packages/gateway", - "models": { - "xai/grok-imagine-video-1.5": { - "id": "xai/grok-imagine-video-1.5", - "name": "Grok Imagine Video 1.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", + }, + "meta-llama/llama-prompt-guard-2-22m": { + "id": "meta-llama/llama-prompt-guard-2-22m", + "name": "Llama Prompt Guard 2 22M", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-06-22", - "last_updated": "2026-06-22", + "temperature": false, + "release_date": "2025-05-29", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 512, + "output": 512 + }, + "status": "beta", + "cost": { + "input": 0.03, + "output": 0.03 } }, - "xai/grok-4.1-fast-reasoning": { - "id": "xai/grok-4.1-fast-reasoning", - "name": "Grok 4.1 Fast Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "canopylabs/orpheus-v1-english": { + "id": "canopylabs/orpheus-v1-english", + "name": "Canopy Labs Orpheus V1 English", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "canopylabs", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-11-19", - "last_updated": "2025-07-09", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-12-19", + "last_updated": "2025-12-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 4000, + "output": 50000 }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } + "status": "beta" }, - "xai/grok-4.20-non-reasoning-beta": { - "id": "xai/grok-4.20-non-reasoning-beta", - "name": "Grok 4.20 Beta Non-Reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "canopylabs/orpheus-arabic-saudi": { + "id": "canopylabs/orpheus-arabic-saudi", + "name": "Canopy Labs Orpheus Arabic Saudi", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "canopylabs", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-13", + "tool_call": false, + "temperature": false, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 4000, + "output": 50000 }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.4 - } + "status": "beta" }, - "xai/grok-4.3": { - "id": "xai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "qwen/qwen3.8-27b": { + "id": "qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -8088,6 +8586,7 @@ "type": "effort", "values": [ "none", + "default", "low", "medium", "high" @@ -8097,144 +8596,73 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-17", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 131042, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 - } - }, - "xai/grok-tts": { - "id": "xai/grok-tts", - "name": "Grok TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "input": 0.8, + "output": 4 } }, - "xai/grok-4.1-fast-non-reasoning": { - "id": "xai/grok-4.1-fast-non-reasoning", - "name": "Grok 4.1 Fast Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": false, - "reasoning": false, + "qwen/qwen3.6-27b": { + "id": "qwen/qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "default" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-11-19", - "last_updated": "2025-07-09", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } - }, - "xai/grok-voice-think-fast-1.0": { - "id": "xai/grok-voice-think-fast-1.0", - "name": "Grok Voice Think Fast 1.0", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "xai/grok-imagine-video": { - "id": "xai/grok-imagine-video", - "name": "Grok Imagine", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-28", - "last_updated": "2026-01-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "video" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "input": 0.6, + "output": 3, + "cache_read": 0.3 } }, - "xai/grok-4.20-multi-agent-beta": { - "id": "xai/grok-4.20-multi-agent-beta", - "name": "Grok 4.20 Multi Agent Beta", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -8243,67 +8671,40 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-13", + "release_date": "2025-08-05", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.075, + "output": 0.3, + "cache_read": 0.0375 } }, - "xai/grok-stt": { - "id": "xai/grok-stt", - "name": "Grok STT", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "grok", + "openai/gpt-oss-safeguard-20b": { + "id": "openai/gpt-oss-safeguard-20b", + "name": "Safety GPT OSS 20B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "xai/grok-4.5": { - "id": "xai/grok-4.5", - "name": "Grok 4.5", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -8318,238 +8719,247 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "release_date": "2025-10-29", + "last_updated": "2026-06-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 131072, + "output": 65536 }, + "status": "beta", "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 0.075, + "output": 0.3 } }, - "xai/grok-4.20-reasoning": { - "id": "xai/grok-4.20-reasoning", - "name": "Grok 4.20 Reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-23", + "release_date": "2025-08-05", + "last_updated": "2025-10-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } - }, - "xai/grok-4.20-reasoning-beta": { - "id": "xai/grok-4.20-reasoning-beta", - "name": "Grok 4.20 Beta Reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + } + } + }, + "blueclaw": { + "id": "blueclaw", + "env": [ + "BLUECLAW_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://openai.blueclaw.network/v1", + "name": "Blue Claw", + "doc": "https://blueclaw.network", + "models": { + "Qwen3.6-27B": { + "id": "Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-13", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 196608, + "output": 65536 }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 - } + "status": "beta" }, - "xai/grok-imagine-video-1.5-preview": { - "id": "xai/grok-imagine-video-1.5-preview", - "name": "Grok Imagine Video 1.5 Preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", + "Qwen/Qwen3.6-35B-A3B-FP8": { + "id": "Qwen/Qwen3.6-35B-A3B-FP8", + "name": "Qwen3.6 35B A3B FP8", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-30", - "last_updated": "2026-05-30", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 - } - }, - "xai/grok-4.20-non-reasoning": { - "id": "xai/grok-4.20-non-reasoning", - "name": "Grok 4.20 Non-Reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "context": 131072, + "output": 65536 + }, + "status": "beta" + } + } + }, + "zai": { + "id": "zai", + "env": [ + "ZHIPU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.z.ai/api/paas/v4", + "name": "Z.AI", + "doc": "https://docs.z.ai/guides/overview/pricing", + "models": { + "glm-4.6v": { + "id": "glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-23", + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.3, + "output": 0.9 } }, - "xai/grok-4.20-multi-agent": { - "id": "xai/grok-4.20-multi-agent", - "name": "Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "glm-4.5": { + "id": "glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "temperature": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-23", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 131072, + "output": 98304 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 - } - }, - "xai/grok-imagine-image": { - "id": "xai/grok-imagine-image", - "name": "Grok Imagine Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-28", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "xai/grok-build-0.1": { - "id": "xai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-05-20", - "last_updated": "2026-04-16", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" @@ -8557,93 +8967,120 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 200000, + "output": 131072 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 + "input": 1.2, + "output": 4, + "cache_read": 0.24, + "cache_write": 0 } }, - "moonshotai/kimi-k2": { - "id": "moonshotai/kimi-k2", - "name": "Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-07-11", - "last_updated": "2025-09-05", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, + "context": 1000000, "output": 131072 }, "cost": { - "input": 0.57, - "output": 2.3 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03, + "cache_write": 0 } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "glm-4.6": { + "id": "glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -8652,22 +9089,23 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 216144, - "output": 216144 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.47, - "output": 2, - "cache_read": 0.141 + "input": 1, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "glm-4.6v-flash": { + "id": "glm-4.6v-flash", + "name": "GLM-4.6V-Flash", + "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -8676,16 +9114,14 @@ } ], "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-26", - "last_updated": "2026-01", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -8693,20 +9129,21 @@ }, "open_weights": true, "limit": { - "context": 262114, - "output": 262114 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "glm-4.5v": { + "id": "glm-4.5v", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -8715,15 +9152,15 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-20", - "last_updated": "2026-04-21", + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -8731,261 +9168,399 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 64000, + "output": 16384 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.6, + "output": 1.8 } }, - "moonshotai/kimi-k2.7-code-highspeed": { - "id": "moonshotai/kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code High Speed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", + "glm-5.3-flashx": { + "id": "glm-5.3-flashx", + "name": "GLM-5.3-FlashX", + "description": "High-speed GLM-5.3-Flash serving option for coding and agent workflows", + "family": "glm-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-15", - "last_updated": "2026-06-12", + "release_date": "2026-09-18", + "last_updated": "2026-09-18", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 + "input": 0.37, + "output": 1.25, + "cache_read": 0.075, + "cache_write": 0 } }, - "klingai/kling-v3.0-motion-control": { - "id": "klingai/kling-v3.0-motion-control", - "name": "Kling v3.0 Motion Control", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ling", + "glm-4.7-flash": { + "id": "glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2026-03-04", - "last_updated": "2026-03-04", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "klingai/kling-v2.6-i2v": { - "id": "klingai/kling-v2.6-i2v", - "name": "Kling v2.6 Image-to-Video", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ling", + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-21", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "klingai/kling-v2.5-turbo-t2v": { - "id": "klingai/kling-v2.5-turbo-t2v", - "name": "Kling v2.5 Turbo Text-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ling", + "glm-4.5-flash": { + "id": "glm-4.5-flash", + "name": "GLM-4.5-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "output": 98304 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "klingai/kling-v3.0-i2v": { - "id": "klingai/kling-v3.0-i2v", - "name": "Kling v3.0 Image-to-Video", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ling", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "klingai/kling-v2.5-turbo-i2v": { - "id": "klingai/kling-v2.5-turbo-i2v", - "name": "Kling v2.5 Turbo Image-to-Video", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ling", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "klingai/kling-v3.0-t2v": { - "id": "klingai/kling-v3.0-t2v", - "name": "Kling v3.0 Text-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ling", + "glm-4.7-flashx": { + "id": "glm-4.7-flashx", + "name": "GLM-4.7-FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0.07, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0 } }, - "klingai/kling-v2.6-motion-control": { - "id": "klingai/kling-v2.6-motion-control", - "name": "Kling v2.6 Motion Control", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ling", + "glm-5-turbo": { + "id": "glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-21", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.24, + "cache_write": 0 } }, - "klingai/kling-v2.6-t2v": { - "id": "klingai/kling-v2.6-t2v", - "name": "Kling v2.6 Text-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ling", + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-21", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "output": 98304 + }, + "cost": { + "input": 0.2, + "output": 1.1, + "cache_read": 0.03, + "cache_write": 0 } }, - "voyage/voyage-4-lite": { - "id": "voyage/voyage-4-lite", - "name": "voyage-4-lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "voyage", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-01-15", - "last_updated": "2026-03-06", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -8994,51 +9569,64 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "output": 0 - } - }, - "voyage/voyage-law-2": { - "id": "voyage/voyage-law-2", - "name": "voyage-law-2", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-04-15", - "last_updated": "2024-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 131072 }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 1536 + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } - }, - "voyage/voyage-4": { - "id": "voyage/voyage-4", - "name": "voyage-4", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, + } + } + }, + "empiriolabs": { + "id": "empiriolabs", + "env": [ + "EMPIRIOLABS_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.empiriolabs.ai/v1", + "name": "EmpirioLabs AI", + "doc": "https://docs.empiriolabs.ai", + "models": { + "qwen3-8-max-0902": { + "id": "qwen3-8-max-0902", + "name": "Qwen3.8 Max 0902", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-01-15", - "last_updated": "2026-03-06", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -9046,49 +9634,38 @@ }, "open_weights": false, "limit": { - "context": 32000, - "output": 0 - } - }, - "voyage/voyage-code-3": { - "id": "voyage/voyage-code-3", - "name": "voyage-code-3", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-12-04", - "last_updated": "2024-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 131072 }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 1536 + "cost": { + "input": 2, + "output": 6, + "cache_read": 2 } }, - "voyage/voyage-4-large": { - "id": "voyage/voyage-4-large", - "name": "voyage-4-large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, + "qwen3-6-plus": { + "id": "qwen3-6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-01-15", - "last_updated": "2026-03-06", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -9096,124 +9673,190 @@ }, "open_weights": false, "limit": { - "context": 32000, - "output": 0 - } - }, - "voyage/rerank-2.5": { - "id": "voyage/rerank-2.5", - "name": "Voyage Rerank 2.5", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-08-11", - "last_updated": "2025-08-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 65536 }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.5, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 2, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 2 + } } }, - "voyage/rerank-2.5-lite": { - "id": "voyage/rerank-2.5-lite", - "name": "Voyage Rerank 2.5 Lite", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, + "qwen3-6-27b": { + "id": "qwen3-6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 80000 + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "output": 32000 + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.412564, + "output": 2.475384, + "cache_read": 0.412564 } }, - "voyage/voyage-code-2": { - "id": "voyage/voyage-code-2", - "name": "voyage-code-2", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, + "glm-5-3-flash": { + "id": "glm-5-3-flash", + "name": "GLM 5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2024-01", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.075, + "output": 0.25, + "cache_read": 0.075 } }, - "voyage/voyage-3.5-lite": { - "id": "voyage/voyage-3.5-lite", - "name": "voyage-3.5-lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "kimi-k2-7-code": { + "id": "kimi-k2-7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 + "context": 256000, + "output": 131072 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.95 } }, - "voyage/voyage-3.5": { - "id": "voyage/voyage-3.5", - "name": "voyage-3.5", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, + "seed-2-0-lite": { + "id": "seed-2-0-lite", + "name": "Seed 2.0 Lite", + "description": "Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -9221,21 +9864,56 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.31, + "output": 2.5, + "cache_read": 0.31, + "tiers": [ + { + "input": 0.62, + "output": 5, + "cache_read": 0.62, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "voyage/voyage-3-large": { - "id": "voyage/voyage-3-large", - "name": "voyage-3-large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "voyage", + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 393216 + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-01-07", - "last_updated": "2024-09", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -9244,52 +9922,87 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 + "context": 1000000, + "output": 393216 + }, + "cost": { + "input": 0.424, + "output": 1.272, + "cache_read": 0.424 } }, - "voyage/voyage-finance-2": { - "id": "voyage/voyage-finance-2", - "name": "voyage-finance-2", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, + "mimo-v2-5": { + "id": "mimo-v2-5", + "name": "MiMo V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-06-03", - "last_updated": "2024-03", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0.7, + "output": 1.4, + "cache_read": 0.014 } }, - "mistral/mistral-nemo": { - "id": "mistral/mistral-nemo", - "name": "Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, + "seed-2-0-mini": { + "id": "seed-2-0-mini", + "name": "Seed 2.0 Mini", + "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-07-18", - "last_updated": "2024-07-01", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -9297,76 +10010,132 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 256000, "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 0.12, + "output": 0.5, + "cache_read": 0.12, + "tiers": [ + { + "input": 0.24, + "output": 1, + "cache_read": 0.24, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "mistral/codestral-embed": { - "id": "mistral/codestral-embed", - "name": "Codestral Embed", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "codestral-embed", - "attachment": false, - "reasoning": false, - "tool_call": false, + "muse-glimmer-30b": { + "id": "muse-glimmer-30b", + "name": "Muse Glimmer 30B", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.2, + "output": 0.8, + "cache_read": 0.05 } }, - "mistral/mistral-embed": { - "id": "mistral/mistral-embed", - "name": "Mistral Embed", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "mistral-embed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2023-12-11", - "last_updated": "2023-12-11", + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 3 } }, - "mistral/devstral-small": { - "id": "mistral/devstral-small", - "name": "Devstral Small 1.1", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", + "glm-4-7-flash": { + "id": "glm-4-7-flash", + "name": "GLM 4.7 Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-05-21", - "last_updated": "2025-05-07", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -9375,52 +10144,60 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 64000 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0, + "output": 0 } }, - "mistral/mistral-large-3": { - "id": "mistral/mistral-large-3", - "name": "Mistral Large 3", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", + "mimo-v2-6-pro": { + "id": "mimo-v2-6-pro", + "name": "MiMo V2.6 Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 2.175, + "output": 4.35, + "cache_read": 2.175 } }, - "mistral/mistral-medium-3.5": { - "id": "mistral/mistral-medium-3.5", - "name": "Mistral Medium Latest", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "qwen3-7-max": { + "id": "qwen3-7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -9428,49 +10205,26 @@ "type": "effort", "values": [ "none", - "high" + "low", + "medium", + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 64000 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-05-21", + "release_date": "2026-05-21", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 1.5, - "output": 7.5 - } - }, - "mistral/mistral-medium": { - "id": "mistral/mistral-medium", - "name": "Mistral Medium 3.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" @@ -9478,62 +10232,82 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 64000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 2 + "input": 2.5, + "output": 7.5, + "cache_read": 2.5 } }, - "mistral/devstral-small-2": { - "id": "mistral/devstral-small-2", - "name": "Devstral Small 2", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", + "step-3-5-flash": { + "id": "step-3-5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-09", - "last_updated": "2025-05-07", + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 256000, - "output": 256000 + "input": 256000, + "output": 131072 }, "cost": { "input": 0.1, - "output": 0.3 + "output": 0.3, + "cache_read": 0.02 } }, - "mistral/ministral-14b": { - "id": "mistral/ministral-14b", - "name": "Ministral 14B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "qwen3-8-max": { + "id": "qwen3-8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-02", - "last_updated": "2025-12-01", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -9541,26 +10315,45 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 2, + "output": 6, + "cache_read": 2 } }, - "mistral/devstral-2": { - "id": "mistral/devstral-2", - "name": "Devstral 2", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", + "deepseek-v3-2": { + "id": "deepseek-v3-2", + "name": "DeepSeek V3.2", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 393216 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -9569,32 +10362,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.4, - "output": 2 + "input": 0.57, + "output": 1.71, + "cache_read": 0.57 } }, - "mistral/mistral-small": { - "id": "mistral/mistral-small", - "name": "Mistral Small (latest)", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, - "reasoning": false, + "minimax-m2-7-highspeed": { + "id": "minimax-m2-7-highspeed", + "name": "MiniMax M2.7 Highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2024-09-17", - "last_updated": "2026-03-16", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -9602,26 +10396,33 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 4000 + "context": 200000, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "mistral/ministral-8b": { - "id": "mistral/ministral-8b", - "name": "Ministral 8B (latest)", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "mimo-v2-5-pro": { + "id": "mimo-v2-5-pro", + "name": "MiMo V2.5 Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-04", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -9632,26 +10433,29 @@ }, "open_weights": true, "limit": { - "context": 128000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 2.175, + "output": 4.35, + "cache_read": 0.018 } }, - "mistral/codestral": { - "id": "mistral/codestral", - "name": "Codestral (latest)", - "description": "Mistral code model for completions, refactors, and developer IDE workflows", - "family": "codestral", + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-05-29", - "last_updated": "2025-01-04", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" @@ -9660,63 +10464,127 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 256000, - "output": 4096 + "output": 65536 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 1.08, + "output": 5.52, + "cache_read": 1.08, + "tiers": [ + { + "input": 2.16, + "output": 11.04, + "cache_read": 2.16, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 2.7, + "output": 13.8, + "cache_read": 2.7, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "mistral/pixtral-12b": { - "id": "mistral/pixtral-12b", - "name": "Pixtral 12B", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", + "seed-2-0-pro": { + "id": "seed-2-0-pro", + "name": "Seed 2.0 Pro", + "description": "Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows", + "family": "seed", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-09-01", - "last_updated": "2024-09-01", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 256000, "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 0.63, + "output": 3.79, + "cache_read": 0.63, + "tiers": [ + { + "input": 1.26, + "output": 7.58, + "cache_read": 1.26, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "mistral/pixtral-large": { - "id": "mistral/pixtral-large", - "name": "Pixtral Large (latest)", - "description": "Mistral's larger vision model for document-heavy image understanding and chat", - "family": "pixtral", + "qwen3-5-4b": { + "id": "qwen3-5-4b", + "name": "Qwen3.5 4B", + "description": "Qwen3.5 4B is a low-cost multimodal reasoning model with 256K context, image and video input, function tools, and structured output.", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 32768 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2024-11-04", + "release_date": "2026-03-02", + "last_updated": "2026-03-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -9724,91 +10592,91 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 2, - "output": 6 + "input": 0.04, + "output": 0.07, + "cache_read": 0.02 } }, - "mistral/ministral-3b": { - "id": "mistral/ministral-3b", - "name": "Ministral 3B (latest)", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, - "reasoning": false, + "step-5-preview": { + "id": "step-5-preview", + "name": "Step 5 Preview", + "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-04", + "release_date": "2026-09-16", + "last_updated": "2026-09-20", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 1024000, + "output": 131072 }, "cost": { - "input": 0.04, - "output": 0.04 - } - }, - "mistral/magistral-small": { - "id": "mistral/magistral-small", - "name": "Magistral Small", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-small", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.5, - "output": 1.5 + "input": 1, + "output": 2.7, + "cache_read": 0.05 } }, - "mistral/magistral-medium": { - "id": "mistral/magistral-medium", - "name": "Magistral Medium (latest)", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-medium", - "attachment": false, + "qwen3-5-27b": { + "id": "qwen3-5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 80000 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-20", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -9816,76 +10684,40 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 256000, + "output": 64000 }, "cost": { - "input": 2, - "output": 5 - } - }, - "google/gemini-embedding-2": { - "id": "google/gemini-embedding-2", - "name": "Gemini Embedding 2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "gemini-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "google/text-multilingual-embedding-002": { - "id": "google/text-multilingual-embedding-002", - "name": "Text Multilingual Embedding 002", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-03-01", - "last_updated": "2024-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" + "input": 0.086, + "output": 0.688, + "cache_read": 0.086, + "tiers": [ + { + "input": 0.258, + "output": 2.064, + "cache_read": 0.258, + "tier": { + "type": "context", + "size": 128000 + } + } ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 1536 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini", + "seed-2-1-turbo": { + "id": "seed-2-1-turbo", + "name": "Seed 2.1 Turbo", + "description": "Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -9895,14 +10727,13 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-06-23", + "last_updated": "2026-06-23", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -9910,154 +10741,156 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65000 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.03 + "input": 0.63, + "output": 3.13, + "cache_read": 0.63 } }, - "google/gemini-3.1-flash-image": { - "id": "google/gemini-3.1-flash-image", - "name": "Gemini 3.1 Flash Image (Nano Banana 2)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini", + "fugu-ultra-v1-0": { + "id": "fugu-ultra-v1-0", + "name": "Fugu Ultra v1.0", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "high" + "high", + "xhigh", + "max" ] } ], - "tool_call": false, - "temperature": true, - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 7.5, + "output": 45, + "cache_read": 1.5, + "tiers": [ + { + "input": 15, + "output": 67.5, + "cache_read": 3, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 15, + "output": 67.5, + "cache_read": 3 + } } }, - "google/gemini-3-flash": { - "id": "google/gemini-3-flash", - "name": "Gemini 3 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "qwen3-5-122b-a10b": { + "id": "qwen3-5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 80000 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 - } - }, - "google/veo-3.1-generate-001": { - "id": "google/veo-3.1-generate-001", - "name": "Veo 3.1", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2026-06-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "video" + "input": 0.115, + "output": 0.917, + "cache_read": 0.115, + "tiers": [ + { + "input": 0.287, + "output": 2.294, + "cache_read": 0.287, + "tier": { + "type": "context", + "size": 128000 + } + } ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini", + "qwen3-5-flash": { + "id": "qwen3-5-flash", + "name": "Qwen3.5 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -10066,107 +10899,79 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 32768 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 + "input": 0.09, + "output": 0.368, + "cache_read": 0.09 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "mimo-v2-6-pro-ultraspeed": { + "id": "mimo-v2-6-pro-ultraspeed", + "name": "MiMo V2.6 Pro UltraSpeed", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "family": "mimo", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, + "context": 1000000, "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.4 - } - }, - "google/veo-3.0-generate-001": { - "id": "google/veo-3.0-generate-001", - "name": "Veo 3.0", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-05-20", - "last_updated": "2026-06-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "video" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "google/veo-3.0-fast-generate-001": { - "id": "google/veo-3.0-fast-generate-001", - "name": "Veo 3.0 Fast Generate", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-07-31", - "last_updated": "2026-06-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "video" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "input": 21.75, + "output": 43.5, + "cache_read": 21.75 } }, - "google/text-embedding-005": { - "id": "google/text-embedding-005", - "name": "Text Embedding 005", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "glm-5-2": { + "id": "glm-5-2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-08-01", - "last_updated": "2024-08", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -10175,156 +10980,136 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 - } - }, - "google/gemini-embedding-001": { - "id": "google/gemini-embedding-001", - "name": "Gemini Embedding 001", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "gemini-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 131072 }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 1536 + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 1.4 } }, - "google/gemini-3.1-flash-lite-image": { - "id": "google/gemini-3.1-flash-lite-image", - "name": "Gemini 3.1 Flash Lite Image (Nano Banana 2 Lite)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini", + "mimo-v2-6-flash": { + "id": "mimo-v2-6-flash", + "name": "MiMo V2.6 Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 4096 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.03 + "input": 0.7, + "output": 1.4, + "cache_read": 0.7 } }, - "google/gemini-2.5-flash-image": { - "id": "google/gemini-2.5-flash-image", - "name": "Nano Banana (Gemini 2.5 Flash Image)", - "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", - "family": "gemini-flash", - "attachment": false, - "reasoning": false, - "tool_call": false, + "muse-spark-1-2": { + "id": "muse-spark-1-2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 - } - }, - "google/imagen-4.0-fast-generate-001": { - "id": "google/imagen-4.0-fast-generate-001", - "name": "Imagen 4 Fast", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-06-01", - "last_updated": "2025-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 480, - "output": 0 + "input": 1.25, + "output": 4.25, + "cache_read": 1 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", + "qwen3-7-flash": { + "id": "qwen3-7-flash", + "name": "Qwen3.7 Flash", + "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] }, { "type": "budget_tokens", - "min": 512, - "max": 24576 + "min": 1, + "max": 131072 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -10332,514 +11117,421 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 1000000, + "input": 991000, "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 + "input": 0.03, + "output": 0.13, + "cache_read": 0.006, + "tiers": [ + { + "input": 0.1, + "output": 0.4, + "cache_read": 0.02, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.2, + "output": 0.8, + "cache_read": 0.04, + "tier": { + "type": "context", + "size": 256000 + } + } + ] } }, - "google/gemini-omni-flash-preview": { - "id": "google/gemini-omni-flash-preview", - "name": "Gemini Omni Flash Preview", - "description": "Omni-modal model for text, vision, audio, and multimodal agent tasks", - "family": "gemini", - "attachment": true, + "minimax-m2-7": { + "id": "minimax-m2-7", + "name": "MiniMax M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 57920 + "context": 200000, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 9 + "input": 0.15, + "output": 0.6, + "cache_read": 0.03 } }, - "google/gemini-3.1-flash-image-preview": { - "id": "google/gemini-3.1-flash-image-preview", - "name": "Gemini 3.1 Flash Image Preview (Nano Banana 2)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini", - "attachment": true, + "qwen3-6-35b-a3b": { + "id": "qwen3-6-35b-a3b", + "name": "Qwen3.6 35B A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "high" + "none", + "low", + "medium", + "high", + "max" ] } ], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, - "output": 32768 + "output": 16384 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 - } - }, - "google/imagen-4.0-ultra-generate-001": { - "id": "google/imagen-4.0-ultra-generate-001", - "name": "Imagen 4 Ultra", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-05-24", - "last_updated": "2025-05-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 480, - "output": 0 + "input": 0.07, + "output": 0.42, + "cache_read": 0.035 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini", + "qwen3-5-9b": { + "id": "qwen3-5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2026-02-19", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 0.09, + "output": 0.13, + "cache_read": 0.045 } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "step-3-5-flash-2603": { + "id": "step-3-5-flash-2603", + "name": "Step 3.5 Flash 2603", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, + "knowledge": "2025-01", "release_date": "2026-04-02", "last_updated": "2026-04-02", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, + "context": 256000, + "input": 256000, "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "google/gemini-3-pro-preview": { - "id": "google/gemini-3-pro-preview", - "name": "Gemini 3 Pro Preview", - "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", - "family": "gemini-pro", + "qwen3-5-35b-a3b": { + "id": "qwen3-5-35b-a3b", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "high" + "medium", + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 80000 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 256000, "output": 64000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, + "input": 0.057, + "output": 0.459, + "cache_read": 0.057, "tiers": [ { - "input": 4, - "output": 18, - "cache_read": 0.4, + "input": 0.229, + "output": 1.835, + "cache_read": 0.229, "tier": { "type": "context", - "size": 200000 + "size": 128000 } } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "google/veo-3.1-fast-generate-001": { - "id": "google/veo-3.1-fast-generate-001", - "name": "Veo 3.1 Fast Generate", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2026-06-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "video" ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 } }, - "google/gemini-3-pro-image": { - "id": "google/gemini-3-pro-image", - "name": "Nano Banana Pro (Gemini 3 Pro Image)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-pro", + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-09-01", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 - } - }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini", - "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 393216 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 65000 + "output": 393216 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.03 - } - }, - "google/imagen-4.0-generate-001": { - "id": "google/imagen-4.0-generate-001", - "name": "Imagen 4", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-05-20", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 480, - "output": 0 + "input": 1.32, + "output": 3.96, + "cache_read": 1.32 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", + "kimi-k2-6": { + "id": "kimi-k2-6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] }, { "type": "budget_tokens", - "min": 0, - "max": 24576 + "min": 1, + "max": 81920 } ], "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-04-21", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "audio", - "video", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 16000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 + "input": 0.8939, + "output": 3.7131, + "cache_read": 0.1788 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-06-01", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "audio", - "video", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 524288 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, + "input": 0.225, + "output": 0.9, + "cache_read": 0.045, "tiers": [ { - "input": 2.5, - "output": 15, - "cache_read": 0.25, + "input": 0.45, + "output": 1.8, + "cache_read": 0.09, "tier": { "type": "context", - "size": 200000 + "size": 512000 } } ], "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 0.45, + "output": 1.8, + "cache_read": 0.09 } } }, - "prodia/flux-fast-schnell": { - "id": "prodia/flux-fast-schnell", - "name": "Flux Schnell", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-08-02", - "last_updated": "2026-06-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 512, - "output": 0 - } - }, - "openai/gpt-oss-safeguard-20b": { - "id": "openai/gpt-oss-safeguard-20b", - "name": "gpt-oss-safeguard-20b", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", + "glm-5-3": { + "id": "glm-5-3", + "name": "GLM 5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -10847,16 +11539,16 @@ "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-10-29", - "last_updated": "2024-12-01", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -10865,33 +11557,43 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "input": 65536, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.075, - "output": 0.3, - "cache_read": 0.037 + "input": 1.4, + "output": 4.4, + "cache_read": 1.4 } }, - "openai/gpt-3.5-turbo-instruct": { - "id": "openai/gpt-3.5-turbo-instruct", - "name": "GPT-3.5 Turbo Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2021-09", - "release_date": "2023-09-18", - "last_updated": "2023-03-01", + "fugu-ultra-v1-1": { + "id": "fugu-ultra-v1-1", + "name": "Fugu Ultra v1.1", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -10899,200 +11601,212 @@ }, "open_weights": false, "limit": { - "context": 8192, - "input": 4096, - "output": 4096 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 2 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "openai/gpt-5.2-chat": { - "id": "openai/gpt-5.2-chat", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "kimi-k2-7-code-highspeed": { + "id": "kimi-k2-7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-11", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 111616, - "output": 16384 + "context": 256000, + "output": 131072 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "openai/text-embedding-3-large": { - "id": "openai/text-embedding-3-large", - "name": "text-embedding-3-large", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "input": 6656, - "output": 1536 + "input": 1.9, + "output": 8, + "cache_read": 1.9 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT 5.2 ", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt", + "deepseek-v4-1-flash": { + "id": "deepseek-v4-1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 393216 } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 21, - "output": 168 - } - }, - "openai/gpt-4o-mini-search-preview": { - "id": "openai/gpt-4o-mini-search-preview", - "name": "GPT 4o Mini Search Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2025-03-12", - "last_updated": "2025-01", - "modalities": { - "input": [ - "text" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 111616, - "output": 16384 + "context": 1000000, + "output": 393216 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.3, + "output": 1.2, + "cache_read": 0.3 } }, - "openai/gpt-5-chat": { - "id": "openai/gpt-5-chat", - "name": "GPT-5 Chat", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "qwen3-7-plus": { + "id": "qwen3-7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 256000 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 111616, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.4, + "output": 1.6, + "cache_read": 0.4, + "tiers": [ + { + "input": 1.2, + "output": 4.8, + "cache_read": 1.2, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.2, + "output": 4.8, + "cache_read": 1.2 + } } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "glm-5-1": { + "id": "glm-5-1", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 38912 + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2021-09", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "release_date": "2026-04-07", + "last_updated": "2026-06-12", "modalities": { "input": [ "text" @@ -11101,87 +11815,92 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16385, - "input": 12289, - "output": 4096 + "context": 202000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 0.825, + "output": 3.301, + "cache_read": 0.165, + "tiers": [ + { + "input": 1.1, + "output": 3.851, + "cache_read": 0.22, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5 pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt", + "glm-4-6v-flash": { + "id": "glm-4-6v-flash", + "name": "GLM 4.6V Flash", + "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 128000, - "output": 272000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 15, - "output": 120 + "input": 0, + "output": 0 } }, - "openai/o3-pro": { - "id": "openai/o3-pro", - "name": "o3 Pro", - "description": "High-effort o3 tier for difficult technical reasoning and careful answers", - "family": "o-pro", + "fugu-ultra-v2-0": { + "id": "fugu-ultra-v2-0", + "name": "Fugu Ultra v2.0", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -11189,49 +11908,69 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 100000, - "output": 100000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 20, - "output": 80 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "openai/gpt-4o-transcribe": { - "id": "openai/gpt-4o-transcribe", - "name": "GPT-4o Transcribe", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "gpt", - "attachment": false, + "mistral-small-4": { + "id": "mistral-small-4", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-03-13", - "last_updated": "2024-03-13", + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "audio" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.15, + "output": 0.6, + "cache_read": 0.15 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT 5.4 Nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt", + "qwen3-8-omni-flash": { + "id": "qwen3-8-omni-flash", + "name": "Qwen3.8 Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -11242,21 +11981,26 @@ "low", "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-09-17", + "last_updated": "2026-09-17", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -11264,32 +12008,38 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.3, + "output": 0.94, + "cache_read": 0.3 } }, - "openai/gpt-5.3-chat": { - "id": "openai/gpt-5.3-chat", - "name": "GPT-5.3 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "qwen3-5-plus": { + "id": "qwen3-5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-06", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -11297,136 +12047,159 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 111616, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.36, + "output": 2.21, + "cache_read": 0.36, + "tiers": [ + { + "input": 1.08, + "output": 6.62, + "cache_read": 1.08, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.08, + "output": 6.62, + "cache_read": 1.08 + } } }, - "openai/gpt-5.1-thinking": { - "id": "openai/gpt-5.1-thinking", - "name": "GPT 5.1 Thinking", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "muse-spark-1-3": { + "id": "muse-spark-1-3", + "name": "Muse Spark 1.3", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-11-12", - "last_updated": "2025-08-07", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { "input": 1.25, - "output": 10, - "cache_read": 0.125 + "output": 4.25, + "cache_read": 1 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1-Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt", - "attachment": true, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 393216 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-11-12", - "last_updated": "2025-11-13", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 393216 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.65, + "output": 3.3, + "cache_read": 1.65 } }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT 5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", + "qwen3-8-flash": { + "id": "qwen3-8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-11-19", - "last_updated": "2025-11-13", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -11434,76 +12207,52 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.16, + "output": 0.47, + "cache_read": 0.16 } }, - "openai/gpt-4o-mini-transcribe": { - "id": "openai/gpt-4o-mini-transcribe", - "name": "GPT-4o mini Transcribe", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "o-mini", - "attachment": false, + "gemma-3-27b": { + "id": "gemma-3-27b", + "name": "Gemma 3 27B", + "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", + "family": "gemma", + "attachment": true, "reasoning": false, "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2024-03-13", - "last_updated": "2024-03-13", + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ - "audio" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 5 - } - }, - "openai/text-embedding-ada-002": { - "id": "openai/text-embedding-ada-002", - "name": "text-embedding-ada-002", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2022-12-15", - "last_updated": "2022-12-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "input": 6656, - "output": 1536 + "input": 0, + "output": 0 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "qwen3-6-flash": { + "id": "qwen3-6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -11514,21 +12263,25 @@ "low", "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 64000 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -11536,45 +12289,61 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.25, + "output": 1.5, + "cache_read": 0.25, + "tiers": [ + { + "input": 1, + "output": 4, + "cache_read": 1, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1, + "output": 4, + "cache_read": 1 + } } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT 5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", + "seed-2-0-code": { + "id": "seed-2-0-code", + "name": "Seed 2.0 Code", + "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -11582,80 +12351,122 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 256000, "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.4, + "output": 2.4, + "cache_read": 0.4, + "tiers": [ + { + "input": 0.8, + "output": 4.8, + "cache_read": 0.8, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "openai/text-embedding-3-small": { - "id": "openai/text-embedding-3-small", - "name": "text-embedding-3-small", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, + "qwen3-8-27b": { + "id": "qwen3-8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "input": 6656, - "output": 1536 + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.17, + "output": 0.5, + "cache_read": 0.08 } }, - "openai/gpt-realtime-1.5": { - "id": "openai/gpt-realtime-1.5", - "name": "GPT-Realtime-1.5", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, + "muse-spark-1-1": { + "id": "muse-spark-1-1", + "name": "Muse Spark 1.1", + "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-04-08", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", + "image", + "video", "audio" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 4, - "output": 16, - "cache_read": 0.4 + "input": 1.25, + "output": 4.25, + "cache_read": 1 } }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT 5.6 Luna", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "qwen3-6-max-preview": { + "id": "qwen3-6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -11665,22 +12476,24 @@ "low", "medium", "high", - "xhigh", "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 393216 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -11688,22 +12501,30 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 256000, + "output": 65536 }, "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 1.31, + "output": 7.88, + "cache_read": 1.31, + "tiers": [ + { + "input": 1.97, + "output": 11.82, + "cache_read": 1.97, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", + "step-3-7-flash": { + "id": "step-3-7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -11719,36 +12540,36 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-11-12", - "last_updated": "2025-11-13", + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "input": 256000, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 } }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT 5.6 Terra", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "qwen3-5-397b-a17b": { + "id": "qwen3-5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -11759,93 +12580,70 @@ "low", "medium", "high", - "xhigh", "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 80000 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125 - } - }, - "openai/gpt-image-1.5": { - "id": "openai/gpt-image-1.5", - "name": "GPT Image 1.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" + "input": 0.172, + "output": 1.032, + "cache_read": 0.172, + "tiers": [ + { + "input": 0.43, + "output": 2.58, + "cache_read": 0.43, + "tier": { + "type": "context", + "size": 128000 + } + } ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - }, - "cost": { - "input": 5, - "output": 32, - "cache_read": 1.25 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "glm-4-5-flash": { + "id": "glm-4-5-flash", + "name": "GLM 4.5 Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -11854,21 +12652,21 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "output": 98304 }, "cost": { - "input": 0.1, - "output": 0.5 + "input": 0, + "output": 0 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT 5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "gemma-4-26b-a4b": { + "id": "gemma-4-26b-a4b", + "name": "Gemma 4 26B-A4B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -11879,44 +12677,47 @@ "low", "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2026-04-02", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 0.05, + "output": 0.29, + "cache_read": 0.025 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT 5.4 Mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt", - "attachment": true, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -11926,109 +12727,123 @@ "low", "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 393216 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 393216 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.14, + "output": 0.28, + "cache_read": 0.14 } - }, - "openai/gpt-realtime-mini": { - "id": "openai/gpt-realtime-mini", - "name": "GPT-Realtime mini", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-10-10", - "last_updated": "2025-10-10", + } + } + }, + "sensenova": { + "id": "sensenova", + "env": [ + "SENSENOVA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://token.sensenova.cn/v1", + "name": "SenseNova (China)", + "doc": "https://platform.sensenova.cn/docs", + "models": { + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "audio" + "image" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06 - } - }, - "openai/tts-1": { - "id": "openai/tts-1", - "name": "TTS-1", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "o", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2023-11-06", - "last_updated": "2023-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "openai/whisper-1": { - "id": "openai/whisper-1", - "name": "Whisper", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, + "sensenova-6.8-flash-lite": { + "id": "sensenova-6.8-flash-lite", + "name": "SenseNova 6.8 Flash Lite", + "description": "SenseNova lightweight multimodal agent model for real-world complex tasks, data analysis, and complex information presentation", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2022-09-21", - "last_updated": "2022-09-21", + "release_date": "2026-08-11", + "last_updated": "2026-08-28", "modalities": { "input": [ - "audio" + "text", + "image" ], "output": [ "text" @@ -12036,126 +12851,177 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 } }, - "openai/tts-1-hd": { - "id": "openai/tts-1-hd", - "name": "TTS-1 HD", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "o", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2023-11-06", - "last_updated": "2023-11-06", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 } }, - "openai/o3-deep-research": { - "id": "openai/o3-deep-research", - "name": "o3-deep-research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "family": "o", - "attachment": true, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "medium" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-06-26", - "last_updated": "2024-06-26", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 100000, - "output": 100000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 10, - "output": 40, - "cache_read": 2.5 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "openai/gpt-image-1": { - "id": "openai/gpt-image-1", - "name": "GPT Image 1", - "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", - "family": "gpt-image", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-03-25", - "last_updated": "2025-04-24", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 5, - "output": 40, - "cache_read": 1.25 + "input": 0, + "output": 0, + "cache_read": 0 } - }, - "openai/gpt-realtime-2": { - "id": "openai/gpt-realtime-2", - "name": "gpt-realtime-2", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", - "attachment": false, + } + } + }, + "alibaba-token-plan": { + "id": "alibaba-token-plan", + "env": [ + "ALIBABA_TOKEN_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://token-plan.ap-southeast-1.maas.aliyuncs.com/compatible-mode/v1", + "name": "Alibaba Token Plan", + "doc": "https://www.alibabacloud.com/help/en/model-studio/token-plan-overview", + "models": { + "happyhorse-1.1-r2v": { + "id": "happyhorse-1.1-r2v", + "name": "HappyHorse 1.1 Reference-to-Video", + "description": "Video model for reference-guided video generation", + "attachment": true, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-07-17", + "last_updated": "2026-07-17", "modalities": { "input": [ - "text", - "audio" + "image", + "text" ], "output": [ - "text", - "audio" + "video" ] }, "open_weights": false, @@ -12164,112 +13030,140 @@ "output": 0 }, "cost": { - "input": 4, - "output": 24, - "cache_read": 0.4 + "input": 0, + "output": 0 } }, - "openai/gpt-image-1-mini": { - "id": "openai/gpt-image-1-mini", - "name": "GPT Image 1 Mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT 5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt", - "attachment": true, + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 30, - "output": 180 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT 5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt", + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "low", "medium", - "high", "xhigh" ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-12-01", - "release_date": "2026-04-24", - "last_updated": "2026-04-23", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ @@ -12279,87 +13173,95 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 872000, - "output": 128000 + "output": 131072 }, "cost": { - "input": 30, - "output": 180 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "happyhorse-1.1-t2v": { + "id": "happyhorse-1.1-t2v", + "name": "HappyHorse 1.1 Text-to-Video", + "description": "Video model for prompt-driven text-to-video generation", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-07-17", + "last_updated": "2026-07-17", "modalities": { "input": [ "text" ], "output": [ + "video" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen-image-2.0": { + "id": "qwen-image-2.0", + "name": "Qwen Image 2.0", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-03-03", + "last_updated": "2026-03-03", + "modalities": { + "input": [ "text" + ], + "output": [ + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 122880, - "output": 8192 + "context": 8192, + "output": 0 }, "cost": { - "input": 0.05, - "output": 0.2 + "input": 0, + "output": 0 } }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT 5.6 Sol", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -12367,80 +13269,101 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 202752, + "output": 16384 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2-Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-18", - "last_updated": "2025-12-11", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-5.1-instant": { - "id": "openai/gpt-5.1-instant", - "name": "GPT-5.1 Instant", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "qwen3.8-max-preview": { + "id": "qwen3.8-max-preview", + "name": "Qwen3.8 Max Preview", + "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-11-12", - "last_updated": "2025-08-07", + "release_date": "2026-07-19", + "last_updated": "2026-07-19", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -12448,27 +13371,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 111616, - "output": 16384 + "context": 1000000, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-image-2": { - "id": "openai/gpt-image-2", - "name": "GPT Image 2", + "qwen-image-2.0-pro": { + "id": "qwen-image-2.0-pro", + "name": "Qwen Image 2.0 Pro", "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text" @@ -12479,235 +13403,220 @@ }, "open_weights": false, "limit": { - "context": 0, + "context": 8192, "output": 0 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 1.25 + "input": 0, + "output": 0 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT 5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-12-01", - "release_date": "2026-04-24", - "last_updated": "2026-04-23", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 872000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "wan2.7-image": { + "id": "wan2.7-image", + "name": "Wan2.7 Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 8192, + "output": 0 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0, + "output": 0 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, + "wan2.7-image-pro": { + "id": "wan2.7-image-pro", + "name": "Wan2.7 Image Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 8192, + "output": 0 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0, + "output": 0 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 98304 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "happyhorse-1.1-i2v": { + "id": "happyhorse-1.1-i2v", + "name": "HappyHorse 1.1 Image-to-Video", + "description": "Video model for image-to-video generation", "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": false, + "tool_call": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2026-07-17", + "last_updated": "2026-07-17", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 0, + "output": 0 }, "cost": { - "input": 10, - "output": 30 + "input": 0, + "output": 0 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 131072 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -12715,122 +13624,135 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", - "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, + "context": 196608, + "input": 196601, "output": 32768 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/o1": { - "id": "openai/o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "xhigh" ] + }, + { + "type": "budget_tokens", + "max": 262144 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -12838,71 +13760,80 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -12911,78 +13842,96 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 202752, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 131072 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -12990,196 +13939,218 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-03", + "last_updated": "2025-12-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0, + "output": 0 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", - "attachment": true, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "zai/glm-4.7": { - "id": "zai/glm-4.7", - "name": "GLM 4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 120000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.12 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "zai/glm-4.5v": { - "id": "zai/glm-4.5v", - "name": "GLM 4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 66000, - "output": 16000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "zai/glm-4.5": { - "id": "zai/glm-4.5", - "name": "GLM 4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, - "interleaved": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -13190,33 +14161,42 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 96000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "zai/glm-4.7-flashx": { - "id": "zai/glm-4.7-flashx", - "name": "GLM 4.7 FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + } + } + }, + "cloudflare-workers-ai": { + "id": "cloudflare-workers-ai", + "env": [ + "CLOUDFLARE_ACCOUNT_ID", + "CLOUDFLARE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.cloudflare.com/client/v4/accounts/${CLOUDFLARE_ACCOUNT_ID}/ai/v1", + "name": "Cloudflare Workers AI", + "doc": "https://developers.cloudflare.com/workers-ai/models/", + "models": { + "@cf/meta/llama-3.3-70b-instruct-fp8-fast": { + "id": "@cf/meta/llama-3.3-70b-instruct-fp8-fast", + "name": "Llama 3.3 70B Instruct fp8 Fast", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -13227,32 +14207,27 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 24000, + "output": 24000 }, "cost": { - "input": 0.06, - "output": 0.4, - "cache_read": 0.01 + "input": 0.293, + "output": 2.253 } }, - "zai/glm-5.1": { - "id": "zai/glm-5.1", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "@cf/meta/llama-guard-3-8b": { + "id": "@cf/meta/llama-guard-3-8b", + "name": "Llama Guard 3 8B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -13261,35 +14236,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202000, - "output": 202000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.3, - "output": 4.3, - "cache_read": 0.26 + "input": 0.484, + "output": 0.03 } }, - "zai/glm-4.6": { - "id": "zai/glm-4.6", - "name": "GLM 4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", + "@cf/meta/llama-3.1-8b-instruct-fp8": { + "id": "@cf/meta/llama-3.1-8b-instruct-fp8", + "name": "Llama 3.1 8B Instruct fp8", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -13300,42 +14269,31 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 96000 + "context": 32000, + "output": 32000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0.152, + "output": 0.287 } }, - "zai/glm-5.2": { - "id": "zai/glm-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, + "@cf/meta/llama-3.2-11b-vision-instruct": { + "id": "@cf/meta/llama-3.2-11b-vision-instruct", + "name": "Llama 3.2 11B Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-06-16", - "last_updated": "2026-06-13", + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -13343,182 +14301,161 @@ }, "open_weights": true, "limit": { - "context": 1040000, + "context": 128000, "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.0485, + "output": 0.676 } }, - "zai/glm-4.6v": { - "id": "zai/glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "@cf/meta/llama-3.2-1b-instruct": { + "id": "@cf/meta/llama-3.2-1b-instruct", + "name": "Llama 3.2 1B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-30", - "last_updated": "2025-12-08", + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 24000 + "context": 60000, + "output": 60000 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.05 + "input": 0.027, + "output": 0.201 } }, - "zai/glm-5.2-fast": { - "id": "zai/glm-5.2-fast", - "name": "GLM 5.2 Fast", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], + "@cf/meta/llama-4-scout-17b-16e-instruct": { + "id": "@cf/meta/llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout 17B 16E Instruct", + "description": "Open Llama with long-context vision for efficient multimodal agents", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131000, + "output": 16384 }, "cost": { - "input": 3, - "output": 10.25, - "cache_read": 0.5 + "input": 0.27, + "output": 0.85 } }, - "zai/glm-5v-turbo": { - "id": "zai/glm-5v-turbo", - "name": "GLM 5V Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "@cf/meta/llama-3.2-3b-instruct": { + "id": "@cf/meta/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 80000, + "output": 80000 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 0.0509, + "output": 0.335 } }, - "zai/glm-4.6v-flash": { - "id": "zai/glm-4.6v-flash", - "name": "GLM-4.6V-Flash", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "@cf/google/gemma-4-26b-a4b-it": { + "id": "@cf/google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 24000 + "context": 256000, + "output": 16384 + }, + "cost": { + "input": 0.1, + "output": 0.3 } }, - "zai/glm-4.5-air": { - "id": "zai/glm-4.5-air", - "name": "GLM 4.5 Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "@cf/aisingapore/gemma-sea-lion-v4-27b-it": { + "id": "@cf/aisingapore/gemma-sea-lion-v4-27b-it", + "name": "Gemma Sea Lion V4 27B It", + "description": "Gemma 3 27B tuned by AI Singapore for Southeast Asian languages and instruction following", + "family": "gemma", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" @@ -13530,31 +14467,25 @@ "open_weights": true, "limit": { "context": 128000, - "output": 96000 + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03 + "input": 0.351, + "output": 0.555 } }, - "zai/glm-4.7-flash": { - "id": "zai/glm-4.7-flash", - "name": "GLM 4.7 Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm", + "@cf/ibm-granite/granite-4.0-h-micro": { + "id": "@cf/ibm-granite/granite-4.0-h-micro", + "name": "Granite 4.0 H Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "granite", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-10-02", + "last_updated": "2025-10-02", "modalities": { "input": [ "text" @@ -13563,32 +14494,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, + "context": 131000, "output": 131000 }, "cost": { - "input": 0.07, - "output": 0.4 + "input": 0.017, + "output": 0.112 } }, - "zai/glm-5": { - "id": "zai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "@cf/deepseek-ai/deepseek-v4-flash-0731": { + "id": "@cf/deepseek-ai/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -13599,32 +14538,38 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 131100 + "context": 1310720, + "output": 1048576 }, "cost": { - "input": 0.95, - "output": 3.15, - "cache_read": 0.2 + "input": 0.44, + "output": 1.32, + "cache_read": 0.014 } }, - "zai/glm-5-turbo": { - "id": "zai/glm-5-turbo", - "name": "GLM 5 Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", + "@cf/deepseek-ai/deepseek-v4-pro-0813": { + "id": "@cf/deepseek-ai/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-03-16", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -13633,110 +14578,105 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202800, - "output": 131100 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 } }, - "bytedance/seedream-5.0-lite": { - "id": "bytedance/seedream-5.0-lite", - "name": "Seedream 5.0 Lite", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "seed", + "@cf/deepseek-ai/deepseek-r1-distill-qwen-32b": { + "id": "@cf/deepseek-ai/deepseek-r1-distill-qwen-32b", + "name": "Deepseek R1 Distill Qwen 32B", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-01-28", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 - } - }, - "bytedance/seedance-2.0-fast": { - "id": "bytedance/seedance-2.0-fast", - "name": "Seedance 2.0 Fast", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-04-14", - "last_updated": "2026-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "video" - ] + "context": 80000, + "output": 80000 }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "cost": { + "input": 0.497, + "output": 4.881 } }, - "bytedance/seedance-v1.0-pro": { - "id": "bytedance/seedance-v1.0-pro", - "name": "Seedance v1.0 Pro", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", + "@cf/mistralai/mistral-small-3.1-24b-instruct": { + "id": "@cf/mistralai/mistral-small-3.1-24b-instruct", + "name": "Mistral Small 3.1 24B Instruct", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-06-11", - "last_updated": "2025-06-11", + "knowledge": "2024-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-17", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.351, + "output": 0.555 } }, - "bytedance/seed-1.6": { - "id": "bytedance/seed-1.6", - "name": "Seed 1.6", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": false, + "@cf/moonshotai/kimi-k2.6": { + "id": "@cf/moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-01", - "last_updated": "2025-09", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -13746,148 +14686,219 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 262144, + "output": 256000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "bytedance/seedance-2.0": { - "id": "bytedance/seedance-2.0", - "name": "Seedance 2.0", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", + "@cf/moonshotai/kimi-k2.7-code": { + "id": "@cf/moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-14", - "last_updated": "2026-04-14", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "bytedance/seedance-v1.5-pro": { - "id": "bytedance/seedance-v1.5-pro", - "name": "Seedance v1.5 Pro", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", - "attachment": false, - "reasoning": false, - "tool_call": false, + "@cf/zai-org/glm-5.3-flash": { + "id": "@cf/zai-org/glm-5.3-flash", + "name": "Glm 5.3 Flash", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 - } - }, - "bytedance/seedance-v1.0-pro-fast": { - "id": "bytedance/seedance-v1.0-pro-fast", - "name": "Seedance v1.0 Pro Fast", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", + "context": 1310720, + "output": 1048576 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 + } + }, + "@cf/zai-org/glm-4.7-flash": { + "id": "@cf/zai-org/glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-10-24", - "last_updated": "2025-10-31", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.0605, + "output": 0.4 } }, - "bytedance/seedream-4.0": { - "id": "bytedance/seedream-4.0", - "name": "Seedream 4.0", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "seed", + "@cf/zai-org/glm-5.2": { + "id": "@cf/zai-org/glm-5.2", + "name": "Glm 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-09-09", - "last_updated": "2025-08-28", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "output": 256000 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "bytedance/seedream-4.5": { - "id": "bytedance/seedream-4.5", - "name": "Seedream 4.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "seed", + "@cf/zai-org/glm-5.3": { + "id": "@cf/zai-org/glm-5.3", + "name": "Glm 5.3", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-11-28", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1310720, + "output": 1048576 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "bytedance/seed-1.8": { - "id": "bytedance/seed-1.8", - "name": "Seed 1.8", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "@cf/nvidia/nemotron-3-120b-a12b": { + "id": "@cf/nvidia/nemotron-3-120b-a12b", + "name": "Nemotron 3 Super 120B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -13897,7 +14908,6 @@ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -13905,136 +14915,119 @@ } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-01", - "last_updated": "2025-10", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 256000, - "output": 64000 + "output": 256000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05 + "input": 0.5, + "output": 1.5 } }, - "morph/morph-v3-large": { - "id": "morph/morph-v3-large", - "name": "Morph v3 Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "morph", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", + "@cf/qwen/qwen3.8-27b": { + "id": "@cf/qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "output": 32000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.9, - "output": 1.9 + "input": 0.45, + "output": 3.2, + "cache_read": 0.05 } }, - "morph/morph-v3-fast": { - "id": "morph/morph-v3-fast", - "name": "Morph v3 Fast", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "morph", + "@cf/qwen/qwq-32b": { + "id": "@cf/qwen/qwq-32b", + "name": "Qwq 32B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16000, - "output": 16000 - }, - "cost": { - "input": 0.8, - "output": 1.2 - } - }, - "sakana/fugu-ultra": { - "id": "sakana/fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "aura", - "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-06-21", - "last_updated": "2026-06-21", + "knowledge": "2024-04", + "release_date": "2025-03-05", + "last_updated": "2025-03-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 24000, + "output": 24000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.66, + "output": 1 } }, - "nvidia/nemotron-3-ultra-550b-a55b": { - "id": "nvidia/nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", + "@cf/qwen/qwen3-30b-a3b-fp8": { + "id": "@cf/qwen/qwen3-30b-a3b-fp8", + "name": "Qwen3 30B A3b fp8", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ "text" @@ -14043,34 +15036,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 + "input": 0.0509, + "output": 0.335 } }, - "nvidia/nemotron-3-nano-30b-a3b": { - "id": "nvidia/nemotron-3-nano-30b-a3b", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", + "@cf/qwen/qwen2.5-coder-32b-instruct": { + "id": "@cf/qwen/qwen2.5-coder-32b-instruct", + "name": "Qwen2.5 Coder 32B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "release_date": "2024-11-12", + "last_updated": "2024-11-12", "modalities": { "input": [ "text" @@ -14079,33 +15066,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.05, - "output": 0.24 + "input": 0.66, + "output": 1 } }, - "nvidia/nemotron-nano-9b-v2": { - "id": "nvidia/nemotron-nano-9b-v2", - "name": "Nvidia Nemotron Nano 9B V2", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", + "@cf/openai/gpt-oss-20b": { + "id": "@cf/openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-08-18", - "last_updated": "2025-08-18", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -14114,32 +15106,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.06, - "output": 0.23 + "input": 0.2, + "output": 0.3 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "NVIDIA Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "@cf/openai/gpt-oss-120b": { + "id": "@cf/openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -14148,22 +15146,34 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.15, - "output": 0.65 + "input": 0.35, + "output": 0.75 } - }, - "nvidia/nemotron-nano-12b-v2-vl": { - "id": "nvidia/nemotron-nano-12b-v2-vl", - "name": "Nvidia Nemotron Nano 12B V2 VL", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", - "attachment": true, + } + } + }, + "poolside": { + "id": "poolside", + "env": [ + "POOLSIDE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.poolside.ai/v1", + "name": "Poolside", + "doc": "https://platform.poolside.ai", + "models": { + "poolside/laguna-s-2.1": { + "id": "poolside/laguna-s-2.1", + "name": "Laguna S 2.1", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -14171,35 +15181,39 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-10-28", - "last_updated": "2025-10-28", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 1048576, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo M2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo-v2.5", - "attachment": true, + "poolside/laguna-xs-2.1": { + "id": "poolside/laguna-xs-2.1", + "name": "Laguna XS 2.1", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -14207,36 +15221,38 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-07-02", + "last_updated": "2026-07-02", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "output": 131100 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "xiaomi/mimo-v2-flash": { - "id": "xiaomi/mimo-v2-flash", - "name": "MiMo V2 Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", + "poolside/laguna-m.1": { + "id": "poolside/laguna-m.1", + "name": "Laguna M.1", + "description": "Poolside's open-weight model for agentic coding and long-horizon work", + "family": "laguna", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -14245,10 +15261,13 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-16", - "last_updated": "2026-02-04", + "release_date": "2026-04-28", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -14257,34 +15276,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "output": 32000 + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "xiaomi/mimo-v2-pro": { - "id": "xiaomi/mimo-v2-pro", - "name": "MiMo V2 Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", + } + } + }, + "nano-gpt": { + "id": "nano-gpt", + "env": [ + "NANO_GPT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://nano-gpt.com/api/v1", + "name": "NanoGPT", + "doc": "https://docs.nano-gpt.com", + "models": { + "deepseek-reasoner-cheaper": { + "id": "deepseek-reasoner-cheaper", + "name": "Deepseek R1 Cheaper", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -14293,38 +15319,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 1, - "output": 3, + "input": 0.4, + "output": 1.7, "cache_read": 0.2 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo V2.5 Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo-v2.5-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "nano-gpt-help": { + "id": "nano-gpt-help", + "name": "NanoGPT Help", + "description": "Text-only NanoGPT support assistant. Questions are processed by the Help inference provider; do not paste secrets or account credentials. Covers the website, models, API, pricing, memory, media generation, and support.", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-06-06", + "last_updated": "2026-06-06", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -14332,51 +15352,57 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "output": 131000 + "context": 6000, + "input": 6000, + "output": 512 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "input": 0, + "output": 0 } }, - "quiverai/arrow-1.1": { - "id": "quiverai/arrow-1.1", - "name": "Arrow 1.1", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "o", + "doubao-seed-2-0-lite-260215": { + "id": "doubao-seed-2-0-lite-260215", + "name": "Doubao Seed 2.0 Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "seed", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": false, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 32000 + }, + "cost": { + "input": 0.1462, + "output": 0.8738, + "cache_read": 0.0731 } }, - "inception/mercury-coder-small": { - "id": "inception/mercury-coder-small", - "name": "Mercury Coder Small Beta", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "mercury", + "doubao-seed-2-0-mini-260215": { + "id": "doubao-seed-2-0-mini-260215", + "name": "Doubao Seed 2.0 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "seed", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-02-26", - "last_updated": "2025-02-26", + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text" @@ -14387,35 +15413,26 @@ }, "open_weights": false, "limit": { - "context": 32000, - "output": 16384 + "context": 256000, + "input": 256000, + "output": 32000 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.0493, + "output": 0.4845, + "cache_read": 0.02465 } }, - "inception/mercury-2": { - "id": "inception/mercury-2", - "name": "Mercury 2", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", + "fastgpt": { + "id": "fastgpt", + "name": "Web Answer", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-24", - "last_updated": "2026-03-06", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-23", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" @@ -14426,32 +15443,29 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.024999999999999998 + "input": 7.5, + "output": 7.5 } }, - "anthropic/claude-3.5-haiku": { - "id": "anthropic/claude-3.5-haiku", - "name": "Claude 3.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "doubao-seed-2-0-pro-260215": { + "id": "doubao-seed-2-0-pro-260215", + "name": "Doubao Seed 2.0 Pro", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "seed", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2024-11-04", - "last_updated": "2024-10-22", + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -14459,34 +15473,32 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 256000, + "input": 256000, + "output": 128000 }, "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "input": 0.782, + "output": 3.876, + "cache_read": 0.391 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "ernie-5.0-thinking-preview": { + "id": "ernie-5.0-thinking-preview", + "name": "Ernie 5.0 Thinking Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "ernie", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "tool_call": false, + "structured_output": false, + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -14494,42 +15506,42 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 1, + "output": 3.5, + "cache_read": 0.5 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "gemini-2.5-flash-lite-preview-09-2025-thinking": { + "id": "gemini-2.5-flash-lite-preview-09-2025-thinking", + "name": "Gemini 2.5 Flash Lite Preview (09/2025) – Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": true, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", "image", + "audio", "pdf" ], "output": [ @@ -14538,95 +15550,63 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "gemma-4-12b-it-station-keeper": { + "id": "gemma-4-12b-it-station-keeper", + "name": "Gemma 4 12B StationKeeper", + "description": "Gemma 4 12B StationKeeper is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 131,072-token context window.", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2026-01-31", - "release_date": "2026-06-29", - "last_updated": "2026-06-30", + "structured_output": false, + "release_date": "2026-09-09", + "last_updated": "2026-09-09", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.05, + "output": 0.25, + "cache_read": 0.025 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "glm-4-air-0111": { + "id": "glm-4-air-0111", + "name": "GLM 4 Air 0111", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "glm", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-11", + "last_updated": "2025-01-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -14634,47 +15614,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.1394, + "output": 0.1394, + "cache_read": 0.0697 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "glm-4.1v-thinking-flashx": { + "id": "glm-4.1v-thinking-flashx", + "name": "GLM 4.1V Thinking FlashX", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "glm-flash", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -14682,194 +15646,127 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 64000, + "input": 64000, + "output": 8192 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.3, + "output": 0.3, + "cache_read": 0.15 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "deepseek-v3-0324": { + "id": "deepseek-v3-0324", + "name": "DeepSeek Chat 0324", + "description": "March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2026-01-31", - "release_date": "2026-07-01", - "last_updated": "2026-06-09", + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.2, + "output": 0.77, + "cache_read": 0.135 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "venice-uncensored": { + "id": "venice-uncensored", + "name": "Venice Uncensored", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "venice", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-10-01", + "last_updated": "2025-02-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, + "context": 128000, + "input": 128000, + "output": 8192 + }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.4, + "output": 1.8, + "cache_read": 0.4 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "Gemma-4-31B-MeroMero-v2:thinking": { + "id": "Gemma-4-31B-MeroMero-v2:thinking", + "name": "Gemma 4 31B MeroMero v2 Thinking", + "description": "Gemma 4 31B MeroMero v2 with thinking enabled for more deliberate emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-08-24", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 0.1, + "output": 0.45, + "cache_read": 0.05 } }, - "anthropic/claude-3-haiku": { - "id": "anthropic/claude-3-haiku", - "name": "Claude Haiku 3", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-haiku", + "gemini-2.0-pro-reasoner": { + "id": "gemini-2.0-pro-reasoner", + "name": "Gemini 2.0 Pro Reasoner", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-03-13", - "last_updated": "2024-03-13", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-02-05", "modalities": { "input": [ "text", - "image" + "audio" ], "output": [ "text" @@ -14877,86 +15774,61 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 4096 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.03, - "cache_write": 0.3 + "input": 1.292, + "output": 4.998, + "cache_read": 0.323 } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "Meta-Llama-3-1-8B-Instruct-FP8": { + "id": "Meta-Llama-3-1-8B-Instruct-FP8", + "name": "Llama 3.1 8B (decentralized)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.02, + "output": 0.03, + "cache_read": 0.01 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "glm-4-plus-0111": { + "id": "glm-4-plus-0111", + "name": "GLM 4 Plus 0111", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "glm", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -14964,110 +15836,92 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 9.996, + "output": 9.996, + "cache_read": 4.998 } }, - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "holo3-35b-a3b:thinking": { + "id": "holo3-35b-a3b:thinking", + "name": "Holo3-35B-A3B Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 65536, + "input": 65536, + "output": 8192 }, - "status": "deprecated", "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.25, + "output": 1.8, + "cache_read": 0.125 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "gemma-4-26b-a4b-it-moonlight": { + "id": "gemma-4-26b-a4b-it-moonlight", + "name": "Moonlight Dusk", + "description": "Moonlight Dusk is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.12, + "output": 0.38, + "cache_read": 0.06 } }, - "cohere/rerank-v3.5": { - "id": "cohere/rerank-v3.5", - "name": "Cohere Rerank 3.5", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "o", + "phi-4-mini-instruct": { + "id": "phi-4-mini-instruct", + "name": "Phi 4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "phi", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-12-02", - "last_updated": "2024-12-02", + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text" @@ -15076,24 +15930,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4096, - "output": 4096 + "context": 128000, + "input": 128000, + "output": 16384 + }, + "cost": { + "input": 0.17, + "output": 0.68, + "cache_read": 0.085 } }, - "cohere/command-a": { - "id": "cohere/command-a", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command", + "auto-model-premium": { + "id": "auto-model-premium", + "name": "Auto model (Premium)", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "structured_output": false, + "release_date": "2025-04-16", + "last_updated": "2024-06-01", "modalities": { "input": [ "text" @@ -15104,94 +15963,54 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 8000 + "context": 1000000, + "input": 1000000, + "output": 1000000 }, "cost": { - "input": 2.5, - "output": 10 - } - }, - "cohere/rerank-v4-fast": { - "id": "cohere/rerank-v4-fast", - "name": "Cohere Rerank 4 Fast", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "o", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 + "input": 9.996, + "output": 19.992, + "cache_read": 4.998 } }, - "cohere/embed-v4.0": { - "id": "cohere/embed-v4.0", - "name": "Embed v4.0", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "gemma-4-31b-it-darkidol": { + "id": "gemma-4-31b-it-darkidol", + "name": "DarkIdol", + "description": "DarkIdol is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 1536 - } - }, - "cohere/rerank-v4-pro": { - "id": "cohere/rerank-v4-pro", - "name": "Cohere Rerank 4 Pro", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "o", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 262144, + "input": 262144, + "output": 32768 }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 + "cost": { + "input": 0.1, + "output": 0.45, + "cache_read": 0.05 } }, - "stepfun/step-3.7-flash": { - "id": "stepfun/step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "family": "step", + "gemini-2.5-flash-lite-preview-06-17": { + "id": "gemini-2.5-flash-lite-preview-06-17", + "name": "Gemini 2.5 Flash Lite Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -15204,15 +16023,15 @@ ] } ], - "tool_call": true, - "temperature": true, - "knowledge": "2026-01-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-29", + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ "text" @@ -15220,22 +16039,22 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "stepfun/step-3.5-flash": { - "id": "stepfun/step-3.5-flash", - "name": "StepFun 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", - "family": "step", - "attachment": false, + "gemini-2.5-flash-preview-09-2025": { + "id": "gemini-2.5-flash-preview-09-2025", + "name": "Gemini 2.5 Flash Preview (09/2025)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -15248,13 +16067,15 @@ } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", + "structured_output": true, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "pdf" ], "output": [ "text" @@ -15262,520 +16083,585 @@ }, "open_weights": false, "limit": { - "context": 262114, - "output": 262114 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.09, - "output": 0.3, - "cache_read": 0.02 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "interfaze/interfaze-beta": { - "id": "interfaze/interfaze-beta", - "name": "Interfaze Beta", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": false, + "gemma-4-26b-a4b-it-opusdistill": { + "id": "gemma-4-26b-a4b-it-opusdistill", + "name": "Opus Distill", + "description": "Opus Distill is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "temperature": true, - "release_date": "2025-10-07", - "last_updated": "2026-04-29", + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32000 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 3.5 + "input": 0.12, + "output": 0.38, + "cache_read": 0.06 } }, - "bfl/flux-kontext-max": { - "id": "bfl/flux-kontext-max", - "name": "FLUX.1 Kontext Max", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", + "auto-model-standard": { + "id": "auto-model-standard", + "name": "Auto model (Standard)", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-05-29", - "last_updated": "2025-06", + "tool_call": true, + "structured_output": false, + "release_date": "2025-04-16", + "last_updated": "2024-06-01", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 512, - "output": 0 + "context": 1000000, + "input": 1000000, + "output": 1000000 + }, + "cost": { + "input": 9.996, + "output": 19.992, + "cache_read": 4.998 } }, - "bfl/flux-2-flex": { - "id": "bfl/flux-2-flex", - "name": "FLUX.2 [flex]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", + "longcat-2.0": { + "id": "longcat-2.0", + "name": "LongCat 2.0", + "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", + "family": "longcat", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-11-25", - "last_updated": "2026-06-08", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1048756, + "input": 1048756, + "output": 262144 + }, + "cost": { + "input": 0.75, + "output": 3, + "cache_read": 0.015 } }, - "bfl/flux-pro-1.1-ultra": { - "id": "bfl/flux-pro-1.1-ultra", - "name": "FLUX1.1 [pro] Ultra", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, + "gemini-2.5-pro-preview-03-25": { + "id": "gemini-2.5-pro-preview-03-25", + "name": "Gemini 2.5 Pro Preview 0325", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": false, - "temperature": true, - "release_date": "2024-11-01", - "last_updated": "2024-11", + "structured_output": false, + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 512, - "output": 0 + "context": 1048576, + "input": 1048576, + "output": 65536 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 0.25 } }, - "bfl/flux-2-max": { - "id": "bfl/flux-2-max", - "name": "FLUX.2 [max]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2026-06-08", + "agnes-3.0-flash": { + "id": "agnes-3.0-flash", + "name": "Agnes 3.0 Flash", + "description": "Agnes 3.0 Flash is a low-cost model for coding, tool use, and multi-turn agent tasks. It supports text and image input, optional thinking, and a 512K-token context window.", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "release_date": "2026-09-09", + "last_updated": "2026-09-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 67300, - "output": 67300 + "context": 524288, + "input": 524288, + "output": 65536 + }, + "cost": { + "input": 0.05, + "output": 0.15, + "cache_read": 0.005 } }, - "bfl/flux-pro-1.1": { - "id": "bfl/flux-pro-1.1", - "name": "FLUX1.1 [pro]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-10-02", - "last_updated": "2024-10", + "Gemma-4-31B-Queen": { + "id": "Gemma-4-31B-Queen", + "name": "Gemma 4 31B Queen", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 512, - "output": 0 + "context": 262144, + "input": 262144, + "output": 16384 + }, + "cost": { + "input": 0.306, + "output": 0.306, + "cache_read": 0.153 } }, - "bfl/flux-pro-1.0-fill": { - "id": "bfl/flux-pro-1.0-fill", - "name": "FLUX.1 Fill [pro]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", + "auto-model": { + "id": "auto-model", + "name": "Auto model", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-10-01", - "last_updated": "2024-10", + "tool_call": true, + "structured_output": false, + "release_date": "2025-04-16", + "last_updated": "2024-06-01", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 512, + "context": 1000000, + "input": 1000000, + "output": 1000000 + }, + "cost": { + "input": 0, "output": 0 } }, - "bfl/flux-2-klein-4b": { - "id": "bfl/flux-2-klein-4b", - "name": "FLUX.2 [klein] 4B", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, + "qvq-max": { + "id": "qvq-max", + "name": "Qwen: QvQ Max", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "qwen", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-01-15", - "last_updated": "2026-06-08", + "structured_output": false, + "release_date": "2025-03-28", + "last_updated": "2025-03-28", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 8192 + }, + "cost": { + "input": 1.2, + "output": 4.8, + "cache_read": 0.6 } }, - "bfl/flux-2-klein-9b": { - "id": "bfl/flux-2-klein-9b", - "name": "FLUX.2 [klein] 9B", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, + "asi1-mini": { + "id": "asi1-mini", + "name": "ASI1 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-01-15", - "last_updated": "2026-06-08", + "structured_output": false, + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 16384 + }, + "cost": { + "input": 1, + "output": 1, + "cache_read": 0.5 } }, - "bfl/flux-kontext-pro": { - "id": "bfl/flux-kontext-pro", - "name": "FLUX.1 Kontext Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, + "gemini-2.5-pro-exp-03-25": { + "id": "gemini-2.5-pro-exp-03-25", + "name": "Gemini 2.5 Pro Experimental 0325", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": false, - "temperature": true, - "release_date": "2025-05-29", - "last_updated": "2025-06", + "structured_output": false, + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 512, - "output": 0 + "context": 1048576, + "input": 1048576, + "output": 65536 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 0.25 } }, - "bfl/flux-2-pro": { - "id": "bfl/flux-2-pro", - "name": "FLUX.2 [pro]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-11-25", - "last_updated": "2026-06-08", + "gemma-4-31b-it-isometry": { + "id": "gemma-4-31b-it-isometry", + "name": "Isometry", + "description": "Isometry is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 67300, - "output": 67300 + "context": 262144, + "input": 262144, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.45, + "cache_read": 0.05 } }, - "recraft/recraft-v4.1-pro": { - "id": "recraft/recraft-v4.1-pro", - "name": "Recraft V4.1 Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", + "celeris-1": { + "id": "celeris-1", + "name": "Celeris 1", + "description": "Celeris 1 is a diffusion language model built for ultra-low-latency classification, extraction, judging, query rewriting, and other short structured responses.", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-05-14", - "last_updated": "2026-05-14", + "structured_output": false, + "release_date": "2026-07-25", + "last_updated": "2026-07-25", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 8192, + "input": 8192, + "output": 8192 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 1 } }, - "recraft/recraft-v4.1": { - "id": "recraft/recraft-v4.1", - "name": "Recraft V4.1", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, + "gemini-3-pro-image-preview": { + "id": "gemini-3-pro-image-preview", + "name": "Gemini 3 Pro Image", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, "reasoning": false, "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-05-14", - "last_updated": "2026-05-14", + "knowledge": "2025-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 65536, + "input": 65536, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "recraft/recraft-v4": { - "id": "recraft/recraft-v4", - "name": "Recraft V4", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "gemma-4-26b-a4b-it-darksoul": { + "id": "gemma-4-26b-a4b-it-darksoul", + "name": "Dark Soul", + "description": "Dark Soul is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "input": 262144, + "output": 32768 + }, + "cost": { + "input": 0.12, + "output": 0.38, + "cache_read": 0.06 } }, - "recraft/recraft-v4-pro": { - "id": "recraft/recraft-v4-pro", - "name": "Recraft V4 Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, + "glm-4.1v-thinking-flash": { + "id": "glm-4.1v-thinking-flash", + "name": "GLM 4.1V Thinking Flash", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "glm-flash", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "structured_output": false, + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 64000, + "input": 64000, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 0.3, + "cache_read": 0.15 } }, - "recraft/recraft-v2": { - "id": "recraft/recraft-v2", - "name": "Recraft V2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-03-13", - "last_updated": "2024-03", + "Gemma-4-31B-GarnetV2": { + "id": "Gemma-4-31B-GarnetV2", + "name": "Gemma 4 31B Garnet V2", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 512, - "output": 0 - } - }, - "recraft/recraft-v3": { - "id": "recraft/recraft-v3", - "name": "Recraft V3", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-10-30", - "last_updated": "2024-10", - "modalities": { - "input": [ "text" - ], - "output": [ - "image" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 512, - "output": 0 - } - }, - "recraft/recraft-v4.1-utility-pro": { - "id": "recraft/recraft-v4.1-utility-pro", - "name": "Recraft V4.1 Utility Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-05-14", - "last_updated": "2026-05-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] + "context": 262144, + "input": 262144, + "output": 16384 }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "cost": { + "input": 0.306, + "output": 0.306, + "cache_read": 0.153 } }, - "recraft/recraft-v4.1-utility": { - "id": "recraft/recraft-v4.1-utility", - "name": "Recraft V4.1 Utility", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, + "gemini-2.5-flash-nothinking": { + "id": "gemini-2.5-flash-nothinking", + "name": "Gemini 2.5 Flash (No Thinking)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-05-14", - "last_updated": "2026-05-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "arcee-ai/trinity-large-preview": { - "id": "arcee-ai/trinity-large-preview", - "name": "Trinity Large Preview", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "trinity", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-01-27", - "last_updated": "2025-01", + "structured_output": false, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -15783,29 +16669,32 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "arcee-ai/trinity-large-thinking": { - "id": "arcee-ai/trinity-large-thinking", - "name": "Trinity Large Thinking", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "trinity", - "attachment": false, + "Gemma-4-31B-MeroMero-v2": { + "id": "Gemma-4-31B-MeroMero-v2", + "name": "Gemma 4 31B MeroMero v2", + "description": "Gemma 4 31B MeroMero v2 is a LoRA finetune for emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-03", + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-08-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -15813,67 +16702,72 @@ }, "open_weights": true, "limit": { - "context": 262100, - "output": 80000 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 0.8999999999999999 + "input": 0.1, + "output": 0.45, + "cache_read": 0.05 } }, - "arcee-ai/trinity-mini": { - "id": "arcee-ai/trinity-mini", - "name": "Trinity Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "trinity", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-01", - "last_updated": "2025-12", + "holo3-35b-a3b": { + "id": "holo3-35b-a3b", + "name": "Holo3-35B-A3B", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 65536, + "input": 65536, + "output": 8192 }, "cost": { - "input": 0.045, - "output": 0.15 + "input": 0.25, + "output": 1.8, + "cache_read": 0.125 } }, - "perplexity/sonar-reasoning-pro": { - "id": "perplexity/sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded reasoning model for multi-step research and cited answers", - "family": "sonar-reasoning", + "claw-medium": { + "id": "claw-medium", + "name": "Claw Medium", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", - "medium", - "high" + "high", + "max" ] } ], - "tool_call": false, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "tool_call": true, + "structured_output": true, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ "text" @@ -15884,22 +16778,28 @@ }, "open_weights": false, "limit": { - "context": 127000, - "output": 8000 + "context": 1048576, + "input": 1048576, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "perplexity/sonar": { - "id": "perplexity/sonar", - "name": "Sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar", + "gemma-4-26b-a4b-it-shadowsiren": { + "id": "gemma-4-26b-a4b-it-shadowsiren", + "name": "Shadow Siren", + "description": "Shadow Siren is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-02", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ "text", @@ -15909,28 +16809,34 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 127000, - "output": 8000 + "context": 262144, + "input": 262144, + "output": 32768 + }, + "cost": { + "input": 0.12, + "output": 0.38, + "cache_read": 0.06 } }, - "perplexity/sonar-pro": { - "id": "perplexity/sonar-pro", - "name": "Sonar Pro", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", + "gemini-2.0-pro-exp-02-05": { + "id": "gemini-2.0-pro-exp-02-05", + "name": "Gemini 2.0 Pro 0205", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-05", + "last_updated": "2025-02-05", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ "text" @@ -15938,65 +16844,39 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8000 - } - }, - "amazon/titan-embed-text-v2": { - "id": "amazon/titan-embed-text-v2", - "name": "Titan Text Embeddings V2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "titan-embed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-04-30", - "last_updated": "2024-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 2097152, + "input": 2097152, + "output": 8192 }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 1536 + "cost": { + "input": 1.989, + "output": 7.956, + "cache_read": 0.49725 } }, - "amazon/nova-2-lite": { - "id": "amazon/nova-2-lite", - "name": "Nova 2 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova", - "attachment": true, + "claw-high": { + "id": "claw-high", + "name": "Claw High", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], - "tool_call": false, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-02", - "last_updated": "2024-12-01", + "tool_call": true, + "structured_output": true, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -16004,163 +16884,134 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "amazon/nova-lite": { - "id": "amazon/nova-lite", - "name": "Nova Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", + "deepseek-chat": { + "id": "deepseek-chat", + "name": "DeepSeek V3/Deepseek Chat", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 8192 - }, - "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.015 - } - }, - "amazon/nova-micro": { - "id": "amazon/nova-micro", - "name": "Nova Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text" + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, + "input": 128000, "output": 8192 }, "cost": { - "input": 0.035, - "output": 0.14, - "cache_read": 0.00875 + "input": 0.1, + "output": 0.425, + "cache_read": 0.05 } }, - "amazon/nova-pro": { - "id": "amazon/nova-pro", - "name": "Nova Pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", + "Gemma-4-26B-A4B-MeroMero:thinking": { + "id": "Gemma-4-26B-A4B-MeroMero:thinking", + "name": "Gemma 4 26B A4B MeroMero Thinking", + "description": "Gemma 4 26B A4B MeroMero with thinking enabled for more deliberate emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 300000, - "output": 8192 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.8, - "output": 3.2, - "cache_read": 0.2 + "input": 0.12, + "output": 0.38, + "cache_read": 0.06 } }, - "alibaba/qwen3-vl-thinking": { - "id": "alibaba/qwen3-vl-thinking", - "name": "Qwen3 VL Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "claw-low": { + "id": "claw-low", + "name": "Claw Low", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-23", - "last_updated": "2025-09-24", + "structured_output": true, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.4, - "output": 4 + "input": 1, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0.08333 } }, - "alibaba/qwen3-coder-plus": { - "id": "alibaba/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", + "doubao-1.5-pro-32k": { + "id": "doubao-1.5-pro-32k", + "name": "Doubao 1.5 Pro 32k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "tool_call": false, + "structured_output": false, + "release_date": "2024-11-20", + "last_updated": "2025-01-22", "modalities": { "input": [ "text" @@ -16169,53 +17020,71 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 32000, + "input": 32000, + "output": 8192 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.2 + "input": 0.1343, + "output": 0.3349, + "cache_read": 0.06715 } }, - "alibaba/wan-v2.6-r2v": { - "id": "alibaba/wan-v2.6-r2v", - "name": "Wan v2.6 Reference-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "o", + "doubao-seed-2-0-code-preview-260215": { + "id": "doubao-seed-2-0-code-preview-260215", + "name": "Doubao Seed 2.0 Code Preview", + "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", + "family": "seed", "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "input": 256000, + "output": 128000 + }, + "cost": { + "input": 0.782, + "output": 3.893, + "cache_read": 0.391 } }, - "alibaba/qwen3-embedding-0.6b": { - "id": "alibaba/qwen3-embedding-0.6b", - "name": "Qwen3 Embedding 0.6B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", + "hermes-low": { + "id": "hermes-low", + "name": "Hermes Low", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "hermes", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ "text" @@ -16226,52 +17095,61 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 1048576, + "input": 1048576, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0.08333 } }, - "alibaba/qwen3-max-preview": { - "id": "alibaba/qwen3-max-preview", - "name": "Qwen3 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gemma-4-31b-it-gemsicle": { + "id": "gemma-4-31b-it-gemsicle", + "name": "Gemsicle", + "description": "Gemsicle is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-05", - "last_updated": "2025-09-23", + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, + "input": 262144, "output": 32768 }, "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24 + "input": 0.1, + "output": 0.45, + "cache_read": 0.05 } }, - "alibaba/qwen3-embedding-8b": { - "id": "alibaba/qwen3-embedding-8b", - "name": "Qwen3 Embedding 8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", + "mercury-coder-small": { + "id": "mercury-coder-small", + "name": "Mercury Coder Small", + "description": "Model by Inception AI. A diffusion large language model that runs incredibly quickly (500+ tokens/second) while matching Claude 3.5 Haiku and GPT-4o-mini. 1st in speed on Copilot arena, and matching 2nd in quality.", + "family": "mercury", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" @@ -16283,66 +17161,44 @@ "open_weights": false, "limit": { "context": 32768, - "output": 32768 - } - }, - "alibaba/qwen3-next-80b-a3b-instruct": { - "id": "alibaba/qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-11", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 + "input": 32768, + "output": 16384 }, "cost": { - "input": 0.15, - "output": 1.2 + "input": 0.25, + "output": 1, + "cache_read": 0.125 } }, - "alibaba/qwen3.7-plus": { - "id": "alibaba/qwen3.7-plus", - "name": "Qwen 3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen3.7-plus", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -16350,93 +17206,116 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "cache_write": 0.5 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "cache_write": 0.375 } }, - "alibaba/wan-v2.7-r2v": { - "id": "alibaba/wan-v2.7-r2v", - "name": "Wan v2.7 Reference-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "o", - "attachment": false, + "ernie-x1.1-preview": { + "id": "ernie-x1.1-preview", + "name": "ERNIE X1.1", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "ernie", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "structured_output": false, + "release_date": "2025-09-10", + "last_updated": "2025-09-10", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 64000, + "input": 64000, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "alibaba/qwen3-vl-instruct": { - "id": "alibaba/qwen3-vl-instruct", - "name": "Qwen3 VL Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-24", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 129024 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 1.6 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "alibaba/qwen3.7-max": { - "id": "alibaba/qwen3.7-max", - "name": "Qwen 3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": true, + "hermes-high": { + "id": "hermes-high", + "name": "Hermes High", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "hermes", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "structured_output": true, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ "text" @@ -16447,53 +17326,59 @@ }, "open_weights": false, "limit": { - "context": 991000, - "output": 64000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 3.75, - "cache_read": 0.25, - "cache_write": 1.5625 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "alibaba/wan-v2.5-t2v-preview": { - "id": "alibaba/wan-v2.5-t2v-preview", - "name": "Wan v2.5 Text-to-Video Preview", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "o", - "attachment": false, + "gemini-2.5-flash-preview-05-20": { + "id": "gemini-2.5-flash-preview-05-20", + "name": "Gemini 2.5 Flash 0520", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "structured_output": false, + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1048000, + "input": 1048000, + "output": 65536 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "alibaba/qwen3-max": { - "id": "alibaba/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", + "glm-4-long": { + "id": "glm-4-long", + "name": "GLM-4 Long", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "glm", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-08-01", "modalities": { "input": [ "text" @@ -16504,33 +17389,27 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 4096 }, "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24 + "input": 0.2006, + "output": 0.2006, + "cache_read": 0.1003 } }, - "alibaba/qwen3-next-80b-a3b-thinking": { - "id": "alibaba/qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B Thinking", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", + "pokee-isaac": { + "id": "pokee-isaac", + "name": "Pokee-Isaac 28B", + "description": "Pokee-Isaac is a 28B agentic model with a roughly 10-million-token context window, function calling, and OpenAI-compatible structured output. Pokee bills in $0.01 increments, rounding each non-zero request up to the next cent.", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1 - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-11", - "last_updated": "2025-09", + "structured_output": true, + "release_date": "2026-08-04", + "last_updated": "2026-08-04", "modalities": { "input": [ "text" @@ -16539,27 +17418,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 10000000, + "input": 10000000, + "output": 60000 }, "cost": { "input": 0.15, - "output": 1.2 + "output": 1, + "cache_read": 0.075 } }, - "alibaba/qwen3-embedding-4b": { - "id": "alibaba/qwen3-embedding-4b", - "name": "Qwen3 Embedding 4B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", + "MiniMax-M1": { + "id": "MiniMax-M1", + "name": "MiniMax M1", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "minimax", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "structured_output": false, + "release_date": "2025-01-08", + "last_updated": "2025-06-16", "modalities": { "input": [ "text" @@ -16568,209 +17449,286 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.1394, + "output": 1.3328, + "cache_read": 0.0697 } }, - "alibaba/qwen3.5-flash": { - "id": "alibaba/qwen3.5-flash", - "name": "Qwen 3.5 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled": { + "id": "Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled", + "name": "Gemma 4 31B Claude 4.6 Opus Reasoning Distilled", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "claude", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "structured_output": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.001, - "cache_write": 0.125 + "input": 0.306, + "output": 0.306, + "cache_read": 0.0306 } }, - "alibaba/qwen3-coder": { - "id": "alibaba/qwen3-coder", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "Qwen3.5-27B-BlueStar-v3-Derestricted": { + "id": "Qwen3.5-27B-BlueStar-v3-Derestricted", + "name": "Qwen3.5 27B BlueStar v3 Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "qwen3.5", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-22", - "last_updated": "2025-04", + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "input": 262144, + "output": 16384 }, "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.3 + "input": 0.306, + "output": 0.306, + "cache_read": 0.153 } }, - "alibaba/qwen-3-235b": { - "id": "alibaba/qwen-3-235b", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "gemma-4-31b-it-novelist": { + "id": "gemma-4-31b-it-novelist", + "name": "Novelist", + "description": "Novelist is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-04", + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "output": 16384 + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.22, - "output": 0.88 - } + "input": 0.1, + "output": 0.45, + "cache_read": 0.05 + } }, - "alibaba/qwen3.5-plus": { - "id": "alibaba/qwen3.5-plus", - "name": "Qwen 3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "ernie-5.1:thinking": { + "id": "ernie-5.1:thinking", + "name": "ERNIE 5.1 Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "ernie", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-10", + "last_updated": "2026-05-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 119000, + "input": 119000, + "output": 64000 + }, + "cost": { + "input": 0.75, + "output": 3, + "cache_read": 0.75 + } + }, + "gemma-4-12b-it-semancer": { + "id": "gemma-4-12b-it-semancer", + "name": "Gemma 4 12B Semancer", + "description": "Gemma 4 12B Semancer is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 131,072-token context window.", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 81920 - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "structured_output": false, + "release_date": "2026-09-09", + "last_updated": "2026-09-09", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.04, - "cache_write": 0.5 + "input": 0.05, + "output": 0.25, + "cache_read": 0.025 } }, - "alibaba/wan-v2.6-t2v": { - "id": "alibaba/wan-v2.6-t2v", - "name": "Wan v2.6 Text-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "o", + "mistral-code-latest": { + "id": "mistral-code-latest", + "name": "Mistral Code Latest", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "mistral", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "tool_call": true, + "structured_output": true, + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "input": 256000, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 0.9, + "cache_read": 0.15 } }, - "alibaba/qwen3-max-thinking": { - "id": "alibaba/qwen3-max-thinking", - "name": "Qwen 3 Max Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "Qwen3.5-27B-Queen-Derestricted": { + "id": "Qwen3.5-27B-Queen-Derestricted", + "name": "Qwen3.5 27B Queen Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "qwen3.5", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-23", - "last_updated": "2025-01", + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "input": 262144, + "output": 16384 + }, + "cost": { + "input": 0.306, + "output": 0.306, + "cache_read": 0.153 + } + }, + "gemma-4-31b-it-fabled": { + "id": "gemma-4-31b-it-fabled", + "name": "Fabled", + "description": "Fabled is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" @@ -16778,80 +17736,170 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24 + "input": 0.1, + "output": 0.45, + "cache_read": 0.05 } }, - "alibaba/wan-v2.6-i2v-flash": { - "id": "alibaba/wan-v2.6-i2v-flash", - "name": "Wan v2.6 Image-to-Video Flash", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "o", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "Gemma-4-26B-A4B-MeroMero": { + "id": "Gemma-4-26B-A4B-MeroMero", + "name": "Gemma 4 26B A4B MeroMero", + "description": "Gemma 4 26B A4B MeroMero is an NVFP4 multimodal mixture-of-experts fine-tune for emotive dialogue, relationship scenes, creative writing, and roleplay.", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-08-26", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "input": 262144, + "output": 32768 + }, + "cost": { + "input": 0.12, + "output": 0.38, + "cache_read": 0.06 + } + }, + "Gemma-4-31B-Cognitive-Unshackled": { + "id": "Gemma-4-31B-Cognitive-Unshackled", + "name": "Gemma 4 31B Cognitive Unshackled", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", + "modalities": { + "input": [ + "text", + "image" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "input": 262144, + "output": 16384 + }, + "cost": { + "input": 0.306, + "output": 0.306, + "cache_read": 0.153 } }, - "alibaba/wan-v2.6-r2v-flash": { - "id": "alibaba/wan-v2.6-r2v-flash", - "name": "Wan v2.6 Reference-to-Video Flash", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "o", + "universal-summarizer": { + "id": "universal-summarizer", + "name": "Universal Summarizer", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "structured_output": false, + "release_date": "2025-12-23", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 32768, + "input": 32768, + "output": 32768 + }, + "cost": { + "input": 30, + "output": 30 } }, - "alibaba/qwen3-coder-next": { - "id": "alibaba/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "gemma-4-12b-it": { + "id": "gemma-4-12b-it", + "name": "Gemma 4 12B Instruct", + "description": "Compact Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-07-22", - "last_updated": "2026-02-19", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "input": 131072, + "output": 32768 + }, + "cost": { + "input": 0.05, + "output": 0.25, + "cache_read": 0.025 + } + }, + "gemini-exp-1206": { + "id": "gemini-exp-1206", + "name": "Gemini 2.0 Pro 1206", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text", + "image", + "audio" ], "output": [ "text" @@ -16859,40 +17907,42 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 2097152, + "input": 2097152, + "output": 8192 }, "cost": { - "input": 0.5, - "output": 1.2 + "input": 1.258, + "output": 4.998, + "cache_read": 0.629 } }, - "alibaba/qwen3.6-27b": { - "id": "alibaba/qwen3.6-27b", - "name": "Qwen 3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "gemini-2.5-flash-preview-09-2025-thinking": { + "id": "gemini-2.5-flash-preview-09-2025-thinking", + "name": "Gemini 2.5 Flash Preview (09/2025) – Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 131072 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", "image", + "audio", "pdf" ], "output": [ @@ -16901,84 +17951,105 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "alibaba/wan-v2.6-i2v": { - "id": "alibaba/wan-v2.6-i2v", - "name": "Wan v2.6 Image-to-Video", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "o", + "doubao-1.5-pro-256k": { + "id": "doubao-1.5-pro-256k", + "name": "Doubao 1.5 Pro 256k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "structured_output": false, + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "input": 256000, + "output": 16384 + }, + "cost": { + "input": 0.799, + "output": 1.445, + "cache_read": 0.3995 } }, - "alibaba/wan-v2.7-t2v": { - "id": "alibaba/wan-v2.7-t2v", - "name": "Wan v2.7 Text-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "o", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "gemma-4-31b-it-garnet": { + "id": "gemma-4-31b-it-garnet", + "name": "Garnet", + "description": "Garnet is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "input": 262144, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.45, + "cache_read": 0.05 } }, - "alibaba/qwen-3-30b": { - "id": "alibaba/qwen-3-30b", - "name": "Qwen3-30B-A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "gemini-2.5-pro-preview-05-06": { + "id": "gemini-2.5-pro-preview-05-06", + "name": "Gemini 2.5 Pro Preview 0506", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-04", + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-06", + "last_updated": "2025-05-06", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -16986,38 +18057,31 @@ }, "open_weights": false, "limit": { - "context": 40960, - "output": 16384 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.12, - "output": 0.5 + "input": 2.5, + "output": 10, + "cache_read": 0.25 } }, - "alibaba/qwen3-235b-a22b-thinking": { - "id": "alibaba/qwen3-235b-a22b-thinking", - "name": "Qwen3 235B A22B Thinking 2507", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "mercury-2": { + "id": "mercury-2", + "name": "Mercury 2", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "mercury", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1, - "max": 81920 - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-04", + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -17025,61 +18089,62 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 50000 }, "cost": { - "input": 0.4, - "output": 4 + "input": 0.25, + "output": 0.75, + "cache_read": 0.025 } }, - "alibaba/qwen3-vl-235b-a22b-instruct": { - "id": "alibaba/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "command-a-reasoning-08-2025": { + "id": "command-a-reasoning-08-2025", + "name": "Cohere Command A (08/2025)", + "description": "Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows", + "family": "command-a", + "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-09-23", - "last_updated": "2026-05-01", + "knowledge": "2024-06-01", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 129024 + "context": 256000, + "input": 256000, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 1.6 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "alibaba/qwen-3-14b": { - "id": "alibaba/qwen-3-14b", - "name": "Qwen3-14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "longcat-2.0:thinking": { + "id": "longcat-2.0:thinking", + "name": "LongCat 2.0 Thinking", + "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", + "family": "longcat", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-04", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text" @@ -17090,36 +18155,27 @@ }, "open_weights": false, "limit": { - "context": 40960, - "output": 16384 + "context": 1048756, + "input": 1048756, + "output": 262144 }, "cost": { - "input": 0.12, - "output": 0.24 + "input": 0.75, + "output": 3, + "cache_read": 0.015 } }, - "alibaba/qwen-3-32b": { - "id": "alibaba/qwen-3-32b", - "name": "Qwen 3.32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "deepseek-r1-sambanova": { + "id": "deepseek-r1-sambanova", + "name": "DeepSeek R1 Fast", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 38912 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-04", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-20", + "last_updated": "2025-02-20", "modalities": { "input": [ "text" @@ -17128,70 +18184,74 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 8192 + "input": 128000, + "output": 4096 }, "cost": { - "input": 0.16, - "output": 0.64 + "input": 4.998, + "output": 6.987, + "cache_read": 2.499 } }, - "alibaba/qwen-3.6-max-preview": { - "id": "alibaba/qwen-3.6-max-preview", - "name": "Qwen 3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "gemini-2.5-flash-preview-04-17": { + "id": "gemini-2.5-flash-preview-04-17", + "name": "Gemini 2.5 Flash Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 131072 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": true, - "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-24", + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-17", + "last_updated": "2025-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 240000, - "output": 64000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 1.3, - "output": 7.8, - "cache_read": 0.26, - "cache_write": 1.625 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "alibaba/qwen3-coder-30b-a3b": { - "id": "alibaba/qwen3-coder-30b-a3b", - "name": "Qwen 3 Coder 30B A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "deepseek-reasoner": { + "id": "deepseek-reasoner", + "name": "DeepSeek Reasoner", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-31", - "last_updated": "2025-04", + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ "text" @@ -17200,42 +18260,44 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 8192 + "context": 64000, + "input": 64000, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.4, + "output": 1.7, + "cache_read": 0.2 } }, - "alibaba/qwen3.6-plus": { - "id": "alibaba/qwen3.6-plus", - "name": "Qwen 3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", + "gemini-2.5-flash-lite-preview-09-2025": { + "id": "gemini-2.5-flash-lite-preview-09-2025", + "name": "Gemini 2.5 Flash Lite Preview (09/2025)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 131072 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "structured_output": true, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", "image", + "audio", "pdf" ], "output": [ @@ -17244,35 +18306,42 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1, - "cache_write": 0.625 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 } }, - "meituan/longcat-flash-thinking-2601": { - "id": "meituan/longcat-flash-thinking-2601", - "name": "LongCat Flash Thinking 2601", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "longcat", - "attachment": false, + "gemini-2.5-flash-preview-04-17:thinking": { + "id": "gemini-2.5-flash-preview-04-17:thinking", + "name": "Gemini 2.5 Flash Preview Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": false, - "temperature": true, - "release_date": "2026-01-15", - "last_updated": "2026-03-13", + "structured_output": false, + "release_date": "2025-04-17", + "last_updated": "2025-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -17280,22 +18349,30 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 1048576, + "input": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.15, + "output": 3.5, + "cache_read": 0.015 } }, - "meituan/longcat-flash-chat": { - "id": "meituan/longcat-flash-chat", - "name": "LongCat Flash Chat", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "longcat", + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek R1", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-01", - "last_updated": "2025-08-30", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" @@ -17304,40 +18381,44 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 100000 + "input": 128000, + "output": 8192 + }, + "cost": { + "input": 0.4, + "output": 1.7, + "cache_read": 0.2 } }, - "meta/muse-spark-1.1": { - "id": "meta/muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "muse", - "attachment": false, + "gemini-2.5-pro-preview-06-05": { + "id": "gemini-2.5-pro-preview-06-05", + "name": "Gemini 2.5 Pro Preview 0605", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -17346,29 +18427,44 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 1048576 + "input": 1048576, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 2.5, + "output": 10, + "cache_read": 0.25 } }, - "meta/llama-3.2-1b": { - "id": "meta/llama-3.2-1b", - "name": "Llama 3.2 1B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-18", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -17376,57 +18472,68 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { "input": 0.1, - "output": 0.1 + "output": 0.4, + "cache_read": 0.01 } }, - "meta/llama-3.2-11b": { - "id": "meta/llama-3.2-11b", - "name": "Llama 3.2 11B Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", + "deepseek-chat-cheaper": { + "id": "deepseek-chat-cheaper", + "name": "DeepSeek V3/Chat Cheaper", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "deepseek", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "structured_output": true, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, + "input": 128000, "output": 8192 }, "cost": { - "input": 0.16, - "output": 0.16 + "input": 0.1, + "output": 0.425, + "cache_read": 0.05 } }, - "meta/llama-3.1-8b": { - "id": "meta/llama-3.1-8b", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "GLM-4.6-Derestricted-v5": { + "id": "GLM-4.6-Derestricted-v5", + "name": "GLM 4.6 Derestricted v5", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "structured_output": false, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -17435,28 +18542,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 131072, + "input": 131072, "output": 8192 }, "cost": { - "input": 0.22, - "output": 0.22 + "input": 0.4, + "output": 1.5, + "cache_read": 0.2 } }, - "meta/llama-3.2-90b": { - "id": "meta/llama-3.2-90b", - "name": "Llama 3.2 90B Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", + "doubao-1.5-vision-pro-32k": { + "id": "doubao-1.5-vision-pro-32k", + "name": "Doubao 1.5 Vision Pro 32k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "tool_call": false, + "structured_output": false, + "release_date": "2024-11-20", + "last_updated": "2025-01-22", "modalities": { "input": [ "text", @@ -17468,26 +18575,27 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 32000, + "input": 32000, "output": 8192 }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 0.459, + "output": 1.377, + "cache_read": 0.2295 } }, - "meta/llama-3.1-70b": { - "id": "meta/llama-3.1-70b", - "name": "Llama 3.1 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "glm-z1-airx": { + "id": "glm-z1-airx", + "name": "GLM Z1 AirX", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "glm", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "structured_output": true, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text" @@ -17498,26 +18606,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 32000, + "input": 32000, + "output": 16384 }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 0.7, + "output": 0.7, + "cache_read": 0.35 } }, - "meta/llama-3.2-3b": { - "id": "meta/llama-3.2-3b", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "auto-model-basic": { + "id": "auto-model-basic", + "name": "Auto model (Basic)", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-18", + "tool_call": true, + "structured_output": false, + "release_date": "2025-04-16", + "last_updated": "2024-06-01", "modalities": { "input": [ "text" @@ -17528,26 +18637,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "input": 1000000, + "output": 1000000 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 9.996, + "output": 19.992, + "cache_read": 4.998 } }, - "meta/llama-4-scout": { - "id": "meta/llama-4-scout", - "name": "Llama-4-Scout-17B-16E-Instruct-FP8", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + "qwen3-vl-235b-a22b-instruct-original": { + "id": "qwen3-vl-235b-a22b-instruct-original", + "name": "Qwen3 VL 235B A22B Instruct Original", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", @@ -17557,28 +18666,72 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.9, + "cache_read": 0.15 } }, - "meta/llama-3.3-70b": { - "id": "meta/llama-3.3-70b", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "gemma-4-26b-a4b-it-luminous": { + "id": "gemma-4-26b-a4b-it-luminous", + "name": "Luminous Mirror", + "description": "Luminous Mirror is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", + "family": "gemma", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "input": 262144, + "output": 32768 + }, + "cost": { + "input": 0.12, + "output": 0.38, + "cache_read": 0.06 + } + }, + "sarvam-105b": { + "id": "sarvam-105b", + "name": "Sarvam 105B", + "description": "Flagship Indian-language reasoning model for enterprise multilingual applications", + "family": "sarvam", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2025-09-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text" @@ -17589,26 +18742,28 @@ }, "open_weights": true, "limit": { - "context": 128000, + "context": 131072, + "input": 131072, "output": 4096 }, "cost": { - "input": 0, - "output": 0 + "input": 0.054, + "output": 0.2124, + "cache_read": 0.0336 } }, - "meta/llama-4-maverick": { - "id": "meta/llama-4-maverick", - "name": "Llama-4-Maverick-17B-128E-Instruct-FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "gemma-4-31b-it-gembrain": { + "id": "gemma-4-31b-it-gembrain", + "name": "Gembrain", + "description": "Gembrain is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ "text", @@ -17620,78 +18775,105 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.45, + "cache_read": 0.05 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek", + "phi-4-multimodal-instruct": { + "id": "phi-4-multimodal-instruct", + "name": "Phi 4 Multimodal", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "phi", "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "input": 128000, + "output": 16384 + }, + "cost": { + "input": 0.07, + "output": 0.11, + "cache_read": 0.035 + } + }, + "gemini-2.5-flash-preview-05-20:thinking": { + "id": "gemini-2.5-flash-preview-05-20:thinking", + "name": "Gemini 2.5 Flash 0520 Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemini", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "xhigh" + "low", + "medium", + "high" ] } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-23", - "last_updated": "2026-04-24", + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048000, + "input": 1048000, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.15, + "output": 3.5, + "cache_read": 0.015 } }, - "deepseek/deepseek-v3.1-terminus": { - "id": "deepseek/deepseek-v3.1-terminus", - "name": "DeepSeek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "gemma-4-26b-a4b-it-musica": { + "id": "gemma-4-26b-a4b-it-musica", + "name": "Musica", + "description": "Musica is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -17699,71 +18881,70 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.27, - "output": 1, - "cache_read": 0.135 + "input": 0.12, + "output": 0.38, + "cache_read": 0.06 } }, - "deepseek/deepseek-v3": { - "id": "deepseek/deepseek-v3", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "gemma-4-26b-a4b-it-chimerax": { + "id": "gemma-4-26b-a4b-it-chimerax", + "name": "Chimera X", + "description": "Chimera X is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-12-26", - "last_updated": "2024-12-26", + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.27, - "output": 1.12, - "cache_read": 0.135 + "input": 0.12, + "output": 0.38, + "cache_read": 0.06 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek", + "hermes-medium": { + "id": "hermes-medium", + "name": "Hermes Medium", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "hermes", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", "high", - "xhigh" + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-23", - "last_updated": "2026-04-24", + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ "text" @@ -17772,31 +18953,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "deepseek/deepseek-v3.2-thinking": { - "id": "deepseek/deepseek-v3.2-thinking", - "name": "DeepSeek V3.2 Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek-thinking", + "doubao-seed-1-6-250615": { + "id": "doubao-seed-1-6-250615", + "name": "Doubao Seed 1.6", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "seed", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-06-15", "modalities": { "input": [ "text" @@ -17807,61 +18986,68 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8000 + "context": 256000, + "input": 256000, + "output": 16384 }, "cost": { - "input": 0.62, - "output": 1.85 + "input": 0.204, + "output": 0.51, + "cache_read": 0.102 } }, - "deepseek/deepseek-v3.1": { - "id": "deepseek/deepseek-v3.1", - "name": "DeepSeek-V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "Gemma-4-31B-DarkIdol": { + "id": "Gemma-4-31B-DarkIdol", + "name": "Gemma 4 31B DarkIdol", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "structured_output": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 1.7 + "input": 0.306, + "output": 0.306, + "cache_read": 0.153 } }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "ernie-5.1": { + "id": "ernie-5.1", + "name": "ERNIE 5.1", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "ernie", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "structured_output": false, + "release_date": "2026-05-10", + "last_updated": "2026-05-10", "modalities": { "input": [ "text" @@ -17872,28 +19058,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8000 + "context": 119000, + "input": 119000, + "output": 64000 }, "cost": { - "input": 0.28, - "output": 0.42, - "cache_read": 0.028 + "input": 0.75, + "output": 3, + "cache_read": 0.75 } }, - "deepseek/deepseek-r1": { - "id": "deepseek/deepseek-r1", - "name": "DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "doubao-seed-1-6-flash-250615": { + "id": "doubao-seed-1-6-flash-250615", + "name": "Doubao Seed 1.6 Flash", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "seed", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-06-15", "modalities": { "input": [ "text" @@ -17904,26 +19089,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 256000, + "input": 256000, + "output": 16384 }, "cost": { - "input": 1.35, - "output": 5.4 + "input": 0.0374, + "output": 0.374, + "cache_read": 0.0187 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax M2.7 High Speed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistral-code-agent-latest": { + "id": "mistral-code-agent-latest", + "name": "Mistral Code Agent Latest", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "mistral", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": true, + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text" @@ -17934,62 +19120,59 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131100 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.4, + "output": 2, + "cache_read": 0.2 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "kimi-k2-instruct-fast": { + "id": "kimi-k2-instruct-fast", + "name": "Kimi K2 0711 Fast", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "kimi-k2", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-15", + "last_updated": "2025-07-15", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 131000 + "context": 131072, + "input": 131072, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.4, + "output": 1.8, + "cache_read": 0.2 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", + "featherless-ai/Qwerky-72B": { + "id": "featherless-ai/Qwerky-72B", + "name": "Qwerky 72B", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "qwerky", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-03-20", "modalities": { "input": [ "text" @@ -17998,38 +19181,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 32000, + "input": 32000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.5, + "output": 0.5, + "cache_read": 0.25 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax-m3", + "unsloth/gemma-3-27b-it": { + "id": "unsloth/gemma-3-27b-it", + "name": "Gemma 3 27B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "unsloth", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-05-31", - "last_updated": "2026-06-01", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-03-10", "modalities": { "input": [ "text", - "image", "pdf" ], "output": [ @@ -18038,32 +19215,31 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 128000, + "input": 128000, + "output": 96000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.2992, + "output": 0.2992, + "cache_read": 0.1496 } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "unsloth/gemma-3-12b-it": { + "id": "unsloth/gemma-3-12b-it", + "name": "Gemma 3 12B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "unsloth", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-03-10", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -18071,31 +19247,31 @@ }, "open_weights": true, "limit": { - "context": 205000, - "output": 205000 + "context": 131072, + "input": 131072, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.272, + "output": 0.272, + "cache_read": 0.136 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "Minimax M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "unsloth/gemma-3-4b-it": { + "id": "unsloth/gemma-3-4b-it", + "name": "Gemma 3 4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "unsloth", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-03-10", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -18103,31 +19279,33 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131000 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.2006, + "output": 0.2006, + "cache_read": 0.1003 } }, - "minimax/minimax-m2.5-highspeed": { - "id": "minimax/minimax-m2.5-highspeed", - "name": "MiniMax M2.5 High Speed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "bytedance-seed/seed-2.0-lite": { + "id": "bytedance-seed/seed-2.0-lite", + "name": "ByteDance Seed 2.0 Lite", + "description": "Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation", + "family": "seed", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-13", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -18135,32 +19313,33 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 131000 + "context": 262144, + "input": 262144, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.25, + "output": 2, + "cache_read": 0.125 } }, - "minimax/minimax-m2.1-lightning": { - "id": "minimax/minimax-m2.1-lightning", - "name": "MiniMax M2.1 Lightning", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, + "bytedance-seed/seed-2-1-turbo": { + "id": "bytedance-seed/seed-2-1-turbo", + "name": "ByteDance Seed 2.1 Turbo", + "description": "ByteDance Seed 2.1 Turbo is a multimodal model for coding and long-horizon agent workflows, including end-to-end software delivery and multi-step task execution. It supports text, image, and video input with a 262k context window.", + "family": "seed", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-23", - "last_updated": "2025-10-27", + "structured_output": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -18168,32 +19347,44 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 235929 }, "cost": { - "input": 0.3, - "output": 2.4, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.5, + "output": 2.5, + "cache_read": 0.25 } }, - "kwaipilot/kat-coder-pro-v1": { - "id": "kwaipilot/kat-coder-pro-v1", - "name": "KAT-Coder-Pro V1", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "kat-coder", - "attachment": false, + "bytedance-seed/seed-2.0-code": { + "id": "bytedance-seed/seed-2.0-code", + "name": "ByteDance Seed 2.0 Code", + "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-11-09", - "last_updated": "2025-10-24", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -18201,27 +19392,29 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 262144, + "input": 262144, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.5, + "output": 3, + "cache_read": 0.25 } }, - "kwaipilot/kat-coder-pro-v2": { - "id": "kwaipilot/kat-coder-pro-v2", - "name": "Kat Coder Pro V2", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "kat-coder", + "poolside/laguna-s-2.1:thinking": { + "id": "poolside/laguna-s-2.1:thinking", + "name": "Laguna S 2.1 Thinking", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-30", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text" @@ -18230,41 +19423,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.1, + "output": 0.2, + "cache_read": 0.01 } - } - } - }, - "submodel": { - "id": "submodel", - "env": [ - "SUBMODEL_INSTAGEN_ACCESS_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://llm.submodel.ai/v1", - "name": "submodel", - "doc": "https://submodel.gitbook.io", - "models": { - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen3 235B A22B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + }, + "poolside/laguna-s-2.1": { + "id": "poolside/laguna-s-2.1", + "name": "Laguna S 2.1", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text" @@ -18275,25 +19458,36 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 1048576, + "input": 1048576, "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 0.1, + "output": 0.2, + "cache_read": 0.01 } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "abliteration-ai/abliterated-model-large-v2": { + "id": "abliteration-ai/abliterated-model-large-v2", + "name": "Abliterated Model Large V2", + "description": "Abliteration.ai's default large text reasoning model is derived from GLM-5.3 for harder reasoning and evaluation workloads, with automatic prompt caching and a one-million-token context window.", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", + "structured_output": true, + "release_date": "2026-08-31", + "last_updated": "2026-08-31", "modalities": { "input": [ "text" @@ -18304,55 +19498,85 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "input": 1000000, + "output": 999990 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 3, + "output": 5, + "cache_read": 0.3 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", + "abliteration-ai/abliterated-model": { + "id": "abliteration-ai/abliterated-model", + "name": "Abliterated Model", + "description": "Abliteration.ai's multimodal reasoning model supports text and image input, structured output, automatic prompt caching, and a 262K-token context window.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": false, + "structured_output": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 131072 + "input": 262144, + "output": 262134 }, "cost": { - "input": 0.2, - "output": 0.3 + "input": 1, + "output": 3, + "cache_read": 0.1 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "abliteration-ai/abliterated-model-large": { + "id": "abliteration-ai/abliterated-model-large", + "name": "Abliterated Model Large", + "description": "Abliteration.ai's large text reasoning model is derived from GLM-5.2 and supports native tool calling, structured output, automatic prompt caching, and a one-million-token context window.", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", + "structured_output": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ "text" @@ -18361,31 +19585,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 999990 }, "cost": { - "input": 0.1, - "output": 0.5 + "input": 3, + "output": 5, + "cache_read": 0.3 } }, - "zai-org/GLM-4.5-FP8": { - "id": "zai-org/GLM-4.5-FP8", - "name": "GLM 4.5 FP8", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "prism-ml/ternary-bonsai-2-27b": { + "id": "prism-ml/ternary-bonsai-2-27b", + "name": "Ternary Bonsai 2 27B", + "description": "Ternary Bonsai 2 27B is a 27B-parameter reasoning model from PrismML derived from Qwen3.8-27B. It supports coding, mathematics, tool calling, and image understanding with a 262K-token context window.", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "structured_output": true, + "release_date": "2026-09-18", + "last_updated": "2026-09-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -18393,57 +19627,93 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0.075, + "output": 0.5, + "cache_read": 0.0375 } }, - "zai-org/GLM-4.5-Air": { - "id": "zai-org/GLM-4.5-Air", - "name": "GLM 4.5 Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.5:thinking": { + "id": "anthropic/claude-opus-4.5:thinking", + "name": "Claude 4.5 Opus Thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.5 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "deepseek-ai/DeepSeek-V3-0324": { - "id": "deepseek-ai/DeepSeek-V3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "anthropic/claude-sonnet-4.5:thinking": { + "id": "anthropic/claude-sonnet-4.5:thinking", + "name": "Claude Sonnet 4.5 Thinking", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -18451,59 +19721,35 @@ }, "open_weights": false, "limit": { - "context": 75000, - "output": 163840 + "context": 200000, + "input": 200000, + "output": 64000 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "anthropic/claude-opus-4.6:thinking:low": { + "id": "anthropic/claude-opus-4.6:thinking:low", + "name": "Claude 4.6 Opus Thinking Low", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 75000, - "output": 163840 - }, - "cost": { - "input": 0.5, - "output": 2.15 - } - }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -18511,293 +19757,356 @@ }, "open_weights": false, "limit": { - "context": 75000, - "output": 163840 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 5, + "output": 25, + "cache_read": 0.5 } - } - } - }, - "huggingface": { - "id": "huggingface", - "env": [ - "HF_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://router.huggingface.co/v1", - "name": "Hugging Face", - "doc": "https://huggingface.co/docs/inference-providers", - "models": { - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, + }, + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude 4.1 Opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 4096 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.59, - "output": 0.79 + "input": 15, + "output": 75, + "cache_read": 1.5 } }, - "moonshotai/Kimi-K2-Thinking": { - "id": "moonshotai/Kimi-K2-Thinking", - "name": "Kimi-K2-Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude 4.5 Opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "input": 200000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "moonshotai/Kimi-K2-Instruct-0905": { - "id": "moonshotai/Kimi-K2-Instruct-0905", - "name": "Kimi-K2-Instruct-0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude 4.6 Opus", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-04", - "last_updated": "2025-09-04", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 3 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "moonshotai/Kimi-K2-Instruct": { - "id": "moonshotai/Kimi-K2-Instruct", - "name": "Kimi-K2-Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude 4.7 Opus", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-07-14", - "last_updated": "2025-07-14", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 3 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi-K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "anthropic/claude-sonnet-4:thinking:64000": { + "id": "anthropic/claude-sonnet-4:thinking:64000", + "name": "Claude 4 Sonnet Thinking (64K)", + "description": "Claude 4 Sonnet with maximum thinking budget (64,000 tokens).", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "input": 1000000, + "output": 64000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "anthropic/claude-opus-latest": { + "id": "anthropic/claude-opus-latest", + "name": "Claude Opus Latest", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-01", - "last_updated": "2026-01-01", + "structured_output": true, + "release_date": "2026-03-29", + "last_updated": "2026-03-29", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "anthropic/claude-opus-4.1:thinking:1024": { + "id": "anthropic/claude-opus-4.1:thinking:1024", + "name": "Claude 4.1 Opus Thinking (1K)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.95, - "output": 4 + "input": 15, + "output": 75, + "cache_read": 1.5 } }, - "stepfun-ai/Step-3.5-Flash": { - "id": "stepfun-ai/Step-3.5-Flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", - "attachment": false, + "anthropic/claude-haiku-latest": { + "id": "anthropic/claude-haiku-latest", + "name": "Claude Haiku Latest", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", + "structured_output": true, + "release_date": "2026-03-29", + "last_updated": "2026-03-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 256000 + "context": 200000, + "input": 200000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 1, + "output": 5, + "cache_read": 0.1 } }, - "stepfun-ai/Step-3.7-Flash": { - "id": "stepfun-ai/Step-3.7-Flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -18806,647 +20115,771 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2026-01-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 256000 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.15 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "anthropic/claude-sonnet-4:thinking:8192": { + "id": "anthropic/claude-sonnet-4:thinking:8192", + "name": "Claude 4 Sonnet Thinking (8K)", + "description": "Claude 4 Sonnet with reduced thinking budget (8,192 tokens).", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 64000 }, "cost": { - "input": 0.14, - "output": 0.4 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "google/gemma-4-26B-A4B-it": { - "id": "google/gemma-4-26B-A4B-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "anthropic/claude-sonnet-4:thinking:1024": { + "id": "anthropic/claude-sonnet-4:thinking:1024", + "name": "Claude 4 Sonnet Thinking (1K)", + "description": "Claude 4 Sonnet with minimal thinking budget (1,024 tokens).", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 64000 }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "anthropic/claude-opus-4.6:thinking": { + "id": "anthropic/claude-opus-4.6:thinking", + "name": "Claude 4.6 Opus Thinking", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.95 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "Qwen/Qwen3-Coder-Next": { - "id": "Qwen/Qwen3-Coder-Next", - "name": "Qwen3-Coder-Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "anthropic/claude-sonnet-latest": { + "id": "anthropic/claude-sonnet-latest", + "name": "Claude Sonnet Latest", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.2, - "output": 1.5 - } - }, - "Qwen/Qwen3-Embedding-8B": { - "id": "Qwen/Qwen3-Embedding-8B", - "name": "Qwen 3 Embedding 8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "structured_output": true, + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 4096 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.01, - "output": 0 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-opus-4.6:thinking:medium": { + "id": "anthropic/claude-opus-4.6:thinking:medium", + "name": "Claude 4.6 Opus Thinking Medium", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.47, - "output": 3.19 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "Qwen/Qwen3-Next-80B-A3B-Instruct": { - "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", - "name": "Qwen3-Next-80B-A3B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-11", - "last_updated": "2025-09-11", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 66536 + "context": 200000, + "input": 200000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 1 + "input": 1, + "output": 5, + "cache_read": 0.1 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen3.5-397B-A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen3-235B-A22B-Thinking-2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "anthropic/claude-fable-latest": { + "id": "anthropic/claude-fable-latest", + "name": "Claude Fable Latest", + "description": "Compatibility alias for Claude Fable.", + "family": "claude-fable", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "structured_output": true, + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 3 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "Qwen/Qwen3-Embedding-4B": { - "id": "Qwen/Qwen3-Embedding-4B", - "name": "Qwen 3 Embedding 4B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "anthropic/claude-sonnet-4:thinking:32768": { + "id": "anthropic/claude-sonnet-4:thinking:32768", + "name": "Claude 4 Sonnet Thinking (32K)", + "description": "Claude 4 Sonnet with extended thinking budget (32,768 tokens).", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 2048 + "context": 1000000, + "input": 1000000, + "output": 64000 }, "cost": { - "input": 0.01, - "output": 0 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "name": "Qwen3-Coder-480B-A35B-Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "anthropic/claude-fable-5.1": { + "id": "anthropic/claude-fable-5.1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 66536 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 2, - "output": 2 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "Qwen/Qwen3.5-122B-A10B": { - "id": "Qwen/Qwen3.5-122B-A10B", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 64000 }, "cost": { - "input": 0.4, - "output": 3.2 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "Qwen/Qwen3-Coder-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.1:thinking:8192": { + "id": "anthropic/claude-opus-4.1:thinking:8192", + "name": "Claude 4.1 Opus Thinking (8K)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.07, - "output": 0.26 + "input": 15, + "output": 75, + "cache_read": 1.5 } }, - "Qwen/Qwen3.5-27B": { - "id": "Qwen/Qwen3.5-27B", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-haiku-4.5:thinking": { + "id": "anthropic/claude-haiku-4.5:thinking", + "name": "Claude Haiku 4.5 Thinking", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 2.4 + "input": 1, + "output": 5, + "cache_read": 0.1 } }, - "Qwen/Qwen3-Next-80B-A3B-Thinking": { - "id": "Qwen/Qwen3-Next-80B-A3B-Thinking", - "name": "Qwen3-Next-80B-A3B-Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": false, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-11", - "last_updated": "2025-09-11", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 2 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "Qwen/Qwen3.5-9B": { - "id": "Qwen/Qwen3.5-9B", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "anthropic/claude-opus-4:thinking:1024": { + "id": "anthropic/claude-opus-4:thinking:1024", + "name": "Claude 4 Opus Thinking (1K)", + "description": "Claude 4 Opus with minimal thinking budget (1,024 tokens).", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.17, - "output": 0.25 + "input": 15, + "output": 75, + "cache_read": 1.5 } }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, + "anthropic/claude-opus-4.8:thinking": { + "id": "anthropic/claude-opus-4.8:thinking", + "name": "Claude Opus 4.8 Thinking", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.29, - "output": 0.59 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "Qwen/Qwen3-235B-A22B": { - "id": "Qwen/Qwen3-235B-A22B", - "name": "Qwen3 235B-A22B", - "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude 4 Sonnet", + "description": "Claude 4 Sonnet by Anthropic. A new generation model with improved capabilities, especially on programming and development. NOTE: Inputs > 200k tokens are charged at 2x input, 1.5x output rate.", + "family": "claude-sonnet", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 16384 + "context": 200000, + "input": 200000, + "output": 64000 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "Qwen/Qwen3.5-35B-A3B": { - "id": "Qwen/Qwen3.5-35B-A3B", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 2 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "anthropic/claude-sonnet-4.6:thinking": { + "id": "anthropic/claude-sonnet-4.6:thinking", + "name": "Claude Sonnet 4.6 Thinking", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -19454,39 +20887,116 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 1000000, + "output": 128000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "anthropic/claude-opus-4:thinking:32768": { + "id": "anthropic/claude-opus-4:thinking:32768", + "name": "Claude 4 Opus Thinking (32K)", + "description": "Claude 4 Opus with extended thinking budget (32,768 tokens).", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "release_date": "2025-08-05", "last_updated": "2025-08-05", "modalities": { "input": [ + "text", + "image", + "pdf" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "input": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5 + } + }, + "anthropic/claude-opus-4.6:thinking:max": { + "id": "anthropic/claude-opus-4.6:thinking:max", + "name": "Claude 4.6 Opus Thinking Max", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 0.69 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "anthropic/claude-sonnet-5:thinking": { + "id": "anthropic/claude-sonnet-5:thinking", + "name": "Claude Sonnet 5 Thinking", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -19494,339 +21004,403 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.5 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "XiaomiMiMo/MiMo-V2.5-Pro": { - "id": "XiaomiMiMo/MiMo-V2.5-Pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 3 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "XiaomiMiMo/MiMo-V2-Flash": { - "id": "XiaomiMiMo/MiMo-V2-Flash", - "name": "MiMo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, + "anthropic/claude-opus-4.7:thinking": { + "id": "anthropic/claude-opus-4.7:thinking", + "name": "Claude 4.7 Opus Thinking", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 4096 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "zai-org/GLM-4.7-Flash": { - "id": "zai-org/GLM-4.7-Flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, + "anthropic/claude-opus-5.5": { + "id": "anthropic/claude-opus-5.5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, + "context": 1000000, + "input": 1000000, "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 } }, - "zai-org/GLM-4.6": { - "id": "zai-org/GLM-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, + "anthropic/claude-opus-4.1:thinking": { + "id": "anthropic/claude-opus-4.1:thinking", + "name": "Claude 4.1 Opus Thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.55, - "output": 2.2 + "input": 15, + "output": 75, + "cache_read": 1.5 } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "anthropic/claude-sonnet-4:thinking": { + "id": "anthropic/claude-sonnet-4:thinking", + "name": "Claude 4 Sonnet Thinking", + "description": "Anthropic's Claude 4 Sonnet with the ability to show its thinking process step by step.", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "structured_output": true, + "release_date": "2025-02-24", + "last_updated": "2025-02-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1000000, + "input": 1000000, + "output": 64000 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "zai-org/GLM-4.7": { - "id": "zai-org/GLM-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "anthropic/claude-opus-4:thinking": { + "id": "anthropic/claude-opus-4:thinking", + "name": "Claude 4 Opus Thinking", + "description": "Anthropic's Claude 4 Opus with the ability to show its thinking process step by step.", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "structured_output": true, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 15, + "output": 75, + "cache_read": 1.5 } }, - "zai-org/GLM-4.5V": { - "id": "zai-org/GLM-4.5V", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "anthropic/claude-opus-4:thinking:8192": { + "id": "anthropic/claude-opus-4:thinking:8192", + "name": "Claude 4 Opus Thinking (8K)", + "description": "Claude 4 Opus with reduced thinking budget (8,192 tokens).", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "structured_output": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.6, - "output": 1.8 + "input": 15, + "output": 75, + "cache_read": 1.5 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Claude 4 Opus", + "description": "Claude 4 Opus by Anthropic. The premium version of the new Claude models. A new generation model with improved capabilities, especially on programming and development.", + "family": "claude-opus", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 1.4, - "output": 4.4 + "input": 15, + "output": 75, + "cache_read": 1.5 } }, - "zai-org/GLM-4.5-Air": { - "id": "zai-org/GLM-4.5-Air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", - "attachment": false, + "anthropic/claude-opus-4.1:thinking:32768": { + "id": "anthropic/claude-opus-4.1:thinking:32768", + "name": "Claude 4.1 Opus Thinking (32K)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.13, - "output": 0.85 + "input": 15, + "output": 75, + "cache_read": 1.5 } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "cohere/command-r-plus-08-2024": { + "id": "cohere/command-r-plus-08-2024", + "name": "Cohere: Command R+", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-04-03", - "last_updated": "2026-04-03", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -19837,28 +21411,39 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 2.856, + "output": 14.246, + "cache_read": 1.428 } }, - "zai-org/GLM-4.5": { - "id": "zai-org/GLM-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", + "deepseek/deepseek-v4-flash:thinking": { + "id": "deepseek/deepseek-v4-flash:thinking", + "name": "DeepSeek V4 Flash (Thinking)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -19869,28 +21454,40 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 1048576, + "input": 1048576, + "output": 384000 }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "deepseek-ai/DeepSeek-R1": { - "id": "deepseek-ai/DeepSeek-R1", - "name": "DeepSeek-R1", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", + "deepseek/deepseek-v4-flash-0731": { + "id": "deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -19901,30 +21498,44 @@ }, "open_weights": true, "limit": { - "context": 64000, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 131072 }, "cost": { - "input": 0.7, - "output": 2.5 + "input": 0.05, + "output": 0.16, + "cache_read": 0.013 } }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", - "name": "DeepSeek-R1-0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "deepseek/deepseek-v4.1-flash": { + "id": "deepseek/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2025-05", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -19932,31 +21543,44 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 1000000, + "input": 1000000, + "output": 384000 }, "cost": { - "input": 3, - "output": 5 + "input": 0.13, + "output": 0.52, + "cache_read": 0.006 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "deepseek/deepseek-v4.1-flash:thinking": { + "id": "deepseek/deepseek-v4.1-flash:thinking", + "name": "DeepSeek V4.1 Flash Thinking", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", "family": "deepseek-flash", - "attachment": false, + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -19964,41 +21588,43 @@ }, "open_weights": true, "limit": { - "context": 1048576, + "context": 1000000, + "input": 1000000, "output": 384000 }, "cost": { - "input": 0.14, - "output": 0.28 + "input": 0.13, + "output": 0.52, + "cache_read": 0.006 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "deepseek/deepseek-v4-flash-vision-exp": { + "id": "deepseek/deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high" + "none", + "low", + "high", + "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -20007,30 +21633,33 @@ "open_weights": true, "limit": { "context": 1048576, - "output": 393216 + "input": 1048576, + "output": 384000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "deepseek/deepseek-v3.2:thinking": { + "id": "deepseek/deepseek-v3.2:thinking", + "name": "DeepSeek V3.2 Thinking", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", "family": "deepseek", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2024-07", "release_date": "2025-12-01", "last_updated": "2025-12-01", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -20038,30 +21667,38 @@ }, "open_weights": true, "limit": { - "context": 163840, + "context": 163000, + "input": 163000, "output": 65536 }, "cost": { "input": 0.28, - "output": 0.4 + "output": 0.42, + "cache_read": 0.14 } }, - "MiniMaxAI/MiniMax-M2.1": { - "id": "MiniMaxAI/MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "deepseek/deepseek-latest": { + "id": "deepseek/deepseek-latest", + "name": "DeepSeek Latest", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-10", - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "structured_output": true, + "release_date": "2026-05-03", + "last_updated": "2026-05-03", "modalities": { "input": [ "text" @@ -20072,26 +21709,39 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 1048576, + "input": 1048576, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 1.1, + "output": 2.5, + "cache_read": 0.04 } }, - "MiniMaxAI/MiniMax-M2": { - "id": "MiniMaxAI/MiniMax-M2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", + "deepseek/deepseek-v4-pro:thinking": { + "id": "deepseek/deepseek-v4-pro:thinking", + "name": "DeepSeek V4 Pro (Thinking)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -20102,29 +21752,39 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 128000 + "context": 1048576, + "input": 1048576, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 1.1, + "output": 2.2, + "cache_read": 0.11 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "deepseek/deepseek-v4-pro-0813:thinking": { + "id": "deepseek/deepseek-v4-pro-0813:thinking", + "name": "DeepSeek V4 Pro 0813 Thinking", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -20135,32 +21795,42 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 1048576, + "input": 1048576, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 1.1, + "output": 2.5, + "cache_read": 0.04 } }, - "MiniMaxAI/MiniMax-M3": { - "id": "MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, + "deepseek/deepseek-v4-pro-0813": { + "id": "deepseek/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -20168,30 +21838,40 @@ }, "open_weights": true, "limit": { - "context": 524288, - "output": 128000 + "context": 1048576, + "input": 1048576, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 1.1, + "output": 2.5, + "cache_read": 0.04 } }, - "MiniMaxAI/MiniMax-M2.7": { - "id": "MiniMaxAI/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "deepseek/deepseek-v4-flash-0731:thinking": { + "id": "deepseek/deepseek-v4-flash-0731:thinking", + "name": "DeepSeek V4 Flash 0731 (Thinking)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -20202,69 +21882,81 @@ }, "open_weights": true, "limit": { - "context": 204800, + "context": 1000000, + "input": 1000000, "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.05, + "output": 0.16, + "cache_read": 0.013 } - } - } - }, - "minimax-coding-plan": { - "id": "minimax-coding-plan", - "env": [ - "MINIMAX_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://api.minimax.io/anthropic/v1", - "name": "MiniMax Token Plan (minimax.io)", - "doc": "https://platform.minimax.io/docs/token-plan/intro", - "models": { - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + }, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ "text" ] }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 1048576, + "input": 1048576, + "output": 384000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.1, + "output": 2.2, + "cache_read": 0.11 } }, - "MiniMax-M2.5-highspeed": { - "id": "MiniMax-M2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "deepseek/deepseek-v4-flash-latest": { + "id": "deepseek/deepseek-v4-flash-latest", + "name": "DeepSeek V4 Flash Latest", + "description": "Compatibility alias that routes to the newest dated DeepSeek V4 Flash release. Currently routes to DeepSeek V4 Flash 0731. ⚠️ This route goes directly to DeepSeek, so privacy and logging guarantees are limited.", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "structured_output": true, + "release_date": "2026-08-02", + "last_updated": "2026-08-02", "modalities": { "input": [ "text" @@ -20275,31 +21967,33 @@ }, "open_weights": true, "limit": { - "context": 204800, + "context": 1000000, + "input": 1000000, "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.05, + "output": 0.16, + "cache_read": 0.013 } }, - "MiniMax-M2.7-highspeed": { - "id": "MiniMax-M2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -20307,28 +22001,39 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 163000, + "input": 163000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.28, + "output": 0.42, + "cache_read": 0.14 } }, - "MiniMax-M2": { - "id": "MiniMax-M2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -20339,26 +22044,27 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 128000 + "context": 1048576, + "input": 1048576, + "output": 384000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "baseten/Kimi-K2-Instruct-FP4": { + "id": "baseten/Kimi-K2-Instruct-FP4", + "name": "Kimi K2 0711 Instruct FP4", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-07-11", "modalities": { "input": [ "text" @@ -20369,69 +22075,62 @@ }, "open_weights": true, "limit": { - "context": 204800, + "context": 131072, + "input": 131072, "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.4, + "output": 1.8, + "cache_read": 0.2 } }, - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + "unbiased/pareto": { + "id": "unbiased/pareto", + "name": "Pareto", + "description": "Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-25", + "release_date": "2026-09-17", + "last_updated": "2026-09-17", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2.5, + "output": 7.5, + "cache_read": 0.25 } }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "lightonai/LightOnOCR-2-1B": { + "id": "lightonai/LightOnOCR-2-1B", + "name": "LightOnOCR 2", + "description": "LightOnOCR 2 hosted by IONOS in Berlin, Germany. Zero data retention.", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": false, + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -20439,40 +22138,26 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.1785, + "output": 0.3465 } - } - } - }, - "novita-ai": { - "id": "novita-ai", - "env": [ - "NOVITA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.novita.ai/openai", - "name": "NovitaAI", - "doc": "https://novita.ai/docs/guides/introduction", - "models": { - "inclusionai/ling-2.6-1t": { - "id": "inclusionai/ling-2.6-1t", - "name": "Ling-2.6-1T", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "ling", + }, + "abacusai/Dracarys-72B-Instruct": { + "id": "abacusai/Dracarys-72B-Instruct", + "name": "Llama 3.1 70B Dracarys 2", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-23", - "last_updated": "2026-06-29", + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-02", + "last_updated": "2025-08-02", "modalities": { "input": [ "text" @@ -20483,28 +22168,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.06 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "inclusionai/ring-2.6-1t": { - "id": "inclusionai/ring-2.6-1t", - "name": "Ring-2.6-1T", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "ring", + "shisa-ai/shisa-v2.1-llama3.3-70b": { + "id": "shisa-ai/shisa-v2.1-llama3.3-70b", + "name": "Shisa V2.1 Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-08", - "last_updated": "2026-05-27", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -20513,29 +22197,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 32768, + "input": 32768, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.06 + "input": 0.5, + "output": 0.5, + "cache_read": 0.25 } }, - "inclusionai/ling-2.6-flash": { - "id": "inclusionai/ling-2.6-flash", - "name": "Ling-2.6-flash", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", + "shisa-ai/shisa-v2-llama3.3-70b": { + "id": "shisa-ai/shisa-v2-llama3.3-70b", + "name": "Shisa V2 Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text" @@ -20546,26 +22230,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 0.5, + "output": 0.5, + "cache_read": 0.25 } }, - "meta-llama/llama-3.1-8b-instruct": { - "id": "meta-llama/llama-3.1-8b-instruct", - "name": "Llama 3.1 8B Instruct", + "Steelskull/L3.3-MS-Nevoria-70b": { + "id": "Steelskull/L3.3-MS-Nevoria-70b", + "name": "Steelskull Nevoria 70b", "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-07-24", - "last_updated": "2024-07-24", + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -20576,26 +22261,27 @@ }, "open_weights": true, "limit": { - "context": 16384, + "context": 32768, + "input": 32768, "output": 16384 }, "cost": { - "input": 0.02, - "output": 0.05 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "meta-llama/llama-3-70b-instruct": { - "id": "meta-llama/llama-3-70b-instruct", - "name": "Llama3 70B Instruct", + "Steelskull/L3.3-Cu-Mai-R1-70b": { + "id": "Steelskull/L3.3-Cu-Mai-R1-70b", + "name": "Llama 3.3 70B Cu Mai", "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2024-04-25", - "last_updated": "2024-04-25", + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -20606,28 +22292,30 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8000 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 0.51, - "output": 0.74 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "meta-llama/llama-4-scout-17b-16e-instruct": { - "id": "meta-llama/llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "attachment": true, + "Steelskull/L3.3-Nevoria-R1-70b": { + "id": "Steelskull/L3.3-Nevoria-R1-70b", + "name": "Steelskull Nevoria R1 70b", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-04-06", - "last_updated": "2025-04-06", + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -20635,26 +22323,27 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 0.18, - "output": 0.59 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "meta-llama/llama-3.3-70b-instruct": { - "id": "meta-llama/llama-3.3-70b-instruct", - "name": "Llama 3.3 70B Instruct", + "Steelskull/L3.3-Electra-R1-70b": { + "id": "Steelskull/L3.3-Electra-R1-70b", + "name": "Steelskull Electra R1 70b", "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-07", - "last_updated": "2024-12-07", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -20665,25 +22354,38 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 120000 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 0.135, - "output": 0.4 + "input": 0.69989, + "output": 0.69989, + "cache_read": 0.349945 } }, - "meta-llama/llama-3-8b-instruct": { - "id": "meta-llama/llama-3-8b-instruct", - "name": "Llama 3 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "tencent/hy3": { + "id": "tencent/hy3", + "name": "Tencent Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-04-25", - "last_updated": "2024-04-25", + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ "text" @@ -20694,25 +22396,38 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 262144, + "input": 262144, + "output": 128000 }, "cost": { - "input": 0.04, - "output": 0.04 + "input": 0.066, + "output": 0.26, + "cache_read": 0.029 } }, - "meta-llama/llama-3.2-3b-instruct": { - "id": "meta-llama/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "tencent/hy4-preview": { + "id": "tencent/hy4-preview", + "name": "Tencent Hy4 Preview", + "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", + "family": "Hy", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2024-09-18", + "release_date": "2026-08-28", + "last_updated": "2026-08-28", "modalities": { "input": [ "text" @@ -20723,28 +22438,31 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32000 + "context": 1048576, + "input": 1048576, + "output": 64000 }, "cost": { - "input": 0.03, - "output": 0.05 + "input": 0.834, + "output": 2.501, + "cache_read": 0.042 } }, - "meta-llama/llama-4-maverick-17b-128e-instruct-fp8": { - "id": "meta-llama/llama-4-maverick-17b-128e-instruct-fp8", - "name": "Llama 4 Maverick Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", + "anthracite-org/magnum-v4-72b": { + "id": "anthracite-org/magnum-v4-72b", + "name": "Magnum v4 72B", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-04-06", - "last_updated": "2025-04-06", + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" @@ -20752,24 +22470,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, + "context": 16384, + "input": 16384, "output": 8192 }, "cost": { - "input": 0.27, - "output": 0.85 + "input": 2.006, + "output": 2.992, + "cache_read": 1.003 } }, - "moonshotai/kimi-k2-instruct": { - "id": "moonshotai/kimi-k2-instruct", - "name": "Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "anthracite-org/magnum-v2-72b": { + "id": "anthracite-org/magnum-v2-72b", + "name": "Magnum V2 72B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-07-11", - "last_updated": "2025-07-11", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-07-01", "modalities": { "input": [ "text" @@ -20780,33 +22501,31 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.57, - "output": 2.3 + "input": 2.006, + "output": 2.992, + "cache_read": 1.003 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "meta-llama/llama-4-scout": { + "id": "meta-llama/llama-4-scout", + "name": "Llama 4 Scout", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2025-11-07", - "last_updated": "2026-06-29", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -20814,41 +22533,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 328000, + "input": 328000, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 0.085, + "output": 0.46, + "cache_read": 0.0425 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "meta-llama/llama-4-maverick": { + "id": "meta-llama/llama-4-maverick", + "name": "Llama 4 Maverick", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -20856,41 +22565,32 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "meta-llama/llama-3.3-70b-instruct": { + "id": "meta-llama/llama-3.3-70b-instruct", + "name": "Llama 3.3 70b Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -20898,28 +22598,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "input": 131072, + "output": 16384 }, "cost": { - "input": 0.8, - "output": 3.4, - "cache_read": 0.16 + "input": 0.05, + "output": 0.23, + "cache_read": 0.025 } }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "meta-llama/llama-3.1-8b-instruct": { + "id": "meta-llama/llama-3.1-8b-instruct", + "name": "Llama 3.1 8b Instruct", + "description": "Compact open Llama model for lightweight chat, drafting, and self-hosting", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -20930,26 +22631,59 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "input": 131072, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.5 + "input": 0.0544, + "output": 0.085, + "cache_read": 0.0272 } }, - "minimaxai/minimax-m1-80k": { - "id": "minimaxai/minimax-m1-80k", - "name": "MiniMax M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "meta-llama/llama-3.2-3b-instruct": { + "id": "meta-llama/llama-3.2-3b-instruct", + "name": "Llama 3.2 3b Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-09-25", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "input": 131072, + "output": 8192 + }, + "cost": { + "input": 0.0306, + "output": 0.0493, + "cache_read": 0.0153 + } + }, + "LLM360/K2-Think": { + "id": "LLM360/K2-Think", + "name": "K2-Think", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text" @@ -20960,29 +22694,30 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 40000 + "context": 128000, + "input": 128000, + "output": 32768 }, "cost": { - "input": 0.55, - "output": 2.2 + "input": 0.17, + "output": 0.68, + "cache_read": 0.085 } }, - "baidu/ernie-4.5-vl-28b-a3b": { - "id": "baidu/ernie-4.5-vl-28b-a3b", - "name": "ERNIE 4.5 VL 28B A3B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2026-06-14", + "deepcogito/cogito-v1-preview-qwen-32B": { + "id": "deepcogito/cogito-v1-preview-qwen-32B", + "name": "Cogito v1 Preview Qwen 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-10", + "last_updated": "2025-05-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -20990,26 +22725,27 @@ }, "open_weights": true, "limit": { - "context": 30000, - "output": 8000 + "context": 128000, + "input": 128000, + "output": 32768 }, "cost": { - "input": 0.14, - "output": 0.56 + "input": 1.8, + "output": 1.8, + "cache_read": 0.9 } }, - "baidu/ernie-4.5-21B-a3b": { - "id": "baidu/ernie-4.5-21B-a3b", - "name": "ERNIE 4.5 21B A3B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "ernie", + "GalrionSoftworks/MN-LooseCannon-12B-v1": { + "id": "GalrionSoftworks/MN-LooseCannon-12B-v1", + "name": "MN-LooseCannon-12B-v1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-07-01", "modalities": { "input": [ "text" @@ -21020,57 +22756,58 @@ }, "open_weights": true, "limit": { - "context": 120000, - "output": 8000 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "baidu/ernie-4.5-vl-424b-a47b": { - "id": "baidu/ernie-4.5-vl-424b-a47b", - "name": "ERNIE 4.5 VL 424B A47B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "meganova-ai/manta-mini-1.0": { + "id": "meganova-ai/manta-mini-1.0", + "name": "Manta Mini 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova", + "attachment": false, + "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "structured_output": false, + "release_date": "2025-09-20", + "last_updated": "2025-12-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 123000, - "output": 16000 + "context": 8192, + "input": 8192, + "output": 8192 }, "cost": { - "input": 0.42, - "output": 1.25 + "input": 0.02, + "output": 0.16, + "cache_read": 0.01 } }, - "baidu/ernie-4.5-21B-a3b-thinking": { - "id": "baidu/ernie-4.5-21B-a3b-thinking", - "name": "ERNIE-4.5-21B-A3B-Thinking", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "ernie", + "meganova-ai/manta-flash-1.0": { + "id": "meganova-ai/manta-flash-1.0", + "name": "Manta Flash 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-09-19", - "last_updated": "2025-09-19", + "structured_output": false, + "release_date": "2025-09-20", + "last_updated": "2025-12-20", "modalities": { "input": [ "text" @@ -21079,27 +22816,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 0.02, + "output": 0.16, + "cache_read": 0.01 } }, - "baidu/ernie-4.5-300b-a47b-paddle": { - "id": "baidu/ernie-4.5-300b-a47b-paddle", - "name": "ERNIE 4.5 300B A47B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "meganova-ai/manta-pro-1.0": { + "id": "meganova-ai/manta-pro-1.0", + "name": "Manta Pro 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "structured_output": false, + "release_date": "2025-09-20", + "last_updated": "2025-12-20", "modalities": { "input": [ "text" @@ -21108,28 +22847,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 123000, - "output": 12000 + "context": 65536, + "input": 65536, + "output": 32768 }, "cost": { - "input": 0.28, - "output": 1.1 + "input": 0.06, + "output": 0.5, + "cache_read": 0.03 } }, - "baidu/ernie-4.5-vl-28b-a3b-thinking": { - "id": "baidu/ernie-4.5-vl-28b-a3b-thinking", - "name": "ERNIE-4.5-VL-28B-A3B-Thinking", + "perceptron/perceptron-mk1": { + "id": "perceptron/perceptron-mk1", + "name": "Perceptron Mk1", "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, "structured_output": true, - "temperature": true, - "release_date": "2025-11-26", - "last_updated": "2025-11-26", + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ "text", @@ -21140,37 +22880,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0.39, - "output": 0.39 + "input": 0.15, + "output": 1.5, + "cache_read": 0.075 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "z-ai/glm-4.6v": { + "id": "z-ai/glm-4.6v", + "name": "GLM 4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -21178,35 +22917,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 24000 }, "cost": { - "input": 0.14, - "output": 0.4 + "input": 0.3, + "output": 0.9, + "cache_read": 0.15 } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "z-ai/glm-4.6-original": { + "id": "z-ai/glm-4.6-original", + "name": "GLM 4.6 Original", + "description": "GLM-4.6, Zhipu's flagship text model with 256K context window and advanced reasoning capabilities. Direct via Z-AI (Zhipu).", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "structured_output": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -21214,29 +22949,32 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 65535 }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 0.35, + "output": 1.4, + "cache_read": 0.175 } }, - "google/gemma-3-12b-it": { - "id": "google/gemma-3-12b-it", - "name": "Gemma 3 12B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "z-ai/glm-4.5": { + "id": "z-ai/glm-4.5", + "name": "GLM 4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -21244,57 +22982,77 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.05, - "output": 0.1 + "input": 0.3, + "output": 1.3, + "cache_read": 0.15 } }, - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma 3 27B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "z-ai/glm-5v-turbo": { + "id": "z-ai/glm-5v-turbo", + "name": "GLM 5V Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 98304, - "output": 16384 + "context": 202800, + "input": 202800, + "output": 131072 }, "cost": { - "input": 0.119, - "output": 0.2 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "microsoft/wizardlm-2-8x22b": { - "id": "microsoft/wizardlm-2-8x22b", - "name": "Wizardlm 2 8x22B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, + "z-ai/glm-5.3-flash": { + "id": "z-ai/glm-5.3-flash", + "name": "GLM 5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-04-24", - "last_updated": "2024-04-24", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -21302,18 +23060,21 @@ }, "open_weights": true, "limit": { - "context": 65535, - "output": 8000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.62, - "output": 0.62 + "input": 0.075, + "output": 0.25, + "cache_read": 0.015 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "OpenAI GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "z-ai/glm-5.3-flash-uncensored": { + "id": "z-ai/glm-5.3-flash-uncensored", + "name": "GLM 5.3 Flash Uncensored", + "description": "GLM 5.3 Flash Uncensored is an uncensored fine-tune of the efficient 320B mixture-of-experts reasoning model, built for unrestricted chat, creative writing, coding, agentic work, tool use, and long-context tasks.", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -21321,16 +23082,14 @@ "type": "effort", "values": [ "low", - "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "release_date": "2026-07-29", + "last_updated": "2026-08-27", "modalities": { "input": [ "text", @@ -21342,19 +23101,22 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 1048576, + "input": 1048576, "output": 32768 }, "cost": { - "input": 0.05, - "output": 0.25 + "input": 0.2, + "output": 0.8, + "cache_read": 0.1 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "OpenAI: GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "attachment": true, + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "GLM 4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -21366,15 +23128,15 @@ ] } ], - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -21382,24 +23144,30 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 200000, + "input": 200000, + "output": 65535 }, "cost": { - "input": 0.04, - "output": 0.15 + "input": 0.35, + "output": 1.4, + "cache_read": 0.175 } }, - "sao10K/l31-70b-euryale-v2.2": { - "id": "sao10K/l31-70b-euryale-v2.2", - "name": "L31 70B Euryale V2.2", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "z-ai/glm-4.7-flash:thinking": { + "id": "z-ai/glm-4.7-flash:thinking", + "name": "GLM 4.7 Flash Thinking", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -21410,25 +23178,28 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 1.48, - "output": 1.48 + "input": 0.07, + "output": 0.4, + "cache_read": 0.035 } }, - "sao10K/L3-8B-stheno-v3.2": { - "id": "sao10K/L3-8B-stheno-v3.2", - "name": "L3 8B Stheno V3.2", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "z-ai/glm-5-original:thinking": { + "id": "z-ai/glm-5-original:thinking", + "name": "GLM 5 Original Thinking", + "description": "GLM-5 original with extended thinking capabilities for complex reasoning.", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2024-11-29", - "last_updated": "2024-11-29", + "structured_output": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -21439,25 +23210,37 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 32000 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.05 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "sao10K/l3-8b-lunaris": { - "id": "sao10K/l3-8b-lunaris", - "name": "Sao10k L3 8B Lunaris\t", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "z-ai/glm-5.3:thinking": { + "id": "z-ai/glm-5.3:thinking", + "name": "GLM 5.3 Thinking", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-11-28", - "last_updated": "2024-11-28", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -21468,24 +23251,29 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.05 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "sao10K/l3-70b-euryale-v2.1": { - "id": "sao10K/l3-70b-euryale-v2.1", - "name": "L3 70B Euryale V2.1\t", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM 5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-06-18", - "last_updated": "2024-06-18", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -21496,27 +23284,29 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 1.48, - "output": 1.48 + "input": 0.5, + "output": 2.55, + "cache_read": 0.13 } }, - "baichuan/baichuan-m2-32b": { - "id": "baichuan/baichuan-m2-32b", - "name": "baichuan-m2-32b", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "baichuan", + "z-ai/GLM-4.5-Air": { + "id": "z-ai/GLM-4.5-Air", + "name": "GLM 4.5 Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-08-13", - "last_updated": "2025-08-13", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -21527,29 +23317,38 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 98304 }, "cost": { - "input": 0.07, - "output": 0.07 + "input": 0.12, + "output": 0.8, + "cache_read": 0.06 } }, - "mistralai/mistral-nemo": { - "id": "mistralai/mistral-nemo", - "name": "Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, + "z-ai/glm-4.5v": { + "id": "z-ai/glm-4.5v", + "name": "GLM 4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": false, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2024-07-30", - "last_updated": "2024-07-30", + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -21557,28 +23356,28 @@ }, "open_weights": true, "limit": { - "context": 60288, - "output": 16000 + "context": 65536, + "input": 65536, + "output": 16384 }, "cost": { - "input": 0.04, - "output": 0.17 + "input": 0.6, + "output": 1.8, + "cache_read": 0.3 } }, - "xiaomimimo/mimo-v2-flash": { - "id": "xiaomimimo/mimo-v2-flash", - "name": "XiaomiMiMo/MiMo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", + "z-ai/GLM-4.6-turbo:thinking": { + "id": "z-ai/GLM-4.6-turbo:thinking", + "name": "GLM 4.6 Turbo (Thinking)", + "description": "GLM 4.6 Turbo with thinking mode enabled for enhanced reasoning; shows internal reasoning and supports long context.", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-19", - "last_updated": "2025-12-19", + "tool_call": false, + "structured_output": false, + "release_date": "2025-10-02", + "last_updated": "2025-10-02", "modalities": { "input": [ "text" @@ -21589,32 +23388,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32000 + "context": 204800, + "input": 204800, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.3 + "input": 1, + "output": 3, + "cache_read": 0.5 } }, - "xiaomimimo/mimo-v2-pro": { - "id": "xiaomimimo/mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", + "z-ai/glm-4.7:thinking": { + "id": "z-ai/glm-4.7:thinking", + "name": "GLM 4.7 Thinking", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-05-27", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -21623,50 +23420,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 65535 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 0.2, + "output": 0.8, + "cache_read": 0.1 } }, - "xiaomimimo/mimo-v2.5-pro": { - "id": "xiaomimimo/mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "z-ai/glm-5.1:thinking": { + "id": "z-ai/glm-5.1:thinking", + "name": "GLM 5.1 Thinking", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-27", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -21677,44 +23463,31 @@ }, "open_weights": true, "limit": { - "context": 1048576, + "context": 200000, + "input": 200000, "output": 131072 }, "cost": { - "input": 0.522, - "output": 1.044, - "cache_read": 0.0043, - "tiers": [ - { - "input": 0.522, - "output": 1.044, - "cache_read": 0.0043, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.522, - "output": 1.044, - "cache_read": 0.0043 - } + "input": 0.75, + "output": 2.6, + "cache_read": 0.15 } }, - "gryphe/mythomax-l2-13b": { - "id": "gryphe/mythomax-l2-13b", - "name": "Mythomax L2 13B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, + "z-ai/glm-4.6v-original": { + "id": "z-ai/glm-4.6v-original", + "name": "GLM 4.6V Original", + "description": "GLM-4.6V scales its context window to 128k tokens in training, and achieves SoTA performance in visual understanding among models of similar parameter scales. Integrates native Function Calling capabilities, bridging 'visual perception' and 'executable action' for multimodal agents. Direct via Z-AI (Zhipu).", + "family": "glm", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-04-25", - "last_updated": "2024-04-25", + "structured_output": false, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -21722,34 +23495,62 @@ }, "open_weights": true, "limit": { - "context": 4096, - "output": 3200 + "context": 128000, + "input": 128000, + "output": 24000 }, "cost": { - "input": 0.09, - "output": 0.09 + "input": 0.6, + "output": 0.9, + "cache_read": 0.3 } }, - "zai-org/glm-4.7": { - "id": "zai-org/glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "z-ai/glm-5v-turbo:thinking": { + "id": "z-ai/glm-5v-turbo:thinking", + "name": "GLM 5V Turbo Thinking", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", "family": "glm", - "attachment": false, + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 202800, + "input": 202800, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.24 + } + }, + "z-ai/GLM-4.6-turbo": { + "id": "z-ai/GLM-4.6-turbo", + "name": "GLM 4.6 Turbo", + "description": "Fast variant of GLM 4.6 for general chat, coding, and analysis with improved latency and strong reasoning.", + "family": "glm", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-10-02", + "last_updated": "2025-10-02", "modalities": { "input": [ "text" @@ -21761,37 +23562,32 @@ "open_weights": true, "limit": { "context": 204800, + "input": 204800, "output": 131072 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 1, + "output": 3, + "cache_read": 0.5 } }, - "zai-org/glm-4.5v": { - "id": "zai-org/glm-4.5v", - "name": "GLM 4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glmv", - "attachment": true, + "z-ai/GLM-4.5-Air:thinking": { + "id": "z-ai/GLM-4.5-Air:thinking", + "name": "GLM 4.5 Air (Thinking)", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "video", - "image" + "text" ], "output": [ "text" @@ -21799,34 +23595,37 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 16384 + "context": 128000, + "input": 128000, + "output": 98304 }, "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11 + "input": 0.12, + "output": 0.8, + "cache_read": 0.06 } }, - "zai-org/glm-4.5": { - "id": "zai-org/glm-4.5", - "name": "GLM-4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "z-ai/glm-5.2:thinking": { + "id": "z-ai/glm-5.2:thinking", + "name": "GLM 5.2 Thinking", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -21837,29 +23636,33 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0.42, + "output": 1.32, + "cache_read": 0.078 } }, - "zai-org/autoglm-phone-9b-multilingual": { - "id": "zai-org/autoglm-phone-9b-multilingual", - "name": "AutoGLM-Phone-9B-Multilingual", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "attachment": true, - "reasoning": false, - "tool_call": false, + "z-ai/glm-4.7-flash": { + "id": "z-ai/glm-4.7-flash", + "name": "GLM 4.7 Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-10", - "last_updated": "2025-12-10", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -21867,34 +23670,30 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 0.035, - "output": 0.138 + "input": 0.07, + "output": 0.4, + "cache_read": 0.035 } }, - "zai-org/glm-5.1": { - "id": "zai-org/glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "GLM 4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -21905,35 +23704,29 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 65535 }, "cost": { - "input": 1.38, - "output": 4.4, - "cache_read": 0.26 + "input": 0.2, + "output": 0.8, + "cache_read": 0.1 } }, - "zai-org/glm-4.6": { - "id": "zai-org/glm-4.6", - "name": "GLM 4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "z-ai/glm-5:thinking": { + "id": "z-ai/glm-5:thinking", + "name": "GLM 5 Thinking", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -21944,18 +23737,19 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 0.55, - "output": 2.2, - "cache_read": 0.11 + "input": 0.5, + "output": 2.55, + "cache_read": 0.13 } }, - "zai-org/glm-5.2": { - "id": "zai-org/glm-5.2", - "name": "GLM-5.2", + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM 5.2", "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, @@ -21971,9 +23765,6 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, "release_date": "2026-06-13", @@ -21989,37 +23780,32 @@ "open_weights": true, "limit": { "context": 1048576, + "input": 1048576, "output": 131072 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.42, + "output": 1.32, + "cache_read": 0.078 } }, - "zai-org/glm-4.6v": { - "id": "zai-org/glm-4.6v", - "name": "GLM 4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glmv", - "attachment": true, + "z-ai/GLM-4.5:thinking": { + "id": "z-ai/GLM-4.5:thinking", + "name": "GLM 4.5 (Thinking)", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "video", - "image" + "text" ], "output": [ "text" @@ -22027,32 +23813,37 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { "input": 0.3, - "output": 0.9, - "cache_read": 0.055 + "output": 1.3, + "cache_read": 0.15 } }, - "zai-org/glm-4.5-air": { - "id": "zai-org/glm-4.5-air", - "name": "GLM 4.5 Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-10-13", - "last_updated": "2025-10-13", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -22063,33 +23854,39 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 200000, + "input": 200000, + "output": 131072 }, "cost": { - "input": 0.13, - "output": 0.85, - "cache_read": 0.025 + "input": 0.75, + "output": 2.6, + "cache_read": 0.15 } }, - "zai-org/glm-4.7-flash": { - "id": "zai-org/glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "z-ai/glm-4.6:thinking": { + "id": "z-ai/glm-4.6:thinking", + "name": "GLM 4.6 Thinking", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -22101,37 +23898,33 @@ "open_weights": true, "limit": { "context": 200000, - "output": 128000 + "input": 200000, + "output": 65535 }, "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.01 + "input": 0.35, + "output": 1.4, + "cache_read": 0.175 } }, - "zai-org/glm-5": { - "id": "zai-org/glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "z-ai/glm-4.5v:thinking": { + "id": "z-ai/glm-4.5v:thinking", + "name": "GLM 4.5V Thinking", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "family": "glm", - "attachment": false, + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-12", + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -22139,29 +23932,31 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 131072 + "context": 65536, + "input": 65536, + "output": 16384 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.6, + "output": 1.8, + "cache_read": 0.3 } }, - "paddlepaddle/paddleocr-vl": { - "id": "paddlepaddle/paddleocr-vl", - "name": "PaddleOCR-VL", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-10-22", - "last_updated": "2025-10-22", + "z-ai/glm-4.7-original": { + "id": "z-ai/glm-4.7-original", + "name": "GLM 4.7 Original", + "description": "GLM-4.7 is a next-gen GLM series text model with stronger reasoning, long-context chat, and reliable tool use. Routed directly via Z-AI (Zhipu).", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -22169,30 +23964,31 @@ }, "open_weights": true, "limit": { - "context": 16384, - "output": 16384 + "context": 200000, + "input": 200000, + "output": 65535 }, "cost": { - "input": 0.02, - "output": 0.02 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "qwen/qwen3-vl-235b-a22b-thinking": { - "id": "qwen/qwen3-vl-235b-a22b-thinking", - "name": "Qwen3 VL 235B A22B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "z-ai/glm-5-original": { + "id": "z-ai/glm-5-original", + "name": "GLM 5 Original", + "description": "GLM-5 is Zhipu's latest flagship model with advanced reasoning and instruction following. Routed directly via Z-AI (Zhipu).", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "tool_call": true, + "structured_output": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -22200,58 +23996,60 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 0.98, - "output": 3.95 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "qwen/qwen3-vl-30b-a3b-thinking": { - "id": "qwen/qwen3-vl-30b-a3b-thinking", - "name": "qwen/qwen3-vl-30b-a3b-thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "GLM 5 Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-11", - "last_updated": "2025-10-11", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 202800, + "input": 202800, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 1 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "qwen/qwen3-235b-a22b-instruct-2507": { - "id": "qwen/qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "z-ai/glm-4.7-original:thinking": { + "id": "z-ai/glm-4.7-original:thinking", + "name": "GLM 4.7 Original Thinking", + "description": "GLM-4.7 original with extended thinking capabilities for complex reasoning.", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -22262,25 +24060,38 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 200000, + "input": 200000, + "output": 65535 }, "cost": { - "input": 0.09, - "output": 0.58 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "qwen/qwen3-coder-30b-a3b-instruct": { - "id": "qwen/qwen3-coder-30b-a3b-instruct", - "name": "Qwen3 Coder 30b A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "z-ai/glm-5.3": { + "id": "z-ai/glm-5.3", + "name": "GLM 5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-09", - "last_updated": "2025-10-09", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -22291,28 +24102,41 @@ }, "open_weights": true, "limit": { - "context": 160000, - "output": 32768 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.07, - "output": 0.27 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "qwen/qwen3-8b-fp8": { - "id": "qwen/qwen3-8b-fp8", - "name": "Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "z-ai/glm-5.3-flash-cybersecurity": { + "id": "z-ai/glm-5.3-flash-cybersecurity", + "name": "GLM 5.3 Flash Cybersecurity", + "description": "GLM 5.3 Flash Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports always-on reasoning, image understanding, tool calling, and a 1,048,576-token context window.", + "family": "glm", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "release_date": "2026-09-19", + "last_updated": "2026-09-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -22320,25 +24144,28 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 20000 + "context": 1048576, + "input": 1048576, + "output": 32768 }, "cost": { - "input": 0.035, - "output": 0.138 + "input": 0.15, + "output": 0.5, + "cache_read": 0.075 } }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "z-ai/glm-latest": { + "id": "z-ai/glm-latest", + "name": "GLM Latest", + "description": "Compatibility alias that routes to the newest thinking GLM model. Currently routes to GLM 5.2 Thinking.", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-09-10", - "last_updated": "2025-09-10", + "release_date": "2026-05-03", + "last_updated": "2026-05-03", "modalities": { "input": [ "text" @@ -22349,162 +24176,173 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "qwen/qwen3-vl-8b-instruct": { - "id": "qwen/qwen3-vl-8b-instruct", - "name": "qwen/qwen3-vl-8b-instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "inference-net/schematron-v2-small": { + "id": "inference-net/schematron-v2-small", + "name": "Schematron V2 Small", + "description": "Inference.net's 3B-parameter HTML-to-JSON extraction model, focused on accuracy for complex schemas and long web pages. It turns HTML into typed, structured data for web scraping and product catalog ingestion, with a 128K-token context window. Supply HTML in the user message and extraction instructions in a JSON schema via response_format; it does not follow ordinary chat or system prompts.", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": true, - "temperature": true, - "release_date": "2025-10-17", - "last_updated": "2025-10-17", + "release_date": "2026-09-12", + "last_updated": "2026-09-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 0.08, - "output": 0.5 + "input": 0.05, + "output": 0.23, + "cache_read": 0.025 } }, - "qwen/qwen3-omni-30b-a3b-thinking": { - "id": "qwen/qwen3-omni-30b-a3b-thinking", - "name": "Qwen3 Omni 30B A3B Thinking", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "inference-net/schematron-v2-turbo": { + "id": "inference-net/schematron-v2-turbo", + "name": "Schematron V2 Turbo", + "description": "Inference.net's 3B-parameter HTML-to-JSON extraction model, optimized for throughput and low cost on high-volume workloads. It turns HTML into typed, structured data for web scraping and product catalog ingestion, with a 128K-token context window. Supply HTML in the user message and extraction instructions in a JSON schema via response_format; it does not follow ordinary chat or system prompts.", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": true, - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "release_date": "2026-09-12", + "last_updated": "2026-09-12", "modalities": { "input": [ - "text", - "audio", - "video", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 0.25, - "output": 0.97, - "input_audio": 2.2, - "output_audio": 1.788 + "input": 0.03, + "output": 0.15, + "cache_read": 0.015 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7-Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "thinkingmachines/inkling:thinking": { + "id": "thinkingmachines/inkling:thinking", + "name": "Inkling Thinking", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-27", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 1048000, + "input": 1048000, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 3.75, - "cache_read": 0.25, - "cache_write": 1.5625 + "input": 1, + "output": 4.05, + "cache_read": 0.17 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "thinkingmachines/Inkling-Small": { + "id": "thinkingmachines/Inkling-Small", + "name": "Inkling Small", + "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", + "family": "ling", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 524288, + "input": 524288, + "output": 32768 }, "cost": { - "input": 2.11, - "output": 8.45 + "input": 0.5, + "output": 1.2, + "cache_read": 0.1 } }, - "qwen/qwen2.5-7b-instruct": { - "id": "qwen/qwen2.5-7b-instruct", - "name": "Qwen2.5 7B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "thinkingmachines/inkling": { + "id": "thinkingmachines/inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -22512,29 +24350,45 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 32000 + "context": 1048000, + "input": 1048000, + "output": 32768 }, "cost": { - "input": 0.07, - "output": 0.07 + "input": 1, + "output": 4.05, + "cache_read": 0.17 } }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "attachment": false, + "thinkingmachines/Inkling-Small:thinking": { + "id": "thinkingmachines/Inkling-Small:thinking", + "name": "Inkling Small Thinking", + "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", + "family": "ling", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-09-10", - "last_updated": "2025-09-10", + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -22542,87 +24396,93 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 524288, + "input": 524288, "output": 32768 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 0.5, + "output": 1.2, + "cache_read": 0.1 } }, - "qwen/qwen3-235b-a22b-thinking-2507": { - "id": "qwen/qwen3-235b-a22b-thinking-2507", - "name": "Qwen3 235B A22b Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "meta/muse-spark-1.3": { + "id": "meta/muse-spark-1.3", + "name": "Muse Spark 1.3", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", + "family": "muse", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 3 - } - }, - "qwen/qwen-mt-plus": { - "id": "qwen/qwen-mt-plus", - "name": "Qwen MT Plus", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-09-03", - "last_updated": "2025-09-03", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 8192 + "context": 1048576, + "input": 1048576, + "output": 943718 }, "cost": { - "input": 0.25, - "output": 0.75 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "qwen/qwen3-32b-fp8": { - "id": "qwen/qwen3-32b-fp8", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "meta/muse-glimmer-30b": { + "id": "meta/muse-glimmer-30b", + "name": "Muse Glimmer 30B", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -22630,153 +24490,216 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 20000 + "context": 131072, + "input": 131072, + "output": 117964 }, "cost": { - "input": 0.1, - "output": 0.45 + "input": 0.35, + "output": 1.5, + "cache_read": 0.04 } }, - "qwen/qwen3-4b-fp8": { - "id": "qwen/qwen3-4b-fp8", - "name": "Qwen3 4B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "meta/muse-spark-1.1": { + "id": "meta/muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", + "family": "muse", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "release_date": "2026-04-08", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 20000 + "context": 1000000, + "input": 1000000, + "output": 65536 }, "cost": { - "input": 0.03, - "output": 0.03 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "qwen/qwen2.5-vl-72b-instruct": { - "id": "qwen/qwen2.5-vl-72b-instruct", - "name": "Qwen2.5 VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "meta/muse-spark-1.2": { + "id": "meta/muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ "text", "image", - "video" + "video", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 65536 }, "cost": { - "input": 0.8, - "output": 0.8 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "qwen/qwen3-30b-a3b-fp8": { - "id": "qwen/qwen3-30b-a3b-fp8", - "name": "Qwen3 30B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "meta/muse-spark-1.2-contributor": { + "id": "meta/muse-spark-1.2-contributor", + "name": "Muse Spark 1.2 Contributor (Data Used for Training)", + "description": "A much cheaper opt-in version of Muse Spark 1.2 with the same multimodal coding and agentic capabilities. Prompts and outputs sent to this Contributor model may be used by Meta for training and to improve its products; use the standard Muse Spark 1.2 model if you do not want your data used for training.", + "family": "muse", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 20000 + "context": 1000000, + "input": 1000000, + "output": 65536 }, "cost": { - "input": 0.09, - "output": 0.45 + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 } }, - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5-27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "meta/muse-spark-1.3-contributor": { + "id": "meta/muse-spark-1.3-contributor", + "name": "Muse Spark 1.3 Contributor", + "description": "Meta's Muse Spark 1.3 Contributor is a frontier multimodal reasoning model for long-horizon coding and agentic workflows, with strong gains in computer use, browsing, professional tool use, codebase understanding, and million-token retrieval. It accepts text, images, audio, video, and files, supports tool calling and structured output, and always reasons before answering. Prompts and outputs may be used by Meta for training and to improve its products.", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1048576, + "input": 1048576, + "output": 943718 }, "cost": { - "input": 0.3, - "output": 2.4 + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 } }, - "qwen/qwen-2.5-72b-instruct": { - "id": "qwen/qwen-2.5-72b-instruct", - "name": "Qwen 2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "mlabonne/NeuralDaredevil-8B-abliterated": { + "id": "mlabonne/NeuralDaredevil-8B-abliterated", + "name": "Neural Daredevil 8B abliterated", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-10-15", - "last_updated": "2024-10-15", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-12-01", "modalities": { "input": [ "text" @@ -22787,26 +24710,27 @@ }, "open_weights": true, "limit": { - "context": 32000, + "context": 8192, + "input": 8192, "output": 8192 }, "cost": { - "input": 0.38, - "output": 0.4 + "input": 0.44, + "output": 0.44, + "cache_read": 0.22 } }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "NeverSleep/Lumimaid-v0.2-70B": { + "id": "NeverSleep/Lumimaid-v0.2-70B", + "name": "Lumimaid v0.2", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-03", - "last_updated": "2026-02-03", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-07-01", "modalities": { "input": [ "text" @@ -22817,64 +24741,58 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 1.5 + "input": 1, + "output": 1.5, + "cache_read": 0.5 } }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5-35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "nanogpt/coding-router:low": { + "id": "nanogpt/coding-router:low", + "name": "Coding Router Low", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 2 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "qwen/qwen3-coder-480b-a35b-instruct": { - "id": "qwen/qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "nanogpt/coding-router:medium": { + "id": "nanogpt/coding-router:medium", + "name": "Coding Router Medium", + "description": "Automatic model router for matching prompts to suitable backends and budgets", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ "text" @@ -22883,137 +24801,125 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.38, - "output": 1.55 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5-397B-A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "nanogpt/coding-router": { + "id": "nanogpt/coding-router", + "name": "Coding Router", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 64000 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 1.1, + "output": 2.2, + "cache_read": 0.11 } }, - "qwen/qwen3-vl-30b-a3b-instruct": { - "id": "qwen/qwen3-vl-30b-a3b-instruct", - "name": "qwen/qwen3-vl-30b-a3b-instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, - "reasoning": false, + "nanogpt/coding-router:max": { + "id": "nanogpt/coding-router:max", + "name": "Coding Router Max", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-10-11", - "last_updated": "2025-10-11", + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ - "text", - "video", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.7 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "qwen/qwen3-omni-30b-a3b-instruct": { - "id": "qwen/qwen3-omni-30b-a3b-instruct", - "name": "Qwen3 Omni 30B A3B Instruct", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, + "nanogpt/coding-router:high": { + "id": "nanogpt/coding-router:high", + "name": "Coding Router High", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ - "text", - "video", - "audio", - "image" + "text" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 0.97, - "input_audio": 2.2, - "output_audio": 1.788 + "input": 1.1, + "output": 2.2, + "cache_read": 0.11 } }, - "qwen/qwen3-vl-235b-a22b-instruct": { - "id": "qwen/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "LatitudeGames/Wayfarer-Large-70B-Llama-3.3": { + "id": "LatitudeGames/Wayfarer-Large-70B-Llama-3.3", + "name": "Llama 3.3 70B Wayfarer", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-20", + "last_updated": "2025-02-20", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -23021,36 +24927,30 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 0.7, + "output": 0.7, + "cache_read": 0.35 } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5-122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "VongolaChouko/Starcannon-Unleashed-12B-v1.0": { + "id": "VongolaChouko/Starcannon-Unleashed-12B-v1.0", + "name": "Mistral Nemo Starcannon 12b v1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-07-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -23058,25 +24958,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 3.2 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "qwen/qwen3-235b-a22b-fp8": { - "id": "qwen/qwen3-235b-a22b-fp8", - "name": "Qwen3 235B A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "arcee-ai/trinity-large-thinking": { + "id": "arcee-ai/trinity-large-thinking", + "name": "Trinity Large Thinking", + "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", + "family": "trinity", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "release_date": "2026-04-01", + "last_updated": "2026-05-28", "modalities": { "input": [ "text" @@ -23087,62 +24991,70 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 20000 + "context": 262144, + "input": 262144, + "output": 80000 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0.25, + "output": 0.9, + "cache_read": 0.125 } }, - "deepseek/deepseek-r1-0528": { - "id": "deepseek/deepseek-r1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "x-ai/grok-4.20-multi-agent": { + "id": "x-ai/grok-4.20-multi-agent", + "name": "Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 32768 + "context": 2000000, + "input": 2000000, + "output": 131072 }, "cost": { - "input": 0.7, - "output": 2.5, - "cache_read": 0.35 + "input": 2, + "output": 6, + "cache_read": 1 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "x-ai/grok-latest": { + "id": "x-ai/grok-latest", + "name": "Grok Latest", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", "low", "medium", "high", @@ -23151,112 +25063,140 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-05-03", + "last_updated": "2026-05-03", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 393216 + "context": 500000, + "input": 500000, + "output": 450000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 2, + "output": 6, + "cache_read": 0.5 } }, - "deepseek/deepseek-v3.1-terminus": { - "id": "deepseek/deepseek-v3.1-terminus", - "name": "Deepseek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "x-ai/grok-4.7": { + "id": "x-ai/grok-4.7", + "name": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 500000, + "input": 500000, + "output": 500000 }, "cost": { - "input": 0.27, - "output": 1, - "cache_read": 0.135 + "input": 1.6, + "output": 4.8, + "cache_read": 0.4 } }, - "deepseek/deepseek-v3-0324": { - "id": "deepseek/deepseek-v3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 1000000, + "input": 1000000, + "output": 900000 }, "cost": { - "input": 0.27, - "output": 1.12, - "cache_read": 0.135 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "deepseek/deepseek-ocr": { - "id": "deepseek/deepseek-ocr", - "name": "DeepSeek-OCR", - "description": "OCR model for extracting structured text from documents and screenshots", + "x-ai/grok-4.5": { + "id": "x-ai/grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-24", - "last_updated": "2025-10-24", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text", @@ -23266,115 +25206,111 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 500000, + "input": 500000, + "output": 450000 }, "cost": { - "input": 0.03, - "output": 0.03 + "input": 2, + "output": 6, + "cache_read": 0.5 } }, - "deepseek/deepseek-r1-distill-llama-70b": { - "id": "deepseek/deepseek-r1-distill-llama-70b", - "name": "DeepSeek R1 Distill LLama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "x-ai/grok-build-0.1": { + "id": "x-ai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-01-27", - "last_updated": "2025-01-27", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 256000, + "input": 256000, + "output": 230400 }, "cost": { - "input": 0.8, - "output": 0.8 + "input": 1, + "output": 2, + "cache_read": 0.2 } }, - "deepseek/deepseek-r1-turbo": { - "id": "deepseek/deepseek-r1-turbo", - "name": "DeepSeek R1 (Turbo)\t", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "attachment": false, + "x-ai/grok-4.20": { + "id": "x-ai/grok-4.20", + "name": "Grok 4.20", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-03-05", - "last_updated": "2025-03-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 64000, - "output": 16000 - }, - "cost": { - "input": 0.7, - "output": 2.5 - } - }, - "deepseek/deepseek-prover-v2-671b": { - "id": "deepseek/deepseek-prover-v2-671b", - "name": "Deepseek Prover V2 671B", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-04-30", + "structured_output": true, + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 160000, - "output": 160000 + "context": 2000000, + "input": 2000000, + "output": 131072 }, "cost": { - "input": 0.7, - "output": 2.5 + "input": 2, + "output": 6, + "cache_read": 1 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "x-ai/grok-4.6": { + "id": "x-ai/grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -23386,49 +25322,43 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 393216 + "context": 500000, + "input": 500000, + "output": 450000 }, "cost": { - "input": 1.6, - "output": 3.2, - "cache_read": 0.135 + "input": 2, + "output": 6, + "cache_read": 0.5 } }, - "deepseek/deepseek-v3.2-exp": { - "id": "deepseek/deepseek-v3.2-exp", - "name": "Deepseek V3.2 Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B": { + "id": "Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B", + "name": "Nemotron Tenyxchat Storybreaker 70b", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-12-01", "modalities": { "input": [ "text" @@ -23439,27 +25369,30 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 65536 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.27, - "output": 0.41 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "deepseek/deepseek-ocr-2": { - "id": "deepseek/deepseek-ocr-2", - "name": "deepseek/deepseek-ocr-2", - "description": "OCR model for extracting structured text from documents and screenshots", - "attachment": true, + "Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B": { + "id": "Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B", + "name": "Llama 3.05 Storybreaker Ministral 70b", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -23467,25 +25400,27 @@ }, "open_weights": true, "limit": { - "context": 8192, + "context": 16384, + "input": 16384, "output": 8192 }, "cost": { - "input": 0.03, - "output": 0.03 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "deepseek/deepseek-r1-distill-qwen-14b": { - "id": "deepseek/deepseek-r1-distill-qwen-14b", - "name": "DeepSeek R1 Distill Qwen 14B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "deepseek-thinking", + "upstage/solar-pro-3": { + "id": "upstage/solar-pro-3", + "name": "Solar Pro 3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "solar-pro", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "structured_output": false, + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text" @@ -23494,33 +25429,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 131072, + "input": 131072, + "output": 117964 }, "cost": { "input": 0.15, - "output": 0.15 + "output": 0.6, + "cache_read": 0.015 } }, - "deepseek/deepseek-v3.1": { - "id": "deepseek/deepseek-v3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "upstage/solar-pro4": { + "id": "upstage/solar-pro4", + "name": "Solar Pro 4", + "description": "Upstage's flagship model, specialized for agentic use", + "family": "solar-pro", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "knowledge": "2026-02", + "release_date": "2026-08-06", + "last_updated": "2026-08-06", "modalities": { "input": [ "text" @@ -23529,28 +25462,40 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 524288, + "input": 524288, + "output": 131072 }, "cost": { - "input": 0.27, - "output": 1, - "cache_read": 0.135 + "input": 0.03, + "output": 0.12, + "cache_read": 0.006 } }, - "deepseek/deepseek-r1-0528-qwen3-8b": { - "id": "deepseek/deepseek-r1-0528-qwen3-8b", - "name": "DeepSeek R1 0528 Qwen3 8B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "upstage/solar-mini4:thinking": { + "id": "upstage/solar-mini4:thinking", + "name": "Solar Mini 4 Thinking", + "description": "Solar Mini 4 with reasoning enabled for agentic tasks and harder analysis across a 524K-token context window.", + "family": "solar", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2025-05-29", - "last_updated": "2025-05-29", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ "text" @@ -23559,27 +25504,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 524288, + "input": 524288, + "output": 131072 }, "cost": { - "input": 0.06, - "output": 0.09 + "input": 0.05, + "output": 0.2, + "cache_read": 0.005 } }, - "deepseek/deepseek-r1-distill-qwen-32b": { - "id": "deepseek/deepseek-r1-distill-qwen-32b", - "name": "DeepSeek R1 Distill Qwen 32B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "deepseek-thinking", + "upstage/solar-pro4:thinking": { + "id": "upstage/solar-pro4:thinking", + "name": "Solar Pro 4 Thinking", + "description": "Upstage's flagship model, specialized for agentic use", + "family": "solar-pro", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "knowledge": "2026-02", + "release_date": "2026-08-06", + "last_updated": "2026-08-06", "modalities": { "input": [ "text" @@ -23588,26 +25548,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 64000, - "output": 32000 + "context": 524288, + "input": 524288, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 0.3 + "input": 0.03, + "output": 0.12, + "cache_read": 0.006 } }, - "deepseek/deepseek-v3-turbo": { - "id": "deepseek/deepseek-v3-turbo", - "name": "DeepSeek V3 (Turbo)\t", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "upstage/solar-mini4": { + "id": "upstage/solar-mini4", + "name": "Solar Mini 4", + "description": "Upstage's compact 35B-parameter mixture-of-experts model with 3B active parameters and a 524K context window. Built for fast, cost-efficient agentic tasks, with strong Korean and English support.", + "family": "solar", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-03-05", - "last_updated": "2025-03-05", + "structured_output": true, + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ "text" @@ -23616,36 +25579,60 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 64000, - "output": 16000 + "context": 524288, + "input": 524288, + "output": 131072 }, "cost": { - "input": 0.4, - "output": 1.3 + "input": 0.05, + "output": 0.2, + "cache_read": 0.005 } }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "Deepseek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "nano/lumen-stealth": { + "id": "nano/lumen-stealth", + "name": "Lumen Stealth", + "description": "Experimental multimodal model focused on reasoning, creative writing, roleplay, and agentic workflows. Available temporarily for evaluation ahead of public release. During this evaluation, prompts and responses are logged and may be reviewed to evaluate and improve the model. Do not send sensitive or confidential information.", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] }, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "open_weights": false, + "limit": { + "context": 200000, + "input": 200000, + "output": 100000 + }, + "cost": { + "input": 0.05, + "output": 0 + } + }, + "bytedance/doubao-seed-2.1-turbo": { + "id": "bytedance/doubao-seed-2.1-turbo", + "name": "Doubao Seed 2.1 Turbo", + "description": "Fast, lower-cost model in the Doubao Seed 2.1 family for everyday chat, coding assistance, document work, and high-throughput productivity tasks. Supports a 256k context window and up to 128k output tokens. Note: privacy and logging guarantees may be limited.", + "family": "seed", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-06-23", + "last_updated": "2026-06-23", "modalities": { "input": [ "text" @@ -23654,30 +25641,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 65536 + "context": 256000, + "input": 256000, + "output": 128000 }, "cost": { - "input": 0.269, - "output": 0.4, - "cache_read": 0.1345 + "input": 0.5, + "output": 2.5, + "cache_read": 0.25 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", + "bytedance/doubao-seed-2.1-pro": { + "id": "bytedance/doubao-seed-2.1-pro", + "name": "Doubao Seed 2.1 Pro", + "description": "Higher-capability model in the Doubao Seed 2.1 family for agentic coding, long-context analysis, complex instruction following, and productivity workflows. Supports a 256k context window and up to 128k output tokens. Note: privacy and logging guarantees may be limited.", + "family": "seed", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-05-27", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-06-23", + "last_updated": "2026-06-23", "modalities": { "input": [ "text" @@ -23686,37 +25672,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 1, + "output": 5, + "cache_read": 0.5 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "bytedance/doubao-seed-character": { + "id": "bytedance/doubao-seed-character", + "name": "Doubao Seed Character", + "description": "ByteDance's character-focused Doubao Seed model for roleplay, persona consistency, dialogue, and creative character interactions. It supports text and image input with a 128k context window. Requests route through ZenMux to ByteDance; ZenMux does not publish a model-API zero-retention or training guarantee, so avoid sensitive data.", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, + "release_date": "2026-07-18", + "last_updated": "2026-07-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -23724,30 +25715,40 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 131100 + "context": 128000, + "input": 128000, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.1179, + "output": 0.2947, + "cache_read": 0.0236, + "cache_write": 0.0025 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "Minimax M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "liquid/lfm-2.5-2.6b": { + "id": "liquid/lfm-2.5-2.6b", + "name": "LFM2.5 2.6B", + "description": "Liquid AI's compact 2.6B reasoning model for agent workflows, data extraction, RAG, and long-context processing. It supports tool calling and structured output, but Liquid advises against using it for agentic coding. Warning: prompts and responses may be logged and used for model training or service improvement; do not send sensitive data.", + "family": "liquid", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text" @@ -23758,68 +25759,87 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.1, + "output": 0.2, + "cache_read": 0.05 } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "google/gemini-flash-latest": { + "id": "google/gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax-m2.7", - "attachment": false, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -23827,34 +25847,33 @@ }, "open_weights": true, "limit": { - "context": 204800, + "context": 262144, + "input": 262144, "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.1, + "output": 0.45, + "cache_read": 0.05 } }, - "minimax/minimax-m2.5-highspeed": { - "id": "minimax/minimax-m2.5-highspeed", - "name": "MiniMax M2.5 Highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax-m2.5", - "attachment": false, + "google/gemini-3.1-pro-preview-low": { + "id": "google/gemini-3.1-pro-preview-low", + "name": "Gemini 3.1 Pro (Preview Low)", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-02-21", + "last_updated": "2026-02-21", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -23862,59 +25881,83 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 131100 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.03 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "kwaipilot/kat-coder-pro": { - "id": "kwaipilot/kat-coder-pro", - "name": "Kat Coder Pro", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "google/gemini-3.5-flash-thinking": { + "id": "google/gemini-3.5-flash-thinking", + "name": "Gemini 3.5 Flash Thinking", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-01-05", - "last_updated": "2026-01-05", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 128000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 0.083333 } }, - "nousresearch/hermes-2-pro-llama-3-8b": { - "id": "nousresearch/hermes-2-pro-llama-3-8b", - "name": "Hermes 2 Pro Llama 3 8B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": false, + "google/gemma-4-26b-a4b-it:thinking": { + "id": "google/gemma-4-26b-a4b-it:thinking", + "name": "Gemma 4 26B A4B Thinking", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-06-27", - "last_updated": "2024-06-27", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -23922,52 +25965,46 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 262144, + "input": 262144, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.14 + "input": 0.13, + "output": 0.4, + "cache_read": 0.065 } - } - } - }, - "xai": { - "id": "xai", - "env": [ - "XAI_API_KEY" - ], - "npm": "@ai-sdk/xai", - "name": "xAI", - "doc": "https://docs.x.ai/docs/models", - "models": { - "grok-4.20-multi-agent-0309": { - "id": "grok-4.20-multi-agent-0309", - "name": "Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + }, + "google/gemini-flash-lite-latest": { + "id": "google/gemini-flash-lite-latest", + "name": "Gemini Flash Lite Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -23976,47 +26013,47 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 1.25, + "input": 0.3, "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "cache_read": 0.03, + "cache_write": 0.08333 } }, - "grok-4.20-0309-non-reasoning": { - "id": "grok-4.20-0309-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "google/gemini-3.6-flash": { + "id": "google/gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -24025,152 +26062,184 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667 } }, - "grok-4.3": { - "id": "grok-4.3", - "name": "Grok 4.3", - "description": "xAI's Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + "google/diffusiongemma": { + "id": "google/diffusiongemma", + "name": "DiffusionGemma", + "description": "DiffusionGemma is a high-speed diffusion-based version of Gemma 4 26B A4B. It supports optional reasoning and a 262,144-token context window.", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "low", - "medium", - "high" + "xhigh" ] } ], + "tool_call": false, + "structured_output": false, + "release_date": "2026-09-19", + "last_updated": "2026-09-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "input": 262144, + "output": 32768 + }, + "cost": { + "input": 0.05, + "output": 0.15, + "cache_read": 0.025 + } + }, + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 30000 + "context": 262144, + "input": 262144, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.12, + "output": 0.38, + "cache_read": 0.06 } }, - "grok-imagine-image-quality": { - "id": "grok-imagine-image-quality", - "name": "Grok Imagine Image Quality", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", + "google/gemini-3.5-flash-lite": { + "id": "google/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-03", - "last_updated": "2026-04-03", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ - "image", - "pdf" + "text" ] }, "open_weights": false, "limit": { - "context": 8000, - "output": 0 + "context": 1048576, + "input": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "cache_write": 0.08333 } }, - "grok-imagine-video": { - "id": "grok-imagine-video", - "name": "Grok Imagine Video", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", + "google/gemma-4-31b-it:thinking": { + "id": "google/gemma-4-31b-it:thinking", + "name": "Gemma 4 31B Thinking", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-01-28", - "last_updated": "2026-01-28", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "video", - "pdf" + "video" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1024, - "output": 0 + "context": 262144, + "input": 262144, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.35, + "cache_read": 0.05 } }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's latest Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro (Preview)", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -24186,13 +26255,15 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -24200,49 +26271,35 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "output": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "grok-4.20-0309-reasoning": { - "id": "grok-4.20-0309-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", + "google/gemini-3.1-pro-preview-high": { + "id": "google/gemini-3.1-pro-preview-high", + "name": "Gemini 3.1 Pro (Preview High)", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "release_date": "2026-02-21", + "last_updated": "2026-02-21", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -24250,77 +26307,78 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 2.5, + "input": 2, + "output": 12, "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "cache_write": 0.375 } }, - "grok-imagine-image": { - "id": "grok-imagine-image", - "name": "Grok Imagine Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", - "attachment": true, + "google/gemma4-31b-splituntied": { + "id": "google/gemma4-31b-splituntied", + "name": "Gemma 4 31B Split-Untied", + "description": "Blazed-Forge's Split-Untied is a text-only Gemma 4 31B community finetune with an untied BF16 output head, built for creative writing, roleplay, expressive dialogue, and tool use.", + "family": "gemma", + "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-01-28", - "last_updated": "2026-01-28", + "tool_call": true, + "structured_output": false, + "release_date": "2026-09-17", + "last_updated": "2026-09-17", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "image", - "pdf" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8000, - "output": 0 + "context": 262144, + "input": 262144, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.3, + "cache_read": 0.05 } }, - "grok-build-0.1": { - "id": "grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -24328,88 +26386,29 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 4, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 4, - "cache_read": 0.4 - } - } - } - } - }, - "privatemode-ai": { - "id": "privatemode-ai", - "env": [ - "PRIVATEMODE_API_KEY", - "PRIVATEMODE_ENDPOINT" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "http://localhost:8080/v1", - "name": "Privatemode AI", - "doc": "https://docs.privatemode.ai/api/overview", - "models": { - "qwen3-embedding-4b": { - "id": "qwen3-embedding-4b", - "name": "Qwen3-Embedding 4B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-06-06", - "last_updated": "2025-06-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32000, - "output": 2560 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 0.083333 } }, - "gemma-3-27b": { - "id": "gemma-3-27b", - "name": "Gemma 3 27B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "google/gemma-4-26b-a4b-it-cybersecurity": { + "id": "google/gemma-4-26b-a4b-it-cybersecurity", + "name": "Gemma 4 26B A4B Cybersecurity", + "description": "Gemma 4 26B A4B Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports optional reasoning, image understanding, tool calling, and a 262,144-token context window.", "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "structured_output": false, + "release_date": "2026-09-19", + "last_updated": "2026-09-19", "modalities": { "input": [ "text", @@ -24421,25 +26420,28 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1056, + "output": 0.3344, + "cache_read": 0.0528 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "google/gemini-3.7-flash": { + "id": "google/gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -24449,183 +26451,203 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-04", - "last_updated": "2025-08-14", + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667 } }, - "whisper-large-v3": { - "id": "whisper-large-v3", - "name": "Whisper large-v3", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro (Preview Custom Tools)", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2023-09-01", - "last_updated": "2023-09-01", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ + "text", + "image", + "video", "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, - "output": 4096 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "qwen3-coder-30b-a3b": { - "id": "qwen3-coder-30b-a3b", - "name": "Qwen3-Coder 30B-A3B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash (Preview)", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - } - } - }, - "drun": { - "id": "drun", - "env": [ - "DRUN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://chat.d.run/v1", - "name": "D.Run (China)", - "doc": "https://www.d.run", - "models": { - "public/deepseek-v3": { - "id": "public/deepseek-v3", - "name": "DeepSeek V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-12-26", - "last_updated": "2024-12-26", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.28, - "output": 1.1 + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "public/deepseek-r1": { - "id": "public/deepseek-r1", - "name": "DeepSeek R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "google/gemini-pro-latest": { + "id": "google/gemini-pro-latest", + "name": "Gemini Pro Latest", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "structured_output": true, + "release_date": "2026-03-29", + "last_updated": "2026-03-29", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.55, - "output": 2.2 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "public/minimax-m25": { - "id": "public/minimax-m25", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "google/gemini-3.8-flash": { + "id": "google/gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, + "structured_output": true, "temperature": true, - "release_date": "2025-03-01", - "last_updated": "2025-03-01", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -24633,91 +26655,95 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.29, - "output": 1.16 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667 } - } - } - }, - "alibaba-token-plan-cn": { - "id": "alibaba-token-plan-cn", - "env": [ - "ALIBABA_TOKEN_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://token-plan.cn-beijing.maas.aliyuncs.com/compatible-mode/v1", - "name": "Alibaba Token Plan (China)", - "doc": "https://www.alibabacloud.com/help/zh/model-studio/token-plan-overview", - "models": { - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + }, + "google/gemini-3-flash-preview-thinking": { + "id": "google/gemini-3-flash-preview-thinking", + "name": "Gemini 3 Flash Thinking", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": false, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -24725,36 +26751,34 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 0.08333 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "TEE/gemma-4-31b-it": { + "id": "TEE/gemma-4-31b-it", + "name": "Gemma 4 31B IT TEE", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 } ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" @@ -24763,41 +26787,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "input": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.15, + "output": 0.46, + "cache_read": 0.075 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "TEE/qwen3.8-27b": { + "id": "TEE/qwen3.8-27b", + "name": "Qwen3.8 27B TEE", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", @@ -24811,35 +26825,38 @@ "open_weights": true, "limit": { "context": 262144, + "input": 262144, "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.4, + "output": 3, + "cache_read": 0.15 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "TEE/nemotron-3.5-lightning": { + "id": "TEE/nemotron-3.5-lightning", + "name": "Nvidia Nemotron 3.5 Lightning TEE", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ "text" @@ -24850,40 +26867,42 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.08, + "output": 0.2, + "cache_read": 0.04 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "TEE/glm-5.3-flash": { + "id": "TEE/glm-5.3-flash", + "name": "GLM 5.3 Flash TEE", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -24891,70 +26910,79 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "wan2.7-image-pro": { - "id": "wan2.7-image-pro", - "name": "Wan2.7 Image Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": false, + "TEE/qwen3.5-27b": { + "id": "TEE/qwen3.5-27b", + "name": "Qwen3.5 27B TEE", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": false, "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 0 + "context": 262144, + "input": 262144, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 2.4, + "cache_read": 0.15 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "TEE/muse-glimmer-30b": { + "id": "TEE/muse-glimmer-30b", + "name": "Muse Glimmer 30B TEE", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", + "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -24962,84 +26990,88 @@ }, "open_weights": true, "limit": { - "context": 1000000, + "context": 131072, + "input": 131072, "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.35, + "output": 1.5, + "cache_read": 0.04 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "TEE/kimi-k3": { + "id": "TEE/kimi-k3", + "name": "Kimi K3 TEE", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 1048576, + "input": 1048576, + "output": 65535 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 3, + "output": 15, + "cache_read": 1.5 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "TEE/deepseek-v4.1-flash": { + "id": "TEE/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash TEE", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -25047,60 +27079,92 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 1048576, + "input": 1048576, + "output": 384000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.65, + "output": 1.45, + "cache_read": 0.13 } }, - "qwen-image-2.0": { - "id": "qwen-image-2.0", - "name": "Qwen Image 2.0", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", + "TEE/llama3-3-70b": { + "id": "TEE/llama3-3-70b", + "name": "Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "structured_output": false, + "release_date": "2025-07-03", + "last_updated": "2025-07-03", "modalities": { "input": [ "text" ], "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "input": 128000, + "output": 16384 + }, + "cost": { + "input": 1.75, + "output": 2.75, + "cache_read": 1.75 + } + }, + "TEE/kimi-k2.6": { + "id": "TEE/kimi-k2.6", + "name": "Kimi K2.6 TEE", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", "image" + ], + "output": [ + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 0 + "context": 262144, + "input": 262144, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.5, + "output": 5.25, + "cache_read": 0.375 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "TEE/gemma-4-26b-a4b-uncensored": { + "id": "TEE/gemma-4-26b-a4b-uncensored", + "name": "Gemma 4 26B A4B Uncensored TEE", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "structured_output": true, + "release_date": "2026-05-23", + "last_updated": "2026-05-23", "modalities": { "input": [ "text" @@ -25111,42 +27175,32 @@ }, "open_weights": true, "limit": { - "context": 196608, - "input": 196601, - "output": 24576 + "context": 65536, + "input": 65536, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.15, + "output": 0.7, + "cache_read": 0.075 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "TEE/gemma4-31b:thinking": { + "id": "TEE/gemma4-31b:thinking", + "name": "Gemma 4 31B Thinking TEE", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "structured_output": true, + "release_date": "2026-05-02", + "last_updated": "2026-05-02", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -25155,91 +27209,103 @@ "open_weights": true, "limit": { "context": 262144, - "output": 16384 + "input": 262144, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.4, + "output": 1, + "cache_read": 0.4 } }, - "qwen-image-2.0-pro": { - "id": "qwen-image-2.0-pro", - "name": "Qwen Image 2.0 Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", - "attachment": false, + "TEE/gemma4-31b": { + "id": "TEE/gemma4-31b", + "name": "Gemma 4 31B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "tool_call": true, + "structured_output": true, + "release_date": "2026-04-04", + "last_updated": "2026-04-04", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 0 + "context": 262144, + "input": 262144, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 1, + "cache_read": 0.4 } }, - "wan2.7-image": { - "id": "wan2.7-image", - "name": "Wan2.7 Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "TEE/qwen3.6-35b-a3b": { + "id": "TEE/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B A3B TEE", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 0 + "context": 262144, + "input": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 1.27, + "cache_read": 0.1 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "TEE/glm-5.2:thinking": { + "id": "TEE/glm-5.2:thinking", + "name": "GLM 5.2 Thinking TEE", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -25248,36 +27314,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202752, - "output": 16384 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.7 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "TEE/qwen3.5-397b-a17b": { + "id": "TEE/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B TEE", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-03", - "last_updated": "2025-12-05", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text" @@ -25288,81 +27348,70 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 262144, + "input": 262144, "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.55, + "output": 3.5, + "cache_read": 0.275 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", + "TEE/glm-5.2": { + "id": "TEE/glm-5.2", + "name": "GLM 5.2 TEE", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.7 } - } - } - }, - "moonshotai": { - "id": "moonshotai", - "env": [ - "MOONSHOT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.moonshot.ai/v1", - "name": "Moonshot AI", - "doc": "https://platform.moonshot.ai/docs/api/chat", - "models": { - "kimi-k2-0905-preview": { - "id": "kimi-k2-0905-preview", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + }, + "TEE/glm-5.1-thinking": { + "id": "TEE/glm-5.1-thinking", + "name": "GLM 5.1 Thinking TEE", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -25373,31 +27422,37 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 202752, + "input": 202752, + "output": 65535 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 1.5, + "output": 5.25, + "cache_read": 0.3 } }, - "kimi-k2-thinking-turbo": { - "id": "kimi-k2-thinking-turbo", - "name": "Kimi K2 Thinking Turbo", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "TEE/glm-5.1": { + "id": "TEE/glm-5.1", + "name": "GLM 5.1 TEE", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -25408,37 +27463,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 202752, + "input": 202752, + "output": 65535 }, "cost": { - "input": 1.15, - "output": 8, - "cache_read": 0.15 + "input": 1.5, + "output": 5.25, + "cache_read": 0.3 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "TEE/qwen2.5-vl-72b-instruct": { + "id": "TEE/qwen2.5-vl-72b-instruct", + "name": "Qwen2.5 VL 72B TEE", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-02-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -25446,34 +27495,33 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 65536, + "input": 65536, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.7, + "output": 0.7, + "cache_read": 0.35 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", - "attachment": false, + "TEE/qwen3.6-27b": { + "id": "TEE/qwen3.6-27b", + "name": "Qwen3.6 27B TEE", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -25482,26 +27530,28 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "input": 262144, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 0.32, + "output": 2.7, + "cache_read": 0.16 } }, - "kimi-k2-0711-preview": { - "id": "kimi-k2-0711-preview", - "name": "Kimi K2 0711", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "TEE/deepseek-v3.2": { + "id": "TEE/deepseek-v3.2", + "name": "DeepSeek V3.2 TEE", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-07-14", - "last_updated": "2025-07-14", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -25512,27 +27562,28 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 164000, + "input": 164000, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 0.5, + "output": 1, + "cache_read": 0.25 } }, - "kimi-k2-turbo-preview": { - "id": "kimi-k2-turbo-preview", - "name": "Kimi K2 Turbo", - "description": "Fast Kimi model for responsive chat, coding help, and agent loops", - "family": "kimi-k2", + "TEE/gpt-oss-120b": { + "id": "TEE/gpt-oss-120b", + "name": "GPT-OSS 120B TEE", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -25543,41 +27594,41 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "input": 131072, + "output": 16384 }, "cost": { - "input": 2.4, - "output": 10, - "cache_read": 0.6 + "input": 2, + "output": 2, + "cache_read": 2 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "TEE/glm-5.3": { + "id": "TEE/glm-5.3", + "name": "GLM 5.3 TEE", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -25585,41 +27636,34 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "TEE/kimi-k2.7-code": { + "id": "TEE/kimi-k2.7-code", + "name": "Kimi K2.7 Code TEE", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, + "structured_output": false, + "temperature": false, "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -25628,36 +27672,29 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "input": 262144, + "output": 65536 }, "cost": { "input": 0.95, "output": 4, - "cache_read": 0.16 + "cache_read": 0.19 } }, - "kimi-k2.7-code-highspeed": { - "id": "kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code HighSpeed", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2": { + "id": "EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2", + "name": "EVA-Qwen2.5-72B-v0.2", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -25665,50 +27702,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 + "input": 0.799, + "output": 0.799, + "cache_read": 0.3995 } - } - } - }, - "fireworks-ai": { - "id": "fireworks-ai", - "env": [ - "FIREWORKS_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.fireworks.ai/inference/v1/", - "name": "Fireworks AI", - "doc": "https://fireworks.ai/docs/", - "models": { - "accounts/fireworks/routers/kimi-k2p6-turbo": { - "id": "accounts/fireworks/routers/kimi-k2p6-turbo", - "name": "Kimi K2.6 Turbo", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + }, + "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0": { + "id": "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0", + "name": "EVA Llama 3.33 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -25716,41 +27733,27 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.3 + "input": 2.006, + "output": 2.006, + "cache_read": 1.003 } }, - "accounts/fireworks/routers/glm-5p2-fast": { - "id": "accounts/fireworks/routers/glm-5p2-fast", - "name": "GLM 5.2 Fast", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", + "EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2": { + "id": "EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2", + "name": "EVA-Qwen2.5-32B-v0.2", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-26", - "last_updated": "2026-06-26", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text" @@ -25761,38 +27764,30 @@ }, "open_weights": true, "limit": { - "context": 1048575, - "output": 131072 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 2.1, - "output": 6.6, - "cache_read": 0.21 + "input": 0.799, + "output": 0.799, + "cache_read": 0.3995 } }, - "accounts/fireworks/routers/kimi-k2p7-code-fast": { - "id": "accounts/fireworks/routers/kimi-k2p7-code-fast", - "name": "Kimi K2.7 Code Fast", - "description": "Kimi coding model for software agents, refactors, and repository reasoning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-16", + "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1": { + "id": "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1", + "name": "EVA-LLaMA-3.33-70B-v0.1", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -25800,34 +27795,27 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 + "input": 2.006, + "output": 2.006, + "cache_read": 1.003 } }, - "accounts/fireworks/routers/glm-5p1-fast": { - "id": "accounts/fireworks/routers/glm-5p1-fast", - "name": "GLM 5.1 Fast", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", + "nothingiisreal/L3.1-70B-Celeste-V0.1-BF16": { + "id": "nothingiisreal/L3.1-70B-Celeste-V0.1-BF16", + "name": "Llama 3.1 70B Celeste v0.1", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -25838,38 +27826,30 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 131072 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 2.8, - "output": 8.8, - "cache_read": 0.52 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "accounts/fireworks/routers/kimi-k2p6-fast": { - "id": "accounts/fireworks/routers/kimi-k2p6-fast", - "name": "Kimi K2.6 Fast", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-06-05", + "MarinaraSpaghetti/NemoMix-Unleashed-12B": { + "id": "MarinaraSpaghetti/NemoMix-Unleashed-12B", + "name": "NemoMix 12B Unleashed", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-07-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -25877,43 +27857,28 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.3 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "accounts/fireworks/models/deepseek-v4-flash": { - "id": "accounts/fireworks/models/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "pamanseau/OpenReasoning-Nemotron-32B": { + "id": "pamanseau/OpenReasoning-Nemotron-32B", + "name": "OpenReasoning Nemotron 32B", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-06-16", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ "text" @@ -25924,43 +27889,28 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 32768, + "input": 32768, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0.1, + "output": 0.4, + "cache_read": 0.05 } }, - "accounts/fireworks/models/deepseek-v4-pro": { - "id": "accounts/fireworks/models/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated": { + "id": "huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated", + "name": "DeepSeek R1 Qwen Abliterated", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -25971,39 +27921,27 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.145 + "input": 1.4, + "output": 1.4, + "cache_read": 0.7 } }, - "accounts/fireworks/models/minimax-m2p7": { - "id": "accounts/fireworks/models/minimax-m2p7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "huihui-ai/Qwen2.5-32B-Instruct-abliterated": { + "id": "huihui-ai/Qwen2.5-32B-Instruct-abliterated", + "name": "Qwen 2.5 32B Abliterated", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-12", - "last_updated": "2026-04-12", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-06", + "last_updated": "2025-01-06", "modalities": { "input": [ "text" @@ -26014,39 +27952,27 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 196608 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.7, + "output": 0.7, + "cache_read": 0.35 } }, - "accounts/fireworks/models/minimax-m3": { - "id": "accounts/fireworks/models/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "huihui-ai/Llama-3.3-70B-Instruct-abliterated": { + "id": "huihui-ai/Llama-3.3-70B-Instruct-abliterated", + "name": "Llama 3.3 70B Instruct abliterated", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ "text" @@ -26057,38 +27983,31 @@ }, "open_weights": true, "limit": { - "context": 512000, - "output": 512000 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.7, + "output": 0.7, + "cache_read": 0.35 } }, - "accounts/fireworks/models/kimi-k2p6": { - "id": "accounts/fireworks/models/kimi-k2p6", - "name": "Kimi K2.6", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": true, + "huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated": { + "id": "huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated", + "name": "DeepSeek R1 Llama 70B Abliterated", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -26096,82 +28015,68 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.7, + "output": 0.7, + "cache_read": 0.35 } }, - "accounts/fireworks/models/qwen3p7-plus": { - "id": "accounts/fireworks/models/qwen3p7-plus", - "name": "Qwen 3.7 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1 - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "Gryphe/MythoMax-L2-13b": { + "id": "Gryphe/MythoMax-L2-13b", + "name": "MythoMax 13B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 4096, + "input": 4096, + "output": 3686 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08 + "input": 0.1003, + "output": 0.1003, + "cache_read": 0.05015 } }, - "accounts/fireworks/models/glm-5p1": { - "id": "accounts/fireworks/models/glm-5p1", - "name": "GLM 5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "ibm-granite/granite-4.2-8b": { + "id": "ibm-granite/granite-4.2-8b", + "name": "Granite 4.2 8B", + "description": "IBM Granite 4.2 8B is an Apache 2.0-licensed dense model with native step-by-step reasoning and specialized training for agentic work. It can plan before acting, sequence tools, navigate codebases, work in terminals, and verify results across coding, search, mathematics, science, and complex instruction-following tasks.", + "family": "granite", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "structured_output": true, + "release_date": "2026-08-31", + "last_updated": "2026-08-31", "modalities": { "input": [ "text" @@ -26182,44 +28087,30 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 131072 + "context": 131072, + "input": 131072, + "output": 117964 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.1, + "output": 0.15, + "cache_read": 0.05 } }, - "accounts/fireworks/models/glm-5p2": { - "id": "accounts/fireworks/models/glm-5p2", - "name": "GLM 5.2", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "TheDrummer/skyfall-36b-v2": { + "id": "TheDrummer/skyfall-36b-v2", + "name": "TheDrummer Skyfall 36B V2", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-03-10", + "last_updated": "2025-03-10", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -26227,39 +28118,31 @@ }, "open_weights": true, "limit": { - "context": 1048575, - "output": 131072 + "context": 32768, + "input": 32768, + "output": 29491 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.14 + "input": 0.55, + "output": 0.8, + "cache_read": 0.25 } }, - "accounts/fireworks/models/gpt-oss-120b": { - "id": "accounts/fireworks/models/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "TheDrummer/Artemis-v1.1": { + "id": "TheDrummer/Artemis-v1.1", + "name": "TheDrummer/Artemis v1.1", + "description": "TheDrummer's Artemis v1.1 is a Gemma 4 31B fine-tune for creative writing, expressive dialogue, and roleplay, with optional thinking and a 262K context window.", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2026-06-16", + "structured_output": false, + "release_date": "2026-09-06", + "last_updated": "2026-09-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -26267,36 +28150,26 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 262144, + "input": 262144, "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 0.1, + "output": 0.45, + "cache_read": 0.05 } }, - "accounts/fireworks/models/gpt-oss-20b": { - "id": "accounts/fireworks/models/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "TheDrummer/Anubis-70B-v1": { + "id": "TheDrummer/Anubis-70B-v1", + "name": "Anubis 70B v1", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" @@ -26307,38 +28180,29 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 65536, + "input": 65536, + "output": 16384 }, "cost": { - "input": 0.07, - "output": 0.3, - "cache_read": 0.035 + "input": 0.31, + "output": 0.31, + "cache_read": 0.155 } }, - "accounts/fireworks/models/kimi-k2p7-code": { - "id": "accounts/fireworks/models/kimi-k2p7-code", - "name": "Kimi K2.7 Code", - "description": "Kimi coding model for software agents, refactors, and repository reasoning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-16", + "TheDrummer/Anubis-70B-v1.1": { + "id": "TheDrummer/Anubis-70B-v1.1", + "name": "Anubis 70B v1.1", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -26346,44 +28210,26 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 32000, + "input": 32000, + "output": 16384 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.31, + "output": 0.31, + "cache_read": 0.155 } - } - } - }, - "vultr": { - "id": "vultr", - "env": [ - "VULTR_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.vultrinference.com/v1", - "name": "Vultr", - "doc": "https://api.vultrinference.com/", - "models": { - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + }, + "TheDrummer/Cydonia-24B-v4.3": { + "id": "TheDrummer/Cydonia-24B-v4.3", + "name": "The Drummer Cydonia 24B v4.3", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-25", + "last_updated": "2025-12-25", "modalities": { "input": [ "text" @@ -26394,33 +28240,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.12, + "output": 0.15, + "cache_read": 0.06 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "TheDrummer/Magidonia-24B-v4.3": { + "id": "TheDrummer/Magidonia-24B-v4.3", + "name": "The Drummer Magidonia 24B v4.3", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-25", + "last_updated": "2025-12-25", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -26428,33 +28270,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 2 + "input": 0.1003, + "output": 0.1207, + "cache_read": 0.05015 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "TheDrummer/Cydonia-24B-v4.1": { + "id": "TheDrummer/Cydonia-24B-v4.1", + "name": "The Drummer Cydonia 24B v4.1", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "structured_output": false, + "release_date": "2025-08-19", + "last_updated": "2025-08-19", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -26462,30 +28300,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "input": 131072, + "output": 117964 }, "cost": { - "input": 0.3, - "output": 2 + "input": 0.35, + "output": 0.55, + "cache_read": 0.16 } }, - "XiaomiMiMo/MiMo-V2.5-Pro": { - "id": "XiaomiMiMo/MiMo-V2.5-Pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "TheDrummer/UnslopNemo-12B-v4.1": { + "id": "TheDrummer/UnslopNemo-12B-v4.1", + "name": "UnslopNemo 12b v4", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -26493,27 +28331,26 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 8192, + "input": 8192, + "output": 26214 }, "cost": { - "input": 0.55, - "output": 1.65 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16": { - "id": "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16", - "name": "NVIDIA Nemotron 3 Nano Omni", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", + "TheDrummer/Cydonia-24B-v2": { + "id": "TheDrummer/Cydonia-24B-v2", + "name": "The Drummer Cydonia 24B v2", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-17", + "last_updated": "2025-02-17", "modalities": { "input": [ "text" @@ -26524,27 +28361,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.13, - "output": 0.38 + "input": 0.1003, + "output": 0.1207, + "cache_read": 0.05015 } }, - "nvidia/DeepSeek-V3.2-NVFP4": { - "id": "nvidia/DeepSeek-V3.2-NVFP4", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "TheDrummer/Rocinante-12B-v1.1": { + "id": "TheDrummer/Rocinante-12B-v1.1", + "name": "Rocinante 12b", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" @@ -26555,27 +28391,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.55, - "output": 1.65 + "input": 0.408, + "output": 0.595, + "cache_read": 0.204 } }, - "nvidia/Nemotron-Cascade-2-30B-A3B": { - "id": "nvidia/Nemotron-Cascade-2-30B-A3B", - "name": "NVIDIA Nemotron Cascade 2", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", + "TheDrummer/Cydonia-24B-v4": { + "id": "TheDrummer/Cydonia-24B-v4", + "name": "The Drummer Cydonia 24B v4", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ "text" @@ -26586,33 +28421,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.2006, + "output": 0.2414, + "cache_read": 0.1003 } }, - "zai-org/GLM-5.2-FP8": { - "id": "zai-org/GLM-5.2-FP8", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -26620,28 +28453,28 @@ }, "open_weights": true, "limit": { - "context": 393216, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.85, - "output": 3.1 + "input": 0.2, + "output": 0.7, + "cache_read": 0.1 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "deepseek-ai/deepseek-v3.2-exp-thinking": { + "id": "deepseek-ai/deepseek-v3.2-exp-thinking", + "name": "DeepSeek V3.2 Exp Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-09-29", "modalities": { "input": [ "text" @@ -26652,26 +28485,28 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 163840, + "input": 163840, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1 + "input": 0.28, + "output": 0.42, + "cache_read": 0.14 } }, - "MiniMaxAI/MiniMax-M2.7": { - "id": "MiniMaxAI/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": false, + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -26682,158 +28517,89 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 163840, + "input": 163840, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.4, + "output": 1.7, + "cache_read": 0.2 } - } - } - }, - "302ai": { - "id": "302ai", - "env": [ - "302AI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.302.ai/v1", - "name": "302.AI", - "doc": "https://doc.302.ai", - "models": { - "gpt-5.4-mini-2026-03-17": { - "id": "gpt-5.4-mini-2026-03-17", - "name": "gpt-5.4-mini-2026-03-17", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + }, + "deepseek-ai/DeepSeek-V3.1-Terminus": { + "id": "deepseek-ai/DeepSeek-V3.1-Terminus", + "name": "DeepSeek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-19", - "last_updated": "2026-03-19", + "release_date": "2025-08-02", + "last_updated": "2025-08-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.75, - "output": 4.5 + "input": 0.25, + "output": 0.7, + "cache_read": 0.125 } }, - "chatgpt-4o-latest": { - "id": "chatgpt-4o-latest", - "name": "chatgpt-4o-latest", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "deepseek-ai/deepseek-v3.2-exp": { + "id": "deepseek-ai/deepseek-v3.2-exp", + "name": "DeepSeek V3.2 Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-08-08", - "last_updated": "2024-08-08", + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 5, - "output": 15 - } - }, - "gpt-5.4-nano-2026-03-17": { - "id": "gpt-5.4-nano-2026-03-17", - "name": "gpt-5.4-nano-2026-03-17", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-19", - "last_updated": "2026-03-19", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 163840, + "input": 163840, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 1.25 + "input": 0.28, + "output": 0.42, + "cache_read": 0.14 } }, - "kimi-k2-0905-preview": { - "id": "kimi-k2-0905-preview", - "name": "kimi-k2-0905-preview", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "deepseek-ai/DeepSeek-V3.1:thinking": { + "id": "deepseek-ai/DeepSeek-V3.1:thinking", + "name": "DeepSeek V3.1 Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek-thinking", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-08-21", "modalities": { "input": [ "text" @@ -26842,88 +28608,113 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.632, - "output": 2.53 + "input": 0.2, + "output": 0.7, + "cache_read": 0.1 } }, - "grok-4.20-beta-0309-non-reasoning": { - "id": "grok-4.20-beta-0309-non-reasoning", - "name": "grok-4.20-beta-0309-non-reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "attachment": true, + "deepseek-ai/DeepSeek-V3.1-Terminus:thinking": { + "id": "deepseek-ai/DeepSeek-V3.1-Terminus:thinking", + "name": "DeepSeek V3.1 Terminus (Thinking)", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek-thinking", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2025-09-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 2, - "output": 6 + "input": 0.25, + "output": 0.7, + "cache_read": 0.125 } }, - "gemini-2.5-flash-nothink": { - "id": "gemini-2.5-flash-nothink", - "name": "gemini-2.5-flash-nothink", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": false, - "tool_call": true, + "stepfun-ai/step-3.5-flash-2603": { + "id": "stepfun-ai/step-3.5-flash-2603", + "name": "Step 3.5 Flash 2603", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, "knowledge": "2025-01", - "release_date": "2025-06-24", - "last_updated": "2025-06-24", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 262144, + "input": 262144, "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 0.1, + "output": 0.3, + "cache_read": 0.05 } }, - "qwen-plus": { - "id": "qwen-plus", - "name": "Qwen-Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "stepfun-ai/step-3.5-flash": { + "id": "stepfun-ai/step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", "attachment": false, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -26932,36 +28723,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 262144, + "input": 262144, + "output": 65536 }, "cost": { - "input": 0.12, - "output": 1.2 + "input": 0.1, + "output": 0.3, + "cache_read": 0.05 } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "glm-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "MiniMaxAI/MiniMax-M1-80k": { + "id": "MiniMaxAI/MiniMax-M1-80k", + "name": "MiniMax M1 80K", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-08", + "last_updated": "2025-06-16", "modalities": { "input": [ "text" @@ -26970,27 +28754,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, + "context": 1000000, + "input": 1000000, "output": 131072 }, "cost": { - "input": 0.286, - "output": 1.142 + "input": 0.6052, + "output": 2.4225, + "cache_read": 0.3026 } }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "qwen3-235b-a22b-instruct-2507", - "description": "Tool-capable chat model for instruction following and agentic application workflows", + "mistralai/ministral-8b-2512": { + "id": "mistralai/ministral-8b-2512", + "name": "Ministral 8B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-30", - "last_updated": "2025-07-30", + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-04", + "last_updated": "2025-12-04", "modalities": { "input": [ "text" @@ -26999,38 +28785,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 65536 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.29, - "output": 1.143 + "input": 0.15, + "output": 0.15, + "cache_read": 0.075 } }, - "glm-4.5v": { - "id": "glm-4.5v", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "mistralai/mistral-small-4-119b-2603:thinking": { + "id": "mistralai/mistral-small-4-119b-2603:thinking", + "name": "Mistral Small 4 119B Thinking", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-12", - "last_updated": "2025-08-12", + "structured_output": true, + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -27038,191 +28828,158 @@ }, "open_weights": true, "limit": { - "context": 64000, + "context": 262144, + "input": 262144, "output": 16384 }, "cost": { - "input": 0.29, - "output": 0.86 + "input": 0.4, + "output": 1.4, + "cache_read": 0.2 } }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "claude-opus-4-5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "mistralai/mistral-medium-3.5": { + "id": "mistralai/mistral-medium-3.5", + "name": "Mistral Medium 3.5", + "description": "Mistral Medium 3.5 is a 128B dense open-weights flagship model for instruction-following, reasoning, coding, long-horizon agentic work, tool use, structured output, and multimodal prompts. It supports a 256k context window and configurable reasoning effort.", + "family": "mistral-medium", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "structured_output": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 5, - "output": 25 + "input": 1.5, + "output": 7.5, + "cache_read": 0.75 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "gemini-2.5-pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "mistralai/mistral-small-24b-instruct-2501": { + "id": "mistralai/mistral-small-24b-instruct-2501", + "name": "Mistral Small 24B", + "description": "Mistral Small 24B hosted by IONOS in Berlin, Germany. Zero data retention.", + "family": "mistral-small", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "structured_output": false, + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.1155, + "output": 0.3465 } }, - "gpt-5": { - "id": "gpt-5", - "name": "gpt-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "mistralai/mixtral-8x22b-instruct-v0.1": { + "id": "mistralai/mixtral-8x22b-instruct-v0.1", + "name": "Mixtral 8x22B", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mixtral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 65536, + "input": 65536, + "output": 52428 }, "cost": { - "input": 1.25, - "output": 10 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "claude-haiku-4-5-20251001", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-16", - "last_updated": "2025-10-16", + "mistralai/mistral-saba": { + "id": "mistralai/mistral-saba", + "name": "Mistral Saba", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-02-17", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 32768, + "input": 32768, + "output": 26214 }, "cost": { - "input": 1, - "output": 5 + "input": 0.1989, + "output": 0.595, + "cache_read": 0.09945 } }, - "kimi-k2-thinking-turbo": { - "id": "kimi-k2-thinking-turbo", - "name": "kimi-k2-thinking-turbo", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "mistralai/mistral-medium-3.1": { + "id": "mistralai/mistral-medium-3.1", + "name": "Mistral Medium 3.1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", + "reasoning": false, + "tool_call": false, + "structured_output": false, "release_date": "2025-09-05", "last_updated": "2025-09-05", "modalities": { @@ -27235,52 +28992,53 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 1.265, - "output": 9.119 + "input": 0.4, + "output": 2, + "cache_read": 0.2 } }, - "claude-3-5-haiku-20241022": { - "id": "claude-3-5-haiku-20241022", - "name": "claude-3-5-haiku-20241022", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "mistralai/mistral-small-3.2-24b-instruct": { + "id": "mistralai/mistral-small-3.2-24b-instruct", + "name": "Mistral Small 3.2 24B (2506)", + "description": "The latest iteration of Mistral Small, version 3.2 (2506) brings enhanced performance and capabilities. With 24 billion parameters, this model delivers state-of-the-art results across text generation tasks with improved efficiency.", + "family": "mistral-small", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.8, - "output": 4 + "input": 0.2, + "output": 0.4, + "cache_read": 0.1 } }, - "glm-4.5": { - "id": "glm-4.5", - "name": "GLM-4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "mistralai/mistral-small-4-119b-2603": { + "id": "mistralai/mistral-small-4-119b-2603", + "name": "Mistral Small 4 119B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -27288,13 +29046,13 @@ } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "structured_output": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -27302,39 +29060,30 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.286, - "output": 1.142 + "input": 0.4, + "output": 1.4, + "cache_read": 0.2 } }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "gpt-5-pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-08", - "last_updated": "2025-10-08", + "mistralai/mistral-large": { + "id": "mistralai/mistral-large", + "name": "Mistral Large 2411", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-02-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -27342,26 +29091,36 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 272000 + "context": 128000, + "input": 128000, + "output": 102400 }, "cost": { - "input": 15, - "output": 120 + "input": 2.006, + "output": 6.001, + "cache_read": 0.2 } }, - "grok-4.20-beta-0309-reasoning": { - "id": "grok-4.20-beta-0309-reasoning", - "name": "grok-4.20-beta-0309-reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "mistralai/mistral-medium-3.5:thinking": { + "id": "mistralai/mistral-medium-3.5:thinking", + "name": "Mistral Medium 3.5 Thinking", + "description": "Mistral Medium 3.5 with reasoning enabled by default (reasoning_effort=high), for complex coding, agentic, and multi-step reasoning prompts.", + "family": "mistral-medium", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "structured_output": true, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", @@ -27371,93 +29130,95 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 2, - "output": 6 + "input": 1.5, + "output": 7.5, + "cache_read": 0.75 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "gemini-2.5-flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "mistralai/devstral-small-2505": { + "id": "mistralai/devstral-small-2505", + "name": "Mistral Devstral Small 2505", + "description": "OpenHands+Devstral is 100% local 100% open, and is SOTA for the category on SWE-Bench Verified: 46.8% accuracy.", + "family": "devstral", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-02", + "last_updated": "2025-08-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 0.06, + "output": 0.06, + "cache_read": 0.03 } }, - "gpt-4o": { - "id": "gpt-4o", - "name": "gpt-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "mistralai/devstral-2-123b-instruct-2512": { + "id": "mistralai/devstral-2-123b-instruct-2512", + "name": "Devstral 2 123B", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "input": 262144, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.4, + "output": 1.4, + "cache_read": 0.2 } }, - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": false, + "mistralai/mistral-medium-3": { + "id": "mistralai/mistral-medium-3", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-12-19", - "last_updated": "2025-12-19", + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -27465,25 +29226,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.4, + "output": 2, + "cache_read": 0.2 } }, - "gemini-2.5-flash-lite-preview-09-2025": { - "id": "gemini-2.5-flash-lite-preview-09-2025", - "name": "gemini-2.5-flash-lite-preview-09-2025", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "mistralai/ministral-14b-instruct-2512": { + "id": "mistralai/ministral-14b-instruct-2512", + "name": "Ministral 3 14B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-09-26", - "last_updated": "2025-09-26", + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", @@ -27493,98 +29256,92 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { "input": 0.1, - "output": 0.4 + "output": 0.4, + "cache_read": 0.05 } }, - "doubao-seed-1-6-vision-250815": { - "id": "doubao-seed-1-6-vision-250815", - "name": "doubao-seed-1-6-vision-250815", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "mistralai/ministral-14b-2512": { + "id": "mistralai/ministral-14b-2512", + "name": "Ministral 14B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-04", + "last_updated": "2025-12-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.114, - "output": 1.143 + "input": 0.2, + "output": 0.2, + "cache_read": 0.1 } }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "claude-opus-4-1-20250805", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "mistralai/mistral-small-3.1-24b-instruct": { + "id": "mistralai/mistral-small-3.1-24b-instruct", + "name": "Mistral Small 3.1 24B (2503)", + "description": "Building upon Mistral Small 3 (2501), Mistral Small 3.1 (2503) adds state-of-the-art vision understanding and enhances long context capabilities up to 128k tokens without compromising text performance. With 24 billion parameters, this model achieves top-tier capabilities in both text and vision tasks.", + "family": "mistral-small", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 128000, + "input": 128000, + "output": 102400 }, "cost": { - "input": 15, - "output": 75 + "input": 0.1, + "output": 0.3, + "cache_read": 0.05 } }, - "qwen3-max-2025-09-23": { - "id": "qwen3-max-2025-09-23", - "name": "qwen3-max-2025-09-23", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "mistralai/ministral-3b-2512": { + "id": "mistralai/ministral-3b-2512", + "name": "Ministral 3B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-04", + "last_updated": "2025-12-04", "modalities": { "input": [ "text" @@ -27593,33 +29350,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 258048, - "output": 65536 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 0.86, - "output": 3.43 + "input": 0.1, + "output": 0.1, + "cache_read": 0.05 } }, - "glm-4.7-flashx": { - "id": "glm-4.7-flashx", - "name": "glm-4.7-flashx", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "mistralai/mistral-nemo-instruct-2407": { + "id": "mistralai/mistral-nemo-instruct-2407", + "name": "Mistral Nemo", + "description": "12B parameter model with multilingual support.", + "family": "mistral-nemo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-20", - "last_updated": "2026-01-20", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" @@ -27630,34 +29383,27 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.0715, - "output": 0.429 + "input": 0.1003, + "output": 0.1207, + "cache_read": 0.05015 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "glm-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "mistralai/codestral-2508": { + "id": "mistralai/codestral-2508", + "name": "Codestral 2508", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "codestral", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-10", - "last_updated": "2026-04-10", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-01", + "last_updated": "2025-08-01", "modalities": { "input": [ "text" @@ -27668,64 +29414,42 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 0.86, - "output": 3.5 + "input": 0.3, + "output": 0.9, + "cache_read": 0.15 } }, - "glm-4.6": { - "id": "glm-4.6", - "name": "glm-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "sakana/fugu-max": { + "id": "sakana/fugu-max", + "name": "Fugu Max", + "description": "Sakana AI's cost-performance Fugu model uses learned multi-agent orchestration to route tasks across expert models for reasoning, coding, and tool use.", + "family": "fugu", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.286, - "output": 1.142 - } - }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "kimi-k2-thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "structured_output": true, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -27733,41 +29457,42 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.575, - "output": 2.3 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "claude-sonnet-4-5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "sakana/fugu-ultra": { + "id": "sakana/fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "structured_output": true, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -27775,29 +29500,41 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "input": 1000000, + "output": 16384 }, "cost": { - "input": 3, - "output": 15 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "glm-4.5-x": { - "id": "glm-4.5-x", - "name": "glm-4.5-x", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": false, + "sakana/fugu-ultra-v1.1": { + "id": "sakana/fugu-ultra-v1.1", + "name": "Fugu Ultra v1.1", + "description": "Sakana AI's upgraded Fugu Ultra release with stronger coding, agentic task execution, and advanced reasoning through dynamic orchestration of frontier models.", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "structured_output": true, + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -27805,26 +29542,27 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 1000000, + "input": 1000000, "output": 16384 }, "cost": { - "input": 1.143, - "output": 2.29 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "deepseek-v3.2-thinking": { - "id": "deepseek-v3.2-thinking", - "name": "DeepSeek-V3.2-Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "inclusionai/ling-3.0-flash": { + "id": "inclusionai/ling-3.0-flash", + "name": "Ling 3.0 Flash", + "description": "Ling-3.0-flash is a 124B-parameter Mixture-of-Experts model with approximately 5.1B parameters active per token. It prioritizes token efficiency and production-scale agentic inference, helping coding and tool-using agents complete more work within constrained latency and serving budgets.", + "family": "ling", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "structured_output": false, + "release_date": "2026-07-23", + "last_updated": "2026-07-23", "modalities": { "input": [ "text" @@ -27835,128 +29573,87 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.29, - "output": 0.43 + "input": 0.075, + "output": 0.22, + "cache_read": 0.015 } }, - "claude-sonnet-4-6-thinking": { - "id": "claude-sonnet-4-6-thinking", - "name": "claude-sonnet-4-6-thinking", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "inclusionai/ling-3.0-flash-vl": { + "id": "inclusionai/ling-3.0-flash-vl", + "name": "Ling 3.0 Flash VL", + "description": "Ling 3.0 Flash VL is inclusionAI's native multimodal Mixture-of-Experts model with 124B total parameters and 5.5B active parameters per token. It combines image and video understanding with reasoning and tool use for document analysis, charts, visual verification, and interface-based agent tasks. Thinking is enabled by default and can be turned off in settings.", + "family": "ling", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-08", - "release_date": "2026-02-18", - "last_updated": "2026-03-13", + "structured_output": false, + "release_date": "2026-09-09", + "last_updated": "2026-09-09", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15 + "input": 0.06, + "output": 0.18, + "cache_read": 0.012 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "claude-opus-4-7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "inclusionai/ling-3.0-flash:thinking": { + "id": "inclusionai/ling-3.0-flash:thinking", + "name": "Ling 3.0 Flash Thinking", + "description": "Ling-3.0-flash Thinking enables visible reasoning on inclusionAI's token-efficient 124B-parameter Mixture-of-Experts model for harder coding, tool use, planning, and production-scale agent workflows.", + "family": "ling", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2026-01-31", - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "structured_output": false, + "release_date": "2026-07-23", + "last_updated": "2026-07-23", "modalities": { "input": [ - "text", - "image", - "pdf" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "grok-4-1-fast-non-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" @@ -27964,71 +29661,74 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 0.5 + "input": 0.075, + "output": 0.22, + "cache_read": 0.015 } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "gpt-5.4-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-19", - "last_updated": "2026-03-19", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "input": 1048576, + "output": 943718 }, "cost": { - "input": 0.2, - "output": 1.25 + "input": 2, + "output": 10, + "cache_read": 0.2 } }, - "claude-opus-4-1-20250805-thinking": { - "id": "claude-opus-4-1-20250805-thinking", - "name": "claude-opus-4-1-20250805-thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "moonshotai/kimi-k2.7-code-highspeed": { + "id": "moonshotai/kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code High-Speed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-05-27", - "last_updated": "2025-05-27", + "structured_output": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", @@ -28038,57 +29738,31 @@ "text" ] }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75 - } - }, - "glm-4.5-airx": { - "id": "glm-4.5-airx", - "name": "glm-4.5-airx", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-29", - "last_updated": "2025-07-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "input": 262144, + "output": 65536 }, "cost": { - "input": 0.572, - "output": 1.714 + "input": 1.9, + "output": 8, + "cache_read": 0.32 } }, - "grok-4.1": { - "id": "grok-4.1", - "name": "grok-4.1", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -28098,27 +29772,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 256000, + "input": 256000, + "output": 65536 }, "cost": { - "input": 2, - "output": 10 + "input": 0.5, + "output": 2.6, + "cache_read": 0.125 } }, - "gemini-2.5-flash-preview-09-2025": { - "id": "gemini-2.5-flash-preview-09-2025", - "name": "gemini-2.5-flash-preview-09-2025", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "moonshotai/kimi-k2.5:thinking": { + "id": "moonshotai/kimi-k2.5:thinking", + "name": "Kimi K2.5 Thinking", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, + "structured_output": false, + "temperature": false, "knowledge": "2025-01", - "release_date": "2025-09-26", - "last_updated": "2025-09-26", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -28128,149 +29807,131 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 256000, + "input": 256000, "output": 65536 }, "cost": { "input": 0.3, - "output": 2.5 + "output": 1.9, + "cache_read": 0.15 } }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "claude-opus-4-5-20251101", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "input": 262144, + "output": 98304 }, "cost": { - "input": 5, - "output": 25 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "claude-opus-4-20250514": { - "id": "claude-opus-4-20250514", - "name": "claude-opus-4-20250514", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - } - ], + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "structured_output": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 256000, + "input": 256000, + "output": 65536 }, "cost": { - "input": 15, - "output": 75 + "input": 0.3, + "output": 1.9, + "cache_read": 0.15 } }, - "gemini-3-pro-image-preview": { - "id": "gemini-3-pro-image-preview", - "name": "gemini-3-pro-image-preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, + "moonshotai/kimi-k2-instruct": { + "id": "moonshotai/kimi-k2-instruct", + "name": "Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "tool_call": true, + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2025-07-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 64000 + "context": 256000, + "input": 256000, + "output": 8192 }, "cost": { - "input": 2, - "output": 120 + "input": 0.4, + "output": 1.8, + "cache_read": 0.2 } }, - "gemini-2.5-flash-image": { - "id": "gemini-2.5-flash-image", - "name": "gemini-2.5-flash-image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "moonshotai/kimi-k2.6:thinking": { + "id": "moonshotai/kimi-k2.6:thinking", + "name": "Kimi K2.6 Thinking", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, "temperature": true, "knowledge": "2025-01", - "release_date": "2025-10-08", - "last_updated": "2025-10-08", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -28280,28 +29941,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 256000, + "input": 256000, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 30 + "input": 0.5, + "output": 2.6, + "cache_read": 0.125 } }, - "glm-for-coding": { - "id": "glm-for-coding", - "name": "glm-for-coding", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "moonshotai/kimi-k2-instruct-0711": { + "id": "moonshotai/kimi-k2-instruct-0711", + "name": "Kimi K2 0711", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2025-07-11", "modalities": { "input": [ "text" @@ -28310,71 +29972,107 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 0.086, - "output": 0.343 + "input": 0.4, + "output": 1.8, + "cache_read": 0.2 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "gpt-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "moonshotai/kimi-latest": { + "id": "moonshotai/kimi-latest", + "name": "Kimi Latest", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-12", - "last_updated": "2025-12-12", + "release_date": "2026-05-03", + "last_updated": "2026-05-03", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "input": 1048576, + "output": 943718 }, "cost": { - "input": 1.75, - "output": 14 - } - }, - "qwen-flash": { - "id": "qwen-flash", - "name": "Qwen-Flash", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "input": 2, + "output": 10, + "cache_read": 0.2 + } + }, + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "input": 262144, + "output": 65536 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + }, + "moonshotai/Kimi-K2-Instruct-0905": { + "id": "moonshotai/Kimi-K2-Instruct-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "structured_output": true, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text" @@ -28383,275 +30081,299 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 262144, + "input": 262144, + "output": 100352 }, "cost": { - "input": 0.022, - "output": 0.22 + "input": 0.4, + "output": 1.8, + "cache_read": 0.2 } }, - "claude-opus-4-6-thinking": { - "id": "claude-opus-4-6-thinking", - "name": "claude-opus-4-6-thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-02-06", - "last_updated": "2026-03-13", + "THUDM/GLM-Z1-9B-0414": { + "id": "THUDM/GLM-Z1-9B-0414", + "name": "GLM Z1 9B 0414", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm-z", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 32000, + "input": 32000, + "output": 8000 }, "cost": { - "input": 5, - "output": 25 + "input": 0.2, + "output": 0.2, + "cache_read": 0.1 } }, - "claude-sonnet-4-20250514": { - "id": "claude-sonnet-4-20250514", - "name": "claude-sonnet-4-20250514", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "THUDM/GLM-4-9B-0414": { + "id": "THUDM/GLM-4-9B-0414", + "name": "GLM 4 9B 0414", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 32000, + "input": 32000, + "output": 8000 }, "cost": { - "input": 3, - "output": 15 + "input": 0.2, + "output": 0.2, + "cache_read": 0.1 } }, - "gpt-5.1-chat-latest": { - "id": "gpt-5.1-chat-latest", - "name": "gpt-5.1-chat-latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "THUDM/GLM-4-32B-0414": { + "id": "THUDM/GLM-4-32B-0414", + "name": "GLM 4 32B 0414", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 16384 + "input": 128000, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.2, + "output": 0.2, + "cache_read": 0.1 } }, - "gpt-5.2-chat-latest": { - "id": "gpt-5.2-chat-latest", - "name": "gpt-5.2-chat-latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, + "nvidia/nemotron-3-nano-30b-a3b": { + "id": "nvidia/nemotron-3-nano-30b-a3b", + "name": "Nvidia Nemotron 3 Nano 30B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-12-15", + "last_updated": "2025-12-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "input": 262144, + "output": 235929 + }, + "cost": { + "input": 0.17, + "output": 0.68, + "cache_read": 0.085 + } + }, + "nvidia/nemotron-3.5-lightning": { + "id": "nvidia/nemotron-3.5-lightning", + "name": "Nvidia Nemotron 3.5 Lightning", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "medium" + "none", + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-12", - "last_updated": "2025-12-12", + "structured_output": false, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "input": 1000000, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.05, + "output": 0.2, + "cache_read": 0.01 } }, - "grok-4-fast-reasoning": { - "id": "grok-4-fast-reasoning", - "name": "grok-4-fast-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "nvidia/Llama-3.3-Nemotron-Super-49B-v1": { + "id": "nvidia/Llama-3.3-Nemotron-Super-49B-v1", + "name": "Nvidia Nemotron Super 49B", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 0.5 + "input": 0.15, + "output": 0.15, + "cache_read": 0.075 } }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "gpt-4.1-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "Nvidia Nemotron 3 Super 120B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 0.05, + "output": 0.25, + "cache_read": 0.025 } }, - "claude-sonnet-4-5-20250929-thinking": { - "id": "claude-sonnet-4-5-20250929-thinking", - "name": "claude-sonnet-4-5-20250929-thinking", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "attachment": true, + "nvidia/nemotron-3-ultra-550b-a55b": { + "id": "nvidia/nemotron-3-ultra-550b-a55b", + "name": "Nvidia Nemotron 3 Ultra 550B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "input": 1000000, + "output": 65536 }, "cost": { - "input": 3, - "output": 15 + "input": 0.5, + "output": 2.5, + "cache_read": 0.25 } }, - "MiniMax-M2.7-highspeed": { - "id": "MiniMax-M2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "nvidia/nemotron-3-super-120b-a12b:thinking": { + "id": "nvidia/nemotron-3-super-120b-a12b:thinking", + "name": "Nvidia Nemotron 3 Super 120B Thinking", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-03-19", - "last_updated": "2026-03-19", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ "text" @@ -28660,26 +30382,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 4.8 + "input": 0.05, + "output": 0.25, + "cache_read": 0.025 } }, - "MiniMax-M2": { - "id": "MiniMax-M2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "nvidia/nemotron-3-ultra-550b-a55b:thinking": { + "id": "nvidia/nemotron-3-ultra-550b-a55b:thinking", + "name": "Nvidia Nemotron 3 Ultra 550B Thinking", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-10-26", - "last_updated": "2025-10-26", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -28688,60 +30415,63 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "input": 1000000, + "output": 65536 }, "cost": { - "input": 0.33, - "output": 1.32 + "input": 0.5, + "output": 2.5, + "cache_read": 0.25 } }, - "gemini-3.1-flash-image-preview": { - "id": "gemini-3.1-flash-image-preview", - "name": "gemini-3.1-flash-image-preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": false, + "nvidia/nemotron-3.5-content-safety": { + "id": "nvidia/nemotron-3.5-content-safety", + "name": "Nemotron 3.5 Content Safety", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-27", - "last_updated": "2026-02-27", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, + "input": 131072, "output": 32768 }, "cost": { - "input": 0.5, - "output": 60 + "input": 0.05, + "output": 0.15, + "cache_read": 0.025 } }, - "qwen3-235b-a22b": { - "id": "qwen3-235b-a22b", - "name": "Qwen3-235B-A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF": { + "id": "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF", + "name": "Nvidia Nemotron 70b", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text" @@ -28750,68 +30480,84 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.29, - "output": 2.86 + "input": 0.357, + "output": 0.408, + "cache_read": 0.1785 } }, - "ministral-14b-2512": { - "id": "ministral-14b-2512", - "name": "ministral-14b-2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "attachment": true, - "reasoning": false, + "nvidia/nemotron-3.5-lightning:thinking": { + "id": "nvidia/nemotron-3.5-lightning:thinking", + "name": "Nvidia Nemotron 3.5 Lightning Thinking", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 1000000, + "input": 1000000, + "output": 65536 }, "cost": { - "input": 0.33, - "output": 0.33 + "input": 0.05, + "output": 0.2, + "cache_read": 0.01 } }, - "glm-4.6v": { - "id": "glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, + "xiaomi/mimo-v2.5-pro:thinking": { + "id": "xiaomi/mimo-v2.5-pro:thinking", + "name": "MiMo V2.5 Pro Thinking", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -28819,61 +30565,67 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.145, - "output": 0.43 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036, + "cache_write": 0 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "claude-haiku-4-5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "xiaomi/mimo-v2.6-pro": { + "id": "xiaomi/mimo-v2.6-pro", + "name": "MiMo V2.6 Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-16", - "last_updated": "2025-10-16", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 1, - "output": 5 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036, + "cache_write": 0 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "gpt-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -28881,91 +30633,90 @@ "type": "effort", "values": [ "none", - "low", - "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 0, - "tiers": [ - { - "input": 5, - "output": 22.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5 - } + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028, + "cache_write": 0 } }, - "doubao-seed-1-6-thinking-250715": { - "id": "doubao-seed-1-6-thinking-250715", - "name": "doubao-seed-1-6-thinking-250715", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "xiaomi/mimo-v2.6-pro-ultraspeed": { + "id": "xiaomi/mimo-v2.6-pro-ultraspeed", + "name": "MiMo V2.6 Pro UltraSpeed", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 16000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.121, - "output": 1.21 + "input": 4.35, + "output": 8.7, + "cache_read": 0.036, + "cache_write": 0 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "gpt-5.4-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "xiaomi/mimo-v2.5:thinking": { + "id": "xiaomi/mimo-v2.5:thinking", + "name": "MiMo V2.5 Thinking", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -28973,163 +30724,171 @@ "type": "effort", "values": [ "none", - "low", - "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-19", - "last_updated": "2026-03-19", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.75, - "output": 4.5 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028, + "cache_write": 0 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "gpt-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo V2.5 Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 2, - "output": 8 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036, + "cache_write": 0 } }, - "doubao-seed-code-preview-251028": { - "id": "doubao-seed-code-preview-251028", - "name": "doubao-seed-code-preview-251028", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "xiaomi/mimo-v2.6-flash": { + "id": "xiaomi/mimo-v2.6-flash", + "name": "MiMo V2.6 Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-11-11", - "last_updated": "2025-11-11", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 0.17, - "output": 1.14 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028, + "cache_write": 0 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "claude-opus-4-6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "undi95/remm-slerp-l2-13b": { + "id": "undi95/remm-slerp-l2-13b", + "name": "ReMM SLERP 13B", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-06", - "last_updated": "2026-03-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", - "image", "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 6144, + "input": 6144, + "output": 4096 }, "cost": { - "input": 5, - "output": 25 + "input": 0.799, + "output": 1.207, + "cache_read": 0.3995 } }, - "qwen3-coder-480b-a35b-instruct": { - "id": "qwen3-coder-480b-a35b-instruct", - "name": "qwen3-coder-480b-a35b-instruct", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "Salesforce/Llama-xLAM-2-70b-fc-r": { + "id": "Salesforce/Llama-xLAM-2-70b-fc-r", + "name": "Llama-xLAM-2 70B fc-r", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-13", + "last_updated": "2025-04-13", "modalities": { "input": [ "text" @@ -29138,71 +30897,68 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.86, - "output": 3.43 + "input": 2.5, + "output": 2.5, + "cache_read": 1.25 } }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "claude-sonnet-4-5-20250929", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "minimax/minimax-latest": { + "id": "minimax/minimax-latest", + "name": "MiniMax Latest", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "structured_output": true, + "release_date": "2026-05-03", + "last_updated": "2026-05-03", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 512000, + "input": 512000, + "output": 80000 }, "cost": { - "input": 3, - "output": 15 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "deepseek-reasoner": { - "id": "deepseek-reasoner", - "name": "Deepseek-Reasoner", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -29211,62 +30967,65 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 204800, + "input": 204800, + "output": 131072 }, "cost": { - "input": 0.29, - "output": 0.43 + "input": 0.3, + "output": 1.2, + "cache_read": 0.15 } }, - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "grok-4-1-fast-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": true, + "minimax/minimax-m2.7-turbo": { + "id": "minimax/minimax-m2.7-turbo", + "name": "MiniMax M2.7 Turbo", + "description": "Efficient MiniMax model for quick assistance, coding, and routine automation", + "family": "minimax-m2.7", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "structured_output": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 204800, + "input": 204800, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.5 + "input": 0.6, + "output": 2.4, + "cache_read": 0.3 } }, - "gemini-3-pro-preview": { - "id": "gemini-3-pro-preview", - "name": "gemini-3-pro-preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "minimax/minimax-01": { + "id": "minimax/minimax-01", + "name": "MiniMax 01", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2025-01-15", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" @@ -29274,30 +31033,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1000192, + "input": 1000192, + "output": 16384 }, "cost": { - "input": 2, - "output": 12 + "input": 0.1394, + "output": 1.122, + "cache_read": 0.0697 } }, - "gpt-5-thinking": { - "id": "gpt-5-thinking", - "name": "gpt-5-thinking", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "minimax/minimax-m2-her": { + "id": "minimax/minimax-m2-her", + "name": "MiniMax M2-her", + "description": "MiniMax M2 variant tuned for conversational and character-driven agent interactions", + "family": "minimax", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "release_date": "2026-01-23", + "last_updated": "2026-01-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -29305,38 +31065,62 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 65532, + "input": 65532, + "output": 2048 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.302, + "output": 1.207, + "cache_read": 0.151 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "gpt-5-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "input": 204800, + "output": 131072 + }, + "cost": { + "input": 0.315, + "output": 1.26, + "cache_read": 0.1575 + } + }, + "minimax/minimax-m3:thinking": { + "id": "minimax/minimax-m3:thinking", + "name": "MiniMax M3 Thinking", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", @@ -29346,167 +31130,159 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 512000, + "input": 512000, + "output": 80000 }, "cost": { - "input": 0.25, - "output": 2 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "gpt-4.1-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 512000, + "input": 512000, + "output": 80000 }, "cost": { - "input": 0.4, - "output": 1.6 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 200000, + "input": 200000, "output": 131072 }, "cost": { - "input": 0.72, - "output": 3.2 + "input": 0.33, + "output": 1.32, + "cache_read": 0.165 } }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "gpt-5.4-pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", - "attachment": true, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "input": 200000, + "output": 131072 }, "cost": { - "input": 30, - "output": 180, - "cache_read": 0, - "cache_write": 0, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } + "input": 0.17, + "output": 1.53, + "cache_read": 0.085 + } + }, + "inflatebot/MN-12B-Mag-Mell-R1": { + "id": "inflatebot/MN-12B-Mag-Mell-R1", + "name": "Mag Mell R1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-07-01", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 16384, + "input": 16384, + "output": 8192 + }, + "cost": { + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "glm-4.5-air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", + "soob3123/GrayLine-Qwen3-8B": { + "id": "soob3123/GrayLine-Qwen3-8B", + "name": "Grayline Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text" @@ -29517,41 +31293,103 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.1143, - "output": 0.286 + "input": 0.3, + "output": 0.3, + "cache_read": 0.15 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "claude-sonnet-4-6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "soob3123/Veiled-Calla-12B": { + "id": "soob3123/Veiled-Calla-12B", + "name": "Veiled Calla 12B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-13", + "last_updated": "2025-04-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "input": 32768, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 0.3, + "cache_read": 0.15 + } + }, + "soob3123/amoral-gemma3-27B-v2": { + "id": "soob3123/amoral-gemma3-27B-v2", + "name": "Amoral Gemma3 27B v2", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-23", + "last_updated": "2025-05-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "input": 32768, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 0.3, + "cache_read": 0.15 + } + }, + "stepfun/step-5-preview": { + "id": "stepfun/step-5-preview", + "name": "Step 5 Preview", + "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-18", - "last_updated": "2026-03-13", + "structured_output": true, + "release_date": "2026-09-16", + "last_updated": "2026-09-20", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -29560,85 +31398,102 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "input": 1000000, + "output": 1000000 }, "cost": { - "input": 3, - "output": 15 + "input": 1, + "output": 2.7, + "cache_read": 0.05 } }, - "grok-4-fast-non-reasoning": { - "id": "grok-4-fast-non-reasoning", - "name": "grok-4-fast-non-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "stepfun/step-3.7-flash:thinking": { + "id": "stepfun/step-3.7-flash:thinking", + "name": "Step 3.7 Flash Thinking", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 262144, + "input": 262144, + "output": 256000 }, "cost": { "input": 0.2, - "output": 0.5 + "output": 1.15, + "cache_read": 0.04 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "gemini-3-flash-preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "attachment": true, + "Sao10K/L3.1-70B-Hanami-x1": { + "id": "Sao10K/L3.1-70B-Hanami-x1", + "name": "Llama 3.1 70B Hanami", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "deepseek-chat": { - "id": "deepseek-chat", - "name": "Deepseek-Chat", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "Sao10K/L3.1-70B-Euryale-v2.2": { + "id": "Sao10K/L3.1-70B-Euryale-v2.2", + "name": "Llama 3.1 70B Euryale", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-11-29", - "last_updated": "2024-11-29", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -29647,27 +31502,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 20480, + "input": 20480, + "output": 16384 }, "cost": { - "input": 0.29, - "output": 0.43 + "input": 0.306, + "output": 0.357, + "cache_read": 0.153 } }, - "MiniMax-M1": { - "id": "MiniMax-M1", - "name": "MiniMax-M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "Sao10K/L3.3-70B-Euryale-v2.3": { + "id": "Sao10K/L3.3-70B-Euryale-v2.3", + "name": "Llama 3.3 70B Euryale", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-06-16", - "last_updated": "2025-06-16", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -29676,41 +31533,74 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 20480, + "input": 20480, + "output": 16384 }, "cost": { - "input": 0.132, - "output": 1.254 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "grok-4.20-multi-agent-beta-0309": { - "id": "grok-4.20-multi-agent-beta-0309", - "name": "grok-4.20-multi-agent-beta-0309", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "attachment": true, + "Sao10K/L3-8B-Stheno-v3.2": { + "id": "Sao10K/L3-8B-Stheno-v3.2", + "name": "Sao10K Stheno 8b", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-11-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 16384, + "input": 16384, + "output": 8192 + }, + "cost": { + "input": 0.2006, + "output": 0.2006, + "cache_read": 0.1003 + } + }, + "inception/mercury-2.5-preview": { + "id": "inception/mercury-2.5-preview", + "name": "Mercury 2.5 Preview", + "description": "Mercury 2.5 Preview is Inception's latest and most intelligent diffusion language model. Instead of generating tokens strictly one at a time, it produces and refines multiple tokens in parallel, reaching up to 1,107 tokens per second on standard GPUs. It delivers a 10+ point intelligence gain over Mercury 2, with tunable reasoning, parallel tool calls, schema-aligned JSON output, and a 260K context window. It is built for latency-sensitive production work such as search agents, voice pipelines, customer support, rapid coding iteration, and coding subagents.", + "family": "mercury", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "structured_output": true, + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -29718,36 +31608,65 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 260000, + "input": 260000, + "output": 65536 }, "cost": { - "input": 2, - "output": 6 + "input": 0.04, + "output": 0.15, + "cache_read": 0.004 } }, - "glm-5": { - "id": "glm-5", - "name": "glm-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "ornith-ai/ornith-1.5-35b-a3b": { + "id": "ornith-ai/ornith-1.5-35b-a3b", + "name": "Ornith 1.5 35B", + "description": "Ornith 1.5 35B A3B is an open-weight mixture-of-experts model for agentic coding, tool use, image understanding, and long-context work. This variant disables thinking for faster direct responses.", + "family": "ornith", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "structured_output": true, + "release_date": "2026-07-29", + "last_updated": "2026-08-20", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "input": 262144, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.05 + } + }, + "ornith-ai/ornith-1.5-35b-a3b:thinking": { + "id": "ornith-ai/ornith-1.5-35b-a3b:thinking", + "name": "Ornith 1.5 35B Thinking", + "description": "Ornith 1.5 35B A3B is an open-weight mixture-of-experts model for agentic coding, reasoning, tool use, image understanding, and long-context work. This variant enables thinking by default.", + "family": "ornith", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "release_date": "2026-07-29", + "last_updated": "2026-08-20", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" @@ -29755,26 +31674,27 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.6 + "input": 0.1, + "output": 0.4, + "cache_read": 0.05 } }, - "qwen-max-latest": { - "id": "qwen-max-latest", - "name": "Qwen-Max-Latest", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "amazon/nova-lite-v1": { + "id": "amazon/nova-lite-v1", + "name": "Amazon Nova Lite 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text" @@ -29785,29 +31705,30 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 300000, + "input": 300000, + "output": 5120 }, "cost": { - "input": 0.343, - "output": 1.372 + "input": 0.0595, + "output": 0.238, + "cache_read": 0.02975 } }, - "mistral-large-2512": { - "id": "mistral-large-2512", - "name": "mistral-large-2512", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "attachment": true, + "amazon/nova-2-lite-v1": { + "id": "amazon/nova-2-lite-v1", + "name": "Amazon Nova 2 Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-12-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -29815,24 +31736,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 262144 + "context": 1000000, + "input": 1000000, + "output": 65535 }, "cost": { - "input": 1.1, - "output": 3.3 + "input": 0.51, + "output": 4.25, + "cache_read": 0.255 } }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "amazon/nova-pro-v1": { + "id": "amazon/nova-pro-v1", + "name": "Amazon Nova Pro 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-03-19", - "last_updated": "2026-03-19", + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text" @@ -29843,58 +31767,57 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 300000, + "input": 300000, + "output": 32000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.799, + "output": 3.196, + "cache_read": 0.3995 } }, - "claude-3-5-haiku-latest": { - "id": "claude-3-5-haiku-latest", - "name": "claude-3-5-haiku-latest", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "Doctor-Shotgun/MS3.2-24B-Magnum-Diamond": { + "id": "Doctor-Shotgun/MS3.2-24B-Magnum-Diamond", + "name": "MS3.2 24B Magnum Diamond", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "tool_call": false, + "structured_output": false, + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.8, - "output": 4 + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "qwen3-30b-a3b": { - "id": "qwen3-30b-a3b", - "name": "Qwen3-30B-A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "aion-labs/aion-2.0": { + "id": "aion-labs/aion-2.0", + "name": "AionLabs: Aion-2.0", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "structured_output": false, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text" @@ -29905,25 +31828,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 0.11, - "output": 1.08 + "input": 0.8, + "output": 1.6, + "cache_read": 0.2 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "deepseek-v3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "aion-labs/aion-rp-llama-3.1-8b": { + "id": "aion-labs/aion-rp-llama-3.1-8b", + "name": "Llama 3.1 8b (uncensored)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -29934,28 +31859,30 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 0.29, - "output": 0.43 + "input": 0.8, + "output": 1.6, + "cache_read": 0.4 } }, - "doubao-seed-1-8-251215": { - "id": "doubao-seed-1-8-251215", - "name": "doubao-seed-1-8-251215", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, - "reasoning": false, + "aion-labs/aion-3.0": { + "id": "aion-labs/aion-3.0", + "name": "AionLabs: Aion 3.0", + "description": "Aion 3.0 is a GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "structured_output": false, + "release_date": "2026-07-07", + "last_updated": "2026-07-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -29963,30 +31890,30 @@ }, "open_weights": false, "limit": { - "context": 224000, - "output": 64000 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 0.114, - "output": 0.286 + "input": 3, + "output": 6, + "cache_read": 0.75 } }, - "claude-opus-4-5-20251101-thinking": { - "id": "claude-opus-4-5-20251101-thinking", - "name": "claude-opus-4-5-20251101-thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "attachment": true, + "aion-labs/aion-3.5": { + "id": "aion-labs/aion-3.5", + "name": "AionLabs: Aion 3.5", + "description": "A GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "structured_output": false, + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -29994,30 +31921,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 5, - "output": 25 + "input": 3, + "output": 6, + "cache_read": 0.75 } }, - "gemini-2.0-flash-lite": { - "id": "gemini-2.0-flash-lite", - "name": "gemini-2.0-flash-lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-06-16", - "last_updated": "2025-06-16", + "aion-labs/aion-3.5-mini": { + "id": "aion-labs/aion-3.5-mini", + "name": "AionLabs: Aion 3.5 Mini", + "description": "A GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -30025,34 +31952,27 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 8192 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 0.7, + "output": 1.4, + "cache_read": 0.18 } }, - "glm-5-turbo": { - "id": "glm-5-turbo", - "name": "glm-5-turbo", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", + "aion-labs/aion-3.0-mini": { + "id": "aion-labs/aion-3.0-mini", + "name": "AionLabs: Aion 3.0 Mini", + "description": "Aion 3.0 Mini is a DeepSeek-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "structured_output": false, + "release_date": "2026-07-07", + "last_updated": "2026-07-07", "modalities": { "input": [ "text" @@ -30063,42 +31983,33 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 0.72, - "output": 3.2 + "input": 0.7, + "output": 1.4, + "cache_read": 0.18 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "gpt-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "qwen/qwen3.5-flash": { + "id": "qwen/qwen3.5-flash", + "name": "Qwen3.5 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -30106,53 +32017,44 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 991808, + "input": 991808, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.1, + "output": 0.4, + "cache_read": 0.05 } - } - } - }, - "trustedrouter": { - "id": "trustedrouter", - "env": [ - "TRUSTEDROUTER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.trustedrouter.com/v1", - "name": "TrustedRouter", - "doc": "https://trustedrouter.com/docs", - "models": { - "zdr": { - "id": "zdr", - "name": "Zero Data Retention", - "description": "TrustedRouter privacy routing alias that prefers zero data retention model endpoints.", + }, + "qwen/qwen3.8-max:thinking": { + "id": "qwen/qwen3.8-max:thinking", + "name": "Qwen3.8 Max Thinking", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-27", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ @@ -30161,75 +32063,75 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 991000, + "input": 991000, "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "e2e": { - "id": "e2e", - "name": "End-to-End Encrypted", - "description": "TrustedRouter privacy routing alias for end-to-end encrypted provider routes where available.", + "qwen/qwen3-vl-235b-a22b-instruct": { + "id": "qwen/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-27", + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "input": 131072, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 1.9, + "cache_read": 0.15 } }, - "synth-code": { - "id": "synth-code", - "name": "Synth Code", - "description": "TrustedRouter code synthesis orchestration alias that combines multiple model responses into one answer.", - "attachment": true, + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "low", - "medium", "high" ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-20", - "last_updated": "2026-06-27", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -30238,150 +32140,136 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "input": 1000000, + "output": 65536 + }, + "cost": { + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "fast": { - "id": "fast", - "name": "Fast", - "description": "TrustedRouter speed routing alias that prefers low-latency healthy model endpoints.", + "qwen/qwen3.8-27b": { + "id": "qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-27", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 0.7, + "cache_read": 0.04 } }, - "synth": { - "id": "synth", - "name": "Synth", - "description": "TrustedRouter synthesis orchestration alias that combines multiple model responses into one answer.", + "qwen/qwen3-vl-235b-a22b-thinking": { + "id": "qwen/qwen3-vl-235b-a22b-thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-06-20", - "last_updated": "2026-06-27", + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 6, + "cache_read": 0.25 } }, - "auto": { - "id": "auto", - "name": "Auto", - "description": "TrustedRouter automatic routing alias that chooses a healthy supported model endpoint for the request.", + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-05-01", - "last_updated": "2026-06-27", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 260096, + "input": 260096, + "output": 65536 + }, + "cost": { + "input": 0.27, + "output": 2.16, + "cache_read": 0.135 } }, - "cheap": { - "id": "cheap", - "name": "Cheap", - "description": "TrustedRouter low-cost routing alias that prefers inexpensive healthy model endpoints.", + "qwen/qwen3.5-omni-plus": { + "id": "qwen/qwen3.5-omni-plus", + "name": "Qwen3.5 Omni Plus", + "description": "Qwen3.5 Omni Plus is Qwen's stronger general multimodal model. We verified live support for text prompts, images, audio files, and direct video URLs on Alibaba's chat-completions-compatible API. Alibaba describes Plus as a comprehensive evolution of Qwen3 Omni with support for over 10 hours of audio input.", + "family": "qwen3.5", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-01", - "last_updated": "2026-06-27", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-30", + "last_updated": "2026-03-30", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -30389,45 +32277,37 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 983616, + "input": 983616, + "output": 65536 } - } - } - }, - "zhipuai": { - "id": "zhipuai", - "env": [ - "ZHIPU_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://open.bigmodel.cn/api/paas/v4", - "name": "Zhipu AI", - "doc": "https://docs.z.ai/guides/overview/pricing", - "models": { - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + }, + "qwen/qwen3.7-flash:thinking": { + "id": "qwen/qwen3.7-flash:thinking", + "name": "Qwen3.7 Flash Thinking", + "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -30435,40 +32315,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 983616, + "input": 983616, + "output": 65536 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "input": 0.03, + "output": 0.13, + "cache_read": 0.006, + "cache_write": 0.038 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "qwen/qwen3-235b-a22b-instruct-2507": { + "id": "qwen/qwen3-235b-a22b-instruct-2507", + "name": "Qwen 3 235b A22B 2507", + "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2025-07-21", + "last_updated": "2025-07-21", "modalities": { "input": [ "text" @@ -30477,43 +32346,34 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "input": 0.13, + "output": 0.5, + "cache_read": 0.065 } }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "qwen/qwen-max": { + "id": "qwen/qwen-max", + "name": "Qwen 2.5 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "knowledge": "2024-04", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -30521,35 +32381,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 32000, + "input": 32000, + "output": 8192 }, "cost": { - "input": 5, - "output": 22, - "cache_read": 1.2, - "cache_write": 0 + "input": 1.5997, + "output": 6.392, + "cache_read": 0.79985 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B (Thinking)", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ "text" @@ -30560,70 +32414,65 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0 + "input": 0.15, + "output": 0.65, + "cache_read": 0.075 } }, - "glm-4.5-flash": { - "id": "glm-4.5-flash", - "name": "GLM-4.5-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen/qwen3.8-max": { + "id": "qwen/qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 991000, + "input": 991000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "glm-4.7-flash": { - "id": "glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ "text" @@ -30634,33 +32483,27 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.2, + "output": 1.5, + "cache_read": 0.1 } }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "qwen/qwen3-coder": { + "id": "qwen/qwen3-coder", + "name": "Qwen 3 Coder 480B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "structured_output": true, + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text" @@ -30671,33 +32514,29 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 262000, + "input": 262000, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03, - "cache_write": 0 + "input": 0.13, + "output": 0.5, + "cache_read": 0.065 } }, - "glm-4.6v": { - "id": "glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "qwen/qwen3.5-35b-a3b:thinking": { + "id": "qwen/qwen3.5-35b-a3b:thinking", + "name": "Qwen3.5 35B A3B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", @@ -30710,71 +32549,68 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 260096, + "input": 260096, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0.225, + "output": 1.8, + "cache_read": 0.1125 } }, - "glm-4.6": { - "id": "glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "qwen/qwen3.5-plus": { + "id": "qwen/qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 983616, + "input": 983616, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 0.4, + "output": 2.4, + "cache_read": 0.04 } }, - "glm-4.7-flashx": { - "id": "glm-4.7-flashx", - "name": "GLM-4.7-FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen/qwen3-32b": { + "id": "qwen/qwen3-32b", + "name": "Qwen 3 32b", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -30782,33 +32618,30 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 41000, + "input": 41000, + "output": 32768 }, "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0 + "input": 0.1, + "output": 0.3, + "cache_read": 0.05 } }, - "glm-4.5": { - "id": "glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", + "qwen/qwen-plus": { + "id": "qwen/qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", "modalities": { "input": [ "text" @@ -30817,35 +32650,39 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 995904, + "input": 995904, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 0.3995, + "output": 1.2002, + "cache_read": 0.19975 } }, - "glm-4.5v": { - "id": "glm-4.5v", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "qwen/qwen3.7-flash": { + "id": "qwen/qwen3.7-flash", + "name": "Qwen3.7 Flash", + "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", @@ -30856,36 +32693,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 64000, - "output": 16384 + "context": 991808, + "input": 991808, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 1.8 + "input": 0.03, + "output": 0.13, + "cache_read": 0.006, + "cache_write": 0.038 } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" @@ -30894,43 +32727,64 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 1.2002, + "output": 6.001, + "cache_read": 0.6001 } - } - } - }, - "cortecs": { - "id": "cortecs", - "env": [ - "CORTECS_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.cortecs.ai/v1", - "name": "Cortecs", - "doc": "https://api.cortecs.ai/v1/models", - "models": { - "deepseek-r1-0528": { - "id": "deepseek-r1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + }, + "qwen/qwen3.8-27b-obliterated:thinking": { + "id": "qwen/qwen3.8-27b-obliterated:thinking", + "name": "Qwen 3.8 27B Obliterated Thinking", + "description": "Qwen 3.8 27B Obliterated with thinking enabled for more deliberate creative work, coding, multimodal analysis, tool use, and long-context problem solving.", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, + "release_date": "2026-07-29", + "last_updated": "2026-08-24", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "input": 524288, + "output": 32768 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.2 + } + }, + "qwen/qwen-turbo": { + "id": "qwen/qwen-turbo", + "name": "Qwen Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "knowledge": "2024-04", + "release_date": "2024-11-01", + "last_updated": "2025-04-28", "modalities": { "input": [ "text" @@ -30939,41 +32793,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 1000000, + "input": 1000000, + "output": 8192 }, "cost": { - "input": 0.585, - "output": 2.307 + "input": 0.04998, + "output": 0.2006, + "cache_read": 0.02499 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "qwen/qwen3-coder-30b-a3b-instruct": { + "id": "qwen/qwen3-coder-30b-a3b-instruct", + "name": "Qwen3 Coder 30B A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -30984,33 +32828,34 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 384000 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.133, - "output": 0.266, - "cache_read": 0.0028 + "input": 0.1, + "output": 0.4, + "cache_read": 0.05 } }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "qwen/qwen3.8-27b:thinking": { + "id": "qwen/qwen3.8-27b:thinking", + "name": "Qwen3.8 27B Thinking", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -31018,29 +32863,32 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 196608 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.32, - "output": 1.18 + "input": 0.15, + "output": 0.7, + "cache_read": 0.04 } }, - "deepseek-v3-0324": { - "id": "deepseek-v3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "qwen/qwen3.8-27b-obliterated": { + "id": "qwen/qwen3.8-27b-obliterated", + "name": "Qwen 3.8 27B Obliterated", + "description": "Qwen 3.8 27B Obliterated is an open-weight multimodal model LoRA-tuned for fewer refusals across chat, coding, reasoning, tool use, and long-context work.", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "structured_output": true, + "release_date": "2026-07-29", + "last_updated": "2026-08-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -31048,41 +32896,42 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 524288, + "input": 524288, + "output": 32768 }, "cost": { - "input": 0.551, - "output": 1.654 + "input": 0.25, + "output": 1.5, + "cache_read": 0.2 } }, - "claude-opus4-7": { - "id": "claude-opus4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen/qwen3.8-omni-flash": { + "id": "qwen/qwen3.8-omni-flash", + "name": "Qwen3.8 Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "release_date": "2026-09-17", + "last_updated": "2026-09-17", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -31090,35 +32939,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 991808, + "input": 991808, + "output": 131072 }, "cost": { - "input": 5.6, - "output": 27.99, - "cache_read": 0.56, - "cache_write": 6.99 + "input": 0.15, + "output": 0.47, + "cache_read": 0.016 } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM 4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "qwen/qwen3.8-27b-uncensored": { + "id": "qwen/qwen3.8-27b-uncensored", + "name": "Qwen 3.8 27B Uncensored", + "description": "Qwen 3.8 27B Uncensored is an NVFP4 open-weight multimodal model LoRA-tuned for fewer refusals across chat, coding, reasoning, tool use, and long-context work.", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "structured_output": true, + "release_date": "2026-07-29", + "last_updated": "2026-08-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -31126,27 +32972,37 @@ }, "open_weights": true, "limit": { - "context": 198000, - "output": 198000 + "context": 524288, + "input": 524288, + "output": 32768 }, "cost": { - "input": 0.45, - "output": 2.23 + "input": 0.15, + "output": 1.2, + "cache_read": 0.125 } }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen/qwen3.7-max:thinking": { + "id": "qwen/qwen3.7-max:thinking", + "name": "Qwen3.7 Max Thinking", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -31155,32 +33011,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "input": 1000000, + "output": 65536 }, "cost": { - "input": 0.062, - "output": 0.408 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3 Coder 30B A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "qwen/qwen3-235b-a22b": { + "id": "qwen/qwen3-235b-a22b", + "name": "Qwen 3 235b A22B", + "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-07-31", - "last_updated": "2025-07-31", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -31188,29 +33048,28 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.053, - "output": 0.222 + "input": 0.3, + "output": 0.5, + "cache_read": 0.15 } }, - "minimax-m2.1": { - "id": "minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text" @@ -31221,29 +33080,32 @@ }, "open_weights": true, "limit": { - "context": 196000, - "output": 196000 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 0.34, - "output": 1.34 + "input": 0.437, + "output": 3.496, + "cache_read": 0.103788 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen/qwen3.8-27b-uncensored:thinking": { + "id": "qwen/qwen3.8-27b-uncensored:thinking", + "name": "Qwen 3.8 27B Uncensored Thinking", + "description": "Qwen 3.8 27B Uncensored with thinking enabled for more deliberate creative work, coding, multimodal analysis, tool use, and long-context problem solving.", "family": "qwen", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "structured_output": true, + "release_date": "2026-07-29", + "last_updated": "2026-08-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -31251,45 +33113,33 @@ }, "open_weights": true, "limit": { - "context": 16384, - "output": 16384 + "context": 524288, + "input": 524288, + "output": 32768 }, "cost": { - "input": 0.099, - "output": 0.33 + "input": 0.15, + "output": 1.2, + "cache_read": 0.125 } }, - "claude-4-6-sonnet": { - "id": "claude-4-6-sonnet", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "qwen/qwen3.5-omni-flash": { + "id": "qwen/qwen3.5-omni-flash", + "name": "Qwen3.5 Omni Flash", + "description": "Qwen3.5 Omni Flash is Qwen's fast multimodal model. We verified live support for text prompts, images, audio files, and direct video URLs on Alibaba's chat-completions-compatible API. Alibaba describes Flash as a fully evolved version of Qwen3 Omni with audio input support across 60+ languages.", + "family": "qwen3.5", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-30", + "last_updated": "2026-03-30", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -31297,62 +33147,61 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 3.59, - "output": 17.92 - } - }, - "claude-sonnet-4": { - "id": "claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": false, + "context": 49152, + "input": 49152, + "output": 16384 + } + }, + "qwen/qwen3.6-35b-a3b": { + "id": "qwen/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 3.307, - "output": 16.536 + "input": 0.112, + "output": 0.8, + "cache_read": 0.056 } }, - "llama-4-maverick": { - "id": "llama-4-maverick", - "name": "Llama 4 Maverick 17B Instruct", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", - "attachment": true, + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B (Instruct)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -31360,32 +33209,30 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 16384 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.124, - "output": 0.603, - "cache_read": 0.03, - "cache_write": 0.151 + "input": 0.15, + "output": 0.65, + "cache_read": 0.075 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "qwen/qwen3-max-2026-01-23": { + "id": "qwen/qwen3-max-2026-01-23", + "name": "Qwen3 Max 2026-01-23", + "description": "Qwen3 Max is Alibaba's flagship Qwen 3 reasoning model with native tool use (web search, web extractor, code interpreter) and a 256K context window.", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-17", + "tool_call": false, + "structured_output": false, + "release_date": "2026-01-26", + "last_updated": "2026-01-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -31393,75 +33240,65 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 1.654, - "output": 11.024 + "input": 1.2002, + "output": 6.001, + "cache_read": 0.6001 } }, - "claude-4-5-sonnet": { - "id": "claude-4-5-sonnet", - "name": "Claude 4.5 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "qwen/qwen3.8-27b-fable": { + "id": "qwen/qwen3.8-27b-fable", + "name": "Qwen 3.8 27B Fable", + "description": "Qwen 3.8 27B Fable is an open-weight multimodal creative finetune for expressive dialogue, long-form storytelling, character work, and roleplay.", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "structured_output": false, + "release_date": "2026-07-29", + "last_updated": "2026-08-28", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 200000 + "context": 524288, + "input": 524288, + "output": 32768 }, "cost": { - "input": 3.259, - "output": 16.296 + "input": 0.25, + "output": 1.5, + "cache_read": 0.125 } }, - "mixtral-8x7B-instruct-v0.1": { - "id": "mixtral-8x7B-instruct-v0.1", - "name": "Mixtral 8x7B Instruct v0.1", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "qwen/qwen3.8-27b-hemmingway": { + "id": "qwen/qwen3.8-27b-hemmingway", + "name": "Qwen 3.8 27B Hemingway", + "description": "Qwen 3.8 27B Hemingway is an open-weight NVFP4 multimodal creative finetune for long-form prose, character dialogue, storytelling, and roleplay.", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2023-12-11", - "last_updated": "2023-12-11", + "tool_call": true, + "structured_output": false, + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -31469,60 +33306,64 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 32000 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.438, - "output": 0.68 + "input": 0.25, + "output": 1.5, + "cache_read": 0.125 } }, - "glm-4.5": { - "id": "glm-4.5", - "name": "GLM 4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 991808, + "input": 991808, + "output": 65536 }, "cost": { - "input": 0.67, - "output": 2.46 + "input": 0.19, + "output": 1.16, + "cache_read": 0.02, + "cache_write": 0.24 } }, - "kimi-k2-instruct": { - "id": "kimi-k2-instruct", - "name": "Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "qwen/qwen3-coder-flash": { + "id": "qwen/qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-07-11", - "last_updated": "2025-09-05", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -31531,29 +33372,67 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.551, - "output": 2.646 + "input": 0.3, + "output": 1.5, + "cache_read": 0.15 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, + "qwen/qwen3.5-plus:thinking": { + "id": "qwen/qwen3.5-plus:thinking", + "name": "Qwen3.5 Plus Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 983616, + "input": 983616, + "output": 65536 + }, + "cost": { + "input": 0.4, + "output": 2.4, + "cache_read": 0.04 + } + }, + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -31562,82 +33441,78 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.089, - "output": 0.275 + "input": 1, + "output": 5, + "cache_read": 0.5 } }, - "claude-opus4-8": { - "id": "claude-opus4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen/qwen3.5-9b": { + "id": "qwen/qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "input": 256000, + "output": 65536 }, "cost": { - "input": 5.64, - "output": 28.198, - "cache_read": 0.563, - "cache_write": 7.049 + "input": 0.05, + "output": 0.15, + "cache_read": 0.025 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "qwen/qwen3.5-122b-a10b:thinking": { + "id": "qwen/qwen3.5-122b-a10b:thinking", + "name": "Qwen3.5 122B A10B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -31645,34 +33520,34 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 1.28, - "output": 4.63, - "cache_read": 0.32 + "input": 0.437, + "output": 3.496, + "cache_read": 0.103788 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "qwen/qwen3.5-397b-a17b:thinking": { + "id": "qwen/qwen3.5-397b-a17b:thinking", + "name": "Qwen3.5 397B A17B Thinking", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-14", - "last_updated": "2026-04-14", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -31680,74 +33555,75 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 258048, + "input": 258048, + "output": 65536 }, "cost": { - "input": 1.31, - "output": 4.1, - "cache_read": 0.24 + "input": 0.6, + "output": 3.6, + "cache_read": 0.3 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "qwen/qwen3.8-flash": { + "id": "qwen/qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 991808, + "input": 991808, + "output": 131072 }, "cost": { - "input": 1.553, - "output": 3.106, - "cache_read": 0.003625 + "input": 0.14, + "output": 0.42, + "cache_read": 0.016, + "cache_write": 0.2 } }, - "qwen3-next-80b-a3b-thinking": { - "id": "qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B Thinking", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "qwen/qwen3.8-27b-cybersecurity": { + "id": "qwen/qwen3.8-27b-cybersecurity", + "name": "Qwen 3.8 27B Cybersecurity", + "description": "Qwen 3.8 27B Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports optional reasoning, image understanding, tool calling, and a 262,144-token context window.", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-11", - "last_updated": "2025-09-11", + "structured_output": false, + "release_date": "2026-09-19", + "last_updated": "2026-09-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -31755,32 +33631,65 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.164, - "output": 1.311 + "input": 0.1, + "output": 0.6, + "cache_read": 0.05 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "qwen/qwen-long": { + "id": "qwen/qwen-long", + "name": "Qwen Long 10M", + "description": "Alibaba's huge context window model. Takes in up to 10 million tokens, which is equivalent to dozens of books.", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-08-01", + "last_updated": "2024-08-01", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 10000000, + "input": 10000000, + "output": 8192 }, + "cost": { + "input": 0.1003, + "output": 0.408, + "cache_read": 0.05015 + } + }, + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5 35B A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -31788,38 +33697,29 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 260096, + "input": 260096, + "output": 65536 }, "cost": { - "input": 0.656, - "output": 2.731 + "input": 0.225, + "output": 1.8, + "cache_read": 0.1125 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "qwen/qwen3.6-max-preview": { + "id": "qwen/qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text" @@ -31828,34 +33728,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 245760, + "input": 245760, + "output": 65536 }, "cost": { - "input": 1.44, - "output": 4.53, - "cache_read": 0.39 + "input": 1.04, + "output": 6.24, + "cache_read": 0.52 } }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "qwen/qwen3-30b-a3b": { + "id": "qwen/qwen3-30b-a3b", + "name": "Qwen3 30B A3B", + "description": "Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -31863,34 +33762,33 @@ }, "open_weights": true, "limit": { - "context": 512000, - "output": 128000 + "context": 41000, + "input": 41000, + "output": 32768 }, "cost": { - "input": 0.355, - "output": 1.775, - "cache_read": 0.089 + "input": 0.1, + "output": 0.3, + "cache_read": 0.05 } }, - "minimax-m2": { - "id": "minimax-m2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -31898,27 +33796,27 @@ }, "open_weights": true, "limit": { - "context": 400000, - "output": 400000 + "context": 258048, + "input": 258048, + "output": 65536 }, "cost": { - "input": 0.39, - "output": 1.57 + "input": 0.6, + "output": 3.6, + "cache_read": 0.3 } }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-m2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "qwen/qwen3-8b": { + "id": "qwen/qwen3-8b", + "name": "Qwen 3 8B", + "description": "Qwen 3 8B is a 8B model. Supports switching between thinking and non thinking: trigger thinking with /think and /no_think anywhere in a prompt or system message to toggle chain-of-thought reasoning.", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" @@ -31929,34 +33827,29 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 196072 + "context": 41000, + "input": 41000, + "output": 32768 }, "cost": { "input": 0.47, - "output": 1.4 + "output": 0.47, + "cache_read": 0.235 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, + "qwen/qwen3.5-27b:thinking": { + "id": "qwen/qwen3.5-27b:thinking", + "name": "Qwen3.5 27B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", @@ -31969,39 +33862,32 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 260096, + "input": 260096, + "output": 65536 }, "cost": { - "input": 0.55, - "output": 2.76 + "input": 0.27, + "output": 2.16, + "cache_read": 0.135 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT Oss 120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "qwen/qwen3.8-27b-queen": { + "id": "qwen/qwen3.8-27b-queen", + "name": "Qwen 3.8 27B Queen", + "description": "Qwen 3.8 27B Queen is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 262,144-token context window.", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-01", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": false, + "release_date": "2026-09-09", + "last_updated": "2026-09-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -32009,26 +33895,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 524288, + "input": 524288, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.25, + "output": 1.5, + "cache_read": 0.125 } }, - "qwen-2.5-72b-instruct": { - "id": "qwen-2.5-72b-instruct", - "name": "Qwen2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen/qwen3-30b-a3b-instruct-2507": { + "id": "qwen/qwen3-30b-a3b-instruct-2507", + "name": "Qwen3 30B A3B Instruct 2507", + "description": "Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference. Significant improvements in general capabilities, including instruction following, logical reasoning, text comprehension, mathematics, science, coding and tool usage.", "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-20", + "last_updated": "2025-02-20", "modalities": { "input": [ "text" @@ -32039,26 +33926,28 @@ }, "open_weights": true, "limit": { - "context": 33000, - "output": 33000 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 0.062, - "output": 0.231 + "input": 0.2, + "output": 0.5, + "cache_read": 0.1 } }, - "hermes-4-70b": { - "id": "hermes-4-70b", - "name": "Hermes 4 70B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "qwen/qwen2.5-coder-32b-instruct": { + "id": "qwen/qwen2.5-coder-32b-instruct", + "name": "Qwen 2.5 Coder 32b", + "description": "Open coding-focused Qwen model for code generation, repair, and repository reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2024-11-12", + "last_updated": "2024-11-12", "modalities": { "input": [ "text" @@ -32069,45 +33958,43 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 32000, + "input": 32000, + "output": 8192 }, "cost": { - "input": 0.116, - "output": 0.358 + "input": 0.2006, + "output": 0.2006, + "cache_read": 0.1003 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "qwen/qwen3.7-plus:thinking": { + "id": "qwen/qwen3.7-plus:thinking", + "name": "Qwen3.7 Plus Thinking", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -32115,81 +34002,66 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 200000 + "context": 983616, + "input": 983616, + "output": 65536 }, "cost": { - "input": 1.09, - "output": 5.43 + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "cache_write": 0.5 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "qwen/qwen-2.5-72b-instruct": { + "id": "qwen/qwen-2.5-72b-instruct", + "name": "Qwen2.5 72B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": false, + "release_date": "2025-07-03", + "last_updated": "2025-07-03", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "output": 128000 + "context": 131072, + "input": 131072, + "output": 8192 }, "cost": { - "input": 3, - "output": 16.13, - "cache_read": 0.25 + "input": 0.357, + "output": 0.408, + "cache_read": 0.1785 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, + "qwen/qwen3.6-35b-a3b:thinking": { + "id": "qwen/qwen3.6-35b-a3b:thinking", + "name": "Qwen3.6 35B A3B Thinking", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, "release_date": "2026-04-17", "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -32197,62 +34069,68 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.81, - "output": 3.54, - "cache_read": 0.2 + "input": 0.112, + "output": 0.8, + "cache_read": 0.056 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT 4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, + "qwen/qwen3.6-27b": { + "id": "qwen/qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 260096, + "input": 260096, + "output": 65536 }, "cost": { - "input": 2.354, - "output": 9.417 + "input": 0.203, + "output": 2.24, + "cache_read": 0.1015 } }, - "nova-pro-v1": { - "id": "nova-pro-v1", - "name": "Nova Pro 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", - "attachment": false, + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen 3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -32260,30 +34138,35 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 5000 + "context": 991808, + "input": 991808, + "output": 65536 }, "cost": { - "input": 1.016, - "output": 4.061 + "input": 0.325, + "output": 1.95, + "cache_read": 0.0325, + "cache_write": 0.40625 } }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "Qwen3 Coder Next 80B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "qwen/qwen3.6-27b:thinking": { + "id": "qwen/qwen3.6-27b:thinking", + "name": "Qwen3.6 27B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-04", - "last_updated": "2026-02-04", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -32291,45 +34174,74 @@ }, "open_weights": true, "limit": { - "context": 256000, + "context": 260096, + "input": 260096, "output": 65536 }, "cost": { - "input": 0.158, - "output": 0.84 + "input": 0.203, + "output": 2.24, + "cache_read": 0.1015 } }, - "claude-opus4-5": { - "id": "claude-opus4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen/qwen3-14b": { + "id": "qwen/qwen3-14b", + "name": "Qwen 3 14b", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 41000, + "input": 41000, + "output": 32768 + }, + "cost": { + "input": 0.08, + "output": 0.24, + "cache_read": 0.04 + } + }, + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -32337,86 +34249,111 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 200000 + "context": 991808, + "input": 991808, + "output": 65536 }, "cost": { - "input": 5.98, - "output": 29.89 + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "cache_write": 0.5 } }, - "qwen3-coder-480b-a35b-instruct": { - "id": "qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "qwen/qwen3.5-flash:thinking": { + "id": "qwen/qwen3.5-flash:thinking", + "name": "Qwen3.5 Flash Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 991808, + "input": 991808, + "output": 65536 }, "cost": { - "input": 0.441, - "output": 1.984 + "input": 0.1, + "output": 0.4, + "cache_read": 0.05 } }, - "devstral-2512": { - "id": "devstral-2512", - "name": "Devstral 2 2512", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "attachment": false, - "reasoning": false, + "qwen/qwen3.8-max-0902": { + "id": "qwen/qwen3.8-max-0902", + "name": "Qwen3.8 Max 0902", + "description": "2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 991808, + "input": 991808, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 6, + "cache_read": 0.17, + "cache_write": 2.5 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5": { + "id": "failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5", + "name": "Llama 3 70B abliterated", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text" @@ -32427,31 +34364,45 @@ }, "open_weights": true, "limit": { - "context": 250000, - "output": 250000 + "context": 8192, + "input": 8192, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.7, + "output": 0.7, + "cache_read": 0.35 } }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT 5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ @@ -32460,29 +34411,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 1.235, - "output": 4.118, - "cache_read": 0.308, - "cache_write": 1.544 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "GLM 4.5 Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-01", - "last_updated": "2025-08-01", + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" @@ -32491,92 +34442,131 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 16385, + "input": 16385, + "output": 4096 }, "cost": { - "input": 0.22, - "output": 1.34 + "input": 0.5, + "output": 1.5 } }, - "glm-4.7-flash": { - "id": "glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, + "openai/gpt-terra-latest": { + "id": "openai/gpt-terra-latest", + "name": "GPT Terra Latest", + "description": "Compatibility alias that routes to GPT 5.6 Terra, the latest supported GPT Terra model.", + "family": "gpt-terra", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "structured_output": true, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 203000, - "output": 203000 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 0.09, - "output": 0.53 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "intellect-3": { - "id": "intellect-3", - "name": "INTELLECT 3", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "openai/gpt-6-astra-pro": { + "id": "openai/gpt-6-astra-pro", + "name": "GPT 6 Astra Pro", + "description": "GPT 6 Astra in Pro reasoning mode. Uses additional model work for difficult tasks, with higher latency and token usage at the same per-token rates. Reasoning effort remains independently configurable.", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-26", - "last_updated": "2025-11-26", + "structured_output": true, + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 1050000, + "input": 1050000, "output": 128000 }, "cost": { - "input": 0.219, - "output": 1.202 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "llama-3.1-405b-instruct": { - "id": "llama-3.1-405b-instruct", - "name": "Llama 3.1 405B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "openai/o3-pro-2025-06-10": { + "id": "openai/o3-pro-2025-06-10", + "name": "OpenAI o3-pro (2025-06-10)", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2025-06-10", "modalities": { "input": [ "text" @@ -32585,221 +34575,267 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 0, - "output": 0 + "input": 22, + "output": 88, + "cache_read": 11 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM 5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT 5.4 Nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 202752 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 1.08, - "output": 3.44 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "mistral-large-2512": { - "id": "mistral-large-2512", - "name": "Mistral Large 3 2512", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT 5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.05 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT 5.1 Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.266, - "output": 0.444 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "codestral-2508": { - "id": "codestral-2508", - "name": "Codestral 2508", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "mistral", - "attachment": false, + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-07-30", - "last_updated": "2025-07-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.03 - } - }, - "qwen3.5-122b-a10b": { - "id": "qwen3.5-122b-a10b", - "name": "Qwen3.5 122B A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.444, - "output": 3.106 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "nemotron-3-super-120b-a12b": { - "id": "nemotron-3-super-120b-a12b", - "name": "Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, + "openai/gpt-chat-latest": { + "id": "openai/gpt-chat-latest", + "name": "GPT Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "structured_output": true, + "release_date": "2026-05-03", + "last_updated": "2026-05-03", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 0.266, - "output": 0.799 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "glm-5-turbo": { - "id": "glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, + "openai/gpt-5.6-sol-pro": { + "id": "openai/gpt-5.6-sol-pro", + "name": "GPT 5.6 Sol Pro", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -32807,47 +34843,45 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 1.235, - "output": 4.118, - "cache_read": 0.308, - "cache_write": 1.544 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "claude-opus4-6": { - "id": "claude-opus4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT 5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "tool_call": false, + "structured_output": false, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -32855,38 +34889,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 5.98, - "output": 29.89 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } - } - } - }, - "nebius": { - "id": "nebius", - "env": [ - "NEBIUS_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.tokenfactory.nebius.com/v1", - "name": "Nebius Token Factory", - "doc": "https://docs.tokenfactory.nebius.com/", - "models": { - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + }, + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "OpenAI o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-12-05", - "last_updated": "2026-02-04", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text" @@ -32895,189 +34930,166 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 120000, - "output": 8192 + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 0.13, - "output": 0.4, - "cache_read": 0.013, - "cache_write": 0.16 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "moonshotai/Kimi-K2.5-fast": { - "id": "moonshotai/Kimi-K2.5-fast", - "name": "Kimi-K2.5-fast", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-k2", - "attachment": true, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "OpenAI o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-12-15", - "last_updated": "2026-02-04", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 8192 + "context": 200000, + "input": 200000, + "output": 100000 }, - "status": "deprecated", "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.05, - "cache_write": 0.625 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi-K2.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-k2", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT 5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-12-15", - "last_updated": "2026-02-04", + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 8192 + "context": 400000, + "input": 400000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.5, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.05, - "cache_write": 0.625 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma-3-27b-it", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT 4.1 Nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-10", - "release_date": "2026-01-20", - "last_updated": "2026-02-04", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 110000, - "input": 100000, - "output": 8192 + "context": 1047576, + "input": 1047576, + "output": 32768 }, "cost": { "input": 0.1, - "output": 0.3, - "cache_read": 0.01, - "cache_write": 0.125 - } - }, - "Qwen/Qwen3-Next-80B-A3B-Thinking-fast": { - "id": "Qwen/Qwen3-Next-80B-A3B-Thinking-fast", - "name": "Qwen3-Next-80B-A3B-Thinking-fast", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-25", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 8000, - "input": 7000, - "output": 8192 - }, - "status": "deprecated", - "cost": { - "input": 0.15, - "output": 1.2, - "cache_read": 0.015, - "cache_write": 0.1875 + "output": 0.4, + "cache_read": 0.025 } }, - "Qwen/Qwen2.5-VL-72B-Instruct": { - "id": "Qwen/Qwen2.5-VL-72B-Instruct", - "name": "Qwen2.5-VL-72B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT 5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-20", - "last_updated": "2026-02-04", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -33087,32 +35099,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 120000, - "output": 8192 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.025, - "cache_write": 0.31 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "Qwen/Qwen3-Embedding-8B": { - "id": "Qwen/Qwen3-Embedding-8B", - "name": "Qwen3-Embedding-8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "openai/o1": { + "id": "openai/o1", + "name": "OpenAI o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": false, "structured_output": false, "temperature": false, - "knowledge": "2025-10", - "release_date": "2026-01-10", - "last_updated": "2026-02-04", + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text" @@ -33121,180 +35142,170 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 0 + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 0.01, - "output": 0 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen3.5-397B-A17B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "openai/gpt-latest": { + "id": "openai/gpt-latest", + "name": "GPT Latest", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-15", - "last_updated": "2026-05-07", + "release_date": "2026-03-29", + "last_updated": "2026-03-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 250000, - "output": 8192 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3.6, - "cache_read": 0.06, - "cache_write": 0.75 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "Qwen/Qwen3.5-397B-A17B-fast": { - "id": "Qwen/Qwen3.5-397B-A17B-fast", - "name": "Qwen3.5-397B-A17B-fast", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": false, + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT 5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "high" + ] } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-15", - "last_updated": "2026-05-07", + "tool_call": false, + "structured_output": false, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8000, - "input": 7000, - "output": 8192 + "context": 400000, + "input": 400000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.6, - "output": 3.6, - "cache_read": 0.06, - "cache_write": 0.75 + "input": 15, + "output": 120, + "cache_read": 1.5 } }, - "Qwen/Qwen3-30B-A3B-Instruct-2507": { - "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen3-30B-A3B-Instruct-2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "openai/gpt-4o-2024-08-06": { + "id": "openai/gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-01-28", - "last_updated": "2026-02-04", + "knowledge": "2023-09", + "release_date": "2024-08-06", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "input": 120000, - "output": 8192 + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01, - "cache_write": 0.125 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "Qwen/Qwen3-Next-80B-A3B-Thinking": { - "id": "Qwen/Qwen3-Next-80B-A3B-Thinking", - "name": "Qwen3-Next-80B-A3B-Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "attachment": false, + "openai/gpt-6-astra": { + "id": "openai/gpt-6-astra", + "name": "GPT 6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-01-28", - "last_updated": "2026-02-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "input": 120000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 1.2, - "reasoning": 1.2, - "cache_read": 0.015, - "cache_write": 0.18 - } - }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-25", - "last_updated": "2025-10-04", + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -33302,136 +35313,125 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 8192 - }, - "cost": { - "input": 0.2, - "output": 0.6 - } - }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen3-32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-01-28", - "last_updated": "2026-02-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "input": 120000, - "output": 8192 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01, - "cache_write": 0.125 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "Qwen/Qwen3-235B-A22B-Thinking-2507-fast": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507-fast", - "name": "Qwen3-235B-A22B-Thinking-2507-fast", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": false, + "openai/gpt-luna-latest": { + "id": "openai/gpt-luna-latest", + "name": "GPT Luna Latest", + "description": "Compatibility alias that routes to GPT 5.6 Luna, the latest supported GPT Luna model.", + "family": "gpt-luna", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-25", - "last_updated": "2026-05-07", + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8000, - "input": 7000, - "output": 8192 + "context": 1050000, + "input": 1050000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.5, - "output": 2, - "cache_read": 0.05, - "cache_write": 0.625 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 } }, - "openai/gpt-oss-120b-fast": { - "id": "openai/gpt-oss-120b-fast", - "name": "gpt-oss-120b-fast", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": false, + "openai/gpt-5.6-luna-pro": { + "id": "openai/gpt-5.6-luna-pro", + "name": "GPT 5.6 Luna Pro", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-06-10", - "last_updated": "2026-05-07", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8000, - "input": 7000, - "output": 8192 + "context": 1050000, + "input": 1050000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "attachment": false, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT 5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -33439,90 +35439,89 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-01-10", - "last_updated": "2026-02-04", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 124000, - "output": 8192 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6, - "reasoning": 0.6, - "cache_read": 0.015, - "cache_write": 0.18 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "NousResearch/Hermes-4-405B": { - "id": "NousResearch/Hermes-4-405B", - "name": "Hermes-4-405B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-11", - "release_date": "2026-01-30", - "last_updated": "2026-02-04", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "input": 120000, - "output": 8192 + "input": 128000, + "output": 16384 }, "cost": { - "input": 1, - "output": 3, - "reasoning": 3, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "NousResearch/Hermes-4-70B": { - "id": "NousResearch/Hermes-4-70B", - "name": "Hermes-4-70B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-11", - "release_date": "2026-01-30", - "last_updated": "2026-02-04", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -33534,30 +35533,36 @@ "open_weights": true, "limit": { "context": 128000, - "input": 120000, - "output": 8192 + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.13, - "output": 0.4, - "reasoning": 0.4, - "cache_read": 0.013, - "cache_write": 0.16 + "input": 0.2, + "output": 0.3 } }, - "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B": { - "id": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", - "name": "Nemotron-3-Nano-30B-A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", + "openai/gpt-oss-safeguard-20b": { + "id": "openai/gpt-oss-safeguard-20b", + "name": "GPT OSS Safeguard 20B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-08-10", - "last_updated": "2026-02-04", + "release_date": "2025-10-29", + "last_updated": "2025-10-29", "modalities": { "input": [ "text" @@ -33568,31 +35573,36 @@ }, "open_weights": true, "limit": { - "context": 32000, - "input": 30000, - "output": 4096 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.006, - "cache_write": 0.075 + "input": 0.075, + "output": 0.3 } }, - "nvidia/Nemotron-3-Nano-Omni": { - "id": "nvidia/Nemotron-3-Nano-Omni", - "name": "Nemotron-3-Nano-Omni", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", + "openai/o3-mini-high": { + "id": "openai/o3-mini-high", + "name": "OpenAI o3-mini (High)", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-20", - "last_updated": "2026-05-07", + "release_date": "2024-01-01", + "last_updated": "2025-01-31", "modalities": { "input": [ "text" @@ -33601,109 +35611,140 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "input": 60000, - "output": 8192 + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.006, - "cache_write": 0.075 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "nvidia/Llama-3_1-Nemotron-Ultra-253B-v1": { - "id": "nvidia/Llama-3_1-Nemotron-Ultra-253B-v1", - "name": "Llama-3.1-Nemotron-Ultra-253B-v1", - "description": "Flagship Nemotron model for high-throughput reasoning and complex agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, + "openai/gpt-sol-latest": { + "id": "openai/gpt-sol-latest", + "name": "GPT Sol Latest", + "description": "Compatibility alias that routes to GPT 5.6 Sol, the latest supported GPT Sol model.", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-15", - "last_updated": "2026-02-04", + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 120000, - "output": 4096 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.06, - "cache_write": 0.75 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "Nemotron-3-Super-120B-A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, + "openai/gpt-6-sol-pro": { + "id": "openai/gpt-6-sol-pro", + "name": "GPT 6 Sol Pro", + "description": "GPT-6 Sol Pro uses the same underlying model as GPT-6 Sol with Pro reasoning mode enabled for higher-quality responses on complex tasks.", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-03-11", - "last_updated": "2026-03-12", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM-5", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": false, + "openai/gpt-astra-latest": { + "id": "openai/gpt-astra-latest", + "name": "GPT Astra Latest", + "description": "Compatibility alias that routes to GPT 6 Astra, the latest supported GPT Astra model.", + "family": "gpt-astra", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-03-01", - "last_updated": "2026-03-10", + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -33711,337 +35752,345 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 16384 + "context": 1050000, + "input": 1050000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.1, - "cache_write": 1 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT 5.4 Mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 432000, - "output": 432000 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.4 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "deepseek-ai/DeepSeek-V3.2-fast": { - "id": "deepseek-ai/DeepSeek-V3.2-fast", - "name": "DeepSeek-V3.2-fast", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": false, + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT 5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-27", - "last_updated": "2026-05-07", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8000, - "input": 7000, - "output": 8192 + "context": 1050000, + "input": 1050000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.04, - "cache_write": 0.5 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT 5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 1.75, - "output": 3.5, - "cache_read": 0.15 + "input": 2.5, + "output": 20, + "cache_read": 0.25 } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "DeepSeek-V3.2", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": false, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT 5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-11", - "release_date": "2026-01-20", - "last_updated": "2026-02-04", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163000, - "input": 160000, - "output": 16384 + "context": 400000, + "input": 400000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.3, - "output": 0.45, - "reasoning": 0.45, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": false, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT 5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-20", - "last_updated": "2026-05-07", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "input": 190000, - "output": 8192 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "MiniMaxAI/MiniMax-M2.5-fast": { - "id": "MiniMaxAI/MiniMax-M2.5-fast", - "name": "MiniMax-M2.5-fast", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT 4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-20", - "last_updated": "2026-05-07", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8000, - "input": 7000, - "output": 8192 + "context": 1047576, + "input": 1047576, + "output": 32768 }, - "status": "deprecated", "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "PrimeIntellect/INTELLECT-3": { - "id": "PrimeIntellect/INTELLECT-3", - "name": "INTELLECT-3", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": false, + "openai/gpt-4o-2024-11-20": { + "id": "openai/gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-10", - "release_date": "2026-01-25", - "last_updated": "2026-02-04", + "knowledge": "2023-09", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "input": 120000, - "output": 8192 + "input": 128000, + "output": 16384 }, - "status": "deprecated", "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.02, - "cache_write": 0.25 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } - } - } - }, - "auriko": { - "id": "auriko", - "env": [ - "AURIKO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.auriko.ai/v1", - "name": "Auriko", - "doc": "https://docs.auriko.ai", - "models": { - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + }, + "openai/o3-mini-low": { + "id": "openai/o3-mini-low", + "name": "OpenAI o3-mini (Low)", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ "text" @@ -34050,49 +36099,35 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT 4.1 Mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -34100,56 +36135,42 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1047576, + "input": 1047576, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "grok-4.3": { - "id": "grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "openai/gpt-5.6-terra-pro": { + "id": "openai/gpt-5.6-terra-pro", + "name": "GPT 5.6 Terra Pro", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -34162,45 +36183,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 2.5, + "input": 2, + "output": 12, "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "cache_write": 2.5 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", + "openai/gpt-6-luna": { + "id": "openai/gpt-6-luna", + "name": "GPT 6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -34211,16 +36216,14 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -34229,31 +36232,38 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 0.05, + "output": 0.25, + "cache_read": 0.005, + "cache_write": 0.0625 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "openai/o4-mini-high": { + "id": "openai/o4-mini-high", + "name": "OpenAI o4-mini high", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-12-04", + "last_updated": "2025-04-16", "modalities": { "input": [ "text" @@ -34262,31 +36272,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 200000, - "output": 131072 + "input": 200000, + "output": 100000 }, "cost": { - "input": 1.4, + "input": 1.1, "output": 4.4, - "cache_read": 0.26 + "cache_read": 0.55 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "openai/gpt-6-luna-pro": { + "id": "openai/gpt-6-luna-pro", + "name": "GPT 6 Luna Pro", + "description": "GPT-6 Luna Pro uses the same underlying model as GPT-6 Luna with Pro reasoning mode enabled for higher-quality responses on complex tasks.", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -34296,47 +36305,44 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0.05, + "output": 0.25, + "cache_read": 0.005, + "cache_write": 0.0625 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT 5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -34346,10 +36352,11 @@ } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -34362,79 +36369,72 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 1050000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } - }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "minimax-m2-7-highspeed": { - "id": "minimax-m2-7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_write": 0.375 + "input": 10, + "output": 30 } }, - "minimax-m2-7": { - "id": "minimax-m2-7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -34445,33 +36445,42 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_write": 0.375 + "input": 0.35, + "output": 0.75 } }, - "qwen-3.6-plus": { - "id": "qwen-3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": false, + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT 5.1 Codex Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -34479,165 +36488,135 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "openai/o3": { + "id": "openai/o3", + "name": "OpenAI o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 0.5, - "output": 2.8 + "input": 2, + "output": 8, + "cache_read": 1 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT 5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT 5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -34646,62 +36625,42 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { "input": 2, - "output": 12, + "output": 10, "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "cache_write": 2.5 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "openai/gpt-5.1-2025-11-13": { + "id": "openai/gpt-5.1-2025-11-13", + "name": "GPT-5.1 (2025-11-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -34709,42 +36668,42 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 400000, "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "openai/gpt-6-sol": { + "id": "openai/gpt-6-sol", + "name": "GPT 6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", @@ -34757,43 +36716,35 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1050000, + "input": 1050000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } - } - } - }, - "stepfun-ai": { - "id": "stepfun-ai", - "env": [ - "STEPFUN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.stepfun.ai/step_plan/v1", - "name": "StepFun AI", - "doc": "https://platform.stepfun.ai/docs/en/step-plan/integrations/open-code", - "models": { - "step-2-16k": { - "id": "step-2-16k", - "name": "Step 2 (16K)", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-01-01", - "last_updated": "2026-02-13", + }, + "openai/o1-pro": { + "id": "openai/o1-pro", + "name": "OpenAI o1 Pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2025-03-19", + "last_updated": "2025-03-19", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -34801,56 +36752,59 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 5.21, - "output": 16.44, - "cache_read": 1.04 + "input": 150, + "output": 600, + "cache_read": 75 } }, - "step-tts-2": { - "id": "step-tts-2", - "name": "Step TTS 2", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "step", - "attachment": false, + "microsoft/wizardlm-2-8x22b": { + "id": "microsoft/wizardlm-2-8x22b", + "name": "WizardLM-2 8x22B", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-03-01", - "last_updated": "2026-07-02", + "structured_output": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 65536, + "input": 65536, + "output": 8192 + }, + "cost": { + "input": 0.493, + "output": 0.493, + "cache_read": 0.2465 } }, - "step-3.5-flash": { - "id": "step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "NousResearch/hermes-3-llama-3.1-70b": { + "id": "NousResearch/hermes-3-llama-3.1-70b", + "name": "Hermes 3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "nousresearch", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-06-15", + "structured_output": false, + "release_date": "2026-01-07", + "last_updated": "2026-01-07", "modalities": { "input": [ "text" @@ -34861,70 +36815,125 @@ }, "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 65536, + "input": 65536, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 0.408, + "output": 0.408, + "cache_read": 0.204 } }, - "stepaudio-2.5-asr": { - "id": "stepaudio-2.5-asr", - "name": "StepAudio 2.5 ASR", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "step", + "NousResearch/hermes-4-405b": { + "id": "NousResearch/hermes-4-405b", + "name": "Hermes 4 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nousresearch", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-04-24", - "last_updated": "2026-07-02", + "structured_output": true, + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.15 } }, - "stepaudio-2.5-tts": { - "id": "stepaudio-2.5-tts", - "name": "StepAudio 2.5 TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "step", + "NousResearch/hermes-4-405b:thinking": { + "id": "NousResearch/hermes-4-405b:thinking", + "name": "Hermes 4 Large (Thinking)", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nousresearch", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-04-16", - "last_updated": "2026-07-02", + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.15 } }, - "step-3.5-flash-2603": { - "id": "step-3.5-flash-2603", - "name": "Step 3.5 Flash 2603", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0": { + "id": "ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0", + "name": "Omega Directive 24B Unslop v2.0", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "input": 32768, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 0.5, + "cache_read": 0.25 + } + } + } + }, + "abliteration-ai": { + "id": "abliteration-ai", + "env": [ + "ABLIT_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.abliteration.ai/v1", + "name": "abliteration.ai", + "doc": "https://docs.abliteration.ai/models", + "models": { + "abliterated-model-large-v2": { + "id": "abliterated-model-large-v2", + "name": "Abliterated Model Large V2", + "description": "GLM-5.3 model abliterated and finetuned for cyber, ML red teaming, and agent testing", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -34932,18 +36941,16 @@ "type": "effort", "values": [ "low", - "high" + "high", + "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-08-29", + "last_updated": "2026-08-31", "modalities": { "input": [ "text" @@ -34952,42 +36959,46 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 1000000, + "input": 1000000, + "output": 999990 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 5, + "output": 5, + "cache_read": 0.5 } }, - "step-3.7-flash": { - "id": "step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "abliterated-model": { + "id": "abliterated-model", + "name": "Abliterated Model", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] + }, + { + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "knowledge": "2026-01-01", - "release_date": "2026-05-29", - "last_updated": "2026-06-29", + "release_date": "2026-01-06", + "last_updated": "2026-07-28", "modalities": { "input": [ "text", @@ -34999,28 +37010,40 @@ }, "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 150000, + "input": 150000, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "input": 3, + "output": 3, + "cache_read": 0.3 } }, - "step-1-32k": { - "id": "step-1-32k", - "name": "Step 1 (32K)", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "abliterated-model-large": { + "id": "abliterated-model-large", + "name": "Abliterated Model Large", + "description": "GLM-5.2 model abliterated and finetuned for cyber, ML red teaming, and agent testing", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + }, + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-01-01", - "last_updated": "2026-02-13", + "release_date": "2026-07-25", + "last_updated": "2026-07-28", "modalities": { "input": [ "text" @@ -35031,34 +37054,34 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 999990 }, "cost": { - "input": 2.05, - "output": 9.59, - "cache_read": 0.41 + "input": 5, + "output": 5, + "cache_read": 0.5 } } } }, - "vivgrid": { - "id": "vivgrid", + "crof": { + "id": "crof", "env": [ - "VIVGRID_API_KEY" + "CROF_API_KEY" ], - "npm": "@ai-sdk/openai", - "api": "https://api.vivgrid.com/v1", - "name": "Vivgrid", - "doc": "https://docs.vivgrid.com/models", + "npm": "@ai-sdk/openai-compatible", + "api": "https://crof.ai/v1", + "name": "CrofAI", + "doc": "https://crof.ai/docs", "models": { - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "gemma-4-31b-it": { + "id": "gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -35072,17 +37095,14 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -35090,23 +37110,23 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 262144, + "output": 262144 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "qwen3.8-27b": { + "id": "qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -35122,46 +37142,46 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "output": 1.5, + "cache_read": 0.03 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "kimi-k3-eco": { + "id": "kimi-k3-eco", + "name": "Kimi K3 Eco", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", "high", "max" ] @@ -35172,12 +37192,14 @@ "field": "reasoning_content" }, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -35186,121 +37208,172 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 1.2, - "output": 4.2, - "cache_read": 0.3 + "input": 1, + "output": 4, + "cache_read": 0.1 } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM 5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.07, + "output": 0.22, + "cache_read": 0.01 } }, - "gpt-5.1-codex-max": { - "id": "gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash (New)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.08, + "output": 0.1, + "cache_read": 0.003 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 2, + "output": 8, + "cache_read": 0.25 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "deepseek-v4-flash-vision-exp": { + "id": "deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -35315,90 +37388,41 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } - }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.08, + "output": 0.2, + "cache_read": 0.007 } }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -35413,80 +37437,51 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.5, + "output": 1.99, + "cache_read": 0.05 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "greg-2-super": { + "id": "greg-2-super", + "name": "Greg 2 Super", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-06-14", + "last_updated": "2026-06-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -35494,23 +37489,20 @@ }, "open_weights": false, "limit": { - "context": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 229376, + "output": 229376 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.03 + "input": 1.5, + "output": 5, + "cache_read": 0.25 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro (0813)", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -35525,10 +37517,13 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -35539,23 +37534,24 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.28, - "output": 0.42 + "input": 0.35, + "output": 0.8, + "cache_read": 0.01 } }, - "gemini-3.1-flash-lite-preview": { - "id": "gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -35569,55 +37565,64 @@ } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 1 + "input": 0.4, + "output": 0.8, + "cache_read": 0.003, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "greg-rp": { + "id": "greg-rp", + "name": "Greg (Roleplay)", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -35625,20 +37630,20 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 229376, + "output": 229376 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -35653,69 +37658,40 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2026-03-13", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.04, + "output": 0.15, + "cache_read": 0.008 } - } - } - }, - "tinfoil": { - "id": "tinfoil", - "env": [ - "TINFOIL_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.tinfoil.sh/v1", - "name": "Tinfoil", - "doc": "https://docs.tinfoil.sh", - "models": { - "kimi-k2-6": { - "id": "kimi-k2-6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + }, + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -35723,26 +37699,23 @@ "type": "effort", "values": [ "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" @@ -35750,55 +37723,30 @@ }, "open_weights": true, "limit": { - "context": 256000, + "context": 262144, "output": 262144 }, - "cost": { - "input": 1.5, - "output": 5.25 - } - }, - "llama3-3-70b": { - "id": "llama3-3-70b", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 1.75, - "output": 2.75 + "input": 0.35, + "output": 1.75, + "cache_read": 0.07 } }, - "gpt-oss-safeguard-120b": { - "id": "gpt-oss-safeguard-120b", - "name": "gpt-oss-safeguard-120b", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -35806,11 +37754,13 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -35821,54 +37771,30 @@ }, "open_weights": true, "limit": { - "context": 131000, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "nomic-embed-text": { - "id": "nomic-embed-text", - "name": "Nomic Embed Text v1.5", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "release_date": "2024-02", - "last_updated": "2024-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 131072 }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 768 + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.05, - "output": 0 + "input": 0.3, + "output": 1.05, + "cache_read": 0.05 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -35876,11 +37802,13 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -35891,19 +37819,24 @@ }, "open_weights": true, "limit": { - "context": 131000, - "output": 32768 + "context": 202752, + "output": 202752 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.45, + "output": 2.15, + "cache_read": 0.08, + "cache_write": 0 } }, - "glm-5-2": { - "id": "glm-5-2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -35911,20 +37844,21 @@ "type": "effort", "values": [ "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -35935,19 +37869,23 @@ }, "open_weights": true, "limit": { - "context": 384000, + "context": 1000000, "output": 131072 }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, "cost": { - "input": 1.5, - "output": 5.25 + "input": 0.35, + "output": 0.8, + "cache_read": 0.003 } }, - "gemma4-31b": { - "id": "gemma4-31b", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "qwen3.6-27b": { + "id": "qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -35955,24 +37893,23 @@ "type": "effort", "values": [ "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" @@ -35980,98 +37917,58 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 1 - } - } - } - }, - "mistral": { - "id": "mistral", - "env": [ - "MISTRAL_API_KEY" - ], - "npm": "@ai-sdk/mistral", - "name": "Mistral", - "doc": "https://docs.mistral.ai/getting-started/models/", - "models": { - "codestral-latest": { - "id": "codestral-latest", - "name": "Codestral (latest)", - "description": "Mistral code model for completions, refactors, and developer IDE workflows", - "family": "codestral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-05-29", - "last_updated": "2025-01-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 262144, + "output": 262144 }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 4096 + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0.2, + "output": 1.5, + "cache_read": 0.04 } }, - "mistral-large-latest": { - "id": "mistral-large-latest", - "name": "Mistral Large (latest)", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, + "greg-2-ultra": { + "id": "greg-2-ultra", + "name": "Greg 2 Ultra", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", + "release_date": "2026-06-14", + "last_updated": "2026-06-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 229376, + "output": 229376 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 3, + "output": 10, + "cache_read": 0.5 } }, - "open-mistral-7b": { - "id": "open-mistral-7b", - "name": "Mistral 7B", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-09-27", - "last_updated": "2023-09-27", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ "text" @@ -36080,28 +37977,43 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8000, - "output": 8000 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0.25, - "output": 0.25 + "input": 0.18, + "output": 0.35, + "cache_read": 0.04 } }, - "devstral-small-2507": { - "id": "devstral-small-2507", - "name": "Devstral Small", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-07-10", - "last_updated": "2025-07-10", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -36112,30 +38024,50 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, - "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.4, + "output": 1.4, + "cache_read": 0.06 } }, - "ministral-3b-latest": { - "id": "ministral-3b-latest", - "name": "Ministral 3B (latest)", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, - "reasoning": false, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-04", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -36143,26 +38075,28 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.04, - "output": 0.04 + "input": 0.55, + "output": 2.25, + "cache_read": 0.05 } }, - "pixtral-large-latest": { - "id": "pixtral-large-latest", - "name": "Pixtral Large (latest)", - "description": "Mistral's larger vision model for document-heavy image understanding and chat", - "family": "pixtral", + "greg-1-mini": { + "id": "greg-1-mini", + "name": "Greg 1 Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2024-11-04", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", @@ -36172,28 +38106,44 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 229376, + "output": 229376 }, "cost": { - "input": 2, - "output": 6 + "input": 0.07, + "output": 0.15, + "cache_read": 0.01 } }, - "mistral-nemo": { - "id": "mistral-nemo", - "name": "Mistral Nemo", - "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", - "family": "mistral-nemo", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -36204,55 +38154,50 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.15, - "output": 0.15 - } - }, - "mistral-embed": { - "id": "mistral-embed", - "name": "Mistral Embed", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "mistral-embed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2023-12-11", - "last_updated": "2023-12-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 131072 }, - "open_weights": false, - "limit": { - "context": 8000, - "output": 3072 + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.1, - "output": 0 + "input": 0.12, + "output": 0.21, + "cache_read": 0.003 } - }, - "mistral-small-2506": { - "id": "mistral-small-2506", - "name": "Mistral Small 3.2", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, - "reasoning": false, + } + } + }, + "standardcompute": { + "id": "standardcompute", + "env": [ + "STANDARDCOMPUTE_API_KEY" + ], + "npm": "@openrouter/ai-sdk-provider", + "api": "https://api.stdcmpt.com/v1", + "name": "Standard Compute", + "doc": "https://standardcompute.com/models", + "models": { + "standardcompute": { + "id": "standardcompute", + "name": "Standard Compute", + "description": "Flat-rate smart-routing gateway: one model id, each request routed across a curated catalog of 1M-context models (DeepSeek, GLM, MiniMax, Qwen, GPT-5.6, Claude 5, Gemini 2.5, Kimi) or pinned to a user-selected model", + "family": "auto", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_details" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", + "release_date": "2026-03-01", + "last_updated": "2026-08-24", "modalities": { "input": [ "text", @@ -36262,28 +38207,40 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 24576 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0, + "output": 0 } - }, - "ministral-8b-latest": { - "id": "ministral-8b-latest", - "name": "Ministral 8B (latest)", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + } + } + }, + "cloudferro-sherlock": { + "id": "cloudferro-sherlock", + "env": [ + "CLOUDFERRO_SHERLOCK_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api-sherlock.cloudferro.com/openai/v1/", + "name": "CloudFerro Sherlock", + "doc": "https://docs.sherlock.cloudferro.com/", + "models": { + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-04", + "knowledge": "2024-10-09", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -36294,26 +38251,28 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 70000, + "output": 70000 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 2.92, + "output": 2.92 } }, - "open-mixtral-8x22b": { - "id": "open-mixtral-8x22b", - "name": "Mixtral 8x22B", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mixtral", + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-17", - "last_updated": "2024-04-17", + "knowledge": "2026-01", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text" @@ -36324,26 +38283,28 @@ }, "open_weights": true, "limit": { - "context": 64000, - "output": 64000 + "context": 196000, + "input": 180000, + "output": 16000 }, "cost": { - "input": 2, - "output": 6 + "input": 0.3, + "output": 1.2 } }, - "mistral-medium-latest": { - "id": "mistral-medium-latest", - "name": "Mistral Medium (latest)", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "OpenAI GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } @@ -36351,12 +38312,11 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -36364,26 +38324,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131000, + "output": 131000 }, "cost": { - "input": 1.5, - "output": 7.5 + "input": 2.92, + "output": 2.92 } }, - "devstral-small-2505": { - "id": "devstral-small-2505", - "name": "Devstral Small 2505", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", + "speakleash/Bielik-11B-v2.6-Instruct": { + "id": "speakleash/Bielik-11B-v2.6-Instruct", + "name": "Bielik 11B v2.6 Instruct", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "knowledge": "2025-03", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ "text" @@ -36394,28 +38354,26 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 32000, + "output": 32000 }, - "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.67, + "output": 0.67 } }, - "magistral-small": { - "id": "magistral-small", - "name": "Magistral Small", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-small", + "speakleash/Bielik-11B-v3.0-Instruct": { + "id": "speakleash/Bielik-11B-v3.0-Instruct", + "name": "Bielik 11B v3.0 Instruct", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-17", + "knowledge": "2025-03", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ "text" @@ -36426,211 +38384,302 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 32000, + "output": 32000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 0.67, + "output": 0.67 } - }, - "mistral-medium-2604": { - "id": "mistral-medium-2604", - "name": "Mistral Medium 3.5", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", + } + } + }, + "anthropic": { + "id": "anthropic", + "env": [ + "ANTHROPIC_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "name": "Anthropic", + "doc": "https://docs.anthropic.com/en/docs/about-claude/models", + "models": { + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.5, - "output": 7.5 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "mistral-small-latest": { - "id": "mistral-small-latest", - "name": "Mistral Small (latest)", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "open-mixtral-8x7b": { - "id": "open-mixtral-8x7b", - "name": "Mixtral 8x7B", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mixtral", - "attachment": false, - "reasoning": false, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-01", - "release_date": "2023-12-11", - "last_updated": "2023-12-11", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 32000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.7, - "output": 0.7 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "devstral-latest": { - "id": "devstral-latest", - "name": "Devstral 2", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", - "attachment": false, - "reasoning": false, + "claude-opus-5-5": { + "id": "claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 8, + "output": 40, + "cache_read": 0.4, + "cache_write": 10 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 2 + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 } }, - "mistral-small-2603": { - "id": "mistral-small-2603", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", + "claude-fable-5-1": { + "id": "claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "high" + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "mistral-medium-2505": { - "id": "mistral-medium-2505", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "release_date": "2025-11-24", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -36638,60 +38687,112 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.4, - "output": 2 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "mistral-large-2411": { - "id": "mistral-large-2411", - "name": "Mistral Large 2.1", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": false, - "reasoning": false, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-18", - "last_updated": "2024-11-18", + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 2, - "output": 6 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "mistral-medium-2508": { - "id": "mistral-medium-2508", - "name": "Mistral Medium 3.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-08-12", - "last_updated": "2025-08-12", + "structured_output": true, + "temperature": false, + "release_date": "2026-06-07", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -36699,288 +38800,387 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 2 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "open-mistral-nemo": { - "id": "open-mistral-nemo", - "name": "Open Mistral Nemo", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 1000000, "output": 128000 }, - "status": "deprecated", + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } + }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "magistral-medium-latest": { - "id": "magistral-medium-latest", - "name": "Magistral Medium (latest)", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-medium", - "attachment": false, + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-20", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 2, - "output": 5 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "devstral-medium-latest": { - "id": "devstral-medium-latest", - "name": "Devstral 2 (latest)", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", - "attachment": false, - "reasoning": false, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-29", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 2 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "devstral-2512": { - "id": "devstral-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", - "attachment": false, - "reasoning": false, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "knowledge": "2025-05-31", + "release_date": "2026-02-04", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 2 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "labs-devstral-small-2512": { - "id": "labs-devstral-small-2512", - "name": "Devstral Small 2", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", - "attachment": false, - "reasoning": false, + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 200000, + "output": 64000 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "pixtral-12b": { - "id": "pixtral-12b", - "name": "Pixtral 12B", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-09-01", - "last_updated": "2024-09-01", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "mistral-large-2512": { - "id": "mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-14", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "devstral-medium-2507": { - "id": "devstral-medium-2507", - "name": "Devstral Medium", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-07-10", - "last_updated": "2025-07-10", - "modalities": { - "input": [ - "text" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 1000000, "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 2 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } } } }, - "cloudflare-workers-ai": { - "id": "cloudflare-workers-ai", + "tinfoil": { + "id": "tinfoil", "env": [ - "CLOUDFLARE_ACCOUNT_ID", - "CLOUDFLARE_API_KEY" + "TINFOIL_API_KEY" ], "npm": "@ai-sdk/openai-compatible", - "api": "https://api.cloudflare.com/client/v4/accounts/${CLOUDFLARE_ACCOUNT_ID}/ai/v1", - "name": "Cloudflare Workers AI", - "doc": "https://developers.cloudflare.com/workers-ai/models/", + "api": "https://inference.tinfoil.sh/v1", + "name": "Tinfoil", + "doc": "https://docs.tinfoil.sh", "models": { - "@cf/ibm-granite/granite-4.0-h-micro": { - "id": "@cf/ibm-granite/granite-4.0-h-micro", - "name": "Granite 4.0 H Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "granite", + "gpt-oss-safeguard-120b": { + "id": "gpt-oss-safeguard-120b", + "name": "gpt-oss-safeguard-120b", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-10-07", - "last_updated": "2025-10-07", + "knowledge": "2024-06", + "release_date": "2025-10-29", + "last_updated": "2025-10-29", "modalities": { "input": [ "text" @@ -36991,44 +39191,42 @@ }, "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.017, - "output": 0.112 + "input": 0.15, + "output": 0.6 } }, - "@cf/moonshotai/kimi-k2.7-code": { - "id": "@cf/moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "glm-5-3-flash": { + "id": "glm-5-3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" @@ -37036,48 +39234,42 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.4, + "output": 1.25, + "cache_read": 0.1 } }, - "@cf/moonshotai/kimi-k2.6": { - "id": "@cf/moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -37086,36 +39278,66 @@ "open_weights": true, "limit": { "context": 262144, - "output": 256000 + "output": 131072 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 4, + "output": 20, + "cache_read": 0.8 } }, - "@cf/google/gemma-4-26b-a4b-it": { - "id": "@cf/google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "llama3-3-70b": { + "id": "llama3-3-70b", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 4096 + }, + "cost": { + "input": 1.75, + "output": 2.75 + } + }, + "gemma4-31b": { + "id": "gemma4-31b", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": true, "release_date": "2026-04-02", @@ -37131,19 +39353,19 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.4, + "output": 1 } }, - "@cf/openai/gpt-oss-120b": { - "id": "@cf/openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "glm-5-3": { + "id": "glm-5-3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -37151,16 +39373,16 @@ "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -37171,30 +39393,46 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.35, - "output": 0.75 + "input": 1.8, + "output": 5.75, + "cache_read": 0.45 } }, - "@cf/openai/gpt-oss-20b": { - "id": "@cf/openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "deepseek-v4-1-flash": { + "id": "deepseek-v4-1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -37202,26 +39440,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 384000 }, + "status": "beta", "cost": { - "input": 0.2, - "output": 0.3 + "input": 0.65, + "output": 1.45, + "cache_read": 0.13 } }, - "@cf/mistralai/mistral-small-3.1-24b-instruct": { - "id": "@cf/mistralai/mistral-small-3.1-24b-instruct", - "name": "Mistral Small 3.1 24B Instruct", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "nomic-embed-text": { + "id": "nomic-embed-text", + "name": "Nomic Embed Text v1.5", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": false, - "temperature": true, - "release_date": "2025-03-18", - "last_updated": "2025-03-18", + "temperature": false, + "release_date": "2024-02", + "last_updated": "2024-02", "modalities": { "input": [ "text" @@ -37232,25 +39471,22 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 8192, + "output": 768 }, "cost": { - "input": 0.351, - "output": 0.555 + "input": 0.05, + "output": 0 } }, - "@cf/nvidia/nemotron-3-120b-a12b": { - "id": "@cf/nvidia/nemotron-3-120b-a12b", - "name": "Nemotron 3 Super 120B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -37261,11 +39497,11 @@ } ], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "knowledge": "2024-06", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -37276,42 +39512,42 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 0.15, + "output": 0.6 } - }, - "@cf/zai-org/glm-5.2": { - "id": "@cf/zai-org/glm-5.2", - "name": "Glm 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + } + } + }, + "llama": { + "id": "llama", + "env": [ + "LLAMA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.llama.com/compat/v1/", + "name": "Llama", + "doc": "https://llama.developer.meta.com/docs/models", + "models": { + "llama-4-maverick-17b-128e-instruct-fp8": { + "id": "llama-4-maverick-17b-128e-instruct-fp8", + "name": "Llama-4-Maverick-17B-128E-Instruct-FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -37319,76 +39555,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0 } }, - "@cf/zai-org/glm-4.7-flash": { - "id": "@cf/zai-org/glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "groq-llama-4-maverick-17b-128e-instruct": { + "id": "groq-llama-4-maverick-17b-128e-instruct", + "name": "Groq-Llama-4-Maverick-17B-128E-Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.0605, - "output": 0.4 - } - }, - "@cf/deepseek-ai/deepseek-r1-distill-qwen-32b": { - "id": "@cf/deepseek-ai/deepseek-r1-distill-qwen-32b", - "name": "Deepseek R1 Distill Qwen 32B", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "knowledge": "2025-01", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text" @@ -37399,30 +39585,30 @@ }, "open_weights": true, "limit": { - "context": 80000, - "output": 80000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.497, - "output": 4.881 + "input": 0, + "output": 0 } }, - "@cf/qwen/qwen3-30b-a3b-fp8": { - "id": "@cf/qwen/qwen3-30b-a3b-fp8", - "name": "Qwen3 30B A3b fp8", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "llama-4-scout-17b-16e-instruct-fp8": { + "id": "llama-4-scout-17b-16e-instruct-fp8", + "name": "Llama-4-Scout-17B-16E-Instruct-FP8", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-04-30", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -37430,57 +39616,26 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.0509, - "output": 0.335 + "input": 0, + "output": 0 } }, - "@cf/qwen/qwen2.5-coder-32b-instruct": { - "id": "@cf/qwen/qwen2.5-coder-32b-instruct", - "name": "Qwen2.5 Coder 32B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "cerebras-llama-4-maverick-17b-128e-instruct": { + "id": "cerebras-llama-4-maverick-17b-128e-instruct", + "name": "Cerebras-Llama-4-Maverick-17B-128E-Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-02-27", - "last_updated": "2025-02-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.66, - "output": 1 - } - }, - "@cf/qwen/qwq-32b": { - "id": "@cf/qwen/qwq-32b", - "name": "Qwq 32B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2025-03-05", - "last_updated": "2025-03-05", + "knowledge": "2025-01", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text" @@ -37491,26 +39646,26 @@ }, "open_weights": true, "limit": { - "context": 24000, - "output": 24000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.66, - "output": 1 + "input": 0, + "output": 0 } }, - "@cf/meta/llama-3.2-1b-instruct": { - "id": "@cf/meta/llama-3.2-1b-instruct", - "name": "Llama 3.2 1B Instruct", + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "family": "llama", - "attachment": false, + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -37521,30 +39676,29 @@ }, "open_weights": true, "limit": { - "context": 60000, - "output": 60000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.027, - "output": 0.201 + "input": 0, + "output": 0 } }, - "@cf/meta/llama-3.2-11b-vision-instruct": { - "id": "@cf/meta/llama-3.2-11b-vision-instruct", - "name": "Llama 3.2 11B Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", + "llama-3.3-8b-instruct": { + "id": "llama-3.3-8b-instruct", + "name": "Llama-3.3-8B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "family": "llama", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -37553,30 +39707,28 @@ "open_weights": true, "limit": { "context": 128000, - "output": 128000 + "output": 4096 }, "cost": { - "input": 0.0485, - "output": 0.676 + "input": 0, + "output": 0 } }, - "@cf/meta/llama-4-scout-17b-16e-instruct": { - "id": "@cf/meta/llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout 17B 16E Instruct", - "description": "Open Llama with long-context vision for efficient multimodal agents", + "cerebras-llama-4-scout-17b-16e-instruct": { + "id": "cerebras-llama-4-scout-17b-16e-instruct", + "name": "Cerebras-Llama-4-Scout-17B-16E-Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", "family": "llama", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2024-08", + "knowledge": "2025-01", "release_date": "2025-04-05", "last_updated": "2025-04-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -37584,26 +39736,37 @@ }, "open_weights": true, "limit": { - "context": 131000, - "output": 16384 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.27, - "output": 0.85 + "input": 0, + "output": 0 } - }, - "@cf/meta/llama-guard-3-8b": { - "id": "@cf/meta/llama-guard-3-8b", - "name": "Llama Guard 3 8B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", + } + } + }, + "cohere": { + "id": "cohere", + "env": [ + "COHERE_API_KEY" + ], + "npm": "@ai-sdk/cohere", + "name": "Cohere", + "doc": "https://docs.cohere.com/docs/models", + "models": { + "command-r-08-2024": { + "id": "command-r-08-2024", + "name": "Command R", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2025-01-22", - "last_updated": "2025-01-22", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -37614,27 +39777,39 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.484, - "output": 0.03 + "input": 0.15, + "output": 0.6 } }, - "@cf/meta/llama-3.3-70b-instruct-fp8-fast": { - "id": "@cf/meta/llama-3.3-70b-instruct-fp8-fast", - "name": "Llama 3.3 70B Instruct fp8 Fast", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "north-mini-code-1-0": { + "id": "north-mini-code-1-0", + "name": "North Mini Code", + "description": "Cohere coding model for practical software engineering and agentic edits", + "family": "north", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-09-23", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text" @@ -37645,26 +39820,30 @@ }, "open_weights": true, "limit": { - "context": 24000, - "output": 24000 + "context": 256000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.cohere.ai/compatibility/v1" }, "cost": { - "input": 0.293, - "output": 2.253 + "input": 0, + "output": 0 } }, - "@cf/meta/llama-3.1-8b-instruct-fp8": { - "id": "@cf/meta/llama-3.1-8b-instruct-fp8", - "name": "Llama 3.1 8B Instruct fp8", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "command-a-translate-08-2025": { + "id": "command-a-translate-08-2025", + "name": "Command A Translate", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "family": "command-a", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2024-07-25", - "last_updated": "2024-07-25", + "knowledge": "2024-06-01", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ "text" @@ -37675,26 +39854,24 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 32000 + "context": 8000, + "output": 8000 }, "cost": { - "input": 0.152, - "output": 0.287 + "input": 2.5, + "output": 10 } }, - "@cf/meta/llama-3.2-3b-instruct": { - "id": "@cf/meta/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "c4ai-aya-expanse-8b": { + "id": "c4ai-aya-expanse-8b", + "name": "Aya Expanse 8B", + "description": "Compact open multilingual model optimized for generation across 23 languages", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "release_date": "2024-10-24", + "last_updated": "2024-10-24", "modalities": { "input": [ "text" @@ -37705,26 +39882,22 @@ }, "open_weights": true, "limit": { - "context": 80000, - "output": 80000 - }, - "cost": { - "input": 0.0509, - "output": 0.335 + "context": 8000, + "output": 4000 } }, - "@cf/aisingapore/gemma-sea-lion-v4-27b-it": { - "id": "@cf/aisingapore/gemma-sea-lion-v4-27b-it", - "name": "Gemma Sea Lion V4 27B It", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "command-r7b-arabic-02-2025": { + "id": "command-r7b-arabic-02-2025", + "name": "Command R7B Arabic", + "description": "Open Command R model optimized for Arabic enterprise chat, RAG, and cultural knowledge", + "family": "command-r", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2024-06-01", + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ "text" @@ -37736,41 +39909,39 @@ "open_weights": true, "limit": { "context": 128000, - "output": 128000 + "output": 4000 }, "cost": { - "input": 0.351, - "output": 0.555 + "input": 0.0375, + "output": 0.15 } - } - } - }, - "bailing": { - "id": "bailing", - "env": [ - "BAILING_API_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.tbox.cn/api/llm/v1/chat/completions", - "name": "Bailing", - "doc": "https://alipaytbox.yuque.com/sxs0ba/ling/intro", - "models": { - "Ring-1T": { - "id": "Ring-1T", - "name": "Ring-1T", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "ring", - "attachment": false, + }, + "command-a-plus-05-2026": { + "id": "command-a-plus-05-2026", + "name": "Command A Plus", + "description": "Cohere's stronger command model for multilingual agents and enterprise workflows", + "family": "command-a", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1 + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-10", - "last_updated": "2025-10", + "knowledge": "2025-04-01", + "release_date": "2026-05-20", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -37779,25 +39950,25 @@ "open_weights": true, "limit": { "context": 128000, - "output": 32000 + "output": 64000 }, "cost": { - "input": 0.57, - "output": 2.29 + "input": 2.5, + "output": 10 } }, - "Ling-1T": { - "id": "Ling-1T", - "name": "Ling-1T", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "ling", + "command-a-03-2025": { + "id": "command-a-03-2025", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-10", - "last_updated": "2025-10", + "knowledge": "2024-06-01", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ "text" @@ -37808,277 +39979,259 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 256000, + "output": 8000 }, "cost": { - "input": 0.57, - "output": 2.29 + "input": 2.5, + "output": 10 } - } - } - }, - "anyapi": { - "id": "anyapi", - "env": [ - "ANYAPI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.anyapi.ai/v1", - "name": "AnyAPI", - "doc": "https://docs.anyapi.ai", - "models": { - "xai/grok-4.3": { - "id": "xai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + }, + "c4ai-aya-vision-8b": { + "id": "c4ai-aya-vision-8b", + "name": "Aya Vision 8B", + "description": "Compact open multilingual vision model for OCR and visual question answering", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2025-03-04", + "last_updated": "2025-05-14", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 30000 + "context": 16000, + "output": 4000 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "c4ai-aya-vision-32b": { + "id": "c4ai-aya-vision-32b", + "name": "Aya Vision 32B", + "description": "Open multilingual vision model for OCR, visual reasoning, and image question answering", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2025-03-04", + "last_updated": "2025-05-14", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 16000, + "output": 4000 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "command-r7b-12-2024": { + "id": "command-r7b-12-2024", + "name": "Command R7B", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-06-01", + "release_date": "2024-12-02", + "last_updated": "2024-12-02", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.0375, + "output": 0.15 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "command-r-plus-08-2024": { + "id": "command-r-plus-08-2024", + "name": "Command R+", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 2.5, + "output": 10 } }, - "google/gemini-3-pro-preview": { - "id": "google/gemini-3-pro-preview", - "name": "Gemini 3 Pro Preview", - "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", - "family": "gemini-pro", + "command-a-vision-07-2025": { + "id": "command-a-vision-07-2025", + "name": "Command A Vision", + "description": "Cohere vision model for multilingual document analysis, OCR, and image understanding", + "family": "command-a", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "knowledge": "2024-06-01", + "release_date": "2025-07-31", + "last_updated": "2025-07-31", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 8000 + }, + "cost": { + "input": 2.5, + "output": 10 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, + "command-a-reasoning-08-2025": { + "id": "command-a-reasoning-08-2025", + "name": "Command A Reasoning", + "description": "Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows", + "family": "command-a", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2024-06-01", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 2.5, + "output": 10 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "c4ai-aya-expanse-32b": { + "id": "c4ai-aya-expanse-32b", + "name": "Aya Expanse 32B", + "description": "Open multilingual model optimized for generation across 23 languages", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-10-24", + "last_updated": "2024-10-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 128000, + "output": 4000 } - }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", + } + } + }, + "deepseek": { + "id": "deepseek", + "env": [ + "DEEPSEEK_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.deepseek.com", + "name": "DeepSeek", + "doc": "https://api-docs.deepseek.com/quick_start/pricing", + "models": { + "deepseek-v4-flash-vision-exp": { + "id": "deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -38088,36 +40241,48 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 384000 + }, + "status": "deprecated", + "cost": { + "input": 0.15, + "output": 0.6, + "reasoning": 0.6, + "cache_read": 0.003 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "deepseek-flash": { + "id": "deepseek-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -38127,35 +40292,46 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "reasoning": 0.6, + "cache_read": 0.003 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -38164,35 +40340,47 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "reasoning": 0.87, + "cache_read": 0.003625 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -38202,121 +40390,109 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 384000 + }, + "status": "deprecated", + "cost": { + "input": 0.15, + "output": 0.6, + "reasoning": 0.6, + "cache_read": 0.003 } - }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + } + } + }, + "baseten": { + "id": "baseten", + "env": [ + "BASETEN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.baseten.co/v1", + "name": "Baseten", + "doc": "https://docs.baseten.co/inference/model-apis/overview", + "models": { + "thinkingmachines/inkling-small": { + "id": "thinkingmachines/inkling-small", + "name": "Inkling Small", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 32768 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - } - }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 + "cost": { + "input": 0.5, + "output": 1.2, + "cache_read": 0.1 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", + "thinkingmachines/inkling": { + "id": "thinkingmachines/inkling", + "name": "Inkling", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", @@ -38326,93 +40502,77 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 32768 + }, + "cost": { + "input": 1, + "output": 4.05 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek V3.1", + "description": "Legacy model retained for compatibility with older integrations", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-08-25", + "last_updated": "2025-08-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 164000, + "output": 131000 + }, + "status": "deprecated", + "cost": { + "input": 0.5, + "output": 1.5 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] + "interleaved": { + "field": "reasoning_content" }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - } - }, - "mistralai/devstral-2512": { - "id": "mistralai/devstral-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -38423,27 +40583,44 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 384000 }, - "status": "deprecated" + "cost": { + "input": 0.13, + "output": 0.26, + "cache_read": 0.028 + } }, - "mistralai/mistral-large-2512": { - "id": "mistralai/mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", - "attachment": true, - "reasoning": false, + "deepseek-ai/DeepSeek-V4-Pro-0813": { + "id": "deepseek-ai/DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -38451,290 +40628,216 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 1048576, "output": 262144 + }, + "cost": { + "input": 1.32, + "output": 3.96 } }, - "anthropic/claude-sonnet-4-5": { - "id": "anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "deepseek-ai/DeepSeek-V4.1-Flash": { + "id": "deepseek-ai/DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", - "medium", - "high" + "high", + "max" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 262144 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "cost": { + "input": 1.74, + "output": 3.48, + "cache_read": 0.145 } }, - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2026-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 204000, + "output": 204000 + }, + "status": "deprecated", + "cost": { + "input": 0.3, + "output": 1.2 } }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262000, + "output": 262000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "moonshotai/Kimi-K3": { + "id": "moonshotai/Kimi-K3", + "name": "Kimi K3", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", - "medium", - "high" + "high", + "max" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] + "interleaved": { + "field": "reasoning_content" }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - } - }, - "cohere/command-r-plus-08-2024": { - "id": "cohere/command-r-plus-08-2024", - "name": "Command R+", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -38742,23 +40845,35 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 1048576, + "output": 262144 + }, + "cost": { + "input": 3, + "output": 15 } }, - "perplexity/sonar-reasoning-pro": { - "id": "perplexity/sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", - "family": "sonar-reasoning", + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", @@ -38768,24 +40883,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262000, + "output": 262000 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "perplexity/sonar-pro": { - "id": "perplexity/sonar-pro", - "name": "Sonar Pro", - "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", - "family": "sonar-pro", + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "knowledge": "2025-12", + "release_date": "2026-01-30", + "last_updated": "2026-02-12", "modalities": { "input": [ "text", @@ -38795,29 +40924,37 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 + "context": 262000, + "output": 262000 + }, + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.12 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -38828,30 +40965,44 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 202800, + "output": 202800 + }, + "cost": { + "input": 1.3, + "output": 4.3, + "cache_read": 0.26 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -38859,26 +41010,36 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048576, + "output": 262144 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.3 } }, - "deepseek/deepseek-r1": { - "id": "deepseek/deepseek-r1", - "name": "DeepSeek Reasoner", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": true, + "zai-org/GLM-4.7": { + "id": "zai-org/GLM-4.7", + "name": "GLM 4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -38889,25 +41050,44 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "output": 200000 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.12 } }, - "deepseek/deepseek-chat": { - "id": "deepseek/deepseek-chat", - "name": "DeepSeek Chat", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "zai-org/GLM-5.2-Fast": { + "id": "zai-org/GLM-5.2-Fast", + "name": "GLM 5.2 Fast", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -38915,261 +41095,192 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048576, + "output": 262144 + }, + "cost": { + "input": 2.1, + "output": 6.6, + "cache_read": 0.21 } - } - } - }, - "google": { - "id": "google", - "env": [ - "GOOGLE_API_KEY", - "GOOGLE_GENERATIVE_AI_API_KEY", - "GEMINI_API_KEY" - ], - "npm": "@ai-sdk/google", - "name": "Google", - "doc": "https://ai.google.dev/gemini-api/docs/models", - "models": { - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + }, + "zai-org/GLM-5.3-Fast": { + "id": "zai-org/GLM-5.3-Fast", + "name": "GLM 5.3 Fast", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 262144 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 2.1, + "output": 6.6 } }, - "gemini-2.5-flash-preview-tts": { - "id": "gemini-2.5-flash-preview-tts", - "name": "Gemini 2.5 Flash Preview TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gemini-flash", + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM 5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-05-01", - "last_updated": "2025-05-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 16384 - }, - "cost": { - "input": 0.5, - "output": 10 - } - }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2026-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 202800, + "output": 202800 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.95, + "output": 3.15, + "cache_read": 0.2 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", + "zai-org/GLM-5.3-Flash": { + "id": "zai-org/GLM-5.3-Flash", + "name": "GLM 5.3 Flash", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 + "input": 0.15, + "output": 0.5 } }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "zai-org/GLM-5.3": { + "id": "zai-org/GLM-5.3", + "name": "GLM 5.3", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 262144 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 + "input": 1.4, + "output": 4.4, + "cache_read": 0.14 } }, - "gemma-4-31b-it": { - "id": "gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, + "nvidia/Nemotron-120B-A12B": { + "id": "nvidia/Nemotron-120B-A12B", + "name": "Nemotron Super", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -39177,14 +41288,17 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2026-02", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -39192,378 +41306,362 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 202800, + "output": 202800 + }, + "cost": { + "input": 0.3, + "output": 0.75, + "cache_read": 0.06 } }, - "gemini-2.0-flash": { - "id": "gemini-2.0-flash", - "name": "Gemini 2.0 Flash", - "description": "Earlier Gemini Flash workhorse for responsive multimodal apps and tool use", - "family": "gemini-flash", - "attachment": true, - "reasoning": false, + "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B": { + "id": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B", + "name": "Nemotron Ultra", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 8192 + "context": 202800, + "output": 202800 }, - "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 } }, - "gemini-embedding-001": { - "id": "gemini-embedding-001", - "name": "Gemini Embedding 001", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "gemini", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "OpenAI GPT 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-05", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2048, - "output": 1 - }, - "cost": { - "input": 0.15, - "output": 0 - } - }, - "gemini-3.1-pro-preview-customtools": { - "id": "gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "knowledge": "2025-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 128072, + "output": 128072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.1, + "output": 0.5 } - }, - "gemini-flash-lite-latest": { - "id": "gemini-flash-lite-latest", - "name": "Gemini Flash-Lite Latest", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + } + } + }, + "nan": { + "id": "nan", + "env": [ + "NAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.nan.builders/v1", + "name": "NaN", + "doc": "https://nan.builders/docs/models", + "models": { + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "audio", - "video", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 0, + "output": 0 } }, - "gemini-3-pro-image-preview": { - "id": "gemini-3-pro-image-preview", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", + "glm5.3-flash": { + "id": "glm5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 2, - "output": 120 + "input": 0, + "output": 0 } }, - "gemini-2.5-flash-image": { - "id": "gemini-2.5-flash-image", - "name": "Nano Banana", - "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", - "family": "gemini-flash", + "gemma4": { + "id": "gemma4", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, + "context": 262144, "output": 32768 }, "cost": { - "input": 0.3, - "output": 30, - "cache_read": 0.075 + "input": 0, + "output": 0 } }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", + "qwen3.6": { + "id": "qwen3.6", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 262144, "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 + "input": 0, + "output": 0 } }, - "gemini-omni-flash-preview": { - "id": "gemini-omni-flash-preview", - "name": "Gemini Omni Flash Preview", - "description": "Video generation and editing model for fast, conversational text- and image-to-video workflows", - "family": "gemini", + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "tool_call": true, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 262144, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 17.5 + "input": 0, + "output": 0 } }, - "gemini-3.1-flash-image-preview": { - "id": "gemini-3.1-flash-image-preview", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, + "glm5.3": { + "id": "glm5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ "text", "image" + ], + "output": [ + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.5, - "output": 60 + "input": 0, + "output": 0 } - }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + } + } + }, + "stepfun-ai-step-plan": { + "id": "stepfun-ai-step-plan", + "env": [ + "STEPFUN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.stepfun.ai/step_plan/v1", + "name": "StepFun Step Plan (Global)", + "doc": "https://platform.stepfun.ai/docs/en/step-plan/integrations/reasoning-api", + "models": { + "step-5-preview": { + "id": "step-5-preview", + "name": "Step 5 Preview", + "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -39577,18 +41675,17 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-09-16", + "last_updated": "2026-09-20", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" @@ -39596,52 +41693,37 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "context": 1000000, + "input": 1000000, + "output": 1000000 } }, - "gemma-4-26b-a4b-it": { - "id": "gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "step-3.5-flash-2603": { + "id": "step-3.5-flash-2603", + "name": "Step 3.5 Flash 2603", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high" + ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, + "knowledge": "2025-01", "release_date": "2026-04-02", "last_updated": "2026-04-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -39649,16 +41731,16 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 256000, + "input": 256000, + "output": 256000 } }, - "gemini-3-pro-preview": { - "id": "gemini-3-pro-preview", - "name": "Gemini 3 Pro Preview", - "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", - "family": "gemini-pro", - "attachment": true, + "step-3.5-flash": { + "id": "step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -39670,63 +41752,38 @@ } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 - }, - "status": "deprecated", - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "context": 256000, + "input": 256000, + "output": 256000 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "step-3.7-flash": { + "id": "step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -39734,70 +41791,90 @@ } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 + "context": 256000, + "input": 256000, + "output": 256000 } - }, - "gemini-2.5-pro-preview-tts": { - "id": "gemini-2.5-pro-preview-tts", - "name": "Gemini 2.5 Pro Preview TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gemini-flash", - "attachment": false, - "reasoning": false, - "tool_call": false, + } + } + }, + "nearai": { + "id": "nearai", + "env": [ + "NEARAI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://cloud-api.near.ai/v1", + "name": "NEAR AI Cloud", + "doc": "https://docs.near.ai/", + "models": { + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-05-01", - "last_updated": "2025-05-01", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 8192, - "output": 16384 + "context": 200000, + "output": 64000 }, "cost": { "input": 1, - "output": 20 + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "gemini-flash-latest": { - "id": "gemini-flash-latest", - "name": "Gemini Flash Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "anthropic/claude-sonnet-4-5": { + "id": "anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -39806,22 +41883,18 @@ }, { "type": "budget_tokens", - "min": 0, - "max": 24576 + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -39830,46 +41903,41 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075, - "input_audio": 1 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "gemini-3.1-flash-lite-preview": { - "id": "gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gemini-flash-lite", + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -39878,36 +41946,41 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-2.0-flash-lite": { - "id": "gemini-2.0-flash-lite", - "name": "Gemini 2.0 Flash-Lite", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gemini-flash-lite", + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -39916,132 +41989,143 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 8192 + "context": 1000000, + "output": 64000 }, - "status": "deprecated", "cost": { - "input": 0.075, - "output": 0.3 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } - } - } - }, - "opencode-go": { - "id": "opencode-go", - "env": [ - "OPENCODE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://opencode.ai/zen/go/v1", - "name": "OpenCode Go", - "doc": "https://opencode.ai/docs/zen", - "models": { - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", - "attachment": false, + }, + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] }, - "structured_output": true, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "black-forest-labs/FLUX.2-klein-4B": { + "id": "black-forest-labs/FLUX.2-klein-4B", + "name": "FLUX.2 Klein 4B", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 1, + "output": 1 } }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax-m2.5", - "attachment": false, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, + "context": 1048576, "output": 65536 }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" - }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.7-plus", + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -40049,60 +42133,52 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1048576, "output": 65536 }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.04, - "cache_write": 0.5, + "input": 1.25, + "output": 10, + "cache_read": 0.125, "tiers": [ { - "input": 1.2, - "output": 4.8, - "cache_read": 0.12, - "cache_write": 1.5, + "input": 2.5, + "output": 15, + "cache_read": 0.25, "tier": { "type": "context", - "size": 256000 + "size": 200000 } } ], "context_over_200k": { - "input": 1.2, - "output": 4.8, - "cache_read": 0.12, - "cache_write": 1.5 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "qwen3.7-max", - "attachment": false, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -40110,117 +42186,103 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1048576, "output": 65536 }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video" + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "input_audio": 0.3 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", + "Qwen/Qwen3-Reranker-0.6B": { + "id": "Qwen/Qwen3-Reranker-0.6B", + "name": "Qwen3 Reranker 0.6B", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-03", + "last_updated": "2025-06-03", "modalities": { "input": [ "text" @@ -40231,42 +42293,30 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 40960, + "output": 1024 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.0145 + "input": 0.01, + "output": 0.01 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "Qwen/Qwen3-VL-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", + "name": "Qwen3-VL 30B-A3B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -40274,37 +42324,28 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 16384, + "output": 8192 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.15, + "output": 0.55 } }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax-m3", + "Qwen/Qwen3-Embedding-0.6B": { + "id": "Qwen/Qwen3-Embedding-0.6B", + "name": "Qwen3 Embedding 0.6B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-31", - "last_updated": "2026-05-31", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-03", + "last_updated": "2025-06-03", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -40312,94 +42353,31 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 32768, + "output": 1024 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "tiers": [ - { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12, - "tier": { - "type": "context", - "size": 512000 - } - } - ], - "context_over_200k": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 - } + "input": 0.01, + "output": 0.01 } }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Legacy model retained for compatibility with older integrations", - "family": "qwen3.5", + "Qwen/Qwen3.6-35B-A3B-FP8": { + "id": "Qwen/Qwen3.6-35B-A3B-FP8", + "name": "Qwen 3.6 35B A3B FP8", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 - } - }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax-m2.7", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text" @@ -40409,117 +42387,39 @@ ] }, "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "output": 8192 }, - "status": "deprecated", "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 0.17, + "output": 1.1, + "cache_read": 0.056 } }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo V2.5", - "description": "MiMo omni model for text, image, video, audio, and agents", - "family": "mimo-v2.5", - "attachment": true, + "zai-org/GLM-5.1-FP8": { + "id": "zai-org/GLM-5.1-FP8", + "name": "GLM-5.1 FP8", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 - } - }, - "mimo-v2-omni": { - "id": "mimo-v2-omni", - "name": "MiMo V2 Omni", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo-v2-omni", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" ], "output": [ "text" @@ -40527,159 +42427,137 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 128000 + "context": 202752, + "output": 16384 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08 + "input": 1.4, + "output": 4.4 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 - } - }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo V2 Pro", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo-v2-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, + "context": 1050000, + "input": 922000, "output": 128000 }, - "status": "deprecated", "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2, + "input": 2.5, + "output": 15, + "cache_read": 0.25, "tiers": [ { - "input": 2, - "output": 6, - "cache_read": 0.4, + "input": 5, + "output": 22.5, + "cache_read": 0.5, "tier": { "type": "context", - "size": 256000 + "size": 272000 } } ], "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 + "input": 5, + "output": 22.5, + "cache_read": 0.5 } } }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo V2.5 Pro", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", - "family": "mimo-v2.5-pro", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.0145 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "glm", + "openai/whisper-large-v3": { + "id": "openai/whisper-large-v3", + "name": "Whisper Large v3", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" @@ -40687,42 +42565,42 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 32768 + "context": 448, + "output": 448 }, - "status": "deprecated", "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.01, + "output": 0.01 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.6", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -40730,54 +42608,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 0.25, + "output": 2, + "cache_read": 0.025 } - } - } - }, - "digitalocean": { - "id": "digitalocean", - "env": [ - "DIGITALOCEAN_ACCESS_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.do-ai.run/v1", - "name": "DigitalOcean", - "doc": "https://docs.digitalocean.com/products/gradient-ai-platform/details/models/", - "models": { - "anthropic-claude-haiku-4.5": { - "id": "anthropic-claude-haiku-4.5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + }, + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -40791,103 +42636,54 @@ } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 1, - "cache_write": 1.25 - } - }, - "openai-gpt-image-1": { - "id": "openai-gpt-image-1", - "name": "GPT Image 1", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, + "structured_output": true, "temperature": false, - "release_date": "2025-04-24", - "last_updated": "2025-04-24", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", "image" ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - }, - "cost": { - "input": 5, - "output": 40, - "cache_read": 1.25 - } - }, - "e5-large-v2": { - "id": "e5-large-v2", - "name": "E5 Large v2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2023-05-19", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text" - ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512, - "output": 1024 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.02, - "output": 0 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "bge-m3": { - "id": "bge-m3", - "name": "BGE M3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "bge", + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2024-01-30", - "last_updated": "2026-04-30", + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ "text" @@ -40896,60 +42692,78 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 1024 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.02, - "output": 0 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "mistral-3-14B": { - "id": "mistral-3-14B", - "name": "Ministral 3 14B Instruct", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2026-04-30", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 128000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "nemotron-3-ultra-550b": { - "id": "nemotron-3-ultra-550b", - "name": "Nemotron 3 Ultra", - "description": "Flagship Nemotron model for high-throughput reasoning and complex agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-12", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -40957,21 +42771,28 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax-m2.5", - "attachment": false, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -40979,34 +42800,37 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08", - "release_date": "2026-02-12", - "last_updated": "2026-04-16", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, + "context": 400000, + "input": 272000, "output": 128000 }, - "status": "beta", "cost": { - "input": 0.3, - "output": 1.2 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openai-gpt-5.4-nano": { - "id": "openai-gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -41023,7 +42847,7 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-08-31", "release_date": "2026-03-17", "last_updated": "2026-03-17", @@ -41039,70 +42863,19 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "deepseek-v3": { - "id": "deepseek-v3", - "name": "DeepSeek V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-12-26", - "last_updated": "2025-03-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 131072 - } - }, - "openai-gpt-image-2": { - "id": "openai-gpt-image-2", - "name": "GPT Image 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-04-24", - "last_updated": "2025-04-24", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "openai-gpt-5.2": { - "id": "openai-gpt-5.2", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", "family": "gpt", "attachment": true, "reasoning": true, @@ -41120,7 +42893,7 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-08-31", "release_date": "2025-12-11", "last_updated": "2025-12-11", @@ -41136,6 +42909,7 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { @@ -41144,12 +42918,12 @@ "cache_read": 0.175 } }, - "deepseek-r1-distill-llama-70b": { - "id": "deepseek-r1-distill-llama-70b", - "name": "DeepSeek R1 Distill Llama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -41159,131 +42933,127 @@ "low", "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-01-30", - "last_updated": "2025-01-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.99, - "output": 0.99 - } - }, - "qwen3-embedding-0.6b": { - "id": "qwen3-embedding-0.6b", - "name": "Qwen3 Embedding 0.6B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, "temperature": false, - "release_date": "2025-06-03", - "last_updated": "2026-04-16", + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8000, - "output": 1024 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "status": "beta", "cost": { - "input": 0.04, - "output": 0 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "gemma-4-31B-it": { - "id": "gemma-4-31B-it", - "name": "Gemma 4 31B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-30", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.18, - "output": 0.5 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "llama-4-maverick": { - "id": "llama-4-maverick", - "name": "Llama 4 Maverick 17B 128E Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2026-04-30", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 0.87 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "anthropic-claude-3.7-sonnet": { - "id": "anthropic-claude-3.7-sonnet", - "name": "Claude 3.7 Sonnet", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-sonnet", + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -41297,14 +43067,16 @@ } ], "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-02-24", - "last_updated": "2025-02-24", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -41313,78 +43085,42 @@ "open_weights": false, "limit": { "context": 200000, - "output": 64000 + "output": 100000 }, - "status": "deprecated", "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "openai-gpt-4o-mini": { - "id": "openai-gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 - } - }, - "anthropic-claude-opus-4.7": { - "id": "anthropic-claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -41392,27 +43128,71 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + } + } + }, + "wandb": { + "id": "wandb", + "env": [ + "WANDB_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.inference.wandb.ai/v1", + "name": "CoreWeave", + "doc": "https://docs.wandb.ai/inference", + "models": { + "JetBrains/Mellum2-12B-A2.5B-Instruct": { + "id": "JetBrains/Mellum2-12B-A2.5B-Instruct", + "name": "Mellum2 12B A2.5B", + "description": "Mellum2-12B-A2.5B-Instruct is a fast MoE model with 131K context built for coding, tool use, and low-latency AI workflows.", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.05, + "output": 0.1, + "cache_read": 0.05 } }, - "deepseek-4-flash": { - "id": "deepseek-4-flash", - "name": "Deepseek V4 Flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek", + "meta-llama/Llama-3.1-70B-Instruct": { + "id": "meta-llama/Llama-3.1-70B-Instruct", + "name": "Llama 3.1 70B", + "description": "Efficient conversational model optimized for responsive multilingual chatbot interactions.", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-27", - "last_updated": "2026-05-29", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -41421,34 +43201,98 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 8192 + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.8, + "output": 0.8, + "cache_read": 0.8 } }, - "anthropic-claude-4.5-haiku": { - "id": "anthropic-claude-4.5-haiku", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "meta-llama/Llama-3.1-8B-Instruct": { + "id": "meta-llama/Llama-3.1-8B-Instruct", + "name": "Llama 3.1 8B", + "description": "Efficient conversational model optimized for responsive multilingual chatbot interactions.", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.22, + "output": 0.22, + "cache_read": 0.22 + } + }, + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama 3.3 70B", + "description": "Multilingual model excelling in conversational tasks, detailed instruction-following, and coding.", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.71, + "output": 0.71, + "cache_read": 0.71 + } + }, + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B", + "description": "Gemma 4 31B Dense is designed for advanced reasoning, agentic workflows, and longer context and is natively trained on 140+ languages.", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -41458,157 +43302,106 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 1, - "cache_write": 1.25 + "input": 0.1, + "output": 0.34, + "cache_read": 0.1 } }, - "anthropic-claude-4.6-sonnet": { - "id": "anthropic-claude-4.6-sonnet", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "google/gemma-4-26B-A4B-it": { + "id": "google/gemma-4-26B-A4B-it", + "name": "Gemma 4 26B A4B", + "description": "Gemma 4 26B A4B is a multimodal MoE model with LoRA support and function calling for agentic workflows.", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.3, - "cache_write": 3.75, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.3, - "cache_write": 3.75 - } + "input": 0.1, + "output": 0.3, + "cache_read": 0.05 } }, - "anthropic-claude-sonnet-4": { - "id": "anthropic-claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "Qwen/Qwen3.8-27B": { + "id": "Qwen/Qwen3.8-27B", + "name": "Qwen3.8 27B", + "description": "Qwen3.8-27B is a dense multimodal model suited for coding, research, vision, and long-running agent tasks.", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.3, - "cache_write": 3.75, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.3, - "cache_write": 3.75 - } + "input": 0.4, + "output": 3, + "cache_read": 0.15 } }, - "ministral-3-8b-instruct-2512": { - "id": "ministral-3-8b-instruct-2512", - "name": "Ministral 3 8B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, + "Qwen/Qwen3-30B-A3B-Instruct-2507": { + "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen3 30B A3B Instruct 2507", + "description": "Qwen3-30B-A3B-Instruct-2507 is a 30.5B MoE instruction-tuned model with enhanced reasoning, coding, and long-context understanding.", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -41618,151 +43411,108 @@ "limit": { "context": 262144, "output": 262144 + }, + "cost": { + "input": 0.1, + "output": 0.3, + "cache_read": 0.1 } }, - "anthropic-claude-4.5-sonnet": { - "id": "anthropic-claude-4.5-sonnet", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B A3B", + "description": "Qwen3.6-35B-A3B is an MoE multimodal model with 262K context optimized for agentic coding workflows.", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-04-15", + "last_updated": "2026-04-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.3, - "cache_write": 3.75, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.3, - "cache_write": 3.75 - } + "input": 0.25, + "output": 1.25, + "cache_read": 0.25 } }, - "anthropic-claude-opus-4.6": { - "id": "anthropic-claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "Qwen/Qwen3.5-35B-A3B": { + "id": "Qwen/Qwen3.5-35B-A3B", + "name": "Qwen3.5-35B-A3B", + "description": "Qwen3.5-35B-A3B is an open-weights multimodal MoE model built for efficient, high-throughput inference across chat, reasoning, and agentic tasks.", + "family": "qwen3.5", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 0.5, - "cache_write": 6.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 0.5, - "cache_write": 6.25 - } + "input": 0.25, + "output": 1.25, + "cache_read": 0.25 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen3.6-27B is a 27B dense multimodal model with 262K context built for flagship-level agentic coding.", + "family": "qwen3.6", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -41770,42 +43520,32 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.975, - "output": 4.3, - "cache_read": 0.26 + "input": 0.6, + "output": 3.6, + "cache_read": 0.12 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", + "ibm-granite/granite-4.2-8b": { + "id": "ibm-granite/granite-4.2-8b", + "name": "Granite 4.2 8B", + "description": "Granite 4.2 8B is an instruct model capable of enhanced tool calling, instruction following, and chat capabilities.", + "family": "granite", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-08-24", + "last_updated": "2026-08-24", "modalities": { "input": [ "text" @@ -41816,37 +43556,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.74, - "output": 3.48 + "input": 0.1, + "output": 0.15, + "cache_read": 0.05 } }, - "openai-gpt-oss-20b": { - "id": "openai-gpt-oss-20b", - "name": "gpt-oss-20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "ibm-granite/granite-4.1-8b": { + "id": "ibm-granite/granite-4.1-8b", + "name": "Granite 4.1 8B", + "description": "Granite 4.1 8B is a long-context instruct model capable of enhanced tool calling, instruction following, and chat capabilities.", + "family": "granite", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-05", - "last_updated": "2026-04-16", + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ "text" @@ -41862,21 +43592,22 @@ }, "cost": { "input": 0.05, - "output": 0.45 + "output": 0.1, + "cache_read": 0.05 } }, - "qwen-2.5-14b-instruct": { - "id": "qwen-2.5-14b-instruct", - "name": "Qwen 2.5 14B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek V3.1", + "description": "A large hybrid model that supports both thinking and non-thinking modes via prompt templates.", + "family": "deepseek", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ "text" @@ -41887,80 +43618,69 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 161000, + "output": 161000 + }, + "cost": { + "input": 0.55, + "output": 1.65, + "cache_read": 0.55 } }, - "anthropic-claude-opus-4.8": { - "id": "anthropic-claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "DeepSeek V4-Flash-0731 is an MoE model great for coding, reasoning, and agentic workloads.", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-28", - "last_updated": "2026-05-29", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.13, + "output": 0.28, + "cache_read": 0.07 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "deepseek-ai/DeepSeek-V4-Pro-0813": { + "id": "deepseek-ai/DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4-Pro-0813 is a 1.6T-parameter MoE model excelling at advanced reasoning, coding, and complex agentic workloads.", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ "text" @@ -41971,115 +43691,108 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 1.05, - "output": 4.4, - "cache_read": 0.21 + "input": 1.31, + "output": 3.96, + "cache_read": 0.044 } }, - "anthropic-claude-opus-4.5": { - "id": "anthropic-claude-opus-4.5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "deepseek-ai/DeepSeek-V4.1-Flash": { + "id": "deepseek-ai/DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash is a multimodal MoE model for coding, reasoning, and agentic workloads with long contexts.", + "family": "deepseek", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.2, + "output": 0.65, + "cache_read": 0.03 } }, - "anthropic-claude-4.1-opus": { - "id": "anthropic-claude-4.1-opus", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "DeepSeek V4-Flash is an MoE model with 1M context length great for coding, reasoning, and agentic workloads.", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.14, + "output": 0.28, + "cache_read": 0.07 } }, - "llama3-8b-instruct": { - "id": "llama3-8b-instruct", - "name": "Llama 3.1 Instruct (8B)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "DeepSeek V4-Pro is a 1.6T-parameter MoE model with 49B active parameters excelling at advanced reasoning, coding, and complex agentic workloads.", + "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -42090,65 +43803,62 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.198, - "output": 0.198 + "input": 1.15, + "output": 2.55, + "cache_read": 0.2 } }, - "stable-diffusion-3.5-large": { - "id": "stable-diffusion-3.5-large", - "name": "Stable Diffusion 3.5 Large", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "stable-diffusion", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-10-22", - "last_updated": "2026-04-30", + "MiniMaxAI/MiniMax-M3": { + "id": "MiniMaxAI/MiniMax-M3", + "name": "MiniMax M3", + "description": "MiniMax M3 is a multimodal MoE model with 23B active parameters optimized for coding and agentic workflows.", + "family": "minimax-m3", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, "open_weights": true, "limit": { - "context": 256, - "output": 1 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.08, - "output": 0 + "input": 0.23, + "output": 0.96, + "cache_read": 0.05 } }, - "openai-gpt-5.4-pro": { - "id": "openai-gpt-5.4-pro", - "name": "GPT-5.4 pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Kimi K2.6 is a multimodal Mixture-of-Experts language model featuring 32 billion activated parameters and a total of 1 trillion parameters.", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", @@ -42158,30 +43868,35 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 30, - "output": 180 + "input": 0.65, + "output": 3.41, + "cache_read": 0.15 } }, - "all-mini-lm-l6-v2": { - "id": "all-mini-lm-l6-v2", - "name": "All-MiniLM-L6-v2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2021-08-30", - "last_updated": "2026-04-16", + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Kimi K2.7 Code is a 1T-parameter MoE model with 32B active parameters purpose-built for long-horizon agentic coding and software engineering.", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -42189,134 +43904,137 @@ }, "open_weights": true, "limit": { - "context": 256, - "output": 384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.009, - "output": 0 + "input": 0.71, + "output": 3.5, + "cache_read": 0.15 } }, - "openai-gpt-image-1.5": { - "id": "openai-gpt-image-1.5", - "name": "GPT Image 1.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM 5.2", + "description": "GLM-5.2 is a Mixture-of-Experts language model featuring 40 billion activated parameters and a total of 744 billion parameters.", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 5, - "output": 10, - "cache_read": 1 + "input": 0.76, + "output": 2.42, + "cache_read": 0.14 } }, - "openai-gpt-4o": { - "id": "openai-gpt-4o", - "name": "GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "zai-org/GLM-5.3-Flash": { + "id": "zai-org/GLM-5.3-Flash", + "name": "GLM 5.3 Flash", + "description": "GLM-5.3-Flash is a natively multimodal model with 320B total parameters and 18B active parameters.", + "family": "glm", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.15, + "output": 0.5, + "cache_read": 0.05 } }, - "anthropic-claude-opus-4": { - "id": "anthropic-claude-opus-4", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B": { + "id": "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B", + "name": "Nemotron 3.5 Lightning", + "description": "Nemotron 3.5 Lightning is an MoE model built for fast, reliable agentic tasks across use cases such as financial services, cybersecurity, telecom, and retail.", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.07, + "output": 0.2, + "cache_read": 0.04 } }, - "bge-reranker-v2-m3": { - "id": "bge-reranker-v2-m3", - "name": "BGE Reranker v2 M3", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "bge", + "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B": { + "id": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B", + "name": "Nemotron 3 Ultra", + "description": "Nemotron 3 Ultra is a powerful MoE model designed for long-running agents across coding, deep research, and enterprise automation.", + "family": "nemotron", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-03-12", - "last_updated": "2026-04-30", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -42327,81 +44045,60 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 1 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.01, - "output": 0 + "input": 0.5, + "output": 2.15, + "cache_read": 0.1 } }, - "openai-gpt-5": { - "id": "openai-gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "gpt-oss-20b", + "description": "Lower latency Mixture-of-Experts model trained on OpenAI's Harmony response format with reasoning capabilities.", + "family": "gpt-oss", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.03, + "output": 0.13, + "cache_read": 0.03 } }, - "openai-o3-mini": { - "id": "openai-o3-mini", - "name": "o3-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Efficient Mixture-of-Experts model designed for high-reasoning, agentic and general-purpose use cases.", + "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -42410,28 +44107,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.03, + "output": 0.17, + "cache_read": 0.03 } }, - "multi-qa-mpnet-base-dot-v1": { - "id": "multi-qa-mpnet-base-dot-v1", - "name": "Multi-QA-mpnet-base-dot-v1", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "OpenPipe/Qwen3-14B-Instruct": { + "id": "OpenPipe/Qwen3-14B-Instruct", + "name": "Qwen3 14B Instruct", + "description": "An efficient multilingual, dense, instruction-tuned model, optimized by OpenPipe for building agents with finetuning.", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2021-08-30", - "last_updated": "2026-04-16", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -42442,41 +44140,46 @@ }, "open_weights": true, "limit": { - "context": 512, - "output": 768 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.009, - "output": 0 + "input": 0.05, + "output": 0.22, + "cache_read": 0.05 } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + } + } + }, + "subconscious": { + "id": "subconscious", + "env": [ + "SUBCONSCIOUS_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://api.subconscious.dev/v1", + "name": "Subconscious", + "doc": "https://docs.subconscious.dev", + "models": { + "subconscious/tim-qwen3.6-27b": { + "id": "subconscious/tim-qwen3.6-27b", + "name": "TIM-Qwen3.6 27B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-04-16", + "temperature": true, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ "text" @@ -42485,28 +44188,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 8192, + "input": 8192, + "output": 5000 }, "cost": { - "input": 0.5, - "output": 2.7 + "input": 0.3, + "output": 3, + "cache_read": 0.15 } }, - "llama3.3-70b-instruct": { - "id": "llama3.3-70b-instruct", - "name": "Llama 3.3 Instruct 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "subconscious/glm-5.2": { + "id": "subconscious/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -42517,72 +44230,106 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.65, - "output": 0.65 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } - }, - "gte-large-en-v1.5": { - "id": "gte-large-en-v1.5", - "name": "GTE Large (v1.5)", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-03-27", - "last_updated": "2026-04-16", + } + } + }, + "zeldoc": { + "id": "zeldoc", + "env": [ + "ZELDOC_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.zeldoc.ai/v1", + "name": "Zeldoc", + "doc": "https://docs.zeldoc.ai", + "models": { + "zdev": { + "id": "zdev", + "name": "ZDev", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 1024 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.09, + "input": 0, "output": 0 } - }, - "openai-gpt-5.4-mini": { - "id": "openai-gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + } + } + }, + "databricks": { + "id": "databricks", + "env": [ + "DATABRICKS_HOST", + "DATABRICKS_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://${DATABRICKS_HOST}/ai-gateway/mlflow/v1", + "name": "Databricks", + "doc": "https://docs.databricks.com/aws/en/machine-learning/foundation-models/", + "models": { + "databricks-claude-opus-4-5": { + "id": "databricks-claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -42590,40 +44337,39 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai-gpt-5.5": { - "id": "openai-gpt-5.5", - "name": "GPT-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "databricks-gpt-5-6-terra": { + "id": "databricks-gpt-5-6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-30", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -42636,18 +44382,19 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, + "input": 2.5, + "output": 15, + "cache_read": 0.25, "tiers": [ { - "input": 10, - "output": 45, - "cache_read": 1, + "input": 5, + "output": 22.5, + "cache_read": 0.5, "tier": { "type": "context", "size": 272000 @@ -42655,38 +44402,40 @@ } ], "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 + "input": 5, + "output": 22.5, + "cache_read": 0.5 } } }, - "nemotron-3-nano-30b": { - "id": "nemotron-3-nano-30b", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, + "databricks-kimi-k2-7-code": { + "id": "databricks-kimi-k2-7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -42696,13 +44445,18 @@ "limit": { "context": 262144, "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "openai-gpt-oss-120b": { - "id": "openai-gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "databricks-glm-5-2": { + "id": "databricks-glm-5-2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -42716,14 +44470,10 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-05", - "last_updated": "2026-04-16", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -42734,26 +44484,26 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 1000000, "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.7 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai-gpt-5-nano": { - "id": "openai-gpt-5-nano", - "name": "GPT-5 nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "databricks-gemini-3-1-flash-lite": { + "id": "databricks-gemini-3-1-flash-lite", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -42762,14 +44512,17 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -42777,135 +44530,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 - } - }, - "wan2-2-t2v-a14b": { - "id": "wan2-2-t2v-a14b", - "name": "Wan2.2-T2V-A14B", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-07-28", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "video" - ] - }, - "open_weights": true, - "limit": { - "context": 100, - "output": 1 - }, - "cost": { - "input": 0.6, - "output": 0 - } - }, - "mistral-7b-instruct-v0.3": { - "id": "mistral-7b-instruct-v0.3", - "name": "Mistral 7B Instruct v0.3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-05-22", - "last_updated": "2024-05-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - } - }, - "qwen3-tts-voicedesign": { - "id": "qwen3-tts-voicedesign", - "name": "Qwen3 TTS VoiceDesign", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 1 - } - }, - "mistral-nemo-instruct-2407": { - "id": "mistral-nemo-instruct-2407", - "name": "Mistral Nemo Instruct", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mistral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 65536 }, - "status": "deprecated", "cost": { - "input": 0.3, - "output": 0.3 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "databricks-gpt-5-mini": { + "id": "databricks-gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", "high" @@ -42913,14 +44559,11 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -42930,28 +44573,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "deepseek-3.2": { - "id": "deepseek-3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "databricks-gemini-3-flash": { + "id": "databricks-gemini-3-flash", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -42959,54 +44603,53 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-02", - "last_updated": "2026-04-30", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { "input": 0.5, - "output": 1.6 + "output": 3, + "cache_read": 0.05, + "input_audio": 1 } }, - "openai-o1": { - "id": "openai-o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "databricks-claude-opus-4-7": { + "id": "databricks-claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -43019,27 +44662,47 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen 3.5 397B A17B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen3.5", + "databricks-gpt-oss-20b": { + "id": "databricks-gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -43049,8 +44712,8 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-04-30", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -43061,73 +44724,103 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 81920 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.55, - "output": 3.5 + "input": 0.05, + "output": 0.2 } }, - "qwen3-coder-flash": { - "id": "qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "databricks-gpt-5-4-mini": { + "id": "databricks-gpt-5-4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2026-04-30", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.45, - "output": 1.7 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "openai-gpt-5.3-codex": { - "id": "openai-gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "databricks-gemini-3-pro": { + "id": "databricks-gemini-3-pro", + "name": "Gemini 3 Pro Preview", + "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -43136,33 +44829,59 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "openai-gpt-4.1": { - "id": "openai-gpt-4.1", - "name": "GPT-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "databricks-gpt-5-2": { + "id": "databricks-gpt-5-2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -43170,30 +44889,34 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "anthropic-claude-3.5-haiku": { - "id": "anthropic-claude-3.5-haiku", - "name": "Claude 3.5 Haiku", - "description": "Legacy model retained for compatibility with older integrations", + "databricks-claude-haiku-4-5": { + "id": "databricks-claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", "family": "claude-haiku", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-11-05", - "last_updated": "2024-11-05", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -43202,45 +44925,44 @@ "open_weights": false, "limit": { "context": 200000, - "output": 8192 + "output": 64000 }, - "status": "deprecated", "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "openai-gpt-5.4": { - "id": "openai-gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "databricks-gemini-3-1-pro": { + "id": "databricks-gemini-3-1-pro", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -43249,61 +44971,96 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "arcee-trinity-large-thinking": { - "id": "arcee-trinity-large-thinking", - "name": "Trinity Large Thinking", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "trinity", - "attachment": false, + "databricks-claude-opus-4-6": { + "id": "databricks-claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-16", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, + "context": 1000000, "output": 128000 }, - "status": "beta", + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } + }, "cost": { - "input": 0.25, - "output": 0.9, - "cache_read": 0.06 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai-o3": { - "id": "openai-o3", - "name": "o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "databricks-gpt-5-6-sol": { + "id": "databricks-gpt-5-6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -43312,16 +45069,17 @@ "values": [ "low", "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -43334,31 +45092,55 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "anthropic-claude-3.5-sonnet": { - "id": "anthropic-claude-3.5-sonnet", - "name": "Claude 3.5 Sonnet", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-sonnet", + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "databricks-claude-opus-4-1": { + "id": "databricks-claude-opus-4-1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-06-20", - "last_updated": "2024-10-22", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -43367,28 +45149,26 @@ "open_weights": false, "limit": { "context": 200000, - "output": 8192 + "output": 32000 }, - "status": "deprecated", "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "openai-gpt-5-mini": { - "id": "openai-gpt-5-mini", - "name": "GPT-5 mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "databricks-gpt-oss-120b": { + "id": "databricks-gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -43397,14 +45177,54 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.072, + "output": 0.28 + } + }, + "databricks-gemini-2-5-pro": { + "id": "databricks-gemini-2-5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -43412,124 +45232,215 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "nvidia-nemotron-3-super-120b": { - "id": "nvidia-nemotron-3-super-120b", - "name": "Nemotron-3-Super-120B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, + "databricks-gemini-2-5-flash": { + "id": "databricks-gemini-2-5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] + "type": "budget_tokens", + "min": 0, + "max": 24576 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-03-11", - "last_updated": "2026-04-16", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 32768 + "context": 1048576, + "output": 65536 }, - "status": "beta", "cost": { "input": 0.3, - "output": 0.65 + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM 5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "databricks-claude-sonnet-4-5": { + "id": "databricks-claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "databricks-gpt-5-4": { + "id": "databricks-gpt-5-4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "release_date": "2026-02-11", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, + "context": 1050000, + "input": 922000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, "cost": { - "input": 1, - "output": 3.2 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "openai-gpt-5.2-pro": { - "id": "openai-gpt-5.2-pro", - "name": "GPT-5.2 pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "databricks-gpt-5-4-nano": { + "id": "databricks-gpt-5-4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "temperature": false, + "structured_output": true, + "temperature": true, "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -43542,29 +45453,38 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 21, - "output": 168 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "anthropic-claude-3-opus": { - "id": "anthropic-claude-3-opus", - "name": "Claude 3 Opus", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-opus", + "databricks-claude-sonnet-4-6": { + "id": "databricks-claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2023-08", - "release_date": "2024-02-29", - "last_updated": "2024-02-29", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -43572,22 +45492,21 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 4096 + "context": 1000000, + "output": 64000 }, - "status": "deprecated", "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "anthropic-claude-fable-5": { - "id": "anthropic-claude-fable-5", - "name": "Anthropic Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "databricks-gpt-5-6-luna": { + "id": "databricks-gpt-5-6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -43596,19 +45515,21 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-06-09", - "last_updated": "2026-06-12", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -43616,61 +45537,78 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 + }, + "cost": { + "input": 1, + "output": 6, + "cache_read": 0.1, + "tiers": [ + { + "input": 2, + "output": 9, + "cache_read": 0.2, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 9, + "cache_read": 0.2 + } } }, - "nemotron-3-nano-omni": { - "id": "nemotron-3-nano-omni", - "name": "Nemotron Nano 3 Omni", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", + "databricks-claude-sonnet-4": { + "id": "databricks-claude-sonnet-4", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-30", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.5, - "output": 0.9 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "alibaba-qwen3-32b": { - "id": "alibaba-qwen3-32b", - "name": "Qwen3-32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "databricks-gpt-5-1": { + "id": "databricks-gpt-5-1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -43679,58 +45617,61 @@ "none", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2026-04-16", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131000, - "output": 40960 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 0.55 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "nemotron-nano-12b-v2-vl": { - "id": "nemotron-nano-12b-v2-vl", - "name": "Nemotron Nano 12B v2 VL", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", + "databricks-gpt-5-nano": { + "id": "databricks-gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-01", - "last_updated": "2026-04-30", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -43740,31 +45681,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "openai-gpt-5.1-codex-max": { - "id": "openai-gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "databricks-gpt-5": { + "id": "databricks-gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], @@ -43772,8 +45715,8 @@ "structured_output": true, "temperature": false, "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -43786,6 +45729,7 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { @@ -43794,140 +45738,125 @@ "cache_read": 0.125 } }, - "fal-ai/fast-sdxl": { - "id": "fal-ai/fast-sdxl", - "name": "Fast SDXL", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "stable-diffusion", - "attachment": false, - "reasoning": false, - "tool_call": false, + "databricks-gpt-5-5": { + "id": "databricks-gpt-5-5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2023-07-26", - "last_updated": "2026-04-16", + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "image" - ] - }, - "open_weights": true, - "limit": { - "context": 0, - "output": 0 - } - }, - "fal-ai/elevenlabs/tts/multilingual-v2": { - "id": "fal-ai/elevenlabs/tts/multilingual-v2", - "name": "ElevenLabs Multilingual TTS v2", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "elevenlabs", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2023-08-22", - "last_updated": "2026-04-16", - "modalities": { - "input": [ "text" - ], - "output": [ - "audio" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 - } - }, - "fal-ai/flux/schnell": { - "id": "fal-ai/flux/schnell", - "name": "FLUX.1 [schnell]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-08-01", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] + "context": 1050000, + "input": 922000, + "output": 128000 }, - "open_weights": true, - "limit": { - "context": 0, - "output": 0 - } - }, - "fal-ai/stable-audio-25/text-to-audio": { - "id": "fal-ai/stable-audio-25/text-to-audio", - "name": "Stable Audio 2.5 (Text-to-Audio)", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-10-08", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } } } }, - "subconscious": { - "id": "subconscious", + "umans-ai-coding-plan": { + "id": "umans-ai-coding-plan", "env": [ - "SUBCONSCIOUS_API_KEY" + "UMANS_AI_CODING_PLAN_API_KEY" ], - "npm": "@ai-sdk/anthropic", - "api": "https://api.subconscious.dev/v1", - "name": "Subconscious", - "doc": "https://docs.subconscious.dev", + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.code.umans.ai/v1", + "name": "Umans AI Coding Plan", + "doc": "https://app.umans.ai/offers/code/docs", "models": { - "subconscious/glm-5.2": { - "id": "subconscious/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "umans-glm-5.3-flash": { + "id": "umans-glm-5.3-flash", + "name": "GLM 5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -43935,88 +45864,92 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048576, + "output": 131071 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "subconscious/tim-qwen3.6-27b": { - "id": "subconscious/tim-qwen3.6-27b", - "name": "TIM-Qwen3.6 27B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "umans-flash": { + "id": "umans-flash", + "name": "Umans Flash", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "temperature": false, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "input": 8192, - "output": 5000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 3, - "cache_read": 0.15 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - } - } - }, - "venice": { - "id": "venice", - "env": [ - "VENICE_API_KEY" - ], - "npm": "venice-ai-sdk-provider", - "name": "Venice AI", - "doc": "https://docs.venice.ai", - "models": { - "z-ai-glm-5-turbo": { - "id": "z-ai-glm-5-turbo", - "name": "GLM 5 Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", + }, + "umans-deepseek-v4-pro-0813": { + "id": "umans-deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-06-11", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -44027,71 +45960,36 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 32768 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 - } - }, - "grok-4-20-multi-agent": { - "id": "grok-4-20-multi-agent", - "name": "Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "release_date": "2026-03-12", - "last_updated": "2026-06-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 128000 + "context": 1048576, + "output": 393215 }, "cost": { - "input": 1.42, - "output": 2.83, - "cache_read": 0.23, - "tiers": [ - { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45 - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", + "umans-deepseek-v4-flash-0731": { + "id": "umans-deepseek-v4-flash-0731", "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" @@ -44099,8 +45997,8 @@ "structured_output": true, "temperature": true, "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-06-11", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -44111,43 +46009,48 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 1048576, + "output": 393215 }, "cost": { - "input": 0.138, - "output": 0.275, - "cache_read": 0.028 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "google-gemma-4-31b-it": { - "id": "google-gemma-4-31b-it", - "name": "Google Gemma 4 31B Instruct", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "umans-kimi-k3": { + "id": "umans-kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "release_date": "2026-04-03", - "last_updated": "2026-06-11", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -44155,32 +46058,44 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 8192 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.12, - "output": 0.36, - "cache_read": 0.09 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "kimi-k2-6": { - "id": "kimi-k2-6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "umans-qwen3.6-35b-a3b": { + "id": "umans-qwen3.6-35b-a3b", + "name": "Qwen3.6 35B A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-20", - "last_updated": "2026-06-11", + "temperature": false, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", @@ -44192,71 +46107,95 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.75, - "output": 3.5, - "cache_read": 0.16 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai-gpt-56-terra-pro": { - "id": "openai-gpt-56-terra-pro", - "name": "GPT-5.6 Terra Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-mini", + "umans-coder": { + "id": "umans-coder", + "name": "Umans Coder", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 131071 }, "cost": { - "input": 3.125, - "output": 18.75, - "cache_read": 0.3125 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "Qwen 3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + } + } + }, + "kuae-cloud-coding-plan": { + "id": "kuae-cloud-coding-plan", + "env": [ + "KUAE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://coding-plan-endpoint.kuaecloud.net/v1", + "name": "KUAE Cloud Coding Plan", + "doc": "https://docs.mthreads.com/kuaecloud/kuaecloud-doc-online/coding_plan/", + "models": { + "GLM-4.7": { + "id": "GLM-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-04-29", - "last_updated": "2026-06-11", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -44267,27 +46206,88 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + } + } + }, + "umans-ai": { + "id": "umans-ai", + "env": [ + "UMANS_AI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.code.umans.ai/v1", + "name": "Umans AI", + "doc": "https://app.umans.ai/offers/code/docs/orgs", + "models": { + "umans-glm-5.3-flash": { + "id": "umans-glm-5.3-flash", + "name": "GLM 5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131071 }, "cost": { "input": 0.15, - "output": 0.75 + "output": 0.5, + "cache_read": 0.03 } }, - "openai-gpt-56-sol-pro": { - "id": "openai-gpt-56-sol-pro", - "name": "GPT-5.6 Sol Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "umans-flash": { + "id": "umans-flash", + "name": "Umans Flash", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", "high" @@ -44295,11 +46295,13 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", @@ -44309,40 +46311,92 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 6.25, - "output": 37.5, - "cache_read": 0.625 + "input": 0.15, + "output": 1, + "cache_read": 0.05 } }, - "nvidia-nemotron-cascade-2-30b-a3b": { - "id": "nvidia-nemotron-cascade-2-30b-a3b", - "name": "Nemotron Cascade 2 30B A3B", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", + "umans-deepseek-v4-pro-0813": { + "id": "umans-deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 393215 + }, + "cost": { + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 + } + }, + "umans-deepseek-v4-flash-0731": { + "id": "umans-deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-03-24", - "last_updated": "2026-06-11", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -44353,28 +46407,43 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 32768 + "context": 1048576, + "output": 393215 }, "cost": { "input": 0.14, - "output": 0.8 + "output": 0.28, + "cache_read": 0.028 } }, - "claude-opus-4-7-fast": { - "id": "claude-opus-4-7-fast", - "name": "Claude Opus 4.7 Fast", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "umans-kimi-k3": { + "id": "umans-kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-05-14", - "last_updated": "2026-06-11", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", @@ -44384,73 +46453,87 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 36, - "output": 180, - "cache_read": 3.6, - "cache_write": 45 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "openai-gpt-55-pro": { - "id": "openai-gpt-55-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt", + "umans-coder": { + "id": "umans-coder", + "name": "Umans Coder", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-24", - "last_updated": "2026-06-11", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 131071 }, "cost": { - "input": 37.5, - "output": 225 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } - }, - "llama-3.3-70b": { - "id": "llama-3.3-70b", - "name": "Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + } + } + }, + "siliconflow": { + "id": "siliconflow", + "env": [ + "SILICONFLOW_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.siliconflow.com/v1", + "name": "SiliconFlow", + "doc": "https://cloud.siliconflow.com/models", + "models": { + "ByteDance-Seed/Seed-OSS-36B-Instruct": { + "id": "ByteDance-Seed/Seed-OSS-36B-Instruct", + "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "seed", "attachment": false, "reasoning": false, "tool_call": true, - "release_date": "2025-04-06", - "last_updated": "2026-06-11", + "structured_output": true, + "temperature": true, + "release_date": "2025-09-04", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -44459,44 +46542,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.7, - "output": 2.8 + "input": 0.21, + "output": 0.57 } }, - "qwen3-5-397b-a17b": { - "id": "qwen3-5-397b-a17b", - "name": "Qwen 3.5 397B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "nex-agi/Nex-N2-Pro": { + "id": "nex-agi/Nex-N2-Pro", + "name": "Nex-N2-Pro", + "description": "Open agentic MoE model (397B total, 17B active) for coding, tool use, and research workflows", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-02-16", - "last_updated": "2026-06-11", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -44504,32 +46584,40 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 262144, + "output": 256000 }, "cost": { - "input": 0.75, - "output": 4.5 + "input": 0.5, + "output": 2.5, + "cache_read": 0.25 } }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "tencent/Hunyuan-A13B-Instruct": { + "id": "tencent/Hunyuan-A13B-Instruct", + "name": "tencent/Hunyuan-A13B-Instruct", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-12-06", - "last_updated": "2026-06-11", + "release_date": "2025-06-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -44537,42 +46625,32 @@ }, "open_weights": false, "limit": { - "context": 198000, - "output": 32768 + "context": 131000, + "output": 131000 }, "cost": { - "input": 6, - "output": 30, - "cache_read": 0.6, - "cache_write": 7.5 + "input": 0.14, + "output": 0.57 } }, - "zai-org-glm-5": { - "id": "zai-org-glm-5", - "name": "GLM 5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "tencent/Hy3": { + "id": "tencent/Hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-06-11", + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ "text" @@ -44583,108 +46661,101 @@ }, "open_weights": true, "limit": { - "context": 198000, - "output": 32000 + "context": 262144, + "input": 192000, + "output": 262144 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.132, + "output": 0.528, + "cache_read": 0.033 } }, - "qwen3-5-35b-a3b": { - "id": "qwen3-5-35b-a3b", - "name": "Qwen 3.5 35B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "meituan-longcat/LongCat-2.0": { + "id": "meituan-longcat/LongCat-2.0", + "name": "LongCat-2.0", + "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", + "family": "longcat", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-02-25", - "last_updated": "2026-06-11", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 16384 + "context": 1049000, + "output": 131072 }, "cost": { - "input": 0.3125, - "output": 1.25, - "cache_read": 0.15625 + "input": 0.75, + "output": 2.95, + "cache_read": 0.015 } }, - "venice-uncensored-role-play": { - "id": "venice-uncensored-role-play", - "name": "Venice Role Play Uncensored", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "venice", - "attachment": true, + "google/gemma-4-12B-it": { + "id": "google/gemma-4-12B-it", + "name": "Gemma 4 12B IT", + "description": "Compact Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-02-20", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.5, - "output": 2 + "input": 0.1, + "output": 0.3 } }, - "qwen-3-6-plus": { - "id": "qwen-3-6-plus", - "name": "Qwen 3.6 Plus Uncensored", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-06", - "last_updated": "2026-06-11", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -44692,60 +46763,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.625, - "output": 3.75, - "cache_read": 0.0625, - "cache_write": 0.78, - "tiers": [ - { - "input": 2.5, - "output": 7.5, - "cache_read": 0.0625, - "cache_write": 0.78, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.0625, - "cache_write": 0.78 - } + "input": 0.13, + "output": 0.4 } }, - "zai-org-glm-4.6": { - "id": "zai-org-glm-4.6", - "name": "GLM 4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", + "google/gemma-4-26B-A4B-it": { + "id": "google/gemma-4-26B-A4B-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2024-04-01", - "last_updated": "2026-06-11", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" @@ -44754,30 +46791,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 198000, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.43, - "output": 1.75, - "cache_read": 0.08 + "input": 0.12, + "output": 0.4 } }, - "openai-gpt-4o-2024-11-20": { - "id": "openai-gpt-4o-2024-11-20", - "name": "GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "Qwen/Qwen3-VL-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", + "name": "Qwen/Qwen3-VL-30B-A3B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2026-02-28", - "last_updated": "2026-06-11", + "release_date": "2025-10-05", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -44789,37 +46824,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 262000, + "output": 262000 }, "cost": { - "input": 3.125, - "output": 12.5 + "input": 0.29, + "output": 1 } }, - "grok-4-3": { - "id": "grok-4-3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "Qwen/Qwen3-VL-8B-Instruct": { + "id": "Qwen/Qwen3-VL-8B-Instruct", + "name": "Qwen/Qwen3-VL-8B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-18", - "last_updated": "2026-06-11", + "release_date": "2025-10-15", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -44831,35 +46855,18 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 1.42, - "output": 2.83, - "cache_read": 0.23, - "tiers": [ - { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45 - } + "input": 0.18, + "output": 0.68 } }, - "qwen-3-7-plus": { - "id": "qwen-3-7-plus", - "name": "Qwen 3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "Qwen/Qwen3-VL-32B-Thinking": { + "id": "Qwen/Qwen3-VL-32B-Thinking", + "name": "Qwen/Qwen3-VL-32B-Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, "reasoning": true, @@ -44867,14 +46874,12 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-11", + "release_date": "2025-10-21", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -44882,56 +46887,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.5, - "output": 2, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 1.5, - "output": 6, - "cache_read": 0.15, - "cache_write": 1.875, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.5, - "output": 6, - "cache_read": 0.15, - "cache_write": 1.875 - } + "input": 0.2, + "output": 1.5 } }, - "minimax-m27": { - "id": "minimax-m27", - "name": "MiniMax M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "Qwen/Qwen3-30B-A3B-Instruct-2507": { + "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-06-11", + "release_date": "2025-07-30", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -44940,77 +46915,58 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 198000, - "output": 32768 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.375, - "output": 1.5, - "cache_read": 0.06875 + "input": 0.09, + "output": 0.3 } }, - "kimi-k2-7-code": { - "id": "kimi-k2-7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "Qwen/Qwen3.5-27B": { + "id": "Qwen/Qwen3.5-27B", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-13", - "last_updated": "2026-06-16", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.75, - "output": 3.5, - "cache_read": 0.16 + "input": 0.25, + "output": 2 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek", + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-06-11", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text" @@ -45019,41 +46975,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.65, - "output": 3.301, - "cache_read": 0.33 + "input": 0.2, + "output": 1.6 } }, - "qwen3-235b-a22b-thinking-2507": { - "id": "qwen3-235b-a22b-thinking-2507", - "name": "Qwen 3 235B A22B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "Qwen/Qwen3-14B": { + "id": "Qwen/Qwen3-14B", + "name": "Qwen/Qwen3-14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "release_date": "2025-04-29", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -45062,30 +47015,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.45, - "output": 3.5 + "input": 0.07, + "output": 0.28 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "Qwen/Qwen3-VL-30B-A3B-Thinking": { + "id": "Qwen/Qwen3-VL-30B-A3B-Thinking", + "name": "Qwen/Qwen3-VL-30B-A3B-Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-01-15", - "last_updated": "2026-06-11", + "release_date": "2025-10-11", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -45097,33 +47049,39 @@ }, "open_weights": false, "limit": { - "context": 198000, - "output": 64000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 3.75, - "output": 18.75, - "cache_read": 0.375, - "cache_write": 4.69 + "input": 0.29, + "output": 1 } }, - "openai-gpt-4o-mini-2024-07-18": { - "id": "openai-gpt-4o-mini-2024-07-18", - "name": "GPT-4o Mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt", - "attachment": true, - "reasoning": false, + "Qwen/Qwen3-8B": { + "id": "Qwen/Qwen3-8B", + "name": "Qwen/Qwen3-8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2026-02-28", - "last_updated": "2026-06-11", + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -45131,33 +47089,29 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.1875, - "output": 0.75, - "cache_read": 0.09375 + "input": 0.06, + "output": 0.06 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen2.5-72B-Instruct": { + "id": "Qwen/Qwen2.5-72B-Instruct", + "name": "Qwen/Qwen2.5-72B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2024-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -45165,34 +47119,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 33000, + "output": 4000 }, "cost": { - "input": 6, - "output": 30, - "cache_read": 0.6, - "cache_write": 7.5 + "input": 0.59, + "output": 0.59 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen/Qwen3-32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-29", - "last_updated": "2026-07-01", + "temperature": true, + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -45200,45 +47159,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.14, + "output": 0.57 } }, - "openai-gpt-53-codex": { - "id": "openai-gpt-53-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "Qwen/Qwen3-Coder-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "name": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-24", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-08-01", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -45246,45 +47189,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 2.19, - "output": 17.5, - "cache_read": 0.219 + "input": 0.07, + "output": 0.28 } }, - "openai-gpt-54": { - "id": "openai-gpt-54", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "Qwen/Qwen2.5-7B-Instruct": { + "id": "Qwen/Qwen2.5-7B-Instruct", + "name": "Qwen/Qwen2.5-7B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2024-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -45292,45 +47219,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, - "output": 131072 + "context": 33000, + "output": 4000 }, "cost": { - "input": 3.13, - "output": 18.8, - "cache_read": 0.313 + "input": 0.05, + "output": 0.05 } }, - "openai-gpt-56-terra": { - "id": "openai-gpt-56-terra", - "name": "GPT-5.6 Terra", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "Qwen/Qwen3.5-35B-A3B": { + "id": "Qwen/Qwen3.5-35B-A3B", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -45338,74 +47249,59 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 3.125, - "output": 18.75, - "cache_read": 0.3125 + "input": 0.24, + "output": 1.8 } }, - "google-gemma-3-27b-it": { - "id": "google-gemma-3-27b-it", - "name": "Google Gemma 3 27B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "Qwen/Qwen3.5-122B-A10B": { + "id": "Qwen/Qwen3.5-122B-A10B", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2025-11-04", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 198000, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.12, - "output": 0.2 + "input": 0.26, + "output": 2.08 } }, - "openai-gpt-56-luna": { - "id": "openai-gpt-56-luna", - "name": "GPT-5.6 Luna", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "Qwen/Qwen3.5-9B": { + "id": "Qwen/Qwen3.5-9B", + "name": "Qwen/Qwen3.5-9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-03-03", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -45413,36 +47309,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 7.5, - "cache_read": 0.125 + "input": 0.1, + "output": 0.15 } }, - "aion-labs-aion-3-0-mini": { - "id": "aion-labs-aion-3-0-mini", - "name": "Aion 3.0 Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text" @@ -45453,61 +47339,64 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.875, - "output": 1.75, - "cache_read": 0.225 + "input": 0.39, + "output": 2.34 } }, - "claude-opus-4-8-fast": { - "id": "claude-opus-4-8-fast", - "name": "Claude Opus 4.8 Fast", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "Qwen/Qwen3.8-2.4T-A95B": { + "id": "Qwen/Qwen3.8-2.4T-A95B", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1049000, + "output": 131000 }, "cost": { - "input": 12, - "output": 60, - "cache_read": 1.2, - "cache_write": 15 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "gemma-4-uncensored": { - "id": "gemma-4-uncensored", - "name": "Gemma 4 Uncensored", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "Qwen/Qwen3-VL-32B-Instruct": { + "id": "Qwen/Qwen3-VL-32B-Instruct", + "name": "Qwen/Qwen3-VL-32B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-04-13", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-10-21", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -45517,27 +47406,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.1625, - "output": 0.5 + "input": 0.2, + "output": 0.6 } }, - "qwen3-next-80b": { - "id": "qwen3-next-80b", - "name": "Qwen 3 Next 80b", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "Qwen/Qwen3-Coder-480B-A35B-Instruct": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "name": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2025-04-29", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-07-31", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -45546,75 +47436,71 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 16384 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.35, - "output": 1.9 + "input": 0.25, + "output": 1 } }, - "mistral-small-2603": { - "id": "mistral-small-2603", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-06-11", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.1875, - "output": 0.75 + "input": 0.3, + "output": 3.2 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "deepseek-ai/DeepSeek-V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-08-25", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -45622,40 +47508,37 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 164000, + "output": 164000 }, "cost": { - "input": 6, - "output": 30, - "cache_read": 0.6, - "cache_write": 7.5 + "input": 0.27, + "output": 1 } }, - "zai-org-glm-4.7-flash": { - "id": "zai-org-glm-4.7-flash", - "name": "GLM 4.7 Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm", + "deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-29", - "last_updated": "2026-06-11", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -45666,165 +47549,158 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.125, - "output": 0.5 + "input": 0.22, + "output": 0.66, + "cache_read": 0.014 } }, - "google-gemma-4-26b-a4b-it": { - "id": "google-gemma-4-26b-a4b-it", - "name": "Google Gemma 4 26B A4B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "deepseek-ai/DeepSeek-R1": { + "id": "deepseek-ai/DeepSeek-R1", + "name": "deepseek-ai/DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-06-11", + "release_date": "2025-05-28", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.13, - "output": 0.4, - "cache_read": 0.05 + "input": 0.5, + "output": 2.18 } }, - "grok-4-5": { - "id": "grok-4-5", - "name": "Grok 4.5", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Pro-0813": { + "id": "deepseek-ai/DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-07-07", - "last_updated": "2026-07-08", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 32000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 2.27, - "output": 6.8, - "cache_read": 0.57, - "tiers": [ - { - "input": 4.53, - "output": 13.6, - "cache_read": 1.13, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4.53, - "output": 13.6, - "cache_read": 1.13 - } + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 } }, - "grok-4-20": { - "id": "grok-4-20", - "name": "Grok 4.20", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-03-12", - "last_updated": "2026-06-11", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1.42, - "output": 2.83, - "cache_read": 0.23, - "tiers": [ - { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45 - } + "input": 0.13, + "output": 0.28, + "cache_read": 0.028 } }, - "hermes-3-llama-3.1-405b": { - "id": "hermes-3-llama-3.1-405b", - "name": "Hermes 3 Llama 3.1 405b", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "hermes", + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2025-09-25", - "last_updated": "2026-06-11", + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -45835,33 +47711,37 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1.1, - "output": 3 + "input": 1.50162, + "output": 3.135, + "cache_read": 0.135 } }, - "aion-labs-aion-2-0": { - "id": "aion-labs-aion-2-0", - "name": "Aion 2.0", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "deepseek-ai/DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], - "tool_call": false, - "release_date": "2026-03-24", - "last_updated": "2026-06-11", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-03", + "last_updated": "2025-12-03", "modalities": { "input": [ "text" @@ -45872,185 +47752,183 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 164000, + "output": 164000 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.25 + "input": 0.27, + "output": 0.42, + "cache_read": 0.135 } }, - "qwen3-vl-235b-a22b": { - "id": "qwen3-vl-235b-a22b", - "name": "Qwen3 VL 235B", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "deepseek-ai/DeepSeek-V3": { + "id": "deepseek-ai/DeepSeek-V3", + "name": "deepseek-ai/DeepSeek-V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-01-16", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2024-12-26", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.21, - "output": 1.9, - "cache_read": 0.1 + "input": 0.25, + "output": 1 } }, - "xiaomi-mimo-v2-5": { - "id": "xiaomi-mimo-v2-5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, + "deepseek-ai/DeepSeek-V3.1-Terminus": { + "id": "deepseek-ai/DeepSeek-V3.1-Terminus", + "name": "deepseek-ai/DeepSeek-V3.1-Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-06-11", - "last_updated": "2026-06-11", + "release_date": "2025-09-29", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.05 + "input": 0.27, + "output": 1 } }, - "venice-uncensored-1-2": { - "id": "venice-uncensored-1-2", - "name": "Venice Uncensored 1.2", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "venice", - "attachment": true, - "reasoning": false, + "deepseek-ai/DeepSeek-V3.2-Exp": { + "id": "deepseek-ai/DeepSeek-V3.2-Exp", + "name": "deepseek-ai/DeepSeek-V3.2-Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, "structured_output": true, - "release_date": "2026-04-01", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-10-10", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.2, - "output": 0.9 + "input": 0.27, + "output": 0.41 } }, - "gemini-3-5-flash": { - "id": "gemini-3-5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini", + "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp": { + "id": "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-22", - "last_updated": "2026-06-11", + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 65536 + "output": 384000 }, "cost": { - "input": 1.55, - "output": 9.45, - "cache_read": 0.155, - "cache_write": 0.086 + "input": 0.44, + "output": 1.32, + "cache_read": 0.028 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, + "stepfun-ai/Step-3.5-Flash": { + "id": "stepfun-ai/Step-3.5-Flash", + "name": "stepfun-ai/Step-3.5-Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "family": "step", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-10", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -46058,42 +47936,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 12, - "output": 60, - "cache_read": 1.2, - "cache_write": 15 + "input": 0.1, + "output": 0.3 } }, - "openai-gpt-oss-120b": { - "id": "openai-gpt-oss-120b", - "name": "OpenAI GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "MiniMaxAI/MiniMax-M3": { + "id": "MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-11-06", - "last_updated": "2026-06-11", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -46101,27 +47976,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 131000 }, "cost": { - "input": 0.07, - "output": 0.3 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "zai-org-glm-5-2": { - "id": "zai-org-glm-5-2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMaxAI/MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "release_date": "2026-02-15", + "last_updated": "2026-06-15", "modalities": { "input": [ "text" @@ -46130,29 +48005,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 197000, + "output": 131000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "nvidia-nemotron-3-nano-30b-a3b": { - "id": "nvidia-nemotron-3-nano-30b-a3b", - "name": "NVIDIA Nemotron 3 Nano 30B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", + "inclusionAI/Ling-flash-2.0": { + "id": "inclusionAI/Ling-flash-2.0", + "name": "inclusionAI/Ling-flash-2.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-06-11", + "release_date": "2025-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -46161,30 +48036,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 0.14, + "output": 0.57 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "moonshotai/Kimi-K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-06-11", + "release_date": "2026-04-21", + "last_updated": "2026-06-15", "modalities": { "input": [ "text", @@ -46194,32 +48077,43 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 6, - "output": 30, - "cache_read": 0.6, - "cache_write": 7.5 + "input": 0.77, + "output": 3.4, + "cache_read": 0.14 } }, - "mistral-small-3-2-24b-instruct": { - "id": "mistral-small-3-2-24b-instruct", - "name": "Mistral Small 3.2 24B Instruct", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, - "reasoning": false, + "moonshotai/Kimi-K3": { + "id": "moonshotai/Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-01-15", - "last_updated": "2026-06-11", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -46227,89 +48121,42 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 1048576, + "output": 262000 }, "cost": { - "input": 0.09375, - "output": 0.25 + "input": 2.7, + "output": 13.5, + "cache_read": 0.27 } }, - "openai-gpt-54-pro": { - "id": "openai-gpt-54-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt", + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-06-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 37.5, - "output": 225, - "tiers": [ - { - "input": 75, - "output": 337.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 75, - "output": 337.5 - } - } - }, - "minimax-m3-preview": { - "id": "minimax-m3-preview", - "name": "MiniMax M3 Preview", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax-m3", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "knowledge": "2025-01", "release_date": "2026-06-12", - "last_updated": "2026-06-13", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -46317,93 +48164,41 @@ }, "open_weights": true, "limit": { - "context": 524288, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.85916, + "output": 3.8, + "cache_read": 0.17993 } }, - "gemini-3-1-pro-preview": { - "id": "gemini-3-1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini", - "attachment": true, + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "moonshotai/Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-06-11", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32768 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.5, - "cache_write": 0.5, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 0.5 - } - } - }, - "qwen3-coder-480b-a35b-instruct-turbo": { - "id": "qwen3-coder-480b-a35b-instruct-turbo", - "name": "Qwen 3 Coder 480B Turbo", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, "structured_output": true, + "temperature": true, "release_date": "2026-01-27", - "last_updated": "2026-06-11", + "last_updated": "2026-01-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -46411,121 +48206,116 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.35, - "output": 1.5, - "cache_read": 0.04 + "input": 0.45, + "output": 2.25, + "cache_read": 0.07 } }, - "qwen3-6-27b": { - "id": "qwen3-6-27b", - "name": "Qwen 3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "zai-org/GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-24", - "last_updated": "2026-06-11", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 205000, + "output": 131000 }, "cost": { - "input": 0.325, - "output": 3.25 + "input": 1.19, + "output": 3.74, + "cache_read": 0.6, + "cache_write": 0 } }, - "openai-gpt-52-codex": { - "id": "openai-gpt-52-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt", - "attachment": true, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", - "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08", - "release_date": "2025-01-15", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 272000, - "output": 65536 + "context": 1049000, + "output": 262000 }, "cost": { - "input": 2.19, - "output": 17.5, - "cache_read": 0.219 + "input": 1.302, + "output": 4.092, + "cache_read": 0.26, + "cache_write": 0 } }, - "qwen-3-7-max": { - "id": "qwen-3-7-max", - "name": "Qwen 3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "zai-org/GLM-4.5-Air": { + "id": "zai-org/GLM-4.5-Air", + "name": "zai-org/GLM-4.5-Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-22", - "last_updated": "2026-06-11", + "release_date": "2025-07-28", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -46533,38 +48323,36 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 131000, + "output": 131000 }, "cost": { - "input": 2.7, - "output": 8.05, - "cache_read": 0.27, - "cache_write": 3.35 + "input": 0.14, + "output": 0.86 } }, - "olafangensan-glm-4.7-flash-heretic": { - "id": "olafangensan-glm-4.7-flash-heretic", - "name": "GLM 4.7 Flash Heretic", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "zai-org/GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-02-04", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-06-15", "modalities": { "input": [ "text" @@ -46575,40 +48363,40 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 24000 + "context": 205000, + "output": 205000 }, "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.035 + "input": 0.95, + "output": 2.55, + "cache_read": 0.2 } }, - "openai-gpt-54-mini": { - "id": "openai-gpt-54-mini", - "name": "GPT-5.4 Mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt", + "zai-org/GLM-5.3-Flash": { + "id": "zai-org/GLM-5.3-Flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-27", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", @@ -46618,31 +48406,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1049000, + "output": 262000 }, "cost": { - "input": 0.9375, - "output": 5.625, - "cache_read": 0.09375 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03, + "cache_write": 0 } }, - "grok-build-0-1": { - "id": "grok-build-0-1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", + "zai-org/GLM-5V-Turbo": { + "id": "zai-org/GLM-5V-Turbo", + "name": "zai-org/GLM-5V-Turbo", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-06-11", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", @@ -46654,35 +48453,20 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 200000, + "output": 131072 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 4, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 4, - "cache_read": 0.4 - } + "input": 1.2, + "output": 4, + "cache_read": 0.24, + "cache_write": 0 } }, - "zai-org-glm-4.7": { - "id": "zai-org-glm-4.7", - "name": "GLM 4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "zai-org/GLM-5.3": { + "id": "zai-org/GLM-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", "family": "glm", "attachment": false, "reasoning": true, @@ -46691,8 +48475,8 @@ "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], @@ -46702,9 +48486,8 @@ }, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-24", - "last_updated": "2026-06-11", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -46715,33 +48498,31 @@ }, "open_weights": true, "limit": { - "context": 198000, - "output": 16384 + "context": 1049000, + "output": 262000 }, "cost": { - "input": 0.55, - "output": 2.65, - "cache_read": 0.11 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "openai/gpt-oss-20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-06-11", + "release_date": "2025-08-13", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -46749,50 +48530,36 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 131000, + "output": 8000 }, "cost": { - "input": 3.6, - "output": 18, - "cache_read": 0.36, - "cache_write": 4.5 + "input": 0.04, + "output": 0.18 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "openai/gpt-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-19", - "last_updated": "2026-06-11", + "release_date": "2025-08-13", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -46800,104 +48567,79 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 131000, + "output": 8000 }, "cost": { - "input": 0.7, - "output": 3.75, - "cache_read": 0.07 + "input": 0.05, + "output": 0.45 } - }, - "openai-gpt-55": { - "id": "openai-gpt-55", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + } + } + }, + "minimax-coding-plan": { + "id": "minimax-coding-plan", + "env": [ + "MINIMAX_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://api.minimax.io/anthropic/v1", + "name": "MiniMax Token Plan (minimax.io)", + "doc": "https://platform.minimax.io/docs/token-plan/intro", + "models": { + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-25", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 922000, - "output": 131072 + "context": 1048576, + "output": 512000 }, "cost": { - "input": 6.25, - "output": 37.5, - "cache_read": 0.625, - "tiers": [ - { - "input": 12.5, - "output": 56.25, - "cache_read": 1.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 12.5, - "output": 56.25, - "cache_read": 1.25 - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "qwen3-5-9b": { - "id": "qwen3-5-9b", - "name": "Qwen 3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": true, + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-05", - "last_updated": "2026-06-11", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -46905,36 +48647,28 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.15 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "mercury-2": { - "id": "mercury-2", - "name": "Mercury 2", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-02-20", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -46943,43 +48677,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 50000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.3125, - "output": 0.9375, - "cache_read": 0.03125 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "zai-org-glm-5-1": { - "id": "zai-org-glm-5-1", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-06-11", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -46990,173 +48711,138 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 24000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.54, - "output": 4.84, - "cache_read": 0.286 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai-gpt-56-luna-pro": { - "id": "openai-gpt-56-luna-pro", - "name": "GPT-5.6 Luna Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-nano", - "attachment": true, + "MiniMax-M2.7-highspeed": { + "id": "MiniMax-M2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 7.5, - "cache_read": 0.125 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "kimi-k2-5": { - "id": "kimi-k2-5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, + "MiniMax-M2": { + "id": "MiniMax-M2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2024-04", - "release_date": "2026-01-27", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.56, - "output": 3.5, - "cache_read": 0.22 + "input": 0, + "output": 0 } }, - "z-ai-glm-5v-turbo": { - "id": "z-ai-glm-5v-turbo", - "name": "GLM 5V Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, + "MiniMax-M2.5-highspeed": { + "id": "MiniMax-M2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-06-11", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 5, - "cache_read": 0.3 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "nvidia-nemotron-3-ultra-550b-a55b": { - "id": "nvidia-nemotron-3-ultra-550b-a55b", - "name": "NVIDIA Nemotron 3 Ultra", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", + } + } + }, + "alibaba-coding-plan": { + "id": "alibaba-coding-plan", + "env": [ + "ALIBABA_CODING_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://coding-intl.dashscope.aliyuncs.com/v1", + "name": "Alibaba Coding Plan", + "doc": "https://www.alibabacloud.com/help/en/model-studio/coding-plan", + "models": { + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-11", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -47165,38 +48851,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.625, - "output": 3.125, - "cache_read": 0.1875 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "aion-labs-aion-3-0": { - "id": "aion-labs-aion-3-0", - "name": "Aion 3.0", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "temperature": true, + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ "text" @@ -47205,45 +48883,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 262144, + "output": 65536 }, "cost": { - "input": 3.75, - "output": 7.5, - "cache_read": 0.9375 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai-gpt-52": { - "id": "openai-gpt-52", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-13", - "last_updated": "2026-06-11", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -47251,38 +48924,35 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 272000, + "context": 1000000, "output": 65536 }, "cost": { - "input": 2.19, - "output": 17.5, - "cache_read": 0.219 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "release_date": "2025-12-04", - "last_updated": "2026-06-11", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -47291,86 +48961,142 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 160000, - "output": 32768 + "context": 202752, + "output": 16384 }, "cost": { - "input": 0.33, - "output": 0.48, - "cache_read": 0.16 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai-gpt-56-sol": { - "id": "openai-gpt-56-sol", - "name": "GPT-5.6 Sol", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 6.25, - "output": 37.5, - "cache_read": 0.625 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "minimax-m25": { - "id": "minimax-m25", - "name": "MiniMax M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "qwen3-max-2026-01-23": { + "id": "qwen3-max-2026-01-23", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-23", + "last_updated": "2026-01-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 }, + "cost": { + "input": 0.1875, + "output": 1.125, + "cache_write": 0.234375 + } + }, + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-06-11", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -47381,25 +49107,31 @@ }, "open_weights": true, "limit": { - "context": 198000, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.27, - "output": 0.95, - "cache_read": 0.03 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "llama-3.2-3b": { - "id": "llama-3.2-3b", - "name": "Llama 3.2 3B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "release_date": "2024-10-03", - "last_updated": "2026-06-11", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -47410,30 +49142,21 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 196608, + "input": 196601, + "output": 24576 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - } - } - }, - "crossmodel": { - "id": "crossmodel", - "env": [ - "CROSSMODEL_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.crossmodel.ai/v1", - "name": "CrossModel", - "doc": "https://www.crossmodel.ai/docs", - "models": { - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", + }, + "glm-4.7": { + "id": "glm-4.7", "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, @@ -47443,6 +49166,9 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, "knowledge": "2025-04", "release_date": "2025-12-22", @@ -47457,33 +49183,21 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 202752, + "output": 16384 }, "cost": { - "input": 0.47, - "output": 2.16, - "cache_read": 0.1, - "cache_write": 0.47, - "tiers": [ - { - "input": 0.62, - "output": 2.47, - "cache_read": 0.13, - "cache_write": 0.62, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -47492,48 +49206,38 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1, - "output": 3.8, - "cache_read": 0.2, - "cache_write": 1, - "tiers": [ - { - "input": 1.2, - "output": 4.4, - "cache_read": 0.3, - "cache_write": 1.2, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -47541,36 +49245,50 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 64000 }, "cost": { - "input": 1.2, - "output": 4.4, - "cache_read": 0.3, - "cache_write": 1.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + } + } + }, + "lilac": { + "id": "lilac", + "env": [ + "LILAC_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.getlilac.com/v1", + "name": "Lilac", + "doc": "https://docs.getlilac.com/inference/models", + "models": { + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -47578,12 +49296,19 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -47591,34 +49316,20 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 262100, + "output": 262100 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.16, - "cache_write": 0.6, - "tiers": [ - { - "input": 0.8, - "output": 3.4, - "cache_read": 0.2, - "cache_write": 0.8, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.11, + "output": 0.35 } }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, + "minimaxai/minimax-m3": { + "id": "minimaxai/minimax-m3", + "name": "MiniMax M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax-m3", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -47628,65 +49339,50 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.9, - "output": 3.7, - "cache_read": 0.18, - "cache_write": 0.9, - "tiers": [ - { - "input": 1.1, - "output": 4.3, - "cache_read": 0.27, - "cache_write": 1.1, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.28, + "output": 1.1, + "cache_read": 0.05 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -47696,292 +49392,259 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02, - "cache_write": 0.2 + "input": 0.7, + "output": 3.5, + "cache_read": 0.2 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, + "zai-org/glm-5.2": { + "id": "zai-org/glm-5.2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 524288, + "output": 524288 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 2.5, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 5 - } + "input": 0.9, + "output": 3, + "cache_read": 0.27 } - }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + } + } + }, + "moonshotai-cn": { + "id": "moonshotai-cn", + "env": [ + "MOONSHOT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.moonshot.cn/v1", + "name": "Moonshot AI (China)", + "doc": "https://platform.moonshot.cn/docs/api/chat", + "models": { + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075, - "cache_write": 0.75 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", + "kimi-k2.7-code-highspeed": { + "id": "kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code HighSpeed", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075, - "cache_write": 0.15 + "input": 1.9, + "output": 8, + "cache_read": 0.38 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 10 - } + "input": 3, + "output": 15, + "cache_read": 0.3 } - }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + } + } + }, + "bothub": { + "id": "bothub", + "env": [ + "BOTHUB_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://openai.bothub.ru/v1", + "name": "Bothub", + "doc": "https://bothub.ru/models", + "models": { + "gemma-4-31b-it:free": { + "id": "gemma-4-31b-it:free", + "name": "Gemma 4 31B IT (free)", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -47990,10 +49653,11 @@ } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -48005,36 +49669,43 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.16, - "output": 0.32, - "cache_read": 0.004, - "cache_write": 0.16 + "input": 0, + "output": 0 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" @@ -48043,164 +49714,149 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 131072 }, "cost": { - "input": 0.47, - "output": 0.94, - "cache_read": 0.005, - "cache_write": 0.47 + "input": 0.12, + "output": 0.44 } }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 384000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.1, + "output": 0.28 } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "nemotron-3-ultra-550b-a55b:free": { + "id": "nemotron-3-ultra-550b-a55b:free", + "name": "Nemotron 3 Ultra (free)", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", + "none", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "interleaved": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, "output": 128000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0, + "output": 0 } }, - "anthropic/claude-opus-4-8": { - "id": "anthropic/claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high", - "xhigh", "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "interleaved": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 384000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.61, + "output": 4.84 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -48210,14 +49866,17 @@ } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -48225,38 +49884,47 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.06, + "output": 0.37 } }, - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "muse-spark-1.3-contributor": { + "id": "muse-spark-1.3-contributor", + "name": "Muse Spark 1.3 Contributor", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf", + "audio" ], "output": [ "text" @@ -48264,88 +49932,92 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.1, + "output": 0.2 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "medium", "high", "max" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 64000 + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 1.72, + "output": 5.41 } - }, - "tencent/hy3-preview": { - "id": "tencent/hy3-preview", - "name": "Hy3 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + } + } + }, + "xiaomi-token-plan-cn": { + "id": "xiaomi-token-plan-cn", + "env": [ + "XIAOMI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://token-plan-cn.xiaomimimo.com/v1", + "name": "Xiaomi Token Plan (China)", + "doc": "https://platform.xiaomimimo.com/#/docs", + "models": { + "mimo-v2.6-pro": { + "id": "mimo-v2.6-pro", + "name": "MiMo-V2.6-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -48353,56 +50025,41 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.19, - "output": 0.63, - "cache_read": 0.063, - "cache_write": 0.19, - "tiers": [ - { - "input": 0.25, - "output": 1, - "cache_read": 0.094, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 16000 - } - }, - { - "input": 0.32, - "output": 1.25, - "cache_read": 0.125, - "cache_write": 0.32, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0, + "output": 0, + "cache_read": 0 } }, - "moonshot/kimi-k2.7-code": { - "id": "moonshot/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -48410,123 +50067,94 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 1, - "output": 4.16, - "cache_read": 0.18, - "cache_write": 1 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "moonshot/kimi-k2.5": { - "id": "moonshot/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "mimo-v2.5-tts-voicedesign": { + "id": "mimo-v2.5-tts-voicedesign", + "name": "MiMo-V2.5-TTS-VoiceDesign", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "audio" ] }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.62, - "output": 3.3, - "cache_read": 0.11, - "cache_write": 0.62 + "input": 0, + "output": 0 } }, - "moonshot/kimi-k2.6": { - "id": "moonshot/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "mimo-v2.5-tts": { + "id": "mimo-v2.5-tts", + "name": "MiMo-V2.5-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "audio" ] }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 1, - "output": 4.16, - "cache_read": 0.18, - "cache_write": 1 + "input": 0, + "output": 0 } }, - "gemini/gemini-2.5-pro": { - "id": "gemini/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -48535,109 +50163,101 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "cache_write": 1.25, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 200000 - } - } + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "mimo-v2-tts": { + "id": "mimo-v2-tts", + "name": "MiMo-V2-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 2.5 - } + "output": [ + "audio" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 } }, - "gemini/gemini-2.5-flash": { - "id": "gemini/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.3 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "gemini/gemini-3.5-flash": { - "id": "gemini/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "mimo-v2.6-flash": { + "id": "mimo-v2.6-flash", + "name": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", @@ -48649,72 +50269,106 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 1.5 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "gemini/gemini-2.5-flash-lite": { - "id": "gemini/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", + "mimo-v2.5-tts-voiceclone": { + "id": "mimo-v2.5-tts-voiceclone", + "name": "MiMo-V2.5-TTS-VoiceClone", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + } + } + }, + "privatemode-ai": { + "id": "privatemode-ai", + "env": [ + "PRIVATEMODE_API_KEY", + "PRIVATEMODE_ENDPOINT" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "http://localhost:8080/v1", + "name": "Privatemode AI", + "doc": "https://docs.privatemode.ai/api/overview", + "models": { + "glm-flash-latest": { + "id": "glm-flash-latest", + "name": "GLM Flash (latest)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 131072 }, + "status": "beta", "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0.1 + "input": 0.8897, + "output": 4.4718, + "cache_read": 0.0924 } }, - "gemini/gemini-3.1-pro-preview": { - "id": "gemini/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -48722,187 +50376,147 @@ "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 131072 }, + "status": "beta", "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 4, - "tier": { - "type": "context", - "size": 200000 - } - } + "input": 0.8897, + "output": 4.4718, + "cache_read": 0.0924 + } + }, + "whisper-large-v3": { + "id": "whisper-large-v3", + "name": "Whisper large-v3", + "description": "Open Whisper checkpoint for robust multilingual transcription and captioning", + "family": "whisper", + "attachment": true, + "reasoning": false, + "tool_call": false, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", + "modalities": { + "input": [ + "audio" ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 4 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 448, + "output": 4096 + }, + "cost": { + "input": 0.01618, + "output": 0 } }, - "gemini/gemini-3-flash-preview": { - "id": "gemini/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] } ], "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.5 + "input": 1.791, + "output": 8.9436, + "cache_read": 0.1733 } }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "deepseek-ocr-2": { + "id": "deepseek-ocr-2", + "name": "DeepSeek OCR 2", + "description": "High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes", + "attachment": true, + "reasoning": false, + "tool_call": false, + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 8192, + "output": 8192 }, + "status": "beta", "cost": { - "input": 0.32, - "output": 1.25, - "cache_read": 0.032, - "cache_write": 0.4, - "tiers": [ - { - "input": 0.96, - "output": 3.75, - "cache_read": 0.096, - "cache_write": 1.2, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.96, - "output": 3.75, - "cache_read": 0.096, - "cache_write": 1.2 - } + "input": 0.8897, + "output": 1.4675, + "cache_read": 0.0924 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "qwen3-embedding-4b": { + "id": "qwen3-embedding-4b", + "name": "Qwen3-Embedding 4B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2025-06", + "release_date": "2025-06-06", + "last_updated": "2025-06-06", "modalities": { "input": [ "text" @@ -48911,163 +50525,107 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 32000, + "output": 2560 }, "cost": { - "input": 1.88, - "output": 5.63, - "cache_read": 0.375, - "cache_write": 2.35 + "input": 0.1502, + "output": 0 } }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "voxtral-mini-3b": { + "id": "voxtral-mini-3b", + "name": "Voxtral Mini 3B", + "description": "Speech-to-text model for audio transcription, translation, and audio understanding", + "family": "voxtral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2025-07", + "last_updated": "2025-07", "modalities": { "input": [ - "text", - "image", - "video" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 32000, + "output": 32000 }, "cost": { - "input": 0.19, - "output": 1.13, - "cache_read": 0.019, - "cache_write": 0.24, - "tiers": [ - { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075, - "cache_write": 0.94, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075, - "cache_write": 0.94 - } + "input": 0.00462, + "output": 0 } }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": false, + "kimi-latest": { + "id": "kimi-latest", + "name": "Kimi (latest)", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 256000, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 0.32, - "output": 1.88, - "cache_read": 0.032, - "cache_write": 0.4, - "tiers": [ - { - "input": 1.25, - "output": 7.5, - "cache_read": 0.124, - "cache_write": 1.57, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.25, - "output": 7.5, - "cache_read": 0.124, - "cache_write": 1.57 - } + "input": 1.791, + "output": 8.9436, + "cache_read": 0.1733 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "xhigh" + "low", + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -49078,41 +50636,37 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 65000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.16, - "output": 0.32, - "cache_read": 0.004, - "cache_write": 0.16 + "input": 0.4969, + "output": 1.9644, + "cache_read": 0.0462 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", "high", - "xhigh" + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -49123,84 +50677,38 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 65000 + "context": 256000, + "output": 131072 }, + "status": "beta", "cost": { - "input": 0.47, - "output": 0.94, - "cache_read": 0.005, - "cache_write": 0.47 + "input": 1.791, + "output": 8.9436, + "cache_read": 0.1733 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, + "glm-latest": { + "id": "glm-latest", + "name": "GLM (latest)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1024000, - "output": 512000 - }, - "cost": { - "input": 0.33, - "output": 1.32, - "cache_read": 0.066, - "cache_write": 0.33, - "tiers": [ - { - "input": 0.66, - "output": 2.63, - "cache_read": 0.132, - "cache_write": 0.66, - "tier": { - "type": "context", - "size": 512000 - } - } - ], - "context_over_200k": { - "input": 0.66, - "output": 2.63, - "cache_read": 0.132, - "cache_write": 0.66 - } - } - }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -49211,49 +50719,51 @@ }, "open_weights": true, "limit": { - "context": 204800, + "context": 256000, "output": 131072 }, + "status": "beta", "cost": { - "input": 0.33, - "output": 1.32, - "cache_read": 0.066, - "cache_write": 0.42 + "input": 1.791, + "output": 8.9436, + "cache_read": 0.1733 } } } }, - "lmstudio": { - "id": "lmstudio", + "llmgateway-providers": { + "id": "llmgateway-providers", "env": [ - "LMSTUDIO_API_KEY" + "LLMGATEWAY_API_KEY" ], "npm": "@ai-sdk/openai-compatible", - "api": "http://127.0.0.1:1234/v1", - "name": "LMStudio", - "doc": "https://lmstudio.ai/models", + "api": "https://api.llmgateway.io/v1", + "name": "LLM Gateway", + "doc": "https://llmgateway.io/docs", "models": { - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "consensusprotocol/gemma-4-31b-it": { + "id": "consensusprotocol/gemma-4-31b-it", + "name": "Gemma 4 31B IT (Consensus Protocol)", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "none", + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" @@ -49264,59 +50774,84 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 262144, "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.25, + "cache_read": 0.01 } }, - "qwen/qwen3-30b-a3b-2507": { - "id": "qwen/qwen3-30b-a3b-2507", - "name": "Qwen3 30B A3B 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "consensusprotocol/Qwen3.8-27B": { + "id": "consensusprotocol/Qwen3.8-27B", + "name": "Qwen3.8 27B (Consensus Protocol)", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-30", - "last_updated": "2025-07-30", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.08, + "output": 0.35, + "cache_read": 0.05 } }, - "qwen/qwen3-coder-30b": { - "id": "qwen/qwen3-coder-30b", - "name": "Qwen3 Coder 30B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "consensusprotocol/glm-5.3-flash": { + "id": "consensusprotocol/glm-5.3-flash", + "name": "GLM-5.3 Flash (Consensus Protocol)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" @@ -49324,45 +50859,43 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.25, + "cache_read": 0.02 } - } - } - }, - "poolside": { - "id": "poolside", - "env": [ - "POOLSIDE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.poolside.ai/v1", - "name": "Poolside", - "doc": "https://platform.poolside.ai", - "models": { - "poolside/laguna-xs.2": { - "id": "poolside/laguna-xs.2", - "name": "Laguna XS.2", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", - "attachment": false, + }, + "consensusprotocol/deepseek-v4.1-flash": { + "id": "consensusprotocol/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (Consensus Protocol)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": false, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-06-13", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -49370,32 +50903,37 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.2, + "output": 0.6, + "cache_read": 0.005 } }, - "poolside/laguna-m.1": { - "id": "poolside/laguna-m.1", - "name": "Laguna M.1", - "description": "Poolside's flagship agentic coding model for long-horizon work", - "family": "laguna", - "attachment": false, + "consensusprotocol/gpt-oss-20b": { + "id": "consensusprotocol/gpt-oss-20b", + "name": "GPT OSS 20B (Consensus Protocol)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": false, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-06-13", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -49406,32 +50944,42 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 65536, "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.04, + "output": 0.19, + "cache_read": 0.01 } }, - "poolside/laguna-xs-2.1": { - "id": "poolside/laguna-xs-2.1", - "name": "Laguna XS 2.1", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", + "consensusprotocol/deepseek-v4-flash": { + "id": "consensusprotocol/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (Consensus Protocol)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-07-02", - "last_updated": "2026-07-02", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -49442,41 +50990,41 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 1050000, + "output": 393216 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.05, + "output": 0.1, + "cache_read": 0.01 } - } - } - }, - "zenifra": { - "id": "zenifra", - "env": [ - "ZENIFRA_AI_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://ai.zenifra.com/v1", - "name": "Zenifra", - "doc": "https://docs.zenifra.com", - "models": { - "alibaba/qwen3.6-35b-a3b": { - "id": "alibaba/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, + }, + "deepinfra/gemma-4-31b-it": { + "id": "deepinfra/gemma-4-31b-it", + "name": "Gemma 4 31B IT (DeepInfra)", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" @@ -49488,73 +51036,82 @@ "open_weights": true, "limit": { "context": 262144, - "output": 65536 - }, - "provider": { - "shape": "completions" + "output": 32768 }, "cost": { - "input": 0.19, - "output": 0.48 + "input": 0.13, + "output": 0.38 } - } - } - }, - "zenmux": { - "id": "zenmux", - "env": [ - "ZENMUX_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://zenmux.ai/api/v1", - "name": "ZenMux", - "doc": "https://docs.zenmux.ai", - "models": { - "inclusionai/ling-1t": { - "id": "inclusionai/ling-1t", - "name": "Ling-1T", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, + }, + "deepinfra/qwen3-vl-235b-a22b-instruct": { + "id": "deepinfra/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct (DeepInfra)", + "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-10-09", - "last_updated": "2025-10-09", + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.56, - "output": 2.24, + "input": 0.2, + "output": 0.88, "cache_read": 0.11 } }, - "inclusionai/ring-2.6-1t": { - "id": "inclusionai/ring-2.6-1t", - "name": "inclusionAI: Ring-2.6-1T", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "deepinfra/mimo-v2.5": { + "id": "deepinfra/mimo-v2.5", + "name": "MiMo V2.5 (DeepInfra)", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-12-31", - "release_date": "2026-05-07", - "last_updated": "2026-05-14", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -49562,65 +51119,101 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 65000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.06 + "input": 0.4, + "output": 2, + "cache_read": 0.08, + "tiers": [ + { + "input": 0.8, + "output": 4, + "cache_read": 0.16, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.8, + "output": 4, + "cache_read": 0.16 + } } }, - "inclusionai/ring-1t": { - "id": "inclusionai/ring-1t", - "name": "Ring-1T", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "deepinfra/qwen3.8-27b": { + "id": "deepinfra/qwen3.8-27b", + "name": "Qwen3.8 27B (DeepInfra)", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-10-12", - "last_updated": "2025-10-12", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 64000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.56, - "output": 2.24, - "cache_read": 0.11 + "input": 0.2, + "output": 2.5, + "cache_read": 0.05 } }, - "moonshotai/kimi-k2.7-code-free": { - "id": "moonshotai/kimi-k2.7-code-free", - "name": "Kimi K2.7 Code (Free)", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "deepinfra/nemotron-3.5-lightning": { + "id": "deepinfra/nemotron-3.5-lightning", + "name": "Nemotron 3.5 Lightning (DeepInfra)", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -49629,26 +51222,38 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.08, + "output": 0.2 } }, - "moonshotai/kimi-k2-thinking-turbo": { - "id": "moonshotai/kimi-k2-thinking-turbo", - "name": "Kimi K2 Thinking Turbo", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "deepinfra/qwen3.8-2.4t-a95b": { + "id": "deepinfra/qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B (DeepInfra)", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text" @@ -49657,36 +51262,47 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 64000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 1.15, - "output": 8, - "cache_read": 0.15 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "deepinfra/muse-glimmer-30b": { + "id": "deepinfra/muse-glimmer-30b", + "name": "Muse Glimmer 30B (DeepInfra)", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -49694,51 +51310,68 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.3, + "output": 1.2, + "cache_read": 0.04 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "attachment": false, + "deepinfra/deepseek-v4.1-flash": { + "id": "deepinfra/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (DeepInfra)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 64000 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 0.2, + "output": 0.6, + "cache_read": 0.006 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "attachment": true, + "deepinfra/ling-3.0-flash": { + "id": "deepinfra/ling-3.0-flash", + "name": "InclusionAI Ling 3.0 Flash (DeepInfra)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -49746,18 +51379,13 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2025-01-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "structured_output": false, + "temperature": true, + "release_date": "2026-08-02", + "last_updated": "2026-08-02", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -49765,66 +51393,58 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.58, - "output": 3.02, - "cache_read": 0.1 + "input": 0.06, + "output": 0.18, + "cache_read": 0.012 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "deepinfra/qwen3-vl-30b-a3b-instruct": { + "id": "deepinfra/qwen3-vl-30b-a3b-instruct", + "name": "Qwen3 VL 30B A3B Instruct (DeepInfra)", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2025-01-01", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "structured_output": false, + "temperature": true, + "release_date": "2025-10-05", + "last_updated": "2025-10-05", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262140, - "output": 262140 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.15, + "output": 0.6 } }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "deepinfra/gemma-4-26b-a4b-it": { + "id": "deepinfra/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT (DeepInfra)", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-09-04", - "last_updated": "2025-09-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" @@ -49833,34 +51453,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 0.07, + "output": 0.34 } }, - "baidu/ernie-5.0-thinking-preview": { - "id": "baidu/ernie-5.0-thinking-preview", - "name": "ERNIE 5.0", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "deepinfra/nemotron-3-ultra-550b": { + "id": "deepinfra/nemotron-3-ultra-550b", + "name": "Nemotron 3 Ultra 550B (DeepInfra)", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -49868,162 +51486,218 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.84, - "output": 3.37 + "input": 0.5, + "output": 2.2, + "cache_read": 0.1 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "deepinfra/inkling-small": { + "id": "deepinfra/inkling-small", + "name": "Inkling Small (DeepInfra)", + "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 524288, + "output": 262144 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025 + "input": 0.45, + "output": 1.2, + "cache_read": 0.1 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "attachment": true, + "deepinfra/mimo-v2.5-pro": { + "id": "deepinfra/mimo-v2.5-pro", + "name": "MiMo V2.5 Pro (DeepInfra)", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], + "tool_call": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 16384 + }, + "cost": { + "input": 1, + "output": 3, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } + } + }, + "deepinfra/qwen3.5-9b": { + "id": "deepinfra/qwen3.5-9b", + "name": "Qwen3.5 9B (DeepInfra)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "pdf", - "image", - "text", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.31, - "cache_write": 4.5 + "input": 0.1, + "output": 0.15 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "deepinfra/inkling": { + "id": "deepinfra/inkling", + "name": "Inkling (DeepInfra)", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "pdf", - "image", "text", + "image", "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048000, - "output": 64000 + "context": 524288, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.07, - "cache_write": 1 + "input": 0.95, + "output": 4.05, + "cache_read": 0.16 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "deepinfra/hy3": { + "id": "deepinfra/hy3", + "name": "Hy3 (DeepInfra)", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", "high" ] } @@ -50031,198 +51705,222 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "input": 192000, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 + "input": 0.14, + "output": 0.58, + "cache_read": 0.035 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "attachment": true, - "reasoning": false, + "deepinfra/glm-5.1": { + "id": "deepinfra/glm-5.1", + "name": "GLM-5.1 (DeepInfra)", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "pdf", - "image", - "text", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048000, - "output": 64000 + "context": 198000, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.03, - "cache_write": 1 + "input": 1.05, + "output": 3.5, + "cache_read": 0.205 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "attachment": true, + "deepinfra/deepseek-v4-pro": { + "id": "deepinfra/deepseek-v4-pro", + "name": "DeepSeek V4 Pro (DeepInfra)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "none", + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2026-02-19", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048000, + "context": 1048576, "output": 64000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 4.5 + "input": 1.3, + "output": 2.6, + "cache_read": 0.1 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "deepinfra/step-3.7-flash": { + "id": "deepinfra/step-3.7-flash", + "name": "Step 3.7 Flash (DeepInfra)", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", "image", - "pdf", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 1 + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 } }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "attachment": true, + "deepinfra/deepseek-v3.2": { + "id": "deepinfra/deepseek-v3.2", + "name": "DeepSeek V3.2 (DeepInfra)", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-03-20", - "last_updated": "2025-03-20", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "output": 65530 + "context": 160000, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 1.5 + "input": 0.26, + "output": 0.38, + "cache_read": 0.13 } }, - "x-ai/grok-code-fast-1": { - "id": "x-ai/grok-code-fast-1", - "name": "Grok Code Fast 1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "deepinfra/glm-5.3": { + "id": "deepinfra/glm-5.3", + "name": "GLM-5.3 (DeepInfra)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -50231,23 +51929,23 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 1.5, - "cache_read": 0.02 + "input": 1.2, + "output": 4, + "cache_read": 0.2 } }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + "deepinfra/deepseek-v4-flash": { + "id": "deepinfra/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (DeepInfra)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -50255,67 +51953,65 @@ "values": [ "none", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 1000000 + "output": 393216 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "cache_write": 0, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "cache_write": 0, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "cache_write": 0 - } + "input": 0.08, + "output": 0.18, + "cache_read": 0.016 } }, - "x-ai/grok-4.1-fast-non-reasoning": { - "id": "x-ai/grok-4.1-fast-non-reasoning", - "name": "Grok 4.1 Fast Non Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "cerebras/gemma-4-31b-it": { + "id": "cerebras/gemma-4-31b-it", + "name": "Gemma 4 31B IT (Cerebras)", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -50325,153 +52021,83 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 64000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.99, + "output": 1.49 } }, - "x-ai/grok-4": { - "id": "x-ai/grok-4", - "name": "Grok 4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "cerebras/qwen3-235b-a22b-instruct-2507": { + "id": "cerebras/qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507 (Cerebras)", + "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "release_date": "2025-07-21", + "last_updated": "2025-07-21", "modalities": { "input": [ - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 262000, + "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 + "input": 0.6, + "output": 1.2 } }, - "x-ai/grok-4-fast": { - "id": "x-ai/grok-4-fast", - "name": "Grok 4 Fast", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "cerebras/llama-3.3-70b-instruct": { + "id": "cerebras/llama-3.3-70b-instruct", + "name": "Llama 3.3 70B Instruct (Cerebras)", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-09-19", - "last_updated": "2025-09-19", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 64000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.85, + "output": 1.2 } }, - "x-ai/grok-4.2-fast-non-reasoning": { - "id": "x-ai/grok-4.2-fast-non-reasoning", - "name": "Grok 4.2 Fast Non Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 30000 - }, - "cost": { - "input": 3, - "output": 9 - } - }, - "x-ai/grok-4.2-fast": { - "id": "x-ai/grok-4.2-fast", - "name": "Grok 4.2 Fast", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 30000 - }, - "cost": { - "input": 3, - "output": 9 - } - }, - "x-ai/grok-4.1-fast": { - "id": "x-ai/grok-4.1-fast", - "name": "Grok 4.1 Fast", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": true, + "cerebras/glm-4.7": { + "id": "cerebras/glm-4.7", + "name": "GLM-4.7 (Cerebras)", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -50479,79 +52105,54 @@ } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 64000 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } - }, - "x-ai/grok-build-0.1": { - "id": "x-ai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 200000, + "output": 128000 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 + "input": 2.25, + "output": 2.75 } }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "GLM 4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "cerebras/gpt-oss-120b": { + "id": "cerebras/gpt-oss-120b", + "name": "GPT OSS 120B (Cerebras)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -50560,33 +52161,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.28, - "output": 1.14, - "cache_read": 0.06 + "input": 0.35, + "output": 0.75 } }, - "z-ai/glm-4.5": { - "id": "z-ai/glm-4.5", - "name": "GLM 4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "groq/gpt-oss-20b": { + "id": "groq/gpt-oss-20b", + "name": "GPT OSS 20B (Groq)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -50595,32 +52201,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 64000 + "context": 131072, + "output": 32766 }, "cost": { - "input": 0.35, - "output": 1.54, - "cache_read": 0.07 + "input": 0.1, + "output": 0.5 } }, - "z-ai/glm-4.7-flashx": { - "id": "z-ai/glm-4.7-flashx", - "name": "GLM 4.7 FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "groq/gpt-oss-120b": { + "id": "groq/gpt-oss-120b", + "name": "GPT OSS 120B (Groq)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -50629,67 +52241,73 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 32766 }, "cost": { - "input": 0.07, - "output": 0.42, - "cache_read": 0.01 + "input": 0.15, + "output": 0.75 } }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, + "scx-ai-gp/glm-5.3-flash": { + "id": "scx-ai-gp/glm-5.3-flash", + "name": "GLM-5.3 Flash (SCX.ai)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-03", - "last_updated": "2026-04-03", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, + "context": 1048576, "output": 131072 }, "cost": { - "input": 0.8781, - "output": 3.5126, - "cache_read": 0.1903 + "input": 0.088, + "output": 0.25, + "cache_read": 0.025 } }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "GLM 4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "scx-ai-gp/glm-5.2-fast": { + "id": "scx-ai-gp/glm-5.2-fast", + "name": "GLM-5.2 Turbo (SCX.ai)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -50698,69 +52316,91 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.35, - "output": 1.54, - "cache_read": 0.07 + "input": 2.2, + "output": 6.5, + "cache_read": 0.45 } }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "scx-ai-gp/qwen3.8-max": { + "id": "scx-ai-gp/qwen3.8-max", + "name": "Qwen3.8 Max (SCX.ai)", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", - "max" + "low", + "medium", + "xhigh" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, "output": 131072 }, "cost": { - "input": 1.4, - "output": 4.5, - "cache_read": 0.26 + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "z-ai/glm-4.6v-flash-free": { - "id": "z-ai/glm-4.6v-flash-free", - "name": "GLM 4.6V Flash (Free)", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "scx-ai-gp/kimi-k3": { + "id": "scx-ai-gp/kimi-k3", + "name": "Kimi K3 (SCX.ai)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", @@ -50771,31 +52411,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0, - "output": 0 + "input": 3.5, + "output": 18, + "cache_read": 0.35 } }, - "z-ai/glm-4.7-flash-free": { - "id": "z-ai/glm-4.7-flash-free", - "name": "GLM 4.7 Flash (Free)", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "scx-ai-gp/deepseek-v4.1-flash": { + "id": "scx-ai-gp/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (SCX.ai)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text" @@ -50804,19 +52443,20 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.6, + "cache_read": 0.01 } }, - "z-ai/glm-5.2-free": { - "id": "z-ai/glm-5.2-free", - "name": "GLM 5.2 (Free)", + "scx-ai-gp/glm-5.2": { + "id": "scx-ai-gp/glm-5.2", + "name": "GLM-5.2 (SCX.ai)", "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, @@ -50825,13 +52465,21 @@ { "type": "effort", "values": [ + "none", + "minimal", + "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, "release_date": "2026-06-13", "last_updated": "2026-06-13", @@ -50849,90 +52497,118 @@ "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.88, + "output": 2.55, + "cache_read": 0.16 } }, - "z-ai/glm-4.6v": { - "id": "z-ai/glm-4.6v", - "name": "GLM 4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "attachment": true, + "scx-ai-gp/glm-5.3": { + "id": "scx-ai-gp/glm-5.3", + "name": "GLM-5.3 (SCX.ai)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.42, - "cache_read": 0.03 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "z-ai/glm-5v-turbo": { - "id": "z-ai/glm-5v-turbo", - "name": "GLM 5V Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "scx-ai-gp/kimi-k2.7-code": { + "id": "scx-ai-gp/kimi-k2.7-code", + "name": "Kimi K2.7 Code (SCX.ai)", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "structured_output": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "video", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.726, - "output": 3.1946, - "cache_read": 0.1743 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "z-ai/glm-4.6v-flash": { - "id": "z-ai/glm-4.6v-flash", - "name": "GLM 4.6V FlashX", + "zai/glm-4.6v": { + "id": "zai/glm-4.6v", + "name": "GLM-4.6V (Z AI)", "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", + "knowledge": "2025-04", "release_date": "2025-12-08", "last_updated": "2025-12-08", "modalities": { @@ -50945,21 +52621,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 16000 }, "cost": { - "input": 0.02, - "output": 0.21, - "cache_read": 0.0043 + "input": 0.3, + "output": 0.9, + "cache_read": 0.05 } }, - "z-ai/glm-4.5-air": { - "id": "z-ai/glm-4.5-air", - "name": "GLM 4.5 Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "zai/glm-4.5": { + "id": "zai/glm-4.5", + "name": "GLM-4.5 (Z AI)", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -50968,44 +52645,11 @@ } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 64000 - }, - "cost": { - "input": 0.11, - "output": 0.56, - "cache_read": 0.02 - } - }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM 5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -51016,115 +52660,116 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 128000, + "output": 98304 }, "cost": { - "input": 0.58, - "output": 2.6, - "cache_read": 0.14 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "GLM 5 Turbo", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "zai/glm-5.3-flash": { + "id": "zai/glm-5.3-flash", + "name": "GLM-5.3 Flash (Z AI)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.88, - "output": 3.48 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "openai/gpt-5.5-instant": { - "id": "openai/gpt-5.5-instant", - "name": "GPT-5.5 Instant", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "zai/glm-4.6": { + "id": "zai/glm-4.6", + "name": "GLM-4.6 (Z AI)", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-12-01", - "release_date": "2026-05-05", - "last_updated": "2026-05-28", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 200000, + "output": 131072 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2-Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "attachment": true, + "zai/glm-4.5-x": { + "id": "zai/glm-4.5-x", + "name": "GLM-4.5 X (Z AI)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "structured_output": false, + "temperature": true, + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -51132,79 +52777,77 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 128000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" - }, "cost": { - "input": 21, - "output": 168 + "input": 2.2, + "output": 8.9, + "cache_read": 0.45 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": true, + "zai/glm-5": { + "id": "zai/glm-5", + "name": "GLM-5 (Z AI)", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 202800, + "output": 131100 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "openai/gpt-5.1-chat": { - "id": "openai/gpt-5.1-chat", - "name": "GPT-5.1 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "attachment": true, + "zai/glm-4.5-airx": { + "id": "zai/glm-4.5-airx", + "name": "GLM-4.5 AirX (Z AI)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "pdf", - "image", "text" ], "output": [ @@ -51214,62 +52857,65 @@ "open_weights": false, "limit": { "context": 128000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12 + "input": 1.1, + "output": 4.5, + "cache_read": 0.22 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, + "zai/glm-4.5v": { + "id": "zai/glm-4.5v", + "name": "GLM-4.5V (Z AI)", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 128000, + "output": 16000 }, "cost": { - "input": 0.2, - "output": 1.25 + "input": 0.6, + "output": 1.8, + "cache_read": 0.11 } }, - "openai/gpt-5.3-chat": { - "id": "openai/gpt-5.3-chat", - "name": "GPT-5.3 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "attachment": true, + "zai/glm-4-32b-0414-128k": { + "id": "zai/glm-4-32b-0414-128k", + "name": "GLM-4 32B (0414-128k) (Z AI)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text" @@ -51281,38 +52927,30 @@ "open_weights": false, "limit": { "context": 128000, - "output": 16380 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "output": 128000 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.1, + "output": 0.1 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "zai/glm-4.6v-flashx": { + "id": "zai/glm-4.6v-flashx", + "name": "GLM-4.6V FlashX (Z AI)", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", @@ -51324,86 +52962,80 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 128000, + "output": 16000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12 + "input": 0.04, + "output": 0.4, + "cache_read": 0.004 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": true, + "zai/glm-4.7": { + "id": "zai/glm-4.7", + "name": "GLM-4.7 (Z AI)", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2025-01-01", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "image", - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 200000, + "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.17 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, + "zai/glm-5.2": { + "id": "zai/glm-5.2", + "name": "GLM-5.2 (Z AI)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "none", + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -51412,120 +53044,115 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, + "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" - }, "cost": { - "input": 1.75, - "output": 14 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1-Codex-Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, + "zai/glm-5.1": { + "id": "zai/glm-5.1", + "name": "GLM-5.1 (Z AI)", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.03 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "attachment": true, + "zai/glm-4.7-flashx": { + "id": "zai/glm-4.7-flashx", + "name": "GLM-4.7 FlashX (Z AI)", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, + "context": 200000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" - }, "cost": { - "input": 3.75, - "output": 18.75 + "input": 0.07, + "output": 0.4, + "cache_read": 0.01 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "zai/glm-4.5-air": { + "id": "zai/glm-4.5-air", + "name": "GLM-4.5 Air (Z AI)", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -51534,87 +53161,83 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 128000, + "output": 98304 }, "cost": { - "input": 0.75, - "output": 4.5 + "input": 0.2, + "output": 1.1, + "cache_read": 0.03 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "attachment": true, + "zai/glm-5.3": { + "id": "zai/glm-5.3", + "name": "GLM-5.3 (Z AI)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "medium", + "low", "high", - "xhigh" + "max" ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, + "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" - }, "cost": { - "input": 45, - "output": 225 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "aws-bedrock/claude-haiku-4-5": { + "id": "aws-bedrock/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (AWS Bedrock)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", @@ -51627,50 +53250,29 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "GPT-5 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "aws-bedrock/llama-4-maverick-17b-instruct": { + "id": "aws-bedrock/llama-4-maverick-17b-instruct", + "name": "Llama 4 Maverick 17B Instruct (AWS Bedrock)", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -51680,42 +53282,46 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 8192, + "output": 2048 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12 + "input": 0.24, + "output": 0.97 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "aws-bedrock/claude-sonnet-4-5": { + "id": "aws-bedrock/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (AWS Bedrock)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2025-01-01", - "release_date": "2026-01-15", - "last_updated": "2026-01-15", + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", @@ -51728,23 +53334,21 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 200000, + "output": 8192 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.17 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "aws-bedrock/claude-fable-5-1": { + "id": "aws-bedrock/claude-fable-5-1", + "name": "Claude Fable 5.1 (AWS Bedrock)", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -51753,19 +53357,22 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": false, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ - "image", "text", + "image", "pdf" ], "output": [ @@ -51774,42 +53381,78 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 64000 + "context": 1000000, + "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "aws-bedrock/llama-4-scout-17b-instruct": { + "id": "aws-bedrock/llama-4-scout-17b-instruct", + "name": "Llama 4 Scout 17B Instruct (AWS Bedrock)", + "description": "Open Llama with long-context vision for efficient multimodal agents", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 2048 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12 + "input": 0.17, + "output": 0.66 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "aws-bedrock/claude-opus-4-1-20250805": { + "id": "aws-bedrock/claude-opus-4-1-20250805", + "name": "Claude Opus 4.1 (AWS Bedrock)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -51822,129 +53465,97 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 200000, + "output": 32000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "aws-bedrock/claude-opus-4-5-20251101": { + "id": "aws-bedrock/claude-opus-4-5-20251101", + "name": "Claude Opus 4.5 (AWS Bedrock)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024, + "max": 31999 + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08, - "tiers": [ - { - "input": 0.8, - "output": 4, - "cache_read": 0.16, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.8, - "output": 4, - "cache_read": 0.16 - } + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "xiaomi/mimo-v2-omni": { - "id": "xiaomi/mimo-v2-omni", - "name": "MiMo V2 Omni", - "description": "MiMo omni model for text, image, video, audio, and agents", - "family": "mimo", + "aws-bedrock/claude-opus-5": { + "id": "aws-bedrock/claude-opus-5", + "name": "Claude Opus 5 (AWS Bedrock)", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": false, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -51953,71 +53564,93 @@ }, "open_weights": false, "limit": { - "context": 265000, - "output": 265000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "xiaomi/mimo-v2-flash": { - "id": "xiaomi/mimo-v2-flash", - "name": "MiMo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, + "aws-bedrock/claude-fable-5": { + "id": "aws-bedrock/claude-fable-5", + "name": "Claude Fable 5 (AWS Bedrock)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2024-12-01", - "release_date": "2025-12-16", - "last_updated": "2026-02-04", + "structured_output": false, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "xiaomi/mimo-v2-pro": { - "id": "xiaomi/mimo-v2-pro", - "name": "MiMo V2 Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", - "attachment": false, + "aws-bedrock/claude-opus-4-8": { + "id": "aws-bedrock/claude-opus-4-8", + "name": "Claude Opus 4.8 (AWS Bedrock)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": false, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -52026,99 +53659,50 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 256000 + "output": 128000 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "aws-bedrock/claude-sonnet-4-5-20250929": { + "id": "aws-bedrock/claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5 (2025-09-29) (AWS Bedrock)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024, + "max": 63999 + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } - } - }, - "sapiens-ai/agnes-1.5-pro": { - "id": "sapiens-ai/agnes-1.5-pro", - "name": "Agnes 1.5 Pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-21", - "last_updated": "2026-03-21", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -52126,28 +53710,46 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.16, - "output": 0.8 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "sapiens-ai/agnes-1.5-lite": { - "id": "sapiens-ai/agnes-1.5-lite", - "name": "Agnes 1.5 Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "aws-bedrock/claude-sonnet-5": { + "id": "aws-bedrock/claude-sonnet-5", + "name": "Claude Sonnet 5 (AWS Bedrock)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-26", - "last_updated": "2026-03-26", + "structured_output": false, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -52155,18 +53757,21 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.12, - "output": 0.6 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "volcengine/doubao-seed-2.0-pro": { - "id": "volcengine/doubao-seed-2.0-pro", - "name": "Doubao-Seed-2.0-pro", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "aws-bedrock/claude-opus-4-6": { + "id": "aws-bedrock/claude-opus-4-6", + "name": "Claude Opus 4.6 (AWS Bedrock)", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -52175,20 +53780,23 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2026-02-14", - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -52196,26 +53804,28 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.45, - "output": 2.24, - "cache_read": 0.09, - "cache_write": 0.0024 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "volcengine/doubao-seed-code": { - "id": "volcengine/doubao-seed-code", - "name": "Doubao-Seed-Code", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "aws-bedrock/grok-4-3": { + "id": "aws-bedrock/grok-4-3", + "name": "Grok 4.3 (AWS Bedrock)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -52223,10 +53833,10 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-11", - "last_updated": "2025-11-11", + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", @@ -52238,19 +53848,36 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.17, - "output": 1.12, - "cache_read": 0.03 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "volcengine/doubao-seed-1.8": { - "id": "volcengine/doubao-seed-1.8", - "name": "Doubao-Seed-1.8", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "aws-bedrock/claude-haiku-4-5-20251001": { + "id": "aws-bedrock/claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5 (2025-10-01) (AWS Bedrock)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -52259,20 +53886,23 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -52280,20 +53910,21 @@ }, "open_weights": false, "limit": { - "context": 256000, + "context": 200000, "output": 64000 }, "cost": { - "input": 0.11, - "output": 0.28, - "cache_read": 0.02, - "cache_write": 0.0024 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "volcengine/doubao-seed-2.0-mini": { - "id": "volcengine/doubao-seed-2.0-mini", - "name": "Doubao-Seed-2.0-mini", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "aws-bedrock/grok-4-6": { + "id": "aws-bedrock/grok-4-6", + "name": "Grok 4.6 (AWS Bedrock)", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -52302,20 +53933,21 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2026-02-14", - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -52323,42 +53955,50 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.03, - "output": 0.28, - "cache_read": 0.01, - "cache_write": 0.0024 + "input": 2, + "output": 6, + "cache_read": 0.5 } }, - "volcengine/doubao-seed-2.0-lite": { - "id": "volcengine/doubao-seed-2.0-lite", - "name": "Doubao-Seed-2.0-lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "aws-bedrock/claude-sonnet-4-6": { + "id": "aws-bedrock/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6 (AWS Bedrock)", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2026-02-14", - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -52366,50 +54006,21 @@ }, "open_weights": false, "limit": { - "context": 256000, + "context": 1000000, "output": 64000 }, "cost": { - "input": 0.09, - "output": 0.51, - "cache_read": 0.02, - "cache_write": 0.0024 - } - }, - "volcengine/doubao-seed-2.0-code": { - "id": "volcengine/doubao-seed-2.0-code", - "name": "Doubao Seed 2.0 Code", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 32000 - }, - "cost": { - "input": 0.9, - "output": 4.48 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "anthropic/claude-sonnet-5-free": { - "id": "anthropic/claude-sonnet-5-free", - "name": "Claude Sonnet 5 (Free)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "aws-bedrock/claude-opus-4-7": { + "id": "aws-bedrock/claude-opus-4-7", + "name": "Claude Opus 4.7 (AWS Bedrock)", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -52418,15 +54029,18 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": false, "temperature": false, "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -52442,32 +54056,28 @@ "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" - }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "anthropic/claude-3.5-haiku": { - "id": "anthropic/claude-3.5-haiku", - "name": "Claude 3.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "attachment": true, + "aws-bedrock/llama-3.1-70b-instruct": { + "id": "aws-bedrock/llama-3.1-70b-instruct", + "name": "Llama 3.1 70B Instruct (AWS Bedrock)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2024-11-04", - "last_updated": "2024-11-04", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -52475,30 +54085,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 128000, + "output": 2048 }, "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "input": 0.72, + "output": 0.72 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "google-ai-studio/gemini-3.6-flash": { + "id": "google-ai-studio/gemini-3.6-flash", + "name": "Gemini 3.6 Flash (Google AI Studio)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -52506,14 +54112,17 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -52522,30 +54131,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.08333 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "google-ai-studio/gemini-3.5-flash-lite": { + "id": "google-ai-studio/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite (Google AI Studio)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -52553,14 +54160,17 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "image", "text", + "image", + "video", + "audio", "pdf" ], "output": [ @@ -52569,35 +54179,46 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "cache_write": 0.08333 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "google-ai-studio/gemini-3.1-pro-preview": { + "id": "google-ai-studio/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro (Preview) (Google AI Studio)", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ + "text", "image", - "text" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -52605,31 +54226,43 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "google-ai-studio/gemini-3.5-flash": { + "id": "google-ai-studio/gemini-3.5-flash", + "name": "Gemini 3.5 Flash (Google AI Studio)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -52637,14 +54270,17 @@ } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -52653,45 +54289,51 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 4 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 0.08333 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "google-ai-studio/gemini-2.5-pro": { + "id": "google-ai-studio/gemini-2.5-pro", + "name": "Gemini 2.5 Pro (Google AI Studio)", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" ] + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -52700,27 +54342,44 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "anthropic/claude-3.7-sonnet": { - "id": "anthropic/claude-3.7-sonnet", - "name": "Claude 3.7 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "google-ai-studio/gemini-2.5-flash": { + "id": "google-ai-studio/gemini-2.5-flash", + "name": "Gemini 2.5 Flash (Google AI Studio)", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1, + "max": 24576 + }, { "type": "effort", "values": [ @@ -52731,14 +54390,17 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-02-24", - "last_updated": "2025-02-24", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -52747,31 +54409,27 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65535 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "google-ai-studio/gemini-3.7-flash": { + "id": "google-ai-studio/gemini-3.7-flash", + "name": "Gemini 3.7 Flash (Google AI Studio)", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -52779,14 +54437,17 @@ } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -52795,31 +54456,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.08333 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "google-ai-studio/gemini-3-flash-preview": { + "id": "google-ai-studio/gemini-3-flash-preview", + "name": "Gemini 3 Flash (Preview) (Google AI Studio)", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -52827,14 +54485,17 @@ } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -52843,24 +54504,20 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65535 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "google-ai-studio/gemini-pro-latest": { + "id": "google-ai-studio/gemini-pro-latest", + "name": "Gemini Pro Latest (Google AI Studio)", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -52874,15 +54531,15 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-02-27", + "last_updated": "2026-02-27", "modalities": { "input": [ - "image", "text", - "pdf" + "image", + "audio" ], "output": [ "text" @@ -52890,24 +54547,20 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "google-ai-studio/gemini-3.8-flash": { + "id": "google-ai-studio/gemini-3.8-flash", + "name": "Gemini 3.8 Flash (Google AI Studio)", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -52921,15 +54574,17 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -52937,45 +54592,36 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.08333 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "google-ai-studio/gemini-2.5-flash-lite": { + "id": "google-ai-studio/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash Lite (Google AI Studio)", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-18", - "last_updated": "2026-02-18", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -52983,30 +54629,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65535 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "google-ai-studio/gemini-3.1-flash-lite": { + "id": "google-ai-studio/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite (Google AI Studio)", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -53014,14 +54657,17 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "image", "text", + "image", + "video", + "audio", "pdf" ], "output": [ @@ -53030,24 +54676,21 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 0.08333 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (Anthropic)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -53056,19 +54699,23 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -53076,77 +54723,98 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 200000, + "output": 64000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "tencent/hy3-preview": { - "id": "tencent/hy3-preview", - "name": "Hy3 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "anthropic/claude-sonnet-4-5": { + "id": "anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (Anthropic)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + }, { "type": "effort", "values": [ - "none", "low", - "high" + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, + "context": 200000, "output": 64000 }, "cost": { - "input": 0.172, - "output": 0.572, - "cache_read": 0.058, - "cache_write": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "stepfun/step-3": { - "id": "stepfun/step-3", - "name": "Step-3", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "anthropic/claude-opus-5-5": { + "id": "anthropic/claude-opus-5-5", + "name": "Claude Opus 5.5 (Anthropic)", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-07-31", - "last_updated": "2025-07-31", + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -53154,18 +54822,21 @@ }, "open_weights": false, "limit": { - "context": 65536, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.21, - "output": 0.57 + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 } }, - "stepfun/step-3.7-flash": { - "id": "stepfun/step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "anthropic/claude-fable-5-1": { + "id": "anthropic/claude-fable-5-1", + "name": "Claude Fable 5.1 (Anthropic)", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -53174,90 +54845,122 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2026-01-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.15 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "stepfun/step-3.7-flash-free": { - "id": "stepfun/step-3.7-flash-free", - "name": "Step 3.7 Flash (Free)", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "anthropic/claude-opus-4-5-20251101": { + "id": "anthropic/claude-opus-4-5-20251101", + "name": "Claude Opus 4.5 (Anthropic)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2026-01-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "stepfun/step-3.5-flash": { - "id": "stepfun/step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5 (Anthropic)", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-02-02", - "last_updated": "2026-02-02", + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -53265,27 +54968,46 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "kuaishou/kat-coder-pro-v2": { - "id": "kuaishou/kat-coder-pro-v2", - "name": "KAT-Coder-Pro-V2", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, - "reasoning": false, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5 (Anthropic)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -53293,29 +55015,46 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 80000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen3-Coder-Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4-8": { + "id": "anthropic/claude-opus-4-8", + "name": "Claude Opus 4.8 (Anthropic)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -53324,36 +55063,50 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": false, + "anthropic/claude-sonnet-4-5-20250929": { + "id": "anthropic/claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5 (2025-09-29) (Anthropic)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024, + "max": 63999 + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -53361,53 +55114,46 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 200000, "output": 64000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "cache_write": 0.5, - "tiers": [ - { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24, - "cache_write": 1.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24, - "cache_write": 1.5 - } + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5 (Anthropic)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -53416,30 +55162,45 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 65536 + "output": 128000 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3-Max-Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "attachment": false, + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6 (Anthropic)", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-01-23", - "last_updated": "2026-01-23", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -53447,29 +55208,46 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.2, - "output": 6 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen/qwen3.5-flash": { - "id": "qwen/qwen3.5-flash", - "name": "Qwen3.5 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "anthropic/claude-haiku-4-5-20251001": { + "id": "anthropic/claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5 (2025-10-01) (Anthropic)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -53477,39 +55255,51 @@ }, "open_weights": false, "limit": { - "context": 1020000, - "output": 1020000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "qwen/qwen3.5-plus": { - "id": "qwen/qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6 (Anthropic)", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -53521,28 +55311,42 @@ "output": 64000 }, "cost": { - "input": 0.8, - "output": 4.8 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6-Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7 (Anthropic)", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -53551,65 +55355,43 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 128000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "deepseek/deepseek-v4.1-flash": { + "id": "deepseek/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (DeepSeek)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", "family": "deepseek-flash", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -53617,32 +55399,29 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 1050000, + "output": 393216 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 } }, "deepseek/deepseek-v4-pro": { "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", + "name": "DeepSeek V4 Pro (DeepSeek)", "description": "Open MoE flagship with million-token context for coding and long agent runs", "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", - "high" + "none", + "high", + "max" ] } ], @@ -53650,7 +55429,7 @@ "interleaved": { "field": "reasoning_content" }, - "structured_output": true, + "structured_output": false, "temperature": true, "knowledge": "2025-05", "release_date": "2026-04-24", @@ -53665,8 +55444,8 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 1050000, + "output": 393216 }, "cost": { "input": 0.435, @@ -53674,222 +55453,315 @@ "cache_read": 0.003625 } }, - "deepseek/deepseek-v3.2-exp": { - "id": "deepseek/deepseek-v3.2-exp", - "name": "DeepSeek-V3.2-Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "runpod/kimi-k3": { + "id": "runpod/kimi-k3", + "name": "Kimi K3 (Runpod)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163000, - "output": 64000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.22, - "output": 0.33 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "deepseek/deepseek-chat": { - "id": "deepseek/deepseek-chat", - "name": "DeepSeek-V3.2 (Non-thinking Mode)", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, - "reasoning": false, + "runware/gemma-4-31b-it": { + "id": "runware/gemma-4-31b-it", + "name": "Gemma 4 31B IT (Runware)", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 64000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.28, - "output": 0.42, - "cache_read": 0.03 + "input": 0.102, + "output": 0.297, + "cache_read": 0.012 } }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "runware/glm-5.3-flash": { + "id": "runware/glm-5.3-flash", + "name": "GLM-5.3 Flash (Runware)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-05", - "last_updated": "2025-12-05", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 64000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.28, - "output": 0.43 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax M2.7 highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "runware/kimi-k3": { + "id": "runware/kimi-k3", + "name": "Kimi K3 (Runware)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 131070 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.611, - "output": 2.4439 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": false, + "runware/deepseek-v4.1-flash": { + "id": "runware/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (Runware)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.15, + "output": 0.6, + "cache_read": 0.01 } }, - "minimax/minimax-m2.5-lightning": { - "id": "minimax/minimax-m2.5-lightning", - "name": "MiniMax M2.5 highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "attachment": false, + "runware/kimi-k2.6": { + "id": "runware/kimi-k2.6", + "name": "Kimi K2.6 (Runware)", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, + "context": 262144, "output": 131072 }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" - }, "cost": { "input": 0.6, - "output": 4.8, - "cache_read": 0.06, - "cache_write": 0.75 + "output": 3.05, + "cache_read": 0.13 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "runware/glm-5.2": { + "id": "runware/glm-5.2", + "name": "GLM-5.2 (Runware)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -53898,43 +55770,47 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1024000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.38 + "input": 0.8, + "output": 2.55, + "cache_read": 0.16 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, + "runware/deepseek-v4-pro": { + "id": "runware/deepseek-v4-pro", + "name": "DeepSeek V4 Pro (Runware)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -53942,30 +55818,37 @@ }, "open_weights": true, "limit": { - "context": 512000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0.6, - "output": 2.4 + "input": 0.961, + "output": 1.922, + "cache_read": 0.079 } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "runware/gpt-oss-120b": { + "id": "runware/gpt-oss-120b", + "name": "GPT OSS 120B (Runware)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -53974,34 +55857,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.38 + "input": 0.032, + "output": 0.14, + "cache_read": 0.032 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": true, + "runware/glm-5.3": { + "id": "runware/glm-5.3", + "name": "GLM-5.3 (Runware)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -54010,35 +55898,20 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 131070 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.3055, - "output": 1.2219 + "input": 1.2, + "output": 4, + "cache_read": 0.2 } - } - } - }, - "kenari": { - "id": "kenari", - "env": [ - "KENARI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://kenari.id/v1", - "name": "Kenari", - "doc": "https://kenari.id/docs", - "models": { - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", + }, + "runware/deepseek-v4-flash": { + "id": "runware/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (Runware)", "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", "family": "deepseek-flash", "attachment": false, @@ -54047,12 +55920,17 @@ { "type": "effort", "values": [ + "none", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, "knowledge": "2025-05", @@ -54068,116 +55946,66 @@ }, "open_weights": true, "limit": { - "context": 1000000, + "context": 1048576, "output": 384000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.076, + "output": 0.153, + "cache_read": 0.014 } }, - "kimi-k2-6": { - "id": "kimi-k2-6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "tencent/glm-5v-turbo": { + "id": "tencent/glm-5v-turbo", + "name": "GLM-5V Turbo (Tencent Cloud)", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "glm-5-1": { - "id": "glm-5-1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gemma-4-31b-it": { - "id": "gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "grok-4-3": { - "id": "grok-4-3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "tencent/kimi-k3": { + "id": "tencent/kimi-k3", + "name": "Kimi K3 (Tencent Cloud)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -54185,78 +56013,113 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 30000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "qwen3-7-plus": { - "id": "qwen3-7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "tencent/glm-5": { + "id": "tencent/glm-5", + "name": "GLM-5 (Tencent Cloud)", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "kimi-k2-7-code": { - "id": "kimi-k2-7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "tencent/kimi-k2.7-code-highspeed": { + "id": "tencent/kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed (Tencent Cloud)", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, "knowledge": "2025-01", @@ -54278,35 +56141,46 @@ "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 1.9, + "output": 8, + "cache_read": 0.38 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "tencent/kimi-k2.6": { + "id": "tencent/kimi-k2.6", + "name": "Kimi K2.6 (Tencent Cloud)", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", + "low", + "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -54314,36 +56188,39 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.858, + "output": 3.566, + "cache_read": 0.145 } }, - "deepseek-v4-flash:free": { - "id": "deepseek-v4-flash:free", - "name": "DeepSeek V4 Flash (Free)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "tencent/minimax-m2.7": { + "id": "tencent/minimax-m2.7", + "name": "MiniMax M2.7 (Tencent Cloud)", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", + "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -54354,67 +56231,98 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "tencent/mimo-v2.5-pro": { + "id": "tencent/mimo-v2.5-pro", + "name": "MiMo V2.5 Pro (Tencent Cloud)", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", + "tencent/minimax-m3": { + "id": "tencent/minimax-m3", + "name": "MiniMax M3 (Tencent Cloud)", "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", "family": "minimax", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, "release_date": "2026-06-01", "last_updated": "2026-06-01", @@ -54430,25 +56338,28 @@ }, "open_weights": true, "limit": { - "context": 512000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "tencent/glm-5.2": { + "id": "tencent/glm-5.2", + "name": "GLM-5.2 (Tencent Cloud)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high", @@ -54458,42 +56369,45 @@ } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "gpt-5-4-mini": { - "id": "gpt-5-4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, + "tencent/hy3": { + "id": "tencent/hy3", + "name": "Hy3 (Tencent Cloud)", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", + "minimal", "low", "medium", "high", @@ -54503,69 +56417,63 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.132, + "output": 0.528, + "cache_read": 0.033 } }, - "mimo-v2-5": { - "id": "mimo-v2-5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "tencent/hy-mt2-plus": { + "id": "tencent/hy-mt2-plus", + "name": "Hy-MT2 Plus (Tencent Cloud)", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 8192, + "output": 4096 }, "cost": { - "input": 0, - "output": 0 + "input": 0.074, + "output": 0.295 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "tencent/glm-5.1": { + "id": "tencent/glm-5.1", + "name": "GLM-5.1 (Tencent Cloud)", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -54574,15 +56482,20 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -54593,44 +56506,42 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "gpt-5-5": { - "id": "gpt-5-5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "tencent/glm-5-turbo": { + "id": "tencent/glm-5-turbo", + "name": "GLM-5 Turbo (Tencent Cloud)", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "structured_output": false, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -54638,36 +56549,45 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 200000, "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "glm-5-2": { - "id": "glm-5-2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "tencent/deepseek-v4-pro": { + "id": "tencent/deepseek-v4-pro", + "name": "DeepSeek V4 Pro (Tencent Cloud)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", + "low", + "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -54679,110 +56599,140 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 131072 + "output": 393216 }, "cost": { - "input": 0, - "output": 0 + "input": 0.435, + "output": 0.87, + "cache_read": 0.00363 } }, - "grok-build-0-1": { - "id": "grok-build-0-1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, + "tencent/hy4-preview": { + "id": "tencent/hy4-preview", + "name": "Hy4 Preview (Tencent Cloud)", + "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", + "family": "Hy", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "release_date": "2026-08-28", + "last_updated": "2026-08-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.834, + "output": 2.501, + "cache_read": 0.042 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "tencent/kimi-k2.7-code": { + "id": "tencent/kimi-k2.7-code", + "name": "Kimi K2.7 Code (Tencent Cloud)", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "tencent/deepseek-v4-flash": { + "id": "tencent/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (Tencent Cloud)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { "input": [ "text" ], @@ -54792,56 +56742,73 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 393216 }, "cost": { - "input": 0, - "output": 0 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "mimo-v2-5-pro": { - "id": "mimo-v2-5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "meta-contributor/muse-spark-1.2-contributor": { + "id": "meta-contributor/muse-spark-1.2-contributor", + "name": "Muse Spark 1.2 Contributor (Meta Contributor)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "muse", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-08-06", + "last_updated": "2026-08-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 } }, - "deepseek-v4-pro:free": { - "id": "deepseek-v4-pro:free", - "name": "DeepSeek V4 Pro (Free)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "meta-contributor/muse-spark-1.3-contributor": { + "id": "meta-contributor/muse-spark-1.3-contributor", + "name": "Muse Spark 1.3 Contributor (Meta Contributor)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", + "low", + "medium", "high", "xhigh" ] @@ -54850,124 +56817,79 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-image-2": { - "id": "gpt-image-2", - "name": "GPT-Image-2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 272000, - "output": 16384 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 } - } - } - }, - "openai": { - "id": "openai", - "env": [ - "OPENAI_API_KEY" - ], - "npm": "@ai-sdk/openai", - "name": "OpenAI", - "doc": "https://platform.openai.com/docs/models", - "models": { - "o3": { - "id": "o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + }, + "novita/glm-4.6v": { + "id": "novita/glm-4.6v", + "name": "GLM-4.6V (NovitaAI)", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 16000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.3, + "output": 0.9, + "cache_read": 0.055 } }, - "text-embedding-3-large": { - "id": "text-embedding-3-large", - "name": "text-embedding-3-large", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "novita/gemma-4-31b-it": { + "id": "novita/gemma-4-31b-it", + "name": "Gemma 4 31B IT (NovitaAI)", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2024-01", - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" @@ -54976,43 +56898,77 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8191, - "output": 3072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.13, - "output": 0 + "input": 0.14, + "output": 0.4 } }, - "gpt-5.2-pro": { - "id": "gpt-5.2-pro", - "name": "GPT-5.2 Pro", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", + "novita/qwen3-vl-235b-a22b-instruct": { + "id": "novita/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct (NovitaAI)", + "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 1.5 + } + }, + "novita/qwen3.7-max": { + "id": "novita/qwen3.7-max", + "name": "Qwen3.7 Max (NovitaAI)", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", + "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -55020,20 +56976,20 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 21, - "output": 168 + "input": 1.25, + "output": 3.75, + "cache_read": 0.25 } }, - "gpt-5.6": { - "id": "gpt-5.6", - "name": "GPT-5.6", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt", + "novita/mimo-v2.5": { + "id": "novita/mimo-v2.5", + "name": "MiMo V2.5 (NovitaAI)", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -55043,250 +56999,260 @@ "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 10, - "output": 60, - "cache_read": 1, - "cache_write": 12.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } + "context": 1048576, + "output": 131072 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25, + "input": 0.168, + "output": 0.336, + "cache_read": 0.0034, "tiers": [ { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5, + "input": 0.8, + "output": 4, + "cache_read": 0.16, "tier": { "type": "context", - "size": 272000 + "size": 256000 } } ], "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.8, + "output": 4, + "cache_read": 0.16 } } }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "novita/qwen3.8-27b": { + "id": "novita/qwen3.8-27b", + "name": "Qwen3.8 27B (NovitaAI)", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": false, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.42, + "output": 3, + "cache_read": 0.085 } }, - "gpt-3.5-turbo": { - "id": "gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, + "novita/qwen3-vl-235b-a22b-thinking": { + "id": "novita/qwen3-vl-235b-a22b-thinking", + "name": "Qwen3 VL 235B A22B Thinking (NovitaAI)", + "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": false, "structured_output": false, "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16385, - "output": 4096 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0 + "input": 0.98, + "output": 3.95 } }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", - "attachment": true, + "novita/qwen3.8-2.4t-a95b": { + "id": "novita/qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B (NovitaAI)", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high" + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 272000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 15, - "output": 120 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "novita/glm-5.3-flash": { + "id": "novita/glm-5.3-flash", + "name": "GLM-5.3 Flash (NovitaAI)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "gpt-4": { - "id": "gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "novita/glm-4.6": { + "id": "novita/glm-4.6", + "name": "GLM-4.6 (NovitaAI)", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.55, + "output": 2.2, + "cache_read": 0.11 + } + }, + "novita/qwen3-235b-a22b-instruct-2507": { + "id": "novita/qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507 (NovitaAI)", + "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": false, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2025-07-21", + "last_updated": "2025-07-21", "modalities": { "input": [ "text" @@ -55295,21 +57261,21 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 131072, + "output": 16384 }, "cost": { - "input": 30, - "output": 60 + "input": 0.09, + "output": 0.58 } }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "novita/qwen3.8-max": { + "id": "novita/qwen3.8-max", + "name": "Qwen3.8 Max (NovitaAI)", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -55318,20 +57284,24 @@ "values": [ "low", "medium", - "high" + "xhigh" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" @@ -55339,98 +57309,75 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "o3-pro": { - "id": "o3-pro", - "name": "o3-pro", - "description": "High-effort o3 tier for difficult technical reasoning and careful answers", - "family": "o-pro", + "novita/kimi-k3": { + "id": "novita/kimi-k3", + "name": "Kimi K3 (NovitaAI)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 20, - "output": 80 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "chatgpt-image-latest": { - "id": "chatgpt-image-latest", - "name": "chatgpt-image-latest", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "novita/llama-4-maverick-17b-instruct": { + "id": "novita/llama-4-maverick-17b-instruct", + "name": "Llama 4 Maverick 17B Instruct (NovitaAI)", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", "attachment": true, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "input": 0, - "output": 0 - } - }, - "gpt-4o-2024-05-13": { - "id": "gpt-4o-2024-05-13", - "name": "GPT-4o (2024-05-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -55440,76 +57387,91 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 8192 }, "cost": { - "input": 5, - "output": 15 + "input": 0.27, + "output": 0.85 } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, + "novita/glm-5": { + "id": "novita/glm-5", + "name": "GLM-5 (NovitaAI)", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", + "minimal", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 202800, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "gpt-5-chat-latest": { - "id": "gpt-5-chat-latest", - "name": "GPT-5 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "novita/deepseek-v4.1-flash": { + "id": "novita/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (NovitaAI)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -55519,24 +57481,23 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.3, + "output": 1.2, + "cache_read": 0.006 } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", - "attachment": true, + "novita/minimax-m2.5": { + "id": "novita/minimax-m2.5", + "name": "MiniMax M2.5 (NovitaAI)", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -55544,42 +57505,41 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 204800, + "output": 131100 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "gpt-5.3-codex-spark": { - "id": "gpt-5.3-codex-spark", - "name": "GPT-5.3 Codex Spark", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex-spark", + "novita/kimi-k2.6": { + "id": "novita/kimi-k2.6", + "name": "Kimi K2.6 (NovitaAI)", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -55587,24 +57547,61 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "pdf" + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.8, + "output": 3.4, + "cache_read": 0.16 + } + }, + "novita/qwen3-max": { + "id": "novita/qwen3-max", + "name": "Qwen3 Max (NovitaAI)", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -55612,74 +57609,66 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 100000, - "output": 32000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.845, + "output": 3.38 } }, - "gpt-5.1-codex-max": { - "id": "gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "novita/glm-4.5v": { + "id": "novita/glm-4.5v", + "name": "GLM-4.5V (NovitaAI)", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 65536, + "output": 16000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.6, + "output": 1.8, + "cache_read": 0.11 } }, - "gpt-5.3-chat-latest": { - "id": "gpt-5.3-chat-latest", - "name": "GPT-5.3 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "novita/llama-4-scout-17b-instruct": { + "id": "novita/llama-4-scout-17b-instruct", + "name": "Llama 4 Scout 17B Instruct (NovitaAI)", + "description": "Open Llama with long-context vision for efficient multimodal agents", + "family": "llama", "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -55689,34 +57678,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.18, + "output": 0.59 } }, - "gpt-4o-2024-08-06": { - "id": "gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "novita/qwen3-235b-a22b-fp8": { + "id": "novita/qwen3-235b-a22b-fp8", + "name": "Qwen3 235B A22B FP8 (NovitaAI)", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-08-06", - "last_updated": "2024-08-06", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -55724,27 +57709,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 40960, + "output": 20000 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.2, + "output": 0.8 } }, - "text-embedding-ada-002": { - "id": "text-embedding-ada-002", - "name": "text-embedding-ada-002", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "novita/qwen3-coder-480b-a35b-instruct": { + "id": "novita/qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct (NovitaAI)", + "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2022-12", - "release_date": "2022-12-15", - "last_updated": "2022-12-15", + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -55753,39 +57738,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0 + "input": 0.38, + "output": 1.55 } }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "novita/qwen3-coder-30b-a3b-instruct": { + "id": "novita/qwen3-coder-30b-a3b-instruct", + "name": "Qwen3 Coder 30B A3B Instruct (NovitaAI)", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -55794,46 +57769,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 160000, + "output": 32768 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.07, + "output": 0.27 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, + "novita/ling-3.0-flash": { + "id": "novita/ling-3.0-flash", + "name": "InclusionAI Ling 3.0 Flash (NovitaAI)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", - "medium", - "high", - "xhigh" + "none" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "structured_output": false, + "temperature": true, + "release_date": "2026-08-02", + "last_updated": "2026-08-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -55841,46 +57809,30 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.06, + "output": 0.18, + "cache_read": 0.012 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "novita/qwen3-vl-30b-a3b-instruct": { + "id": "novita/qwen3-vl-30b-a3b-instruct", + "name": "Qwen3 VL 30B A3B Instruct (NovitaAI)", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "structured_output": false, + "temperature": true, + "release_date": "2025-10-05", + "last_updated": "2025-10-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -55888,28 +57840,56 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.2, + "output": 0.7 } }, - "text-embedding-3-small": { - "id": "text-embedding-3-small", - "name": "text-embedding-3-small", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "novita/gemma-4-26b-a4b-it": { + "id": "novita/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT (NovitaAI)", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.13, + "output": 0.4 + } + }, + "novita/llama-3-70b-instruct": { + "id": "novita/llama-3-70b-instruct", + "name": "Llama 3 70B Instruct (NovitaAI)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "knowledge": "2024-01", - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "structured_output": false, + "temperature": true, + "release_date": "2024-04-18", + "last_updated": "2024-04-18", "modalities": { "input": [ "text" @@ -55920,19 +57900,79 @@ }, "open_weights": false, "limit": { - "context": 8191, - "output": 1536 + "context": 8192, + "output": 8000 }, "cost": { - "input": 0.02, - "output": 0 + "input": 0.51, + "output": 0.74 } }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-nano", + "novita/qwen3-235b-a22b-thinking-2507": { + "id": "novita/qwen3-235b-a22b-thinking-2507", + "name": "Qwen3 235B A22B Thinking 2507 (NovitaAI)", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-07-25", + "last_updated": "2025-07-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 3 + } + }, + "novita/llama-3.3-70b-instruct": { + "id": "novita/llama-3.3-70b-instruct", + "name": "Llama 3.3 70B Instruct (NovitaAI)", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 120000 + }, + "cost": { + "input": 0.135, + "output": 0.4 + } + }, + "novita/qwen3.6-35b-a3b": { + "id": "novita/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B A3B (NovitaAI)", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -55940,6 +57980,7 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", "high", @@ -55949,84 +57990,68 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "structured_output": false, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } + "context": 262144, + "output": 64000 }, "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1, - "cache_write": 1.25, - "tiers": [ - { - "input": 2, - "output": 9, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 272000 - } - } + "input": 0.248, + "output": 1.485 + } + }, + "novita/qwen3-next-80b-a3b-instruct": { + "id": "novita/qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct (NovitaAI)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 2, - "output": 9, - "cache_read": 0.2, - "cache_write": 2.5 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 1.5 } }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "novita/minimax-m2.7": { + "id": "novita/minimax-m2.7", + "name": "MiniMax M2.7 (NovitaAI)", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -56034,155 +58059,167 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": false, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 204800, + "output": 131100 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-mini", - "attachment": true, + "novita/mimo-v2.5-pro": { + "id": "novita/mimo-v2.5-pro", + "name": "MiMo V2.5 Pro (NovitaAI)", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125, + "input": 0.522, + "output": 1.044, + "cache_read": 0.0043, "tiers": [ { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25, + "input": 2, + "output": 6, + "cache_read": 0.4, "tier": { "type": "context", - "size": 272000 + "size": 256000 } } ], "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2, + "output": 6, + "cache_read": 0.4 } } }, - "gpt-5.1-chat-latest": { - "id": "gpt-5.1-chat-latest", - "name": "GPT-5.1 Chat", - "description": "Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations", - "family": "gpt-codex", - "attachment": true, + "novita/glm-4.7": { + "id": "novita/glm-4.7", + "name": "GLM-4.7 (NovitaAI)", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 128000 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 + } + }, + "novita/qwen3.8-flash": { + "id": "novita/qwen3.8-flash", + "name": "Qwen3.8 Flash (NovitaAI)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "medium" + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -56190,36 +58227,41 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.15, + "output": 0.47, + "cache_read": 0.016 } }, - "gpt-5.2-chat-latest": { - "id": "gpt-5.2-chat-latest", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "novita/qwen35-397b-a17b": { + "id": "novita/qwen35-397b-a17b", + "name": "Qwen3.5 397B A17B (NovitaAI)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.5", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "medium" + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", @@ -56231,133 +58273,151 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 64000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.6, + "output": 3.6 } }, - "o4-mini-deep-research": { - "id": "o4-mini-deep-research", - "name": "o4-mini-deep-research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "family": "o-mini", - "attachment": true, + "novita/glm-5.2": { + "id": "novita/glm-5.2", + "name": "GLM-5.2 (NovitaAI)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "medium" + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-06-26", - "last_updated": "2024-06-26", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "gpt-image-1.5": { - "id": "gpt-image-1.5", - "name": "gpt-image-1.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "novita/ernie-4.5-vl-424b-a47b": { + "id": "novita/ernie-4.5-vl-424b-a47b", + "name": "ERNIE 4.5 VL 424B A47B (NovitaAI)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ernie", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "temperature": false, - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "structured_output": false, + "temperature": true, + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "input": 0, - "output": 0 + "context": 123000, + "output": 16000 + }, + "cost": { + "input": 0.42, + "output": 1.25 } }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, + "novita/hy3": { + "id": "novita/hy3", + "name": "Hy3 (NovitaAI)", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "input": 192000, + "output": 262144 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 0.14, + "output": 0.58, + "cache_read": 0.035 } }, - "gpt-4o-2024-11-20": { - "id": "gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "novita/kimi-k2": { + "id": "novita/kimi-k2", + "name": "Kimi K2 (NovitaAI)", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", + "release_date": "2025-07-11", + "last_updated": "2025-07-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -56365,86 +58425,123 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.57, + "output": 2.3 } }, - "o1": { - "id": "o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, + "novita/glm-5.1": { + "id": "novita/glm-5.1", + "name": "GLM-5.1 (NovitaAI)", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 1.38, + "output": 4.4, + "cache_read": 0.26 } }, - "o1-pro": { - "id": "o1-pro", - "name": "o1-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", + "novita/step-3.7-flash": { + "id": "novita/step-3.7-flash", + "name": "Step 3.7 Flash (NovitaAI)", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2025-03-19", - "last_updated": "2025-03-19", + "structured_output": false, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", - "image" + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 256000 + }, + "cost": { + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 + } + }, + "novita/llama-3.2-3b-instruct": { + "id": "novita/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct (NovitaAI)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2024-09-18", + "last_updated": "2024-09-18", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -56452,467 +58549,584 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 32768, + "output": 32000 }, "cost": { - "input": 150, - "output": 600 + "input": 0.03, + "output": 0.05 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, + "novita/deepseek-v3.2": { + "id": "novita/deepseek-v3.2", + "name": "DeepSeek V3.2 (NovitaAI)", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 163840, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0.269, + "output": 0.4, + "cache_read": 0.1345 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, + "novita/glm-5.3": { + "id": "novita/glm-5.3", + "name": "GLM-5.3 (NovitaAI)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "structured_output": false, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", + "novita/kimi-k2.7-code": { + "id": "novita/kimi-k2.7-code", + "name": "Kimi K2.7 Code (NovitaAI)", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "o3-deep-research": { - "id": "o3-deep-research", - "name": "o3-deep-research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "family": "o", - "attachment": true, + "novita/minimax-m2.1": { + "id": "novita/minimax-m2.1", + "name": "MiniMax M2.1 (NovitaAI)", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "medium" + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-06-26", - "last_updated": "2024-06-26", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 10, - "output": 40, - "cache_read": 2.5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, + "novita/deepseek-v4-flash": { + "id": "novita/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (NovitaAI)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1050000, + "output": 393216 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "gpt-image-1": { - "id": "gpt-image-1", - "name": "gpt-image-1", - "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", - "family": "gpt-image", - "attachment": true, + "embercloud/glm-4.5": { + "id": "embercloud/glm-4.5", + "name": "GLM-4.5 (EmberCloud)", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-04-24", - "last_updated": "2025-04-24", + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "input": 0, - "output": 0 + "context": 131000, + "output": 96000 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, + "embercloud/qwen3-coder-next": { + "id": "embercloud/qwen3-coder-next", + "name": "Qwen3 Coder Next (EmberCloud)", + "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.108, + "output": 0.675, + "cache_read": 0.06 } }, - "gpt-4-turbo": { - "id": "gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, + "embercloud/glm-5": { + "id": "embercloud/glm-5", + "name": "GLM-5 (EmberCloud)", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 203000, + "output": 131000 }, "cost": { - "input": 10, - "output": 30 + "input": 0.72, + "output": 2.3, + "cache_read": 0.144 } }, - "gpt-image-1-mini": { - "id": "gpt-image-1-mini", - "name": "gpt-image-1-mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, + "embercloud/kimi-k2.5": { + "id": "embercloud/kimi-k2.5", + "name": "Kimi K2.5 (EmberCloud)", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": false, "temperature": false, - "release_date": "2025-09-26", - "last_updated": "2025-09-26", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "input": 0, - "output": 0 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.405, + "output": 1.98, + "cache_read": 0.225 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, + "embercloud/glm-4.7-flash": { + "id": "embercloud/glm-4.7-flash", + "name": "GLM-4.7 Flash (EmberCloud)", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131000 + }, + "cost": { + "input": 0.06, + "output": 0.4, + "cache_read": 0.01 + } + }, + "embercloud/glm-4.7": { + "id": "embercloud/glm-4.7", + "name": "GLM-4.7 (EmberCloud)", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131000 + }, + "cost": { + "input": 0.38, + "output": 1.98, + "cache_read": 0.19 + } + }, + "embercloud/glm-5.2": { + "id": "embercloud/glm-5.2", + "name": "GLM-5.2 (EmberCloud)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 203000, + "output": 131000 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 1.26, + "output": 3.96, + "cache_read": 0.234 } }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", - "attachment": true, + "embercloud/glm-5.1": { + "id": "embercloud/glm-5.1", + "name": "GLM-5.1 (EmberCloud)", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image" + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 203000, + "output": 131000 + }, + "cost": { + "input": 0.931, + "output": 2.93, + "cache_read": 0.173 + } + }, + "embercloud/glm-4.5-air": { + "id": "embercloud/glm-4.5-air", + "name": "GLM-4.5 Air (EmberCloud)", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131000, + "output": 96000 + }, + "cost": { + "input": 0.13, + "output": 0.85, + "cache_read": 0.025 + } + }, + "perplexity/sonar-pro": { + "id": "perplexity/sonar-pro", + "name": "Sonar Pro (Perplexity)", + "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", + "family": "sonar-pro", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -56920,57 +59134,102 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 8192 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } + "input": 3, + "output": 15 + } + }, + "perplexity/sonar": { + "id": "perplexity/sonar", + "name": "Sonar (Perplexity)", + "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", + "family": "sonar", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 130000, + "output": 4096 + }, + "cost": { + "input": 1, + "output": 1 } }, - "gpt-5.5-pro": { - "id": "gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "perplexity/sonar-reasoning-pro": { + "id": "perplexity/sonar-reasoning-pro", + "name": "Sonar Reasoning Pro (Perplexity)", + "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", + "family": "sonar-reasoning", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 2, + "output": 8 + } + }, + "azure-ai-foundry/grok-4-1-fast-reasoning": { + "id": "azure-ai-foundry/grok-4-1-fast-reasoning", + "name": "Grok 4.1 Fast Reasoning (Azure AI Foundry)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "structured_output": false, + "temperature": true, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -56978,47 +59237,30 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 0.2, + "output": 0.5 } }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", + "azure-ai-foundry/grok-4-1-fast-non-reasoning": { + "id": "azure-ai-foundry/grok-4-1-fast-non-reasoning", + "name": "Grok 4.1 Fast Non-Reasoning (Azure AI Foundry)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -57026,21 +59268,20 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.2, + "output": 0.5 } }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt", - "attachment": true, + "azure-ai-foundry/grok-4-3": { + "id": "azure-ai-foundry/grok-4-3", + "name": "Grok 4.3 (Azure AI Foundry)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -57049,23 +59290,18 @@ "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "structured_output": false, + "temperature": true, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -57073,88 +59309,62 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 10, - "output": 60, - "cache_read": 1, - "cache_write": 12.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } + "context": 20000, + "output": 8192 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25, + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, "tiers": [ { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5, + "input": 2.5, + "output": 5, + "cache_read": 0.4, "tier": { "type": "context", - "size": 272000 + "size": 200000 } } ], "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5 + "input": 2.5, + "output": 5, + "cache_read": 0.4 } } }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "meta/muse-spark-1.3": { + "id": "meta/muse-spark-1.3", + "name": "Muse Spark 1.3 (Meta)", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf", + "audio" ], "output": [ "text" @@ -57162,27 +59372,27 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 1048576 }, "cost": { "input": 1.25, - "output": 10, - "cache_read": 0.125 + "output": 4.25, + "cache_read": 0.15 } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", + "meta/muse-spark-1.1": { + "id": "meta/muse-spark-1.1", + "name": "Muse Spark 1.1 (Meta)", + "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high", @@ -57192,15 +59402,15 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-04-08", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "pdf" + "pdf", + "video" ], "output": [ "text" @@ -57208,76 +59418,46 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "gpt-image-2": { - "id": "gpt-image-2", - "name": "gpt-image-2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "input": 0, - "output": 0 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 1.25 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", + "meta/muse-spark-1.2": { + "id": "meta/muse-spark-1.2", + "name": "Muse Spark 1.2 (Meta)", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf", + "audio" ], "output": [ "text" @@ -57285,20 +59465,19 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { "input": 1.25, - "output": 10, - "cache_read": 0.125 + "output": 4.25, + "cache_read": 0.15 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "azure/gpt-5.4": { + "id": "azure/gpt-5.4", + "name": "GPT-5.4 (Azure)", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", "family": "gpt", "attachment": true, "reasoning": true, @@ -57316,10 +59495,10 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -57336,73 +59515,39 @@ "input": 922000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", + "azure/gpt-5.4-pro": { + "id": "azure/gpt-5.4-pro", + "name": "GPT-5.4 Pro (Azure)", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": false, - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -57414,83 +59559,24 @@ "input": 922000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 20, - "output": 100, - "cache_read": 2, - "cache_write": 25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } - }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } + "input": 30, + "output": 180 } - } - } - }, - "berget": { - "id": "berget", - "env": [ - "BERGET_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.berget.ai/v1", - "name": "Berget.AI", - "doc": "https://api.berget.ai", - "models": { - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + }, + "azure/gpt-3.5-turbo": { + "id": "azure/gpt-3.5-turbo", + "name": "GPT-3.5 Turbo (Azure)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-04-27", - "last_updated": "2025-04-27", + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" @@ -57499,99 +59585,114 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 16385, + "output": 4096 }, "cost": { - "input": 0.99, - "output": 0.99 + "input": 0.5, + "output": 1.5 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "azure/gpt-5.4-nano": { + "id": "azure/gpt-5.4-nano", + "name": "GPT-5.4 Nano (Azure)", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.83, - "output": 3.85, - "cache_read": 0.16 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "azure/gpt-5.2-codex": { + "id": "azure/gpt-5.2-codex", + "name": "GPT-5.2 Codex (Azure)", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "audio", - "image", "text", - "video" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.275, - "output": 0.55 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT-OSS-120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "azure/gpt-5.1-codex": { + "id": "azure/gpt-5.1-codex", + "name": "GPT-5.1 Codex (Azure)", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -57604,215 +59705,212 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": false, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 272000 }, "cost": { - "input": 0.22, - "output": 0.83 + "input": 1.25, + "output": 10 } }, - "mistralai/Mistral-Medium-3.5-128B": { - "id": "mistralai/Mistral-Medium-3.5-128B", - "name": "Mistral Medium 3.5 128B", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "azure/gpt-4o": { + "id": "azure/gpt-4o", + "name": "GPT-4o (Azure)", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2026-04", - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.65, - "output": 5.5 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "mistralai/Mistral-Small-3.2-24B-Instruct-2506": { - "id": "mistralai/Mistral-Small-3.2-24B-Instruct-2506", - "name": "Mistral Small 3.2 24B Instruct 2506", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, + "azure/gpt-5-mini": { + "id": "azure/gpt-5-mini", + "name": "GPT-5 Mini (Azure)", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-10-01", - "last_updated": "2025-10-01", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.33, - "output": 0.33 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "zai-org/GLM-4.7": { - "id": "zai-org/GLM-4.7", - "name": "GLM 4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "azure/gpt-5.2-pro": { + "id": "azure/gpt-5.2-pro", + "name": "GPT-5.2 Pro (Azure)", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 272000 }, "cost": { - "input": 0.77, - "output": 2.75 + "input": 21, + "output": 168 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "azure/o4-mini": { + "id": "azure/o4-mini", + "name": "o4 Mini (Azure)", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.54, - "output": 4.84 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } - } - } - }, - "snowflake-cortex": { - "id": "snowflake-cortex", - "env": [ - "SNOWFLAKE_ACCOUNT", - "SNOWFLAKE_CORTEX_PAT" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://${SNOWFLAKE_ACCOUNT}.snowflakecomputing.com/api/v2/cortex/v1", - "name": "Snowflake Cortex", - "doc": "https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-rest-api", - "models": { - "openai-gpt-5.1": { - "id": "openai-gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + }, + "azure/o3-mini": { + "id": "azure/o3-mini", + "name": "o3 Mini (Azure)", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -57820,23 +59918,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "snowflake-llama3.3-70b": { - "id": "snowflake-llama3.3-70b", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": true, + "azure/gpt-4": { + "id": "azure/gpt-4", + "name": "GPT-4 (Azure)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ "text" @@ -57845,41 +59948,45 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 30, + "output": 60 } }, - "openai-gpt-5.2": { - "id": "openai-gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "azure/gpt-5.3-codex": { + "id": "azure/gpt-5.3-codex", + "name": "GPT-5.3 Codex (Azure)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, + "structured_output": false, + "temperature": true, "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -57890,30 +59997,30 @@ "context": 400000, "input": 272000, "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "azure/gpt-4.1-nano": { + "id": "azure/gpt-4.1-nano", + "name": "GPT-4.1 Nano (Azure)", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -57921,28 +60028,42 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 16384 + "context": 1000000, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "azure/gpt-5-nano": { + "id": "azure/gpt-5-nano", + "name": "GPT-5 Nano (Azure)", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -57950,71 +60071,85 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, - "status": "beta", - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek-R1", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", - "attachment": false, + "azure/o1": { + "id": "azure/o1", + "name": "o1 (Azure)", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "azure/gpt-6-astra": { + "id": "azure/gpt-6-astra", + "name": "GPT-6 Astra (Azure)", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ "text", @@ -58027,14 +60162,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "openai-gpt-5": { - "id": "openai-gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "azure/gpt-5.1": { + "id": "azure/gpt-5.1", + "name": "GPT-5.1 (Azure)", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", "family": "gpt", "attachment": true, "reasoning": true, @@ -58046,16 +60188,17 @@ "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -58071,13 +60214,17 @@ "input": 272000, "output": 128000 }, - "status": "beta" + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } }, - "openai-gpt-5.5": { - "id": "openai-gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "azure/gpt-5.4-mini": { + "id": "azure/gpt-5.4-mini", + "name": "GPT-5.4 Mini (Azure)", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -58085,24 +60232,23 @@ "type": "effort", "values": [ "none", - "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -58110,25 +60256,42 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, + "input": 272000, "output": 128000 }, - "status": "beta" + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "azure/gpt-5.6-luna": { + "id": "azure/gpt-5.6-luna", + "name": "GPT-5.6 Luna (Azure)", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -58141,15 +60304,22 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 } }, - "openai-gpt-5-nano": { - "id": "openai-gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "azure/gpt-5.2": { + "id": "azure/gpt-5.2", + "name": "GPT-5.2 (Azure)", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -58157,19 +60327,19 @@ "type": "effort", "values": [ "none", - "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -58185,25 +60355,37 @@ "input": 272000, "output": 128000 }, - "status": "beta" + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "azure/gpt-5.5": { + "id": "azure/gpt-5.5", + "name": "GPT-5.5 (Azure)", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -58216,40 +60398,35 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 16384 - } - }, - "mistral-large2": { - "id": "mistral-large2", - "name": "Mistral Large (latest)", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] + "context": 1050000, + "input": 922000, + "output": 128000 }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "openai-gpt-4.1": { - "id": "openai-gpt-4.1", - "name": "GPT-4.1", + "azure/gpt-4.1": { + "id": "azure/gpt-4.1", + "name": "GPT-4.1 (Azure)", "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", "family": "gpt", "attachment": true, @@ -58272,27 +60449,28 @@ }, "open_weights": false, "limit": { - "context": 1047576, + "context": 1000000, "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "azure/gpt-4.1-mini": { + "id": "azure/gpt-4.1-mini", + "name": "GPT-4.1 Mini (Azure)", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -58306,14 +60484,19 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 16384 + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "openai-gpt-5.4": { - "id": "openai-gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "azure/gpt-5.6-terra": { + "id": "azure/gpt-5.6-terra", + "name": "GPT-5.6 Terra (Azure)", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -58321,19 +60504,20 @@ "type": "effort", "values": [ "none", - "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -58350,45 +60534,30 @@ "input": 922000, "output": 128000 }, - "status": "beta", - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "gemini-3.1-pro": { - "id": "gemini-3.1-pro", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "azure/gpt-4-turbo": { + "id": "azure/gpt-4-turbo", + "name": "GPT-4 Turbo (Azure)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -58396,23 +60565,25 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 10, + "output": 30 } }, - "openai-gpt-5-mini": { - "id": "openai-gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, + "azure/gpt-oss-120b": { + "id": "azure/gpt-oss-120b", + "name": "GPT OSS 120B (Azure)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", "high" @@ -58420,51 +60591,36 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 272000, - "input": 272000, - "output": 8192 + "context": 131072, + "output": 32768 }, - "status": "beta" - } - } - }, - "tencent-token-plan": { - "id": "tencent-token-plan", - "env": [ - "TENCENT_TOKEN_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.lkeap.cloud.tencent.com/plan/v3", - "name": "Tencent Token Plan", - "doc": "https://cloud.tencent.com/document/product/1823/130060", - "models": { - "hy3": { - "id": "hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "azure/gpt-5.1-codex-mini": { + "id": "azure/gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini (Azure)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -58475,57 +60631,50 @@ } ], "tool_call": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "structured_output": false, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } - } - } - }, - "github-models": { - "id": "github-models", - "env": [ - "GITHUB_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://models.github.ai/inference", - "name": "GitHub Models", - "doc": "https://docs.github.com/en/github-models", - "models": { - "ai21-labs/ai21-jamba-1.5-mini": { - "id": "ai21-labs/ai21-jamba-1.5-mini", - "name": "AI21 Jamba 1.5 Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "jamba", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-08-29", - "last_updated": "2024-08-29", + }, + "azure/o3": { + "id": "azure/o3", + "name": "o3 (Azure)", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -58533,30 +60682,43 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 4096 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "ai21-labs/ai21-jamba-1.5-large": { - "id": "ai21-labs/ai21-jamba-1.5-large", - "name": "AI21 Jamba 1.5 Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "jamba", - "attachment": false, + "azure/gpt-5": { + "id": "azure/gpt-5", + "name": "GPT-5 (Azure)", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-08-29", - "last_updated": "2024-08-29", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -58564,58 +60726,78 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "core42/jais-30b-chat": { - "id": "core42/jais-30b-chat", - "name": "JAIS 30b Chat", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "jais", - "attachment": false, + "azure/gpt-5.6-sol": { + "id": "azure/gpt-5.6-sol", + "name": "GPT-5.6 Sol (Azure)", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-03", - "release_date": "2023-08-30", - "last_updated": "2023-08-30", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 2048 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5 } }, - "xai/grok-3-mini": { - "id": "xai/grok-3-mini", - "name": "Grok 3 Mini", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "alibaba/qwen-flash": { + "id": "alibaba/qwen-flash", + "name": "Qwen Flash (Alibaba Cloud)", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-09", - "last_updated": "2024-12-09", + "knowledge": "2024-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -58626,27 +60808,42 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.05, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0.0625 } }, - "xai/grok-3": { - "id": "xai/grok-3", - "name": "Grok 3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "alibaba/qwen3.7-max": { + "id": "alibaba/qwen3.7-max", + "name": "Qwen3.7 Max (Alibaba Cloud)", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-09", - "last_updated": "2024-12-09", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -58657,27 +60854,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "microsoft/phi-3.5-moe-instruct": { - "id": "microsoft/phi-3.5-moe-instruct", - "name": "Phi-3.5-MoE instruct (128k)", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", + "alibaba/qwen-coder-plus": { + "id": "alibaba/qwen-coder-plus", + "name": "Qwen Coder Plus (Alibaba Cloud)", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-08-20", - "last_updated": "2024-08-20", + "release_date": "2024-09-18", + "last_updated": "2024-09-18", "modalities": { "input": [ "text" @@ -58686,60 +60884,63 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.502, + "output": 1.004 } }, - "microsoft/phi-3-small-128k-instruct": { - "id": "microsoft/phi-3-small-128k-instruct", - "name": "Phi-3-small instruct (128k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "alibaba/qwen3-vl-plus": { + "id": "alibaba/qwen3-vl-plus", + "name": "Qwen3 VL Plus (Alibaba Cloud)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 1.6, + "cache_read": 0.04, + "cache_write": 0.25 } }, - "microsoft/phi-3.5-mini-instruct": { - "id": "microsoft/phi-3.5-mini-instruct", - "name": "Phi-3.5-mini instruct (128k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "alibaba/qwen-max": { + "id": "alibaba/qwen-max", + "name": "Qwen Max (Alibaba Cloud)", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-08-20", - "last_updated": "2024-08-20", + "knowledge": "2024-04", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", "modalities": { "input": [ "text" @@ -58748,63 +60949,95 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 1.6, + "output": 6.4 } }, - "microsoft/phi-3-medium-128k-instruct": { - "id": "microsoft/phi-3-medium-128k-instruct", - "name": "Phi-3-medium instruct (128k)", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", - "attachment": false, + "alibaba/qwen3.8-max": { + "id": "alibaba/qwen3.8-max", + "name": "Qwen3.8 Max (Alibaba Cloud)", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "microsoft/phi-3-small-8k-instruct": { - "id": "microsoft/phi-3-small-8k-instruct", - "name": "Phi-3-small instruct (8k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, + "alibaba/kimi-k3": { + "id": "alibaba/kimi-k3", + "name": "Kimi K3 (Alibaba Cloud)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -58812,27 +61045,43 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 2048 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "microsoft/phi-4-reasoning": { - "id": "microsoft/phi-4-reasoning", - "name": "Phi-4-Reasoning", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", + "alibaba/glm-5": { + "id": "alibaba/glm-5", + "name": "GLM-5 (Alibaba Cloud)", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -58843,58 +61092,74 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 202752, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.573, + "output": 2.58 } }, - "microsoft/mai-ds-r1": { - "id": "microsoft/mai-ds-r1", - "name": "MAI-DS-R1", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mai", - "attachment": false, + "alibaba/deepseek-v4.1-flash": { + "id": "alibaba/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (Alibaba Cloud)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 8192 + "context": 1000000, + "output": 393216 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "microsoft/phi-4-mini-instruct": { - "id": "microsoft/phi-4-mini-instruct", - "name": "Phi-4-mini-instruct", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "alibaba/qwen-plus": { + "id": "alibaba/qwen-plus", + "name": "Qwen Plus (Alibaba Cloud)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", "modalities": { "input": [ "text" @@ -58903,128 +61168,167 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 131072, + "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 1.2, + "cache_read": 0.08, + "cache_write": 0.5 } }, - "microsoft/phi-4": { - "id": "microsoft/phi-4", - "name": "Phi-4", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", - "attachment": false, + "alibaba/qwen3.7-flash": { + "id": "alibaba/qwen3.7-flash", + "name": "Qwen3.7 Flash (Alibaba Cloud)", + "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16000, - "output": 4096 + "context": 1000000, + "input": 991000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.03, + "output": 0.13, + "cache_read": 0.006, + "cache_write": 0.0375 } }, - "microsoft/phi-3.5-vision-instruct": { - "id": "microsoft/phi-3.5-vision-instruct", - "name": "Phi-3.5-vision instruct (128k)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "phi", - "attachment": false, + "alibaba/qwen3-max": { + "id": "alibaba/qwen3-max", + "name": "Qwen3 Max (Alibaba Cloud)", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-08-20", - "last_updated": "2024-08-20", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 256000, + "output": 32800 }, "cost": { - "input": 0, - "output": 0 + "input": 1.2, + "output": 6, + "cache_read": 0.24, + "cache_write": 1.5 } }, - "microsoft/phi-3-medium-4k-instruct": { - "id": "microsoft/phi-3-medium-4k-instruct", - "name": "Phi-3-medium instruct (4k)", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "alibaba/qwen3-vl-flash": { + "id": "alibaba/qwen3-vl-flash", + "name": "Qwen3 VL Flash (Alibaba Cloud)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 4096, - "output": 1024 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.05, + "output": 0.4, + "cache_read": 0.01 } }, - "microsoft/phi-4-multimodal-instruct": { - "id": "microsoft/phi-4-multimodal-instruct", - "name": "Phi-4-multimodal-instruct", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "phi", - "attachment": false, + "alibaba/kimi-k2.5": { + "id": "alibaba/kimi-k2.5", + "name": "Kimi K2.5 (Alibaba Cloud)", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "audio" + "video" ], "output": [ "text" @@ -59032,30 +61336,46 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 98304 }, "cost": { - "input": 0, - "output": 0 + "input": 0.574, + "output": 3.011 } }, - "microsoft/phi-4-mini-reasoning": { - "id": "microsoft/phi-4-mini-reasoning", - "name": "Phi-4-mini-reasoning", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, + "alibaba/qwen3.6-35b-a3b": { + "id": "alibaba/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B A3B (Alibaba Cloud)", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -59063,58 +61383,75 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.375, + "output": 2.25 } }, - "microsoft/phi-3-mini-128k-instruct": { - "id": "microsoft/phi-3-mini-128k-instruct", - "name": "Phi-3-mini instruct (128k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, + "alibaba/qwen3.6-flash": { + "id": "alibaba/qwen3.6-flash", + "name": "Qwen3.6 Flash (Alibaba Cloud)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.25, + "output": 1.5, + "cache_read": 0.05, + "cache_write": 0.3125 } }, - "microsoft/phi-3-mini-4k-instruct": { - "id": "microsoft/phi-3-mini-4k-instruct", - "name": "Phi-3-mini instruct (4k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "alibaba/qwen3-coder-flash": { + "id": "alibaba/qwen3-coder-flash", + "name": "Qwen3 Coder Flash (Alibaba Cloud)", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -59123,61 +61460,31 @@ "text" ] }, - "open_weights": true, - "limit": { - "context": 4096, - "output": 1024 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "openai/o3": { - "id": "openai/o3", - "name": "OpenAI o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": false, - "knowledge": "2024-04", - "release_date": "2025-01-31", - "last_updated": "2025-01-31", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.5, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "openai/o1-mini": { - "id": "openai/o1-mini", - "name": "OpenAI o1-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "alibaba/qwen3-coder-plus": { + "id": "alibaba/qwen3-coder-plus", + "name": "Qwen3 Coder Plus (Alibaba Cloud)", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": false, - "knowledge": "2023-10", - "release_date": "2024-09-12", - "last_updated": "2024-12-17", + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -59188,31 +61495,46 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 65536 + "context": 1000000, + "output": 66000 }, "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 5, + "cache_read": 0.2, + "cache_write": 1.25 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "alibaba/qwen3.8-flash": { + "id": "alibaba/qwen3.8-flash", + "name": "Qwen3.8 Flash (Alibaba Cloud)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", "image", - "audio" + "video" ], "output": [ "text" @@ -59220,27 +61542,42 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.47, + "cache_read": 0.016, + "cache_write": 0.2 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "OpenAI o4-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, + "alibaba/qwen35-397b-a17b": { + "id": "alibaba/qwen35-397b-a17b", + "name": "Qwen3.5 397B A17B (Alibaba Cloud)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.5", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": false, - "knowledge": "2024-04", - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", @@ -59252,27 +61589,28 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.6, + "output": 3.6 } }, - "openai/o1-preview": { - "id": "openai/o1-preview", - "name": "OpenAI o1-preview", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "alibaba/qwen3.6-max-preview": { + "id": "alibaba/qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview (Alibaba Cloud)", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "temperature": false, - "knowledge": "2023-10", - "release_date": "2024-09-12", - "last_updated": "2024-09-12", + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text" @@ -59283,27 +61621,44 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.3, + "output": 7.8, + "cache_read": 0.13 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "OpenAI o3-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "alibaba/glm-5.2": { + "id": "alibaba/glm-5.2", + "name": "GLM-5.2 (Alibaba Cloud)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": false, - "knowledge": "2024-04", - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -59312,95 +61667,133 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.28 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "alibaba/qwen-omni-turbo": { + "id": "alibaba/qwen-omni-turbo", + "name": "Qwen Omni Turbo (Alibaba Cloud)", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-01-19", + "last_updated": "2025-03-26", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 0.8 } }, - "openai/o1": { - "id": "openai/o1", - "name": "OpenAI o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "alibaba/deepseek-v4-pro": { + "id": "alibaba/deepseek-v4-pro", + "name": "DeepSeek V4 Pro (Alibaba Cloud)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": false, - "knowledge": "2023-10", - "release_date": "2024-09-12", - "last_updated": "2024-12-17", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 393216 }, "cost": { - "input": 0, - "output": 0 + "input": 2.4, + "output": 4.8, + "cache_read": 0.2 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "alibaba/qwen3.6-plus": { + "id": "alibaba/qwen3.6-plus", + "name": "Qwen3.6 Plus (Alibaba Cloud)", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -59408,30 +61801,49 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "alibaba/qwen-plus-latest": { + "id": "alibaba/qwen-plus-latest", + "name": "Qwen Plus Latest (Alibaba Cloud)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2024-09-09", + "last_updated": "2024-09-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -59439,63 +61851,89 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 1.2, + "cache_read": 0.08, + "cache_write": 0.5 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + "alibaba/glm-5.3": { + "id": "alibaba/glm-5.3", + "name": "GLM-5.3 (Alibaba Cloud)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.28 } }, - "mistral-ai/mistral-small-2503": { - "id": "mistral-ai/mistral-small-2503", - "name": "Mistral Small 3.1", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, + "alibaba/qwen3.7-plus": { + "id": "alibaba/qwen3.7-plus", + "name": "Qwen3.7 Plus (Alibaba Cloud)", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-09", - "release_date": "2025-03-01", - "last_updated": "2025-03-01", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -59503,27 +61941,46 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "cache_write": 0.5 } }, - "mistral-ai/mistral-nemo": { - "id": "mistral-ai/mistral-nemo", - "name": "Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + "alibaba/deepseek-v4-flash": { + "id": "alibaba/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (Alibaba Cloud)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -59534,93 +61991,145 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 393216 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 0.4, + "cache_read": 0.04 } }, - "mistral-ai/mistral-medium-2505": { - "id": "mistral-ai/mistral-medium-2505", - "name": "Mistral Medium 3 (25.05)", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "baidu/glm-5": { + "id": "baidu/glm-5", + "name": "GLM-5 (Baidu)", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-09", - "release_date": "2025-05-01", - "last_updated": "2025-05-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "mistral-ai/mistral-large-2411": { - "id": "mistral-ai/mistral-large-2411", - "name": "Mistral Large 24.11", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": false, + "baidu/deepseek-v4.1-flash": { + "id": "baidu/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (Baidu)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-11-01", - "last_updated": "2024-11-01", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.2, + "cache_read": 0.006 } }, - "mistral-ai/ministral-3b": { - "id": "mistral-ai/ministral-3b", - "name": "Ministral 3B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, + "baidu/kimi-k2.6": { + "id": "baidu/kimi-k2.6", + "name": "Kimi K2.6 (Baidu)", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -59628,27 +62137,44 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "mistral-ai/codestral-2501": { - "id": "mistral-ai/codestral-2501", - "name": "Codestral 25.01", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "codestral", + "baidu/glm-5.2": { + "id": "baidu/glm-5.2", + "name": "GLM-5.2 (Baidu)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-03", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -59657,29 +62183,46 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "output": 8192 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "cohere/cohere-command-a": { - "id": "cohere/cohere-command-a", - "name": "Cohere Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", + "baidu/glm-5.1": { + "id": "baidu/glm-5.1", + "name": "GLM-5.1 (Baidu)", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-11-01", - "last_updated": "2024-11-01", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -59688,28 +62231,47 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "cohere/cohere-command-r-08-2024": { - "id": "cohere/cohere-command-r-08-2024", - "name": "Cohere Command R 08-2024", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", + "baidu/deepseek-v4-pro": { + "id": "baidu/deepseek-v4-pro", + "name": "DeepSeek V4 Pro (Baidu)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-08-01", - "last_updated": "2024-08-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -59718,28 +62280,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.32, + "output": 3.96, + "cache_read": 0.042 } }, - "cohere/cohere-command-r-plus": { - "id": "cohere/cohere-command-r-plus", - "name": "Cohere Command R+", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", + "baidu/glm-5.3": { + "id": "baidu/glm-5.3", + "name": "GLM-5.3 (Baidu)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-04-04", - "last_updated": "2024-08-01", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -59748,28 +62321,47 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "cohere/cohere-command-r-plus-08-2024": { - "id": "cohere/cohere-command-r-plus-08-2024", - "name": "Cohere Command R+ 08-2024", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", + "baidu/deepseek-v4-flash": { + "id": "baidu/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (Baidu)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-08-01", - "last_updated": "2024-08-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -59778,29 +62370,43 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.44, + "output": 1.32, + "cache_read": 0.014 } }, - "cohere/cohere-command-r": { - "id": "cohere/cohere-command-r", - "name": "Cohere Command R", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", + "gonka24/glm-5.3-flash": { + "id": "gonka24/glm-5.3-flash", + "name": "GLM-5.3 Flash (Gonka24)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-03-11", - "last_updated": "2024-08-01", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text" @@ -59809,29 +62415,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 200000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.3, + "cache_read": 0.035 } }, - "meta/meta-llama-3-8b-instruct": { - "id": "meta/meta-llama-3-8b-instruct", - "name": "Meta-Llama-3-8B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "gonka24/minimax-m2.7": { + "id": "gonka24/minimax-m2.7", + "name": "MiniMax M2.7 (Gonka24)", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-04-18", - "last_updated": "2024-04-18", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -59842,32 +62460,45 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 2048 + "context": 204800, + "output": 131100 }, "cost": { - "input": 0, - "output": 0 + "input": 0.08, + "output": 0.32, + "cache_read": 0.017 } }, - "meta/llama-3.2-11b-vision-instruct": { - "id": "meta/llama-3.2-11b-vision-instruct", - "name": "Llama-3.2-11B-Vision-Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", + "gonka24/deepseek-v4-flash": { + "id": "gonka24/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (Gonka24)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -59875,58 +62506,81 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 390000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.065, + "output": 0.116, + "cache_read": 0.012 } }, - "meta/meta-llama-3.1-405b-instruct": { - "id": "meta/meta-llama-3.1-405b-instruct", - "name": "Meta-Llama-3.1-405B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "bytedance/seed-1-6-flash-250715": { + "id": "bytedance/seed-1-6-flash-250715", + "name": "Seed 1.6 Flash (250715) (ByteDance)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.07, + "output": 0.3, + "cache_read": 0.015 } }, - "meta/llama-4-scout-17b-16e-instruct": { - "id": "meta/llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout 17B 16E Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": false, + "bytedance/seed-1-6-250915": { + "id": "bytedance/seed-1-6-250915", + "name": "Seed 1.6 (250915) (ByteDance)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", @@ -59936,60 +62590,81 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.25, + "output": 2, + "cache_read": 0.05 } }, - "meta/llama-3.3-70b-instruct": { - "id": "meta/llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "bytedance/seed-1-8-251228": { + "id": "bytedance/seed-1-8-251228", + "name": "Seed 1.8 (251228) (ByteDance)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.25, + "output": 2, + "cache_read": 0.05 } }, - "meta/meta-llama-3.1-70b-instruct": { - "id": "meta/meta-llama-3.1-70b-instruct", - "name": "Meta-Llama-3.1-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "bytedance/glm-4.7": { + "id": "bytedance/glm-4.7", + "name": "GLM-4.7 (ByteDance)", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -60000,63 +62675,90 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "meta/meta-llama-3.1-8b-instruct": { - "id": "meta/meta-llama-3.1-8b-instruct", - "name": "Meta-Llama-3.1-8B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "bytedance/seed-1-6-250615": { + "id": "bytedance/seed-1-6-250615", + "name": "Seed 1.6 (250615) (ByteDance)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2025-06-25", + "last_updated": "2025-06-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.25, + "output": 2, + "cache_read": 0.05 } }, - "meta/llama-3.2-90b-vision-instruct": { - "id": "meta/llama-3.2-90b-vision-instruct", - "name": "Llama-3.2-90B-Vision-Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", + "bytedance/glm-5.2": { + "id": "bytedance/glm-5.2", + "name": "GLM-5.2 (ByteDance)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -60064,31 +62766,48 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 1024000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "meta/llama-4-maverick-17b-128e-instruct-fp8": { - "id": "meta/llama-4-maverick-17b-128e-instruct-fp8", - "name": "Llama 4 Maverick 17B 128E Instruct FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "bytedance/deepseek-v4-pro": { + "id": "bytedance/deepseek-v4-pro", + "name": "DeepSeek V4 Pro (ByteDance)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -60096,27 +62815,33 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0, - "output": 0 + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 } }, - "meta/meta-llama-3-70b-instruct": { - "id": "meta/meta-llama-3-70b-instruct", - "name": "Meta-Llama-3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "bytedance/deepseek-v3.2": { + "id": "bytedance/deepseek-v3.2", + "name": "DeepSeek V3.2 (ByteDance)", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-04-18", - "last_updated": "2024-04-18", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -60127,27 +62852,37 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 2048 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.28, + "output": 0.42, + "cache_read": 0.056 } }, - "deepseek/deepseek-r1-0528": { - "id": "deepseek/deepseek-r1-0528", - "name": "DeepSeek-R1-0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "bytedance/gpt-oss-120b": { + "id": "bytedance/gpt-oss-120b", + "name": "GPT OSS 120B (ByteDance)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -60158,27 +62893,45 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 8192 + "context": 128000, + "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.5, + "cache_read": 0.02 } }, - "deepseek/deepseek-v3-0324": { - "id": "deepseek/deepseek-v3-0324", - "name": "DeepSeek-V3-0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "bytedance/deepseek-v4-flash": { + "id": "bytedance/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (ByteDance)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -60189,68 +62942,79 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0, - "output": 0 + "input": 0.44, + "output": 1.32, + "cache_read": 0.014 } }, - "deepseek/deepseek-r1": { - "id": "deepseek/deepseek-r1", - "name": "DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "xai/grok-4-5": { + "id": "xai/grok-4-5", + "name": "Grok 4.5 (xAI)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 8192 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 6, + "cache_read": 0.3 } - } - } - }, - "neuralwatt": { - "id": "neuralwatt", - "env": [ - "NEURALWATT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.neuralwatt.com/v1", - "name": "Neuralwatt", - "doc": "https://portal.neuralwatt.com/docs", - "models": { - "kimi-k2.5-fast": { - "id": "kimi-k2.5-fast", - "name": "Kimi K2.5 Fast", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + }, + "xai/grok-build-0-1": { + "id": "xai/grok-build-0-1", + "name": "Grok Build 0.1 (xAI)", + "description": "Grok coding model for agentic engineering, edits, and codebase workflows", + "family": "grok-build", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-05-20", + "last_updated": "2026-05-20", "modalities": { "input": [ "text", @@ -60260,36 +63024,46 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262128, - "output": 262128 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.52, - "output": 2.59, - "cache_read": 0.13 + "input": 1, + "output": 2, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 4, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 4, + "cache_read": 0.4 + } } }, - "kimi-k2.6-flex": { - "id": "kimi-k2.6-flex", - "name": "Kimi K2.6 Flex", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "xai/grok-4": { + "id": "xai/grok-4", + "name": "Grok 4 (xAI)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text", @@ -60299,218 +63073,249 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262128, - "output": 262128 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.345, - "output": 1.61, - "cache_read": 0.08625 + "input": 3, + "output": 15, + "cache_read": 0.75 } }, - "glm-5.2-short-fast-flex": { - "id": "glm-5.2-short-fast-flex", - "name": "GLM 5.2 Short Fast Flex", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, + "xai/grok-4-20-beta-0309-non-reasoning": { + "id": "xai/grok-4-20-beta-0309-non-reasoning", + "name": "Grok 4.20 Beta Non-Reasoning (0309) (xAI)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 199984, - "output": 199984 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.725, - "output": 2.25, - "cache_read": 0.18125 + "input": 2, + "output": 6, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "glm-5.2-flex": { - "id": "glm-5.2-flex", - "name": "GLM 5.2 Flex", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "xai/grok-4-3": { + "id": "xai/grok-4-3", + "name": "Grok 4.3 (xAI)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048560, - "output": 1048560 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.725, - "output": 2.25, - "cache_read": 0.18125 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM 5.2", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "xai/grok-4-20-beta-0309-reasoning": { + "id": "xai/grok-4-20-beta-0309-reasoning", + "name": "Grok 4.20 Beta Reasoning (0309) (xAI)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048560, - "output": 1048560 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 1.45, - "output": 4.5, - "cache_read": 0.3625 - } - }, - "glm-5.2-short-fast": { - "id": "glm-5.2-short-fast", - "name": "GLM 5.2 Short Fast", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", - "modalities": { - "input": [ - "text" + "input": 2, + "output": 6, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 199984, - "output": 199984 - }, - "cost": { - "input": 1.45, - "output": 4.5, - "cache_read": 0.3625 + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "qwen3.5-397b-fast": { - "id": "qwen3.5-397b-fast", - "name": "Qwen3.5 397B Fast", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, - "reasoning": false, + "xai/grok-4-6": { + "id": "xai/grok-4-6", + "name": "Grok 4.6 (xAI)", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262128, - "output": 262128 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.69, - "output": 4.14, - "cache_read": 0.1725 + "input": 2, + "output": 6, + "cache_read": 0.5 } }, - "kimi-k2.6-fast": { - "id": "kimi-k2.6-fast", - "name": "Kimi K2.6 Fast", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "xai/grok-4-7": { + "id": "xai/grok-4-7", + "name": "Grok 4.7 (xAI)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ "text", @@ -60520,32 +63325,50 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262128, - "output": 262128 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.69, - "output": 3.22, - "cache_read": 0.1725 - } + "input": 2, + "output": 6, + "cache_read": 0.5 + } }, - "qwen3.6-35b-fast": { - "id": "qwen3.6-35b-fast", - "name": "Qwen3.6 35B Fast", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "qwen3.6", - "attachment": true, - "reasoning": false, + "ranoai/deepseek-v4-flash": { + "id": "ranoai/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (RanoAI)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -60553,26 +63376,23 @@ }, "open_weights": true, "limit": { - "context": 131056, - "output": 131056 + "context": 1000000, + "output": 393216 }, "cost": { - "input": 0.29, - "output": 1.15, - "cache_read": 0.0725 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "glm-5.2-short-flex": { - "id": "glm-5.2-short-flex", - "name": "GLM 5.2 Short Flex", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "canopywave/kimi-k3": { + "id": "canopywave/kimi-k3", + "name": "Kimi K3 (CanopyWave)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -60584,18 +63404,18 @@ "xhigh", "max" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, - "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "structured_output": false, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -60603,29 +63423,50 @@ }, "open_weights": true, "limit": { - "context": 199984, - "output": 199984 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.725, - "output": 2.25, - "cache_read": 0.18125 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "glm-5.2-fast": { - "id": "glm-5.2-fast", - "name": "GLM 5.2 Fast", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, - "reasoning": false, + "canopywave/kimi-k2.6": { + "id": "canopywave/kimi-k2.6", + "name": "Kimi K2.6 (CanopyWave)", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -60633,26 +63474,23 @@ }, "open_weights": true, "limit": { - "context": 1048560, - "output": 1048560 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.45, - "output": 4.5, - "cache_read": 0.3625 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "glm-5.2-short": { - "id": "glm-5.2-short", - "name": "GLM 5.2 Short", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "canopywave/glm-5.2": { + "id": "canopywave/glm-5.2", + "name": "GLM-5.2 (CanopyWave)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -60664,15 +63502,16 @@ "xhigh", "max" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -60683,33 +63522,48 @@ }, "open_weights": true, "limit": { - "context": 199984, - "output": 199984 + "context": 200000, + "output": 32768 }, "cost": { - "input": 1.45, - "output": 4.5, - "cache_read": 0.3625 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "kimi-k2.7-code-flex": { - "id": "kimi-k2.7-code-flex", - "name": "Kimi K2.7 Code Flex", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "canopywave/deepseek-v4-pro": { + "id": "canopywave/deepseek-v4-pro", + "name": "DeepSeek V4 Pro (CanopyWave)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -60717,38 +63571,48 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 393216 }, "cost": { - "input": 0.475, - "output": 2, - "cache_read": 0.11875 + "input": 1.74, + "output": 3.48, + "cache_read": 0.01 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, + "canopywave/deepseek-v4-flash": { + "id": "canopywave/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (CanopyWave)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": false, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -60756,38 +63620,45 @@ }, "open_weights": true, "limit": { - "context": 262128, - "output": 262128 + "context": 1000000, + "output": 393216 }, "cost": { - "input": 0.69, - "output": 3.22, - "cache_read": 0.1725 + "input": 0.14, + "output": 0.28, + "cache_read": 0.03 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "fireworks/kimi-k3": { + "id": "fireworks/kimi-k3", + "name": "Kimi K3 (Fireworks AI)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "structured_output": false, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -60795,29 +63666,41 @@ }, "open_weights": true, "limit": { - "context": 262128, - "output": 262128 + "context": 1040384, + "output": 1040384 }, "cost": { - "input": 0.52, - "output": 2.59, - "cache_read": 0.13 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "fireworks/deepseek-v4.1-flash": { + "id": "fireworks/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (Fireworks AI)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -60829,38 +63712,45 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.2375 + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 } }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "fireworks/kimi-k3-fast": { + "id": "fireworks/kimi-k3-fast", + "name": "Kimi K3 Fast (Fireworks AI)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "structured_output": false, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -60868,34 +63758,44 @@ }, "open_weights": true, "limit": { - "context": 131056, - "output": 131056 + "context": 1040384, + "output": 1040384 }, "cost": { - "input": 0.29, - "output": 1.15, - "cache_read": 0.0725 + "input": 4.5, + "output": 22.5, + "cache_read": 0.45 } }, - "Qwen/Qwen3.5-397B-A17B-FP8": { - "id": "Qwen/Qwen3.5-397B-A17B-FP8", - "name": "Qwen3.5 397B A17B FP8", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "fireworks/deepseek-v4-pro": { + "id": "fireworks/deepseek-v4-pro", + "name": "DeepSeek V4 Pro (Fireworks AI)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -60906,39 +63806,44 @@ }, "open_weights": true, "limit": { - "context": 262128, - "output": 262128 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.69, - "output": 4.14, - "cache_read": 0.1725 + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 } - } - } - }, - "siliconflow-cn": { - "id": "siliconflow-cn", - "env": [ - "SILICONFLOW_CN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.siliconflow.cn/v1", - "name": "SiliconFlow (China)", - "doc": "https://cloud.siliconflow.com/models", - "models": { - "baidu/ERNIE-4.5-300B-A47B": { - "id": "baidu/ERNIE-4.5-300B-A47B", - "name": "baidu/ERNIE-4.5-300B-A47B", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "ernie", + }, + "fireworks/deepseek-v4-flash": { + "id": "fireworks/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (Fireworks AI)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-07-02", - "last_updated": "2025-11-25", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -60947,31 +63852,43 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.28, - "output": 1.1 + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 } }, - "ByteDance-Seed/Seed-OSS-36B-Instruct": { - "id": "ByteDance-Seed/Seed-OSS-36B-Instruct", - "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "seed", - "attachment": false, - "reasoning": false, + "sakana/fugu-max": { + "id": "sakana/fugu-max", + "name": "Fugu Max (Sakana AI)", + "description": "Multi-agent model for routing expert agents across complex analytical tasks", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-09-04", - "last_updated": "2025-11-25", + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -60979,30 +63896,41 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.21, - "output": 0.57 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "stepfun-ai/Step-3.5-Flash": { - "id": "stepfun-ai/Step-3.5-Flash", - "name": "stepfun-ai/Step-3.5-Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "family": "step", - "attachment": false, + "sakana/fugu-ultra-v2.0": { + "id": "sakana/fugu-ultra-v2.0", + "name": "Fugu Ultra v2.0 (Sakana AI)", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -61010,29 +63938,41 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "inclusionAI/Ling-flash-2.0": { - "id": "inclusionAI/Ling-flash-2.0", - "name": "inclusionAI/Ling-flash-2.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", - "attachment": false, - "reasoning": false, + "sakana/fugu-ultra": { + "id": "sakana/fugu-ultra", + "name": "Fugu Ultra (Sakana AI)", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-18", - "last_updated": "2025-11-25", + "structured_output": false, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -61040,34 +63980,30 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "Pro/moonshotai/Kimi-K2.6": { - "id": "Pro/moonshotai/Kimi-K2.6", - "name": "Pro/moonshotai/Kimi-K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", + "scx-ai/gemma-4-31b-it": { + "id": "scx-ai/gemma-4-31b-it", + "name": "Gemma 4 31B IT (SCX.ai (Turbo))", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -61075,35 +64011,30 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.3, + "output": 0.91 } }, - "Pro/moonshotai/Kimi-K2.5": { - "id": "Pro/moonshotai/Kimi-K2.5", - "name": "Pro/moonshotai/Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", + "scx-ai/llama-4-maverick-17b-instruct": { + "id": "scx-ai/llama-4-maverick-17b-instruct", + "name": "Llama 4 Maverick 17B Instruct (SCX.ai (Turbo))", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -61111,19 +64042,19 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.45, - "output": 2.25 + "input": 0.53, + "output": 1.62 } }, - "Pro/zai-org/GLM-5": { - "id": "Pro/zai-org/GLM-5", - "name": "Pro/zai-org/GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "scx-ai/qwen3-32b": { + "id": "scx-ai/qwen3-32b", + "name": "Qwen3 32B (SCX.ai (Turbo))", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -61132,13 +64063,11 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -61149,30 +64078,38 @@ }, "open_weights": true, "limit": { - "context": 205000, - "output": 205000 + "context": 32768, + "output": 8192 }, "cost": { - "input": 1, - "output": 3.2 + "input": 0.36, + "output": 0.87 } }, - "Pro/zai-org/GLM-5.1": { - "id": "Pro/zai-org/GLM-5.1", - "name": "Pro/zai-org/GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "scx-ai/minimax-m2.7": { + "id": "scx-ai/minimax-m2.7", + "name": "MiniMax M2.7 (SCX.ai (Turbo))", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -61183,34 +64120,37 @@ }, "open_weights": true, "limit": { - "context": 205000, - "output": 205000 + "context": 196608, + "output": 196608 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_write": 0 + "input": 0.48, + "output": 1.79, + "cache_read": 0.05 } }, - "Pro/deepseek-ai/DeepSeek-R1": { - "id": "Pro/deepseek-ai/DeepSeek-R1", - "name": "Pro/deepseek-ai/DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "scx-ai/gpt-oss-120b": { + "id": "scx-ai/gpt-oss-120b", + "name": "GPT OSS 120B (SCX.ai (Turbo))", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-11-25", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -61219,33 +64159,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 164000, - "output": 164000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 2.18 + "input": 0.17, + "output": 0.55 } }, - "Pro/deepseek-ai/DeepSeek-V3.1-Terminus": { - "id": "Pro/deepseek-ai/DeepSeek-V3.1-Terminus", - "name": "Pro/deepseek-ai/DeepSeek-V3.1-Terminus", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", + "together-ai/qwen3.8-2.4t-a95b": { + "id": "together-ai/qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B (Together AI)", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-11-25", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text" @@ -61254,131 +64200,178 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 164000, - "output": 164000 + "context": 1010000, + "output": 909000 }, "cost": { - "input": 0.27, - "output": 1 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "Pro/deepseek-ai/DeepSeek-V3.2": { - "id": "Pro/deepseek-ai/DeepSeek-V3.2", - "name": "Pro/deepseek-ai/DeepSeek-V3.2", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", - "attachment": false, + "together-ai/glm-5.3-flash": { + "id": "together-ai/glm-5.3-flash", + "name": "GLM-5.3 Flash (Together AI)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 164000, - "output": 164000 + "context": 1048576, + "output": 943717 }, "cost": { - "input": 0.27, - "output": 0.42 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "Pro/deepseek-ai/DeepSeek-V3": { - "id": "Pro/deepseek-ai/DeepSeek-V3", - "name": "Pro/deepseek-ai/DeepSeek-V3", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "together-ai/muse-glimmer-30b": { + "id": "together-ai/muse-glimmer-30b", + "name": "Muse Glimmer 30B (Together AI)", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2025-11-25", + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 164000, - "output": 164000 + "context": 131072, + "output": 117964 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.35, + "output": 1.5, + "cache_read": 0.04 } }, - "Pro/MiniMaxAI/MiniMax-M2.5": { - "id": "Pro/MiniMaxAI/MiniMax-M2.5", - "name": "Pro/MiniMaxAI/MiniMax-M2.5", - "description": "Frontier MiniMax model for engineering, office tasks, and agentic reasoning", - "family": "minimax", - "attachment": false, - "reasoning": false, + "together-ai/kimi-k3": { + "id": "together-ai/kimi-k3", + "name": "Kimi K3 (Together AI)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "structured_output": false, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 192000, - "output": 131000 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.3, - "output": 1.22 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen/Qwen3.6-35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "together-ai/deepseek-v4.1-flash": { + "id": "together-ai/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (Together AI)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -61386,36 +64379,46 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.23, - "output": 1.86 + "input": 0.3, + "output": 1.2, + "cache_read": 0.006 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen/Qwen3.5-397B-A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "together-ai/inkling": { + "id": "together-ai/inkling", + "name": "Inkling (Together AI)", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", "image", - "video" + "audio" ], "output": [ "text" @@ -61423,33 +64426,36 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 524288, + "output": 471859 }, "cost": { - "input": 0.29, - "output": 1.74 + "input": 1, + "output": 4.05, + "cache_read": 0.17 } }, - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "together-ai/glm-4.7": { + "id": "together-ai/glm-4.7", + "name": "GLM-4.7 (Together AI)", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-11-25", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -61458,21 +64464,21 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 202752, + "output": 128000 }, "cost": { - "input": 0.13, - "output": 0.6 + "input": 0.45, + "output": 2 } }, - "Qwen/Qwen3.5-122B-A10B": { - "id": "Qwen/Qwen3.5-122B-A10B", - "name": "Qwen/Qwen3.5-122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "together-ai/minimax-m3": { + "id": "together-ai/minimax-m3", + "name": "MiniMax M3 (Together AI)", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -61480,16 +64486,14 @@ "type": "toggle" } ], - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -61497,36 +64501,44 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 524288, + "output": 131072 }, "cost": { - "input": 0.29, - "output": 2.32 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "Qwen/Qwen3.5-27B": { - "id": "Qwen/Qwen3.5-27B", - "name": "Qwen/Qwen3.5-27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "together-ai/deepseek-v4-pro": { + "id": "together-ai/deepseek-v4-pro", + "name": "DeepSeek V4 Pro (Together AI)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-25", - "last_updated": "2026-02-25", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -61534,36 +64546,37 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 1048576, + "output": 163840 }, "cost": { - "input": 0.26, - "output": 2.09 + "input": 1.32, + "output": 3.96, + "cache_read": 0.13 } }, - "Qwen/Qwen3-8B": { - "id": "Qwen/Qwen3-8B", - "name": "Qwen/Qwen3-8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "together-ai/gpt-oss-120b": { + "id": "together-ai/gpt-oss-120b", + "name": "GPT OSS 120B (Together AI)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -61572,38 +64585,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.06, - "output": 0.06 + "input": 0.15, + "output": 0.6 } }, - "Qwen/Qwen3.5-4B": { - "id": "Qwen/Qwen3.5-4B", - "name": "Qwen/Qwen3.5-4B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "together-ai/glm-5.3": { + "id": "together-ai/glm-5.3", + "name": "GLM-5.3 (Together AI)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -61611,36 +64628,44 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 1048576, + "output": 943717 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "Qwen/Qwen3.5-9B": { - "id": "Qwen/Qwen3.5-9B", - "name": "Qwen/Qwen3.5-9B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "together-ai/deepseek-v4-flash": { + "id": "together-ai/deepseek-v4-flash", + "name": "DeepSeek V4 Flash (Together AI)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "xhigh", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -61648,116 +64673,155 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0.22, - "output": 1.74 + "input": 0.14, + "output": 0.28, + "cache_read": 0.03 } }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen/Qwen3-32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "xiaomi/mimo-v2.6-pro": { + "id": "xiaomi/mimo-v2.6-pro", + "name": "MiMo V2.6 Pro (Xiaomi)", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } }, - "Qwen/Qwen3-14B": { - "id": "Qwen/Qwen3-14B", - "name": "Qwen/Qwen3-14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo V2.5 (Xiaomi)", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028, + "tiers": [ + { + "input": 0.8, + "output": 4, + "cache_read": 0.16, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.8, + "output": 4, + "cache_read": 0.16 + } } }, - "Qwen/Qwen3.5-35B-A3B": { - "id": "Qwen/Qwen3.5-35B-A3B", - "name": "Qwen/Qwen3.5-35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo V2.5 Pro (Xiaomi)", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-25", - "last_updated": "2026-02-25", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -61765,62 +64829,106 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.23, - "output": 1.86 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "Qwen/Qwen3-VL-32B-Thinking": { - "id": "Qwen/Qwen3-VL-32B-Thinking", - "name": "Qwen/Qwen3-VL-32B-Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "xiaomi/mimo-v2.6-flash": { + "id": "xiaomi/mimo-v2.6-flash", + "name": "MiMo V2.6 Flash (Xiaomi)", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-10-21", - "last_updated": "2025-11-25", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 1.5 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "Qwen/Qwen3-VL-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", - "name": "Qwen/Qwen3-VL-30B-A3B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "quartz/gemini-3.1-pro-preview": { + "id": "quartz/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro (Preview) (Quartz)", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-05", - "last_updated": "2025-11-25", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -61828,26 +64936,55 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.29, - "output": 1 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "Qwen/Qwen2.5-72B-Instruct": { - "id": "Qwen/Qwen2.5-72B-Instruct", - "name": "Qwen/Qwen2.5-72B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax M2.5 (MiniMax)", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2025-11-25", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -61856,32 +64993,44 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 33000, - "output": 4000 + "context": 204800, + "output": 131100 }, "cost": { - "input": 0.59, - "output": 0.59 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "Qwen/Qwen3-VL-32B-Instruct": { - "id": "Qwen/Qwen3-VL-32B-Instruct", - "name": "Qwen/Qwen3-VL-32B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, + "minimax/minimax-m2.1-lightning": { + "id": "minimax/minimax-m2.1-lightning", + "name": "MiniMax M2.1 Lightning (MiniMax)", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-10-21", - "last_updated": "2025-11-25", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -61889,31 +65038,30 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 196608, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 0.12, + "output": 0.48 } }, - "Qwen/Qwen3-VL-235B-A22B-Thinking": { - "id": "Qwen/Qwen3-VL-235B-A22B-Thinking", - "name": "Qwen/Qwen3-VL-235B-A22B-Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "minimax/minimax-text-01": { + "id": "minimax/minimax-text-01", + "name": "MiniMax Text 01 (MiniMax)", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-10-04", - "last_updated": "2025-11-25", + "release_date": "2025-01-15", + "last_updated": "2025-01-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -61921,89 +65069,128 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.45, - "output": 3.5 + "input": 0.2, + "output": 1.1 } }, - "Qwen/Qwen3-VL-8B-Instruct": { - "id": "Qwen/Qwen3-VL-8B-Instruct", - "name": "Qwen/Qwen3-VL-8B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax M2.7 (MiniMax)", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-11-25", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 204800, + "output": 131100 }, "cost": { - "input": 0.18, - "output": 0.68 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "Qwen/Qwen3-VL-30B-A3B-Thinking": { - "id": "Qwen/Qwen3-VL-30B-A3B-Thinking", - "name": "Qwen/Qwen3-VL-30B-A3B-Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax M3 (MiniMax)", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-10-11", - "last_updated": "2025-11-25", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 512000, + "output": 131072 }, "cost": { - "input": 0.29, - "output": 1 + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 } }, - "Qwen/Qwen3-30B-A3B-Instruct-2507": { - "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax M2.7 Highspeed (MiniMax)", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-07-30", - "last_updated": "2025-11-25", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -62012,28 +65199,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 204800, + "output": 131100 }, "cost": { - "input": 0.09, - "output": 0.3 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06 } }, - "Qwen/Qwen3-Coder-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "name": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "minimax/minimax-m2.5-highspeed": { + "id": "minimax/minimax-m2.5-highspeed", + "name": "MiniMax M2.5 Highspeed (MiniMax)", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-08-01", - "last_updated": "2025-11-25", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -62042,28 +65242,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 204800, + "output": 131100 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 0.6, + "output": 2.4, + "cache_read": 0.03 } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "name": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax M2.1 (MiniMax)", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-07-31", - "last_updated": "2025-11-25", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -62072,59 +65285,71 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 196608, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.27, + "output": 1.1 } }, - "Qwen/Qwen3-VL-235B-A22B-Instruct": { - "id": "Qwen/Qwen3-VL-235B-A22B-Instruct", - "name": "Qwen/Qwen3-VL-235B-A22B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax M2 (MiniMax)", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-10-04", - "last_updated": "2025-11-25", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 196608, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 0.2, + "output": 1, + "cache_read": 0.03 } }, - "Qwen/Qwen2.5-7B-Instruct": { - "id": "Qwen/Qwen2.5-7B-Instruct", - "name": "Qwen/Qwen2.5-7B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "inference.net/llama-3.2-11b-instruct": { + "id": "inference.net/llama-3.2-11b-instruct", + "name": "Llama 3.2 11B Instruct (Inference.net)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2025-11-25", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text" @@ -62135,63 +65360,91 @@ }, "open_weights": false, "limit": { - "context": 33000, - "output": 4000 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.05 + "input": 0.07, + "output": 0.33 } }, - "PaddlePaddle/PaddleOCR-VL-1.5": { - "id": "PaddlePaddle/PaddleOCR-VL-1.5", - "name": "PaddlePaddle/PaddleOCR-VL-1.5", - "description": "Multimodal model for analyzing text, images, documents, and rich media", + "azure-anthropic/claude-opus-5": { + "id": "azure-anthropic/claude-opus-5", + "name": "Claude Opus 5 (Azure Anthropic)", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-29", - "last_updated": "2026-01-29", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "tencent/Hunyuan-A13B-Instruct": { - "id": "tencent/Hunyuan-A13B-Instruct", - "name": "tencent/Hunyuan-A13B-Instruct", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", - "attachment": false, + "azure-anthropic/claude-fable-5": { + "id": "azure-anthropic/claude-fable-5", + "name": "Claude Fable 5 (Azure Anthropic)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -62199,72 +65452,93 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "azure-anthropic/claude-opus-4-8": { + "id": "azure-anthropic/claude-opus-4-8", + "name": "Claude Opus 4.8 (Azure Anthropic)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1049000, - "output": 262000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_write": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "zai-org/GLM-4.5-Air": { - "id": "zai-org/GLM-4.5-Air", - "name": "zai-org/GLM-4.5-Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", - "attachment": false, - "reasoning": false, + "azure-anthropic/claude-sonnet-5": { + "id": "azure-anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5 (Azure Anthropic)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -62272,36 +65546,45 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.86 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "deepseek-ai/DeepSeek-R1": { - "id": "deepseek-ai/DeepSeek-R1", - "name": "deepseek-ai/DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "azure-anthropic/claude-opus-4-6": { + "id": "azure-anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6 (Azure Anthropic)", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-11-25", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -62309,34 +65592,46 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 2.18 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-ai/DeepSeek-V3.1-Terminus": { - "id": "deepseek-ai/DeepSeek-V3.1-Terminus", - "name": "deepseek-ai/DeepSeek-V3.1-Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "azure-anthropic/claude-opus-4-7": { + "id": "azure-anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7 (Azure Anthropic)", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -62344,37 +65639,41 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 1 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "atria/atria-dawn-preview": { + "id": "atria/atria-dawn-preview", + "name": "Atria Dawn Preview (Atria)", + "description": "Preview model for early access evaluation, prototyping, and compatibility testing", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-09-12", + "last_updated": "2026-09-12", "modalities": { "input": [ "text" @@ -62383,30 +65682,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.003 + "input": 0, + "output": 0 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "deepseek-ai/DeepSeek-V4-Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", + "mistral/mistral-large-latest": { + "id": "mistral/mistral-large-latest", + "name": "Mistral Large Latest (Mistral AI)", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": false, "temperature": true, - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", "modalities": { "input": [ "text" @@ -62417,35 +65715,30 @@ }, "open_weights": true, "limit": { - "context": 1049000, - "output": 393000 + "context": 128000, + "output": 262144 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.145 + "input": 4, + "output": 12 } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "deepseek-ai/DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", + "mistral/ministral-8b-2512": { + "id": "mistral/ministral-8b-2512", + "name": "Ministral 8B (Mistral AI)", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -62453,28 +65746,30 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.27, - "output": 0.42 + "input": 0.15, + "output": 0.15 } }, - "deepseek-ai/DeepSeek-OCR": { - "id": "deepseek-ai/DeepSeek-OCR", - "name": "deepseek-ai/DeepSeek-OCR", - "description": "OCR model for extracting structured text from documents and screenshots", - "attachment": true, + "mistral/devstral-2512": { + "id": "mistral/devstral-2512", + "name": "Devstral 2 (Mistral AI)", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-10-20", - "last_updated": "2025-10-20", + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -62482,73 +65777,62 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 2 } }, - "deepseek-ai/DeepSeek-V3": { - "id": "deepseek-ai/DeepSeek-V3", - "name": "deepseek-ai/DeepSeek-V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "mistral/mistral-large-2512": { + "id": "mistral/mistral-large-2512", + "name": "Mistral Large 3 (Mistral AI)", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", + "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2025-11-25", + "knowledge": "2024-11", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 164000, - "output": 164000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.5, + "output": 1.5 } - } - } - }, - "merge-gateway": { - "id": "merge-gateway", - "env": [ - "MERGE_GATEWAY_API_KEY" - ], - "npm": "merge-gateway-ai-sdk-provider", - "name": "Merge Gateway", - "doc": "https://docs.merge.dev/merge-gateway", - "models": { - "xai/grok-4.3": { - "id": "xai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + }, + "mistral/ministral-14b-2512": { + "id": "mistral/ministral-14b-2512", + "name": "Ministral 14B (Mistral AI)", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -62556,152 +65840,89 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.2, + "output": 0.2 } }, - "xai/grok-4.20-0309-reasoning": { - "id": "xai/grok-4.20-0309-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", + "mistral/mistral-small-2506": { + "id": "mistral/mistral-small-2506", + "name": "Mistral Small 3.2 (Mistral AI)", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 30000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.1, + "output": 0.3 } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "mistral/ministral-3b-2512": { + "id": "mistral/ministral-3b-2512", + "name": "Ministral 3B (Mistral AI)", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 32768 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.95, - "output": 4 + "input": 0.1, + "output": 0.1 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", + "mistral/codestral-2508": { + "id": "mistral/codestral-2508", + "name": "Codestral (Mistral AI)", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "codestral", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 32768 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ "text" @@ -62710,294 +65931,369 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.6, - "output": 2.5 + "input": 0.3, + "output": 0.9 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": false, + "vertex-anthropic/claude-haiku-4-5": { + "id": "vertex-anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (Vertex AI (Anthropic))", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 3 + "input": 1, + "output": 5, + "cache_read": 0.1 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "vertex-anthropic/claude-sonnet-4-5": { + "id": "vertex-anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (Vertex AI (Anthropic))", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024, + "max": 63999 }, { - "type": "budget_tokens", - "min": 1, - "max": 262144 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.95, - "output": 4 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "moonshotai/kimi-k2.7-code-highspeed": { - "id": "moonshotai/kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", + "vertex-anthropic/claude-opus-4-5-20251101": { + "id": "vertex-anthropic/claude-opus-4-5-20251101", + "name": "Claude Opus 4.5 (Vertex AI (Anthropic))", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024, + "max": 31999 }, { - "type": "budget_tokens", - "min": 1, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1.9, - "output": 8 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "mistral/codestral-latest": { - "id": "mistral/codestral-latest", - "name": "Codestral (latest)", - "description": "Mistral code model for completions, refactors, and developer IDE workflows", - "family": "codestral", - "attachment": false, - "reasoning": false, + "vertex-anthropic/claude-sonnet-5": { + "id": "vertex-anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5 (Vertex AI (Anthropic))", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-05-29", - "last_updated": "2025-01-04", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 4096 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "mistral/mistral-large-latest": { - "id": "mistral/mistral-large-latest", - "name": "Mistral Large (latest)", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", + "vertex-anthropic/claude-opus-4-6": { + "id": "vertex-anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6 (Vertex AI (Anthropic))", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "mistral/devstral-small-2507": { - "id": "mistral/devstral-small-2507", - "name": "Devstral Small", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, + "vertex-anthropic/claude-sonnet-4-6": { + "id": "vertex-anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6 (Vertex AI (Anthropic))", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-07-10", - "last_updated": "2025-07-10", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 1000000, + "output": 64000 }, - "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.3 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "mistral/pixtral-large-latest": { - "id": "mistral/pixtral-large-latest", - "name": "Pixtral Large (latest)", - "description": "Mistral's larger vision model for document-heavy image understanding and chat", - "family": "pixtral", + "vertex-anthropic/claude-opus-4-7": { + "id": "vertex-anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7 (Vertex AI (Anthropic))", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2024-11-04", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 2, - "output": 6 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "mistral/mistral-medium-latest": { - "id": "mistral/mistral-medium-latest", - "name": "Mistral Medium (latest)", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "vertex-openai/qwen3-235b-a22b-instruct-2507": { + "id": "vertex-openai/qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507 (Vertex AI (OpenAI-compatible))", + "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "release_date": "2025-07-21", + "last_updated": "2025-07-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -63006,30 +66302,42 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "output": 32768 }, "cost": { - "input": 0.4, - "output": 2 + "input": 0.22, + "output": 0.88 } }, - "mistral/mistral-small-latest": { - "id": "mistral/mistral-small-latest", - "name": "Mistral Small (latest)", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, + "vertex-openai/qwen3-next-80b-a3b-thinking": { + "id": "vertex-openai/qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B Thinking (Vertex AI (OpenAI-compatible))", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -63037,57 +66345,75 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 131072, + "output": 32768 }, "cost": { "input": 0.15, - "output": 0.6 + "output": 1.2 } }, - "mistral/mistral-medium-2505": { - "id": "mistral/mistral-medium-2505", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, + "vertex-openai/glm-5": { + "id": "vertex-openai/glm-5", + "name": "GLM-5 (Vertex AI (OpenAI-compatible))", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 202752, + "output": 32768 }, "cost": { - "input": 0.4, - "output": 2 + "input": 1, + "output": 3.2, + "cache_read": 0.1 } }, - "mistral/mistral-large-2411": { - "id": "mistral/mistral-large-2411", - "name": "Mistral Large 2.1", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", + "vertex-openai/qwen3-coder-480b-a35b-instruct": { + "id": "vertex-openai/qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct (Vertex AI (OpenAI-compatible))", + "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-18", - "last_updated": "2024-11-18", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -63098,27 +66424,32 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 2, - "output": 6 + "input": 0.22, + "output": 1.8, + "cache_read": 0.022 } }, - "mistral/magistral-medium-latest": { - "id": "mistral/magistral-medium-latest", - "name": "Magistral Medium (latest)", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-medium", + "vertex-openai/kimi-k2-thinking": { + "id": "vertex-openai/kimi-k2-thinking", + "name": "Kimi K2 Thinking (Vertex AI (OpenAI-compatible))", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-20", + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ "text" @@ -63127,28 +66458,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 2, - "output": 5 + "input": 0.6, + "output": 2.5, + "cache_read": 0.06 } }, - "mistral/devstral-medium-latest": { - "id": "mistral/devstral-medium-latest", - "name": "Devstral 2 (latest)", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", + "vertex-openai/qwen3-next-80b-a3b-instruct": { + "id": "vertex-openai/qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct (Vertex AI (OpenAI-compatible))", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ "text" @@ -63159,27 +66492,35 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 32768 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 2 + "input": 0.15, + "output": 1.2 } }, - "mistral/devstral-2512": { - "id": "mistral/devstral-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", + "vertex-openai/glm-4.7": { + "id": "vertex-openai/glm-4.7", + "name": "GLM-4.7 (Vertex AI (OpenAI-compatible))", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -63190,27 +66531,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 202752, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 2 + "input": 0.6, + "output": 2.2 } }, - "mistral/mistral-large-2512": { - "id": "mistral/mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", + "vertex-openai/grok-4-20-non-reasoning": { + "id": "vertex-openai/grok-4-20-non-reasoning", + "name": "Grok 4.20 Non-Reasoning (Vertex AI (OpenAI-compatible))", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "grok", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", @@ -63220,52 +66560,38 @@ "text" ] }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "mistral/devstral-medium-2507": { - "id": "mistral/devstral-medium-2507", - "name": "Devstral Medium", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-07-10", - "last_updated": "2025-07-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 2000000, + "output": 30000 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 2 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "vertex-openai/grok-4-6": { + "id": "vertex-openai/grok-4-6", + "name": "Grok 4.6 (Vertex AI (OpenAI-compatible))", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -63282,16 +66608,13 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -63299,43 +66622,41 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 2, + "output": 6, + "cache_read": 0.5 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "vertex-openai/grok-4-20-reasoning": { + "id": "vertex-openai/grok-4-20-reasoning", + "name": "Grok 4.20 Reasoning (Vertex AI (OpenAI-compatible))", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -63343,18 +66664,18 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 2000000, + "output": 30000 }, "cost": { "input": 1.25, - "output": 10, - "cache_read": 0.125, + "output": 2.5, + "cache_read": 0.2, "tiers": [ { "input": 2.5, - "output": 15, - "cache_read": 0.25, + "output": 5, + "cache_read": 0.4, "tier": { "type": "context", "size": 200000 @@ -63363,40 +66684,73 @@ ], "context_over_200k": { "input": 2.5, - "output": 15, - "cache_read": 0.25 + "output": 5, + "cache_read": 0.4 } } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "vertex-openai/deepseek-v3.2": { + "id": "vertex-openai/deepseek-v3.2", + "name": "DeepSeek V3.2 (Vertex AI (OpenAI-compatible))", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 65536 + }, + "cost": { + "input": 0.56, + "output": 1.68, + "cache_read": 0.056 + } + }, + "google-vertex/gemini-3.6-flash": { + "id": "google-vertex/gemini-3.6-flash", + "name": "Gemini 3.6 Flash (Google Vertex AI)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", "image", - "audio", "video", + "audio", "pdf" ], "output": [ @@ -63409,17 +66763,17 @@ "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.08333 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", + "google-vertex/gemini-3.5-flash-lite": { + "id": "google-vertex/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite (Google Vertex AI)", "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -63436,9 +66790,9 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", @@ -63457,48 +66811,16 @@ "output": 65536 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "cache_write": 0.08333 } }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "google-vertex/gemini-3.1-pro-preview": { + "id": "google-vertex/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro (Preview) (Google Vertex AI)", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", "family": "gemini-pro", "attachment": true, "reasoning": true, @@ -63557,11 +66879,11 @@ } } }, - "google/gemini-flash-lite-latest": { - "id": "google/gemini-flash-lite-latest", - "name": "Gemini Flash-Lite Latest", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "google-vertex/gemini-3.5-flash": { + "id": "google-vertex/gemini-3.5-flash", + "name": "Gemini 3.5 Flash (Google Vertex AI)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -63579,14 +66901,14 @@ "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", "image", - "audio", "video", + "audio", "pdf" ], "output": [ @@ -63599,26 +66921,33 @@ "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 0.08333 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", + "google-vertex/gemini-2.5-pro": { + "id": "google-vertex/gemini-2.5-pro", + "name": "Gemini 2.5 Pro (Google Vertex AI)", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] }, { "type": "budget_tokens", - "min": 512, - "max": 24576 + "min": 128, + "max": 32768 } ], "tool_call": true, @@ -63645,20 +66974,40 @@ "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "google-vertex/gemini-2.5-flash": { + "id": "google-vertex/gemini-2.5-flash", + "name": "Gemini 2.5 Flash (Google Vertex AI)", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1, + "max": 24576 + }, { "type": "effort", "values": [ @@ -63672,14 +67021,14 @@ "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video", "audio", + "video", "pdf" ], "output": [ @@ -63689,74 +67038,28 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65536 + "output": 65535 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "google/gemini-3-pro-preview": { - "id": "google/gemini-3-pro-preview", - "name": "Gemini 3 Pro Preview", - "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", - "family": "gemini-pro", + "google-vertex/gemini-3.7-flash": { + "id": "google-vertex/gemini-3.7-flash", + "name": "Gemini 3.7 Flash (Google Vertex AI)", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", + "medium", "high" ] } @@ -63764,9 +67067,9 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ "text", @@ -63785,30 +67088,15 @@ "output": 65536 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.08333 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", + "google-vertex/gemini-3-flash-preview": { + "id": "google-vertex/gemini-3-flash-preview", + "name": "Gemini 3 Flash (Preview) (Google Vertex AI)", "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", "family": "gemini-flash", "attachment": true, @@ -63845,27 +67133,25 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65536 + "output": 65535 }, "cost": { "input": 0.5, "output": 3, - "cache_read": 0.05, - "input_audio": 1 + "cache_read": 0.05 } }, - "google/gemini-flash-latest": { - "id": "google/gemini-flash-latest", - "name": "Gemini Flash Latest", + "google-vertex/gemini-3.8-flash": { + "id": "google-vertex/gemini-3.8-flash", + "name": "Gemini 3.8 Flash (Google Vertex AI)", "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -63875,9 +67161,43 @@ "tool_call": true, "structured_output": true, "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.08333 + } + }, + "google-vertex/gemini-2.5-flash-lite": { + "id": "google-vertex/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash Lite (Google Vertex AI)", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, "knowledge": "2025-01", - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", @@ -63893,18 +67213,17 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65536 + "output": 65535 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075, - "input_audio": 1 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 } }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", + "google-vertex/gemini-3.1-flash-lite": { + "id": "google-vertex/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite (Google Vertex AI)", "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", "family": "gemini-flash-lite", "attachment": true, @@ -63924,8 +67243,8 @@ "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", @@ -63947,14 +67266,14 @@ "input": 0.25, "output": 1.5, "cache_read": 0.025, - "input_audio": 0.5 + "cache_write": 0.08333 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "aws-mantle/gpt-6-astra": { + "id": "aws-mantle/gpt-6-astra", + "name": "GPT-6 Astra (AWS Mantle)", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -63963,16 +67282,18 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ "text", @@ -63985,43 +67306,48 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", + "aws-mantle/gpt-5.6-luna": { + "id": "aws-mantle/gpt-5.6-luna", + "name": "GPT-5.6 Luna (AWS Mantle)", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -64029,29 +67355,42 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 921600, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.22, + "output": 1.32, + "cache_read": 0.022, + "cache_write": 0.275 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "aws-mantle/gpt-5.6-terra": { + "id": "aws-mantle/gpt-5.6-terra", + "name": "GPT-5.6 Terra (AWS Mantle)", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -64064,33 +67403,47 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 921600, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 2.2, + "output": 13.2, + "cache_read": 0.22, + "cache_write": 2.75 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "aws-mantle/gpt-5.6-sol": { + "id": "aws-mantle/gpt-5.6-sol", + "name": "GPT-5.6 Sol (AWS Mantle)", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -64098,32 +67451,46 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 921600, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 4.4, + "output": 22, + "cache_read": 0.44, + "cache_write": 5.5 } }, - "openai/gpt-4o-2024-05-13": { - "id": "openai/gpt-4o-2024-05-13", - "name": "GPT-4o (2024-05-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4 (OpenAI)", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -64131,27 +67498,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 5, - "output": 15 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro (OpenAI)", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", "medium", "high", "xhigh" @@ -64159,11 +67526,11 @@ } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": false, "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -64175,34 +67542,62 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 30, + "output": 180 } }, - "openai/gpt-5-chat-latest": { - "id": "openai/gpt-5-chat-latest", - "name": "GPT-5 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5 Turbo (OpenAI)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16385, + "output": 4096 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "openai/gpt-4o-transcribe": { + "id": "openai/gpt-4o-transcribe", + "name": "GPT-4o Transcribe (OpenAI)", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "gpt", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2025-03-20", + "last_updated": "2025-03-20", "modalities": { "input": [ "text", - "image" + "audio" ], "output": [ "text" @@ -64210,29 +67605,97 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 16000, + "output": 2000 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro (OpenAI)", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "openai/gpt-5.3-chat-latest": { - "id": "openai/gpt-5.3-chat-latest", - "name": "GPT-5.3 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 Nano (OpenAI)", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-08-31", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -64244,19 +67707,20 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "openai/gpt-4o-2024-08-06": { - "id": "openai/gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o (OpenAI)", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", "family": "gpt", "attachment": true, "reasoning": false, @@ -64264,12 +67728,13 @@ "structured_output": true, "temperature": true, "knowledge": "2023-09", - "release_date": "2024-08-06", + "release_date": "2024-05-13", "last_updated": "2024-08-06", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -64286,17 +67751,18 @@ "cache_read": 1.25 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": false, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini (OpenAI)", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -64306,12 +67772,13 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -64319,28 +67786,27 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2 Pro (OpenAI)", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", "medium", "high", "xhigh" @@ -64348,7 +67814,7 @@ } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": false, "knowledge": "2025-08-31", "release_date": "2025-12-11", @@ -64366,28 +67832,36 @@ "limit": { "context": 400000, "input": 272000, - "output": 128000 + "output": 272000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 21, + "output": 168 } }, - "openai/gpt-5.1-chat-latest": { - "id": "openai/gpt-5.1-chat-latest", - "name": "GPT-5.1 Chat", - "description": "Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations", - "family": "gpt-codex", + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4 Mini (OpenAI)", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", @@ -64399,33 +67873,107 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "openai/gpt-5.2-chat-latest": { - "id": "openai/gpt-5.2-chat-latest", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3 Mini (OpenAI)", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 + } + }, + "openai/gpt-4": { + "id": "openai/gpt-4", + "name": "GPT-4 (OpenAI)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 30, + "output": 60 + } + }, + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex (OpenAI)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, + "structured_output": false, + "temperature": true, "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -64433,8 +67981,9 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { "input": 1.75, @@ -64444,7 +67993,7 @@ }, "openai/gpt-4.1-nano": { "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", + "name": "GPT-4.1 Nano (OpenAI)", "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", "family": "gpt-nano", "attachment": true, @@ -64466,7 +68015,7 @@ }, "open_weights": false, "limit": { - "context": 1047576, + "context": 1000000, "output": 32768 }, "cost": { @@ -64475,23 +68024,33 @@ "cache_read": 0.025 } }, - "openai/gpt-4o-2024-11-20": { - "id": "openai/gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano (OpenAI)", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -64499,18 +68058,19 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, "openai/o1": { "id": "openai/o1", - "name": "o1", + "name": "o1 (OpenAI)", "description": "O-series reasoning model for hard analysis, math, coding, and planning", "family": "o", "attachment": true, @@ -64525,7 +68085,7 @@ ] } ], - "tool_call": true, + "tool_call": false, "structured_output": true, "temperature": false, "knowledge": "2023-09", @@ -64552,36 +68112,31 @@ "cache_read": 7.5 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5 Pro (OpenAI)", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", - "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -64589,73 +68144,39 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 400000, + "input": 272000, + "output": 272000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 15, + "output": 120 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "openai/gpt-6-astra": { + "id": "openai/gpt-6-astra", + "name": "GPT-6 Astra (OpenAI)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-09-03", + "last_updated": "2026-09-03", "modalities": { "input": [ "text", @@ -64667,50 +68188,46 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1 (OpenAI)", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -64718,73 +68235,29 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o Mini (OpenAI)", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", "family": "gpt-mini", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", @@ -64797,39 +68270,40 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 Mini (OpenAI)", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -64846,24 +68320,37 @@ "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT-5.6 Luna (OpenAI)", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -64876,19 +68363,21 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2 (OpenAI)", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", "family": "gpt", "attachment": true, "reasoning": true, @@ -64899,16 +68388,17 @@ "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -64925,14 +68415,14 @@ "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, "openai/gpt-5.5": { "id": "openai/gpt-5.5", - "name": "GPT-5.5", + "name": "GPT-5.5 (OpenAI)", "description": "Default frontier GPT for coding, computer use, research, and knowledge work", "family": "gpt", "attachment": true, @@ -64971,22 +68461,6 @@ "input": 922000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, "cost": { "input": 5, "output": 30, @@ -65009,339 +68483,345 @@ } } }, - "zai/glm-4.7": { - "id": "zai/glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1 (OpenAI)", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "zai/glm-4.5": { - "id": "zai/glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 Mini (OpenAI)", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "zai/glm-4.7-flashx": { - "id": "zai/glm-4.7-flashx", - "name": "GLM-4.7-FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, + "openai/gpt-6-luna": { + "id": "openai/gpt-6-luna", + "name": "GPT-6 Luna (OpenAI)", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.07, - "output": 0.4, + "input": 0.1, + "output": 0.5, "cache_read": 0.01, - "cache_write": 0 + "cache_write": 0.125 } }, - "zai/glm-5.1": { - "id": "zai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "openai/gpt-4o-mini-transcribe": { + "id": "openai/gpt-4o-mini-transcribe", + "name": "GPT-4o Mini Transcribe (OpenAI)", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-03-20", + "last_updated": "2025-03-20", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 16000, + "output": 2000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "input": 1.25, + "output": 5 } }, - "zai/glm-4.6": { - "id": "zai/glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra (OpenAI)", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "zai/glm-5.2": { - "id": "zai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 50000 - } - ], + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo (OpenAI)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.4, - "output": 4.4 + "input": 10, + "output": 30 } }, - "zai/glm-4.5-air": { - "id": "zai/glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "openai/o3": { + "id": "openai/o3", + "name": "o3 (OpenAI)", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03, - "cache_write": 0 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "zai/glm-5": { - "id": "zai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5 (OpenAI)", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "zai/glm-5-turbo": { - "id": "zai/glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol (OpenAI)", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -65349,38 +68829,43 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24, - "cache_write": 0 + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5 } }, - "anthropic/claude-haiku-4-5-20251001": { - "id": "anthropic/claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "openai/gpt-6-sol": { + "id": "openai/gpt-6-sol", + "name": "GPT-6 Sol (OpenAI)", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", @@ -65393,71 +68878,68 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "anthropic/claude-opus-4-1-20250805": { - "id": "anthropic/claude-opus-4-1-20250805", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "moonshot/kimi-k3": { + "id": "moonshot/kimi-k3", + "name": "Kimi K3 (Moonshot AI)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": false, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "moonshot/kimi-k2.7-code-highspeed": { + "id": "moonshot/kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed (Moonshot AI)", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -65470,334 +68952,309 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.9, + "output": 8, + "cache_read": 0.38 } }, - "anthropic/claude-opus-4-5-20251101": { - "id": "anthropic/claude-opus-4-5-20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "moonshot/kimi-k2.6": { + "id": "moonshot/kimi-k2.6", + "name": "Kimi K2.6 (Moonshot AI)", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "anthropic/claude-opus-4-8": { - "id": "anthropic/claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "moonshot/kimi-k2.5": { + "id": "moonshot/kimi-k2.5", + "name": "Kimi K2.5 (Moonshot AI)", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 128000 + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 5, - "output": 25 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "anthropic/claude-opus-4-20250514": { - "id": "anthropic/claude-opus-4-20250514", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "moonshot/kimi-k2.7-code": { + "id": "moonshot/kimi-k2.7-code", + "name": "Kimi K2.7 Code (Moonshot AI)", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } - }, - "anthropic/claude-sonnet-4-20250514": { - "id": "anthropic/claude-sonnet-4-20250514", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + } + } + }, + "above": { + "id": "above", + "env": [ + "ABOVE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.above.dev/v1", + "name": "above.dev", + "doc": "https://above.dev/docs", + "models": { + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM 5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.165, + "output": 0.55, + "cache_read": 0.0319 } }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "glm-5.2-fast": { + "id": "glm-5.2-fast", + "name": "GLM 5.2 Fast", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", + "none", "high", "max" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2.31, + "output": 7.26, + "cache_read": 0.231 } }, - "anthropic/claude-sonnet-4-5-20250929": { - "id": "anthropic/claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen 3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -65805,21 +69262,20 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 2.2, + "output": 6.6, + "cache_read": 0.275 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "deepseek-v4-flash-vision-exp": { + "id": "deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision (Exp)", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -65830,86 +69286,60 @@ "type": "effort", "values": [ "low", - "medium", "high", "max" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 64000 + "output": 384000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.242, + "output": 0.726, + "reasoning": 0.726, + "cache_read": 0.0077 } }, - "cohere/command-a-03-2025": { - "id": "cohere/command-a-03-2025", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo V2.5 Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8000 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 2.5, - "output": 10 - } - }, - "cohere/command-r-08-2024": { - "id": "cohere/command-r-08-2024", - "name": "Command R", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -65920,56 +69350,40 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.5077, + "output": 1.0154, + "cache_read": 0.0042 } }, - "cohere/command-r7b-12-2024": { - "id": "cohere/command-r7b-12-2024", - "name": "Command R7B", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-12-02", - "last_updated": "2024-12-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4000 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.0375, - "output": 0.15 - } - }, - "cohere/command-r-plus-08-2024": { - "id": "cohere/command-r-plus-08-2024", - "name": "Command R+", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -65980,19 +69394,20 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 10 + "input": 1.54, + "output": 4.84, + "cache_read": 0.154 } }, - "alibaba/qwen3.7-max": { - "id": "alibaba/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -66000,15 +69415,21 @@ "type": "toggle" }, { - "type": "budget_tokens", - "min": 1, - "max": 250000 + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -66017,78 +69438,51 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 65536 + "output": 384000 }, "cost": { - "input": 1.65, - "output": 4.95 + "input": 0.726, + "output": 2.178, + "reasoning": 2.178, + "cache_read": 0.0242 } }, - "alibaba/qwen3.6-plus": { - "id": "alibaba/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": false, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens", - "min": 1, - "max": 250000 + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3 - } - }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -66100,28 +69494,36 @@ "output": 384000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.165, + "output": 0.66, + "reasoning": 0.66, + "cache_read": 0.0033 } - }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + } + } + }, + "kilo": { + "id": "kilo", + "env": [ + "KILO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.kilo.ai/api/gateway", + "name": "Kilo Gateway", + "doc": "https://kilo.ai", + "models": { + "sao10k/l3-lunaris-8b": { + "id": "sao10k/l3-lunaris-8b", + "name": "Sao10K: Llama 3 8B Lunaris", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2024-08-13", + "last_updated": "2024-08-13", "modalities": { "input": [ "text" @@ -66130,29 +69532,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 8192, + "output": 7372 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0.04, + "output": 0.05 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", + "sao10k/l3.3-euryale-70b": { + "id": "sao10k/l3.3-euryale-70b", + "name": "Sao10K: Llama 3.3 Euryale 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2024-12-18", + "last_updated": "2024-12-18", "modalities": { "input": [ "text" @@ -66161,30 +69562,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.65, + "output": 0.75 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "sao10k/l3.1-euryale-70b": { + "id": "sao10k/l3.1-euryale-70b", + "name": "Sao10K: Llama 3.1 Euryale 70B v2.2", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2024-08-28", + "last_updated": "2024-08-28", "modalities": { "input": [ "text" @@ -66193,30 +69592,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.85, + "output": 0.85 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", + "kwaipilot/kat-coder-pro-v2.5": { + "id": "kwaipilot/kat-coder-pro-v2.5", + "name": "Kwaipilot: KAT-Coder-Pro V2.5", + "description": "KAT-Coder-Pro V2.5 is a flagship-level Agentic Coding model that can directly hand over an entire issue or an entire business workflow to it, allowing it to autonomously locate and make...", + "family": "kat-coder", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2026-07-10", + "last_updated": "2026-07-10", "modalities": { "input": [ "text" @@ -66225,176 +69622,186 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.74, + "output": 2.96, + "cache_read": 0.15 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "stealth/claude-opus-4.6": { + "id": "stealth/claude-opus-4.6", + "name": "Stealth: Claude Opus 4.6 (20% off)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 128000 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.4 + "input": 4, + "output": 20, + "reasoning": 0, + "cache_read": 0.4, + "cache_write": 5 } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, + "stealth/claude-opus-4.7": { + "id": "stealth/claude-opus-4.7", + "name": "Stealth: Claude Opus 4.7 (20% off)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 4, + "output": 20, + "reasoning": 0, + "cache_read": 0.4, + "cache_write": 5 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, + "stealth/claude-opus-4.8": { + "id": "stealth/claude-opus-4.8", + "name": "Stealth: Claude Opus 4.8 (20% off)", + "description": "Your prompts and completions may be retained and used to train or improve the provider's services. This third-party-served variant of Claude Opus 4.8 is offered at 20% lower cost than standard Claude Opus 4.8 pricing and is not served by Anthropic or Kilo Code.", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 4, + "output": 20, + "reasoning": 0, + "cache_read": 0.4, + "cache_write": 5 } }, - "minimax/minimax-m2.5-highspeed": { - "id": "minimax/minimax-m2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 - } - } - } - }, - "qihang-ai": { - "id": "qihang-ai", - "env": [ - "QIHANG_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.qhaigc.net/v1", - "name": "QiHang", - "doc": "https://www.qhaigc.net/docs", - "models": { - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "stealth/claude-sonnet-4.6": { + "id": "stealth/claude-sonnet-4.6", + "name": "Stealth: Claude Sonnet 4.6 (20% off)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-10-01", - "last_updated": "2025-10-01", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text", @@ -66407,35 +69814,45 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 1000000, "output": 64000 }, "cost": { - "input": 0.14, - "output": 0.71 + "input": 2.4, + "output": 12, + "reasoning": 0, + "cache_read": 0.24, + "cache_write": 3 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "stealth/qwen3.6-plus": { + "id": "stealth/qwen3.6-plus", + "name": "Stealth: Qwen3.6 Plus (50% off)", + "description": "Your prompts and completions may be retained and used to train or improve the provider's services. This third-party-served variant of Qwen3.6 Plus is offered at 50% lower cost than standard Qwen3.6 Plus pricing and is not served by Alibaba or Kilo Code. Note: a surcharge applies to long-context workloads exceeding 256K input tokens.", + "family": "qwen3.6", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -66443,45 +69860,43 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.09, - "output": 0.71, - "tiers": [ - { - "input": 0.09, - "output": 0.71, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 0.09, - "output": 0.71 - } + "input": 0.25, + "output": 1.5, + "reasoning": 0, + "cache_read": 0.025, + "cache_write": 0.3125 } }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "bytedance-seed/seed-1.6-flash": { + "id": "bytedance-seed/seed-1.6-flash", + "name": "ByteDance Seed: Seed 1.6 Flash", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ + "image", "text", - "image" + "video" ], "output": [ "text" @@ -66489,32 +69904,43 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.71, - "output": 3.57 + "input": 0.075, + "output": 0.3 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "bytedance-seed/seed-2.0-lite": { + "id": "bytedance-seed/seed-2.0-lite", + "name": "Seed 2.0 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -66522,33 +69948,40 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { "input": 0.25, "output": 2 } }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "bytedance-seed/seed-2-1-turbo": { + "id": "bytedance-seed/seed-2-1-turbo", + "name": "ByteDance Seed: Seed 2.1 Turbo", + "description": "Seed 2.1 Turbo is a multimodal model from ByteDance Seed for coding and long-horizon agent workflows. It is suited for end-to-end software delivery, multi-step task execution, and understanding visual and...", + "family": "seed", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -66556,32 +69989,42 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.43, - "output": 2.14 + "input": 0.5, + "output": 2.5 } }, - "gemini-3-pro-preview": { - "id": "gemini-3-pro-preview", - "name": "Gemini 3 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "bytedance-seed/seed-2.0-mini": { + "id": "bytedance-seed/seed-2.0-mini", + "name": "Seed 2.0 Mini", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ @@ -66590,31 +70033,42 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.57, - "output": 3.43 + "input": 0.1, + "output": 0.4 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5-Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "bytedance-seed/seed-2.0-code": { + "id": "bytedance-seed/seed-2.0-code", + "name": "Seed 2.0 Code", + "description": "Seed 2.0 Code is a model from ByteDance Seed optimized for agentic coding. It is suited for frontend development, multilingual programming tasks, and coding-agent workflows in tools such as Claude...", + "family": "seed", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -66622,35 +70076,40 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.04, - "output": 0.29 + "input": 0.5, + "output": 3 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "bytedance-seed/seed-1.6": { + "id": "bytedance-seed/seed-1.6", + "name": "ByteDance Seed: Seed 1.6", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", "image", - "video", - "audio", - "pdf" + "text", + "video" ], "output": [ "text" @@ -66658,46 +70117,42 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.07, - "output": 0.43, - "tiers": [ - { - "input": 0.07, - "output": 0.43, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 0.07, - "output": 0.43 - } + "input": 0.25, + "output": 2 } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "~moonshotai/kimi-latest": { + "id": "~moonshotai/kimi-latest", + "name": "MoonshotAI: Kimi Latest", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -66705,75 +70160,75 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 0.14, - "output": 1.14 + "input": 1.4989, + "output": 10.758, + "cache_read": 0.3 } - } - } - }, - "xiaomi-token-plan-ams": { - "id": "xiaomi-token-plan-ams", - "env": [ - "XIAOMI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://token-plan-ams.xiaomimimo.com/v1", - "name": "Xiaomi Token Plan (Europe)", - "doc": "https://platform.xiaomimimo.com/#/docs", - "models": { - "mimo-v2.5-tts": { - "id": "mimo-v2.5-tts", - "name": "MiMo-V2.5-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", + }, + "poolside/laguna-s-2.1:free": { + "id": "poolside/laguna-s-2.1:free", + "name": "Poolside: Laguna S 2.1 (free)", + "description": "Laguna S 2.1 is the latest coding agent model from [Poolside](). Laguna S 2.1 is a 118B total parameter model with 8B active parameters, scoring 70.2% on Terminal-Bench 2.1 and...", + "family": "laguna-s", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 262144, + "output": 32768 }, "cost": { "input": 0, "output": 0 } }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "poolside/laguna-s-2.1": { + "id": "poolside/laguna-s-2.1", + "name": "Poolside: Laguna S 2.1", + "description": "Laguna S 2.1 is the latest coding agent model from [Poolside](). Laguna S 2.1 is a 118B total parameter model with 8B active parameters, scoring 70.2% on Terminal-Bench 2.1 and...", + "family": "laguna-s", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text" @@ -66782,37 +70237,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.1, + "output": 0.2, + "cache_read": 0.01 } }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", + "poolside/laguna-xs-2.1:free": { + "id": "poolside/laguna-xs-2.1:free", + "name": "Poolside: Laguna XS 2.1 (free)", + "description": "Laguna XS 2.1 is the latest coding agent model in the 33B-A3B category from [Poolside](https://poolside.ai/) and a step forward from their Laguna XS.2 model (released in April 2026). It combines...", + "family": "laguna", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-07-02", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" @@ -66823,570 +70279,796 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 32768 }, - "status": "deprecated", "cost": { "input": 0, - "output": 0, - "cache_read": 0 + "output": 0 } }, - "mimo-v2-tts": { - "id": "mimo-v2-tts", - "name": "MiMo-V2-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", + "poolside/laguna-xs-2.1": { + "id": "poolside/laguna-xs-2.1", + "name": "Poolside: Laguna XS 2.1", + "description": "Laguna XS 2.1 is the latest coding agent model in the 33B-A3B category from [Poolside](https://poolside.ai/) and a step forward from their Laguna XS.2 model (released in April 2026). It combines...", + "family": "laguna", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-02", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.2, + "cache_read": 0.05 } }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "prism-ml/ternary-bonsai-2-27b": { + "id": "prism-ml/ternary-bonsai-2-27b", + "name": "PrismML: Ternary Bonsai 2 27B", + "description": "Bonsai 2 27B is a 27B-parameter reasoning model from PrismML derived from Qwen3.8-27B. It supports coding, mathematics, tool calling, and image understanding with a 262K-token context window. Ternary compression shrinks...", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "medium", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-09-18", + "last_updated": "2026-09-18", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.075, + "output": 0.5 } }, - "mimo-v2.5-tts-voicedesign": { - "id": "mimo-v2.5-tts-voicedesign", - "name": "MiMo-V2.5-TTS-VoiceDesign", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "mimo-v2.5-tts-voiceclone": { - "id": "mimo-v2.5-tts-voiceclone", - "name": "MiMo-V2.5-TTS-VoiceClone", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } - } - } - }, - "modelscope": { - "id": "modelscope", - "env": [ - "MODELSCOPE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api-inference.modelscope.cn/v1", - "name": "ModelScope", - "doc": "https://modelscope.cn/docs/model-service/API-Inference/intro", - "models": { - "Qwen/Qwen3-30B-A3B-Thinking-2507": { - "id": "Qwen/Qwen3-30B-A3B-Thinking-2507", - "name": "Qwen3 30B A3B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + }, + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-30", - "last_updated": "2025-07-30", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen3-235B-A22B-Thinking-2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "Qwen/Qwen3-Coder-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "name": "Qwen3 Coder 30B A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "anthropic/claude-3-haiku": { + "id": "anthropic/claude-3-haiku", + "name": "Anthropic: Claude 3 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-31", - "last_updated": "2025-07-31", + "release_date": "2024-03-13", + "last_updated": "2024-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 4096 }, "cost": { - "input": 0, - "output": 0 + "input": 0.25, + "output": 1.25, + "cache_read": 0.03, + "cache_write": 0.3 } }, - "Qwen/Qwen3-30B-A3B-Instruct-2507": { - "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen3 30B A3B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Claude Opus 4.8 is Anthropic's most capable generally available model in the Opus family. It supports text, image, and file inputs with text output, with reasoning support and a 1M-token...", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-30", - "last_updated": "2025-07-30", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-07-21", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "ZhipuAI/GLM-4.6": { - "id": "ZhipuAI/GLM-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5", + "description": "Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work. It is particularly strong at end-to-end software tasks, code review and bug finding, visual analysis...", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 98304 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "ZhipuAI/GLM-4.5": { - "id": "ZhipuAI/GLM-4.5", - "name": "GLM-4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "anthropic/claude-fable-5.1": { + "id": "anthropic/claude-fable-5.1", + "name": "Claude Fable 5.1", + "description": "Claude Fable 5.1 improves on Claude Fable 5 across the board, with the biggest gains in agentic coding, long-running agentic workflows, and knowledge work: long code refactors, front-end and visual...", + "family": "claude-fable", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } - } - } - }, - "groq": { - "id": "groq", - "env": [ - "GROQ_API_KEY" - ], - "npm": "@ai-sdk/groq", - "name": "Groq", - "doc": "https://console.groq.com/docs/models", - "models": { - "llama-3.3-70b-versatile": { - "id": "llama-3.3-70b-versatile", - "name": "Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + }, + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.59, - "output": 0.79 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "llama-3.1-8b-instant": { - "id": "llama-3.1-8b-instant", - "name": "Llama 3.1 8B", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", - "attachment": false, - "reasoning": false, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude Fable 5 is a Mythos-class model from Anthropic, built for autonomous knowledge work and coding. It supports text, image, and file inputs with text output, with reasoning support and...", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.08 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "whisper-large-v3-turbo": { - "id": "whisper-large-v3-turbo", - "name": "Whisper Large V3 Turbo", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Anthropic: Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "audio" + "image", + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "whisper-large-v3": { - "id": "whisper-large-v3", - "name": "Whisper", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2023-09-01", - "last_updated": "2025-09-05", + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Sonnet 5 is Anthropic's most capable Sonnet-class model, with frontier performance across coding, agents, and professional work. It supports adaptive thinking with selectable reasoning effort levels (low, medium, high, max,...", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "audio" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "meta-llama/llama-prompt-guard-2-86m": { - "id": "meta-llama/llama-prompt-guard-2-86m", - "name": "Prompt Guard 2 86M", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-05-29", - "last_updated": "2025-05-29", + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512, - "output": 512 + "context": 1000000, + "output": 128000 }, - "status": "beta", "cost": { - "input": 0.04, - "output": 0.04 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "meta-llama/llama-prompt-guard-2-22m": { - "id": "meta-llama/llama-prompt-guard-2-22m", - "name": "Llama Prompt Guard 2 22M", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-05-29", - "last_updated": "2025-05-29", + "anthropic/claude-opus-5.5": { + "id": "anthropic/claude-opus-5.5", + "name": "Claude Opus 5.5", + "description": "Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code...", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512, - "output": 512 + "context": 1000000, + "output": 128000 }, - "status": "beta", "cost": { - "input": 0.03, - "output": 0.03 + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 } }, - "meta-llama/llama-4-scout-17b-16e-instruct": { - "id": "meta-llama/llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout 17B 16E", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + "nex-agi/nex-n2.5-mini": { + "id": "nex-agi/nex-n2.5-mini", + "name": "Nex AGI: Nex-N2.5-Mini", + "description": "Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...", + "family": "agi", "attachment": true, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "medium", + "high" + ] + } + ], + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ "text", @@ -67396,29 +71078,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 262144, + "output": 235929 }, - "status": "beta", "cost": { - "input": 0.11, - "output": 0.34 + "input": 0.025, + "output": 0.1, + "cache_read": 0.0025 } }, - "openai/gpt-oss-safeguard-20b": { - "id": "openai/gpt-oss-safeguard-20b", - "name": "Safety GPT OSS 20B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, + "nex-agi/nex-n2.5-mini:free": { + "id": "nex-agi/nex-n2.5-mini:free", + "name": "Nex AGI: Nex-N2.5-Mini (free)", + "description": "Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...", + "family": "agi", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", + "none", "medium", "high" ] @@ -67427,39 +71109,39 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2026-06-29", + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 262144, + "output": 235929 }, - "status": "beta", "cost": { - "input": 0.075, - "output": 0.3 + "input": 0, + "output": 0 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "nex-agi/nex-n2.5-pro:free": { + "id": "nex-agi/nex-n2.5-pro:free", + "name": "Nex AGI: Nex-N2.5-Pro (free)", + "description": "Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...", + "family": "agi", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", + "none", "medium", "high" ] @@ -67468,39 +71150,39 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-10-21", + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0, + "output": 0 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "nex-agi/nex-n2.5-pro": { + "id": "nex-agi/nex-n2.5-pro", + "name": "Nex AGI: Nex-N2.5-Pro", + "description": "Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...", + "family": "agi", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", + "none", "medium", "high" ] @@ -67509,93 +71191,84 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-09-25", + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 262144, + "output": 235929 }, "cost": { "input": 0.075, - "output": 0.3, - "cache_read": 0.0375 + "output": 0.25, + "cache_read": 0.015 } }, - "canopylabs/orpheus-v1-english": { - "id": "canopylabs/orpheus-v1-english", - "name": "Canopy Labs Orpheus V1 English", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "canopylabs", + "cohere/command-r-08-2024": { + "id": "cohere/command-r-08-2024", + "name": "Command R", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-12-19", - "last_updated": "2025-12-19", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" ], "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 4000, - "output": 50000 - }, - "status": "beta" - }, - "canopylabs/orpheus-arabic-saudi": { - "id": "canopylabs/orpheus-arabic-saudi", - "name": "Canopy Labs Orpheus Arabic Saudi", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "canopylabs", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", - "modalities": { - "input": [ "text" - ], - "output": [ - "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4000, - "output": 50000 + "context": 128000, + "output": 4000 }, - "status": "beta" + "cost": { + "input": 0.15, + "output": 0.6 + } }, - "groq/compound": { - "id": "groq/compound", - "name": "Compound", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "groq", - "attachment": false, - "reasoning": false, - "tool_call": false, + "cohere/command-a-plus": { + "id": "cohere/command-a-plus", + "name": "Cohere: Command A+", + "description": "Command A+ is Cohere's flagship model for enterprise agentic workflows. It accepts text and image inputs with a 192K context window, supports native tool calling with strict tool schemas, structured...", + "family": "command-a", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-09-04", - "last_updated": "2025-09-04", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -67603,21 +71276,27 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 192000, + "output": 64000 + }, + "cost": { + "input": 0.3, + "output": 1.5, + "cache_read": 0.15 } }, - "groq/compound-mini": { - "id": "groq/compound-mini", - "name": "Compound Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "groq", + "cohere/command-a": { + "id": "cohere/command-a", + "name": "Cohere: Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-09-04", - "last_updated": "2025-09-04", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ "text" @@ -67628,15 +71307,19 @@ }, "open_weights": false, "limit": { - "context": 131072, + "context": 256000, "output": 8192 + }, + "cost": { + "input": 2.5, + "output": 10 } }, - "qwen/qwen3-32b": { - "id": "qwen/qwen3-32b", - "name": "Qwen3-32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "cohere/north-mini-code:free": { + "id": "cohere/north-mini-code:free", + "name": "Cohere: North Mini Code (free)", + "description": "North Mini Code is Cohere's first agentic coding model and the debut of its North family. A sparse mixture-of-experts model with 30B total parameters and 3B active, it is optimized...", + "family": "north", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -67644,14 +71327,15 @@ "type": "effort", "values": [ "none", - "default" + "high" ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-06-11", - "last_updated": "2025-06-12", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ "text" @@ -67660,45 +71344,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 40960 + "context": 256000, + "output": 64000 }, - "status": "beta", "cost": { - "input": 0.29, - "output": 0.59 + "input": 0, + "output": 0 } - } - } - }, - "mixlayer": { - "id": "mixlayer", - "env": [ - "MIXLAYER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://models.mixlayer.ai/v1", - "name": "Mixlayer", - "doc": "https://docs.mixlayer.com", - "models": { - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + }, + "cohere/command-r7b-12-2024": { + "id": "cohere/command-r7b-12-2024", + "name": "Command R7B", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-r", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-06-01", + "release_date": "2024-12-02", + "last_updated": "2024-12-02", "modalities": { "input": [ "text" @@ -67709,30 +71377,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.3, - "output": 2.4 + "input": 0.0375, + "output": 0.15 } }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5 35B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "cohere/command-r-plus-08-2024": { + "id": "cohere/command-r-plus-08-2024", + "name": "Command R+", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -67743,30 +71408,35 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.25, - "output": 1.3 + "input": 2.5, + "output": 10 } }, - "qwen/qwen3.5-9b": { - "id": "qwen/qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "deepseek/deepseek-chat-v3.1": { + "id": "deepseek/deepseek-chat-v3.1", + "name": "DeepSeek: DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ "text" @@ -67775,32 +71445,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 163840, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 0.27, + "output": 1, + "cache_read": 0.135 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "deepseek/deepseek-v4-flash-0731": { + "id": "deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -67811,33 +71490,43 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.44, + "output": 1.32, + "cache_read": 0.028 } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B A10B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "deepseek/deepseek-v4.1-flash": { + "id": "deepseek/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the cost-efficient tier of the V4.1 family. DeepSeek reports that it exceeds V4 Pro on performance, speed, and task...", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -67845,180 +71534,109 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 0.4, - "output": 3.2 + "input": 0.3, + "output": 1.2, + "cache_read": 0.006 } - } - } - }, - "orcarouter": { - "id": "orcarouter", - "env": [ - "ORCAROUTER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.orcarouter.ai/v1", - "name": "OrcaRouter", - "doc": "https://docs.orcarouter.ai", - "models": { - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + }, + "deepseek/deepseek-v4-flash-vision-exp": { + "id": "deepseek/deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "DeepSeek V4 Flash Vision Exp is an experimental vision-enabled version of [DeepSeek V4 Flash 0731](https://openrouter.ai/deepseek/deepseek-v4-flash-0731) from DeepSeek, adding image understanding while matching the base model on text capabilities including agents,...", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 943718 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.44, + "output": 1.32, + "cache_read": 0.028 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, + "deepseek/deepseek-r1-distill-llama-70b": { + "id": "deepseek/deepseek-r1-distill-llama-70b", + "name": "DeepSeek: R1 Distill Llama 70B (retires Sep 28)", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2025-01-23", + "last_updated": "2025-01-23", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 - } - }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 8192, + "output": 7372 }, "cost": { - "input": 0.13, - "output": 0.38 + "input": 0.8, + "output": 0.8 } }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "deepseek/deepseek-v3.1-terminus": { + "id": "deepseek/deepseek-v3.1-terminus", + "name": "DeepSeek: DeepSeek V3.1 Terminus (retires Sep 28)", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } @@ -68026,16 +71644,11 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -68043,128 +71656,100 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 4, - "output": 18, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.27, + "output": 1, + "cache_read": 0.135 } }, - "google/gemini-flash-lite-latest": { - "id": "google/gemini-flash-lite-latest", - "name": "Gemini Flash-Lite Latest", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "deepseek/deepseek-chat": { + "id": "deepseek/deepseek-chat", + "name": "DeepSeek Chat", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 163840, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025 + "input": 0.32, + "output": 0.89 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, + "deepseek/deepseek-v4-pro-0813": { + "id": "deepseek/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 1024000, + "output": 384000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, + "deepseek/deepseek-v3.2-exp": { + "id": "deepseek/deepseek-v3.2-exp", + "name": "DeepSeek: DeepSeek V3.2 Exp (retires Sep 28)", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } @@ -68172,16 +71757,11 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -68189,193 +71769,148 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 163840, "output": 65536 }, "cost": { - "input": 4, - "output": 18, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.27, + "output": 0.41 } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "deepseek/deepseek-chat-v3-0324": { + "id": "deepseek/deepseek-chat-v3-0324", + "name": "DeepSeek: DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 163840, + "output": 147456 }, "cost": { - "input": 0.06, - "output": 0.33 + "input": 0.25, + "output": 1 } }, - "google/gemini-3-pro-preview": { - "id": "google/gemini-3-pro-preview", - "name": "Gemini 3 Pro Preview", - "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", - "family": "gemini-pro", - "attachment": true, + "deepseek/deepseek-r1": { + "id": "deepseek/deepseek-r1", + "name": "DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", "high" ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 64000, + "output": 16000 }, "cost": { - "input": 4, - "output": 18, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.7, + "output": 2.5 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "none", + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 1024000, + "output": 384000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 + "input": 1.6, + "output": 3.2, + "cache_read": 0.135 } }, - "google/gemini-flash-latest": { - "id": "google/gemini-flash-latest", - "name": "Gemini Flash Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "deepseek/deepseek-r1-0528": { + "id": "deepseek/deepseek-r1-0528", + "name": "DeepSeek: R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -68383,29 +71918,27 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 163840, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.075, - "input_audio": 1 + "input": 0.7, + "output": 2.5, + "cache_read": 0.35 } }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } @@ -68413,49 +71946,51 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 163840, "output": 65536 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 0.269, + "output": 0.4, + "cache_read": 0.1345 } }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -68466,65 +72001,68 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 1024000, + "output": 384000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "z-ai/glm-4.5": { - "id": "z-ai/glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "unbiased/pareto": { + "id": "unbiased/pareto", + "name": "Pareto", + "description": "Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-09-17", + "last_updated": "2026-09-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 2.5, + "output": 7.5, + "cache_read": 0.25 } }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "tencent/hy3-preview": { + "id": "tencent/hy3-preview", + "name": "Hy3 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text" @@ -68535,29 +72073,36 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 262144, + "output": 235929 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "input": 0.18, + "output": 0.6, + "cache_read": 0.06 } }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", + "tencent/hunyuan-a13b-instruct": { + "id": "tencent/hunyuan-a13b-instruct", + "name": "Tencent: Hunyuan A13B Instruct", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2025-07-08", + "last_updated": "2025-07-08", "modalities": { "input": [ "text" @@ -68566,31 +72111,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 117964 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 0.14, + "output": 0.57 } }, - "z-ai/glm-4.5-air": { - "id": "z-ai/glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "tencent/hy-mt2-30b-a3b": { + "id": "tencent/hy-mt2-30b-a3b", + "name": "Tencent: Hy-MT2-30B-A3B", + "description": "Hy-MT2-30B-A3B is Tencent's flagship translation model in the Hy-MT2 family. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and...", + "family": "Hy", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-08-20", + "last_updated": "2026-08-20", "modalities": { "input": [ "text" @@ -68599,33 +72141,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 8192, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03, - "cache_write": 0 + "input": 0.074, + "output": 0.295 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "tencent/hy-mt2-1.8b": { + "id": "tencent/hy-mt2-1.8b", + "name": "Tencent: Hy-MT2-1.8B", + "description": "Hy-MT2-1.8B is a compact 1.8B-parameter translation model from Tencent. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and style-guided...", + "family": "Hy", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-08-20", + "last_updated": "2026-08-20", "modalities": { "input": [ "text" @@ -68634,109 +72171,70 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 8192, + "output": 4096 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0 + "input": 0.044, + "output": 0.177 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2 Pro", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", - "attachment": true, + "tencent/hy3": { + "id": "tencent/hy3", + "name": "Hy3", + "description": "Hy3 is a 295B-parameter Mixture-of-Experts model from Tencent (21B active, 192 experts with top-8 routing) built for reasoning, agentic workflows, and real-world production use. It supports a configurable reasoning effort:...", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 21, - "output": 168 - } - }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, + "context": 262144, + "input": 192000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.0825, + "output": 0.33, + "cache_read": 0.020625 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "tencent/hy-mt2-7b": { + "id": "tencent/hy-mt2-7b", + "name": "Tencent: Hy-MT2-7B", + "description": "Hy-MT2-7B is a 7B-parameter translation model from Tencent. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and style-guided translation.", + "family": "Hy", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "release_date": "2026-08-19", + "last_updated": "2026-08-19", "modalities": { "input": [ "text" @@ -68747,76 +72245,69 @@ }, "open_weights": false, "limit": { - "context": 16385, + "context": 8192, "output": 4096 }, "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0 + "input": 0.074, + "output": 0.295 } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", - "attachment": true, + "tencent/hy4-preview": { + "id": "tencent/hy4-preview", + "name": "Hy4 preview", + "description": "Tencent: Hy4 preview is a mixture-of-experts model from Tencent, with 49B active parameters out of 770B total. It is designed for coding agents, complex tool-use workflows, and productivity tasks that...", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "temperature": true, + "release_date": "2026-08-28", + "last_updated": "2026-08-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 272000 + "context": 1048576, + "output": 64000 }, "cost": { - "input": 15, - "output": 120 + "input": 0.834, + "output": 2.501, + "cache_read": 0.042 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, + "anthracite-org/magnum-v4-72b": { + "id": "anthracite-org/magnum-v4-72b", + "name": "Magnum v4 72B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -68824,31 +72315,30 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 32768, + "output": 4096 }, "cost": { "input": 2.5, - "output": 10, - "cache_read": 1.25 + "output": 5 } }, - "openai/gpt-4": { - "id": "openai/gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "meta-llama/llama-4-scout": { + "id": "meta-llama/llama-4-scout", + "name": "Meta: Llama 4 Scout", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -68856,31 +72346,30 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 8192 - }, + "context": 327680, + "output": 16384 + }, "cost": { - "input": 30, - "output": 60 + "input": 0.1, + "output": 0.3 } }, - "openai/gpt-4o-2024-05-13": { - "id": "openai/gpt-4o-2024-05-13", - "name": "GPT-4o (2024-05-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "meta-llama/llama-guard-4-12b": { + "id": "meta-llama/llama-guard-4-12b", + "name": "Meta: Llama Guard 4 12B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "release_date": "2025-04-30", + "last_updated": "2025-04-30", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" @@ -68888,28 +72377,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 163840, + "output": 16384 }, "cost": { - "input": 5, - "output": 15 + "input": 0.18, + "output": 0.18 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "meta-llama/llama-4-maverick": { + "id": "meta-llama/llama-4-maverick", + "name": "Meta: Llama 4 Maverick", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -68921,69 +72408,91 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.1875, + "output": 0.6525 } }, - "openai/gpt-5-chat-latest": { - "id": "openai/gpt-5-chat-latest", - "name": "GPT-5 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "meta-llama/llama-3.3-70b-instruct": { + "id": "meta-llama/llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.1, + "output": 0.32 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "meta-llama/llama-3.1-8b-instruct": { + "id": "meta-llama/llama-3.1-8b-instruct", + "name": "Llama-3.1-8B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image" + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 117964 + }, + "cost": { + "input": 0.02, + "output": 0.04 + } + }, + "meta-llama/llama-3.2-1b-instruct": { + "id": "meta-llama/llama-3.2-1b-instruct", + "name": "Meta: Llama 3.2 1B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2024-09-25", + "last_updated": "2024-09-25", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -68991,34 +72500,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 60000, + "output": 54000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.027, + "output": 0.201 } }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "meta-llama/llama-3.2-3b-instruct": { + "id": "meta-llama/llama-3.2-3b-instruct", + "name": "Meta: Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -69026,66 +72530,74 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 117964 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.05, + "output": 0.33 } }, - "openai/gpt-5.3-chat-latest": { - "id": "openai/gpt-5.3-chat-latest", - "name": "GPT-5.3 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "meta-llama/llama-3.1-70b-instruct": { + "id": "meta-llama/llama-3.1-70b-instruct", + "name": "Llama-3.1-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 131072, "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.4, + "output": 0.4 } }, - "openai/gpt-4o-2024-08-06": { - "id": "openai/gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "~google/gemini-flash-latest": { + "id": "~google/gemini-flash-latest", + "name": "Google: Gemini Flash Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-08-06", - "last_updated": "2024-08-06", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf", + "audio" ], "output": [ "text" @@ -69093,33 +72605,46 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "~google/gemini-pro-latest": { + "id": "~google/gemini-pro-latest", + "name": "Google: Gemini Pro Latest", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ + "audio", + "pdf", + "image", "text", - "image" + "video" ], "output": [ "text" @@ -69127,35 +72652,42 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "perceptron/perceptron-mk1": { + "id": "perceptron/perceptron-mk1", + "name": "Perceptron: Perceptron Mk1", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -69163,103 +72695,124 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 32768, + "output": 8192 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.15, + "output": 1.5 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "z-ai/glm-4.6v": { + "id": "z-ai/glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.3, + "output": 0.9, + "cache_read": 0.055 } }, - "openai/gpt-5.1-chat-latest": { - "id": "openai/gpt-5.1-chat-latest", - "name": "GPT-5.1 Chat", - "description": "Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations", - "family": "gpt-codex", - "attachment": true, + "z-ai/glm-4.5": { + "id": "z-ai/glm-4.5", + "name": "GLM-4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 98304 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "openai/gpt-5.2-chat-latest": { - "id": "openai/gpt-5.2-chat-latest", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "z-ai/glm-5v-turbo": { + "id": "z-ai/glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "structured_output": false, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ + "image", "text", - "image" + "video" ], "output": [ "text" @@ -69267,163 +72820,161 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 202752, + "output": 131072 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", + "z-ai/glm-5.3-flash": { + "id": "z-ai/glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks. Its hybrid sparse and linear attention architecture maintains accurate long-context behavior while...", + "family": "glm-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "openai/gpt-4o-2024-11-20": { - "id": "openai/gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 198000, "output": 16384 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.43, + "output": 1.75, + "cache_read": 0.08 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, + "context": 198000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, "cost": { - "input": 5, - "output": 22.5, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0.6, + "output": 1.92, + "cache_read": 0.12 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "z-ai/glm-4.5v": { + "id": "z-ai/glm-4.5v", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ "text", @@ -69433,52 +72984,44 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 65536, + "output": 16384 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.6, + "output": 1.8, + "cache_read": 0.11 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", + "z-ai/glm-5.3-flashx": { + "id": "z-ai/glm-5.3-flashx", + "name": "Z.ai: GLM 5.3 FlashX", + "description": "GLM-5.3-FlashX is the high-speed variant of Z.ai's GLM-5.3-Flash, a native multimodal model delivering inference speeds of up to 200 tokens/s. Built on the same hybrid sparse and linear attention architecture...", + "family": "glm", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-09-18", + "last_updated": "2026-09-18", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -69486,178 +73029,201 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.37, + "output": 1.25, + "cache_read": 0.075 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, + "z-ai/glm-4.7-flash": { + "id": "z-ai/glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 117964 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.0605, + "output": 0.4 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 202752, + "output": 131072 }, "cost": { "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "output": 1.75, + "cache_read": 0.08 } }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "GLM 5.2 is a large-scale reasoning model from Z.ai. It supports text input and output with a 1M-token context window, and is suited for long-horizon agent workflows, project-level software engineering,...", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 10, - "output": 30 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, + "context": 200000, "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", - "attachment": true, + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -69665,175 +73231,151 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 60, - "output": 270, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, + "z-ai/glm-5.2:free": { + "id": "z-ai/glm-5.2:free", + "name": "Z.ai: GLM 5.2 (free)", + "description": "GLM 5.2 is a large-scale reasoning model from Z.ai. It supports text input and output with a 1M-token context window, and is suited for long-horizon agent workflows, project-level software engineering,...", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "none", + "high", + "xhigh" ] } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 32768, + "output": 29491 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 0, + "output": 0 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + "z-ai/glm-4.5-air": { + "id": "z-ai/glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.13, + "output": 0.85, + "cache_read": 0.025 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "z-ai/glm-5.3": { + "id": "z-ai/glm-5.3", + "name": "GLM-5.3", + "description": "GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks. It supports text input and output with a 1M-token context window, and improves...", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "inference-net/schematron-v2-small": { + "id": "inference-net/schematron-v2-small", + "name": "Inference.net: Schematron V2 Small", + "description": "Schematron V2 Small is a 3B-parameter HTML-to-JSON extraction model from Inference.net. It prioritizes extraction quality for complex schemas and long pages. Extraction instructions must be supplied through a JSON schema...", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-09-12", + "last_updated": "2026-09-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -69841,34 +73383,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.05, + "output": 0.23, + "cache_read": 0.05 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "inference-net/schematron-v2-turbo": { + "id": "inference-net/schematron-v2-turbo", + "name": "Inference.net: Schematron V2 Turbo", + "description": "Schematron V2 Turbo is a 3B-parameter HTML-to-JSON extraction model from Inference.net. It prioritizes throughput for high-volume extraction workloads. Extraction instructions must be supplied through a JSON schema in response_format rather...", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2026-09-12", + "last_updated": "2026-09-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -69876,106 +73413,90 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.03, + "output": 0.15, + "cache_read": 0.03 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "thinkingmachines/inkling-small:free": { + "id": "thinkingmachines/inkling-small:free", + "name": "Thinking Machines: Inkling Small (free)", + "description": "Inkling Small is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 12B active parameters out of 276B total. It is positioned as the smaller, more efficient member of...", + "family": "ling", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 1048576, + "output": 262144 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0, + "output": 0 } }, - "kimi/kimi-k2.5": { - "id": "kimi/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": false, + "thinkingmachines/inkling-small": { + "id": "thinkingmachines/inkling-small", + "name": "Inkling Small", + "description": "Inkling Small is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 12B active parameters out of 276B total. It is positioned as the smaller, more efficient member of...", + "family": "ling", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", "image", - "video" + "audio" ], "output": [ "text" @@ -69983,37 +73504,45 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 524288, "output": 262144 }, "cost": { - "input": 0.6, - "output": 3, + "input": 0.45, + "output": 1.2, "cache_read": 0.1 } }, - "kimi/kimi-k2.6": { - "id": "kimi/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "thinkingmachines/inkling": { + "id": "thinkingmachines/inkling", + "name": "Inkling", + "description": "Inkling is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 41B active parameters out of 975B total. It is designed for general-purpose reasoning, coding, agentic and tool-use systems,...", + "family": "ling", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", "image", - "video" + "audio" ], "output": [ "text" @@ -70021,31 +73550,67 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 524288, + "output": 471859 }, "cost": { "input": 0.95, - "output": 4, + "output": 4.05, "cache_read": 0.16 } }, - "orcarouter/auto": { - "id": "orcarouter/auto", - "name": "OrcaRouter Auto", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": true, - "reasoning": false, + "meituan/longcat-2.0": { + "id": "meituan/longcat-2.0", + "name": "Meituan: LongCat 2.0", + "description": "LongCat 2.0 is a sparse mixture-of-experts language model from Meituan, with 48B active parameters out of 1.6T total. It is suited for coding, repository-level changes, long-horizon problem solving, and agentic...", + "family": "longcat", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2026-05-14", + "release_date": "2026-07-20", + "last_updated": "2026-07-20", "modalities": { "input": [ - "text", - "image" + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048756, + "output": 262144 + }, + "cost": { + "input": 0.75, + "output": 3, + "cache_read": 0.015 + } + }, + "openrouter/bodybuilder": { + "id": "openrouter/bodybuilder", + "name": "Body Builder (beta)", + "description": "Preview model for early access evaluation, prototyping, and compatibility testing", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-03-15", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -70054,31 +73619,30 @@ "open_weights": false, "limit": { "context": 128000, - "output": 16384 + "output": 32768 }, + "status": "beta", "cost": { "input": 0, "output": 0 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "openrouter/free": { + "id": "openrouter/free", + "name": "OpenRouter Free Models Router", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -70087,68 +73651,91 @@ "open_weights": false, "limit": { "context": 200000, - "output": 64000 + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude Sonnet 4 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "openrouter/pareto-code": { + "id": "openrouter/pareto-code", + "name": "Pareto Code Router", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-05-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "openrouter/auto": { + "id": "openrouter/auto", + "name": "Auto Router", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-03-15", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", + "audio", "image", - "pdf" + "pdf", + "text", + "video" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 2000000, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "perplexity/sonar-pro": { + "id": "perplexity/sonar-pro", + "name": "Perplexity: Sonar Pro", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2025-03-07", + "last_updated": "2025-03-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -70157,42 +73744,37 @@ "open_weights": false, "limit": { "context": 200000, - "output": 64000 + "output": 8000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 3, + "output": 15 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "perplexity/sonar-deep-research": { + "id": "perplexity/sonar-deep-research", + "name": "Perplexity: Sonar Deep Research", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar-deep-research", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-03-07", + "last_updated": "2025-03-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -70200,54 +73782,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "context": 128000, + "output": 115200 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2, + "output": 8, + "reasoning": 3 } }, - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "perplexity/sonar": { + "id": "perplexity/sonar", + "name": "Perplexity: Sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -70255,34 +73814,39 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 127072, + "output": 114364 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 1, + "output": 1 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "perplexity/sonar-reasoning-pro": { + "id": "perplexity/sonar-reasoning-pro", + "name": "Perplexity: Sonar Reasoning Pro", + "description": "Web-grounded reasoning model for multi-step research and cited answers", + "family": "sonar-reasoning", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2025-03-07", + "last_updated": "2025-03-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -70290,43 +73854,38 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 115200 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2, + "output": 8 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "perplexity/sonar-pro-search": { + "id": "perplexity/sonar-pro-search", + "name": "Perplexity: Sonar Pro Search", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", "high" ] } ], - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2025-10-30", + "last_updated": "2025-10-30", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -70334,34 +73893,46 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 200000, + "output": 8000 }, "cost": { "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "output": 15 } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Claude Opus 4 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "meta/muse-spark-1.3": { + "id": "meta/muse-spark-1.3", + "name": "Muse Spark 1.3", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It is designed to keep track of information across extended tasks, work through...", + "family": "muse", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "pdf", + "audio" ], "output": [ "text" @@ -70369,21 +73940,20 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "meta/muse-glimmer-30b": { + "id": "meta/muse-glimmer-30b", + "name": "Muse Glimmer 30B", + "description": "Muse Glimmer 30B is a dense, open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark and optimized for autonomous agents on consumer hardware. It is suited for long-horizon...", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -70392,15 +73962,17 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", "modalities": { "input": [ "text", @@ -70411,56 +73983,48 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "context": 131072, + "output": 16384 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.3, + "output": 1.1, + "cache_read": 0.04 } }, - "grok/grok-4.3": { - "id": "grok/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "meta/muse-spark-1.1": { + "id": "meta/muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Muse Spark 1.1 is a multimodal reasoning model from Meta, built for agentic tasks. It accepts text, images, video, audio, and PDF documents and returns text, with a 1M-token context...", + "family": "muse", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-04-08", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "pdf", + "audio" ], "output": [ "text" @@ -70468,80 +74032,93 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 1048576, + "output": 943718 }, "cost": { "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "output": 4.25, + "cache_read": 0.15 } }, - "qwen/qwen3.6-35b-a3b": { - "id": "qwen/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "meta/muse-spark-1.2": { + "id": "meta/muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a reasoning model from Meta, designed for complex agentic tasks. It accepts text, images, video, audio, and PDF documents, returns text, and offers a 1M-token context...", + "family": "muse", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ "text", "image", "video", + "pdf", "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 0.248, - "output": 1.485 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "meta/muse-spark-1.2-contributor": { + "id": "meta/muse-spark-1.2-contributor", + "name": "Meta: Muse Spark 1.2 Contributor", + "description": "Muse Spark 1.2 contributor tier is a reasoning model from Meta designed for developers who want to start building at an even lower cost. It’s meaningfully cheaper than Muse Spark...", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-08-21", + "last_updated": "2026-08-21", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf", + "audio" ], "output": [ "text" @@ -70549,32 +74126,47 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 0.359, - "output": 1.434 + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 } }, - "qwen/qwen3.5-plus": { - "id": "qwen/qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "meta/muse-spark-1.3-contributor": { + "id": "meta/muse-spark-1.3-contributor", + "name": "Meta: Muse Spark 1.3 Contributor", + "description": "Muse Spark 1.3 Contributor is the cost-efficient contributor tier of Meta’s multimodal reasoning model for experimentation, learning, and early-stage agentic, multi-agent, and coding workflows. It is designed to track information...", + "family": "muse", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", "image", - "video" + "video", + "pdf", + "audio" ], "output": [ "text" @@ -70582,169 +74174,182 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 0.115, - "output": 0.688, - "reasoning": 2.4 + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 } }, - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "nousresearch/hermes-3-llama-3.1-70b": { + "id": "nousresearch/hermes-3-llama-3.1-70b", + "name": "Nous: Hermes 3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "nousresearch", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2024-08-18", + "last_updated": "2024-08-18", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.086, - "output": 0.688 + "input": 0.7, + "output": 0.7 } }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "nousresearch/hermes-3-llama-3.1-405b": { + "id": "nousresearch/hermes-3-llama-3.1-405b", + "name": "Nous: Hermes 3 405B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "nousresearch", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2024-08-16", + "last_updated": "2024-08-16", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.057, - "output": 0.459 + "input": 1, + "output": 1 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, + "nousresearch/hermes-4-405b": { + "id": "nousresearch/hermes-4-405b", + "name": "Nous: Hermes 4 405B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "nousresearch", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "output": 117964 }, "cost": { - "input": 0.172, - "output": 1.032 + "input": 1, + "output": 3 } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "~z-ai/glm-flash-latest": { + "id": "~z-ai/glm-flash-latest", + "name": "Z.ai: GLM Flash Latest", + "description": "This model always redirects to the latest model in the GLM Flash family.", + "family": "glm-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ "text", "image", - "video", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.115, - "output": 0.917 + "input": 0.075, + "output": 0.25, + "cache_read": 0.015 } }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", + "~z-ai/glm-latest": { + "id": "~z-ai/glm-latest", + "name": "Z.ai: GLM Latest", + "description": "This model always redirects to the latest GLM model from Z.ai.", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-08-19", + "last_updated": "2026-08-19", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -70752,51 +74357,35 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 0.5614, + "output": 1.7644, + "cache_read": 0.10426 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "arcee-ai/trinity-large-thinking": { + "id": "arcee-ai/trinity-large-thinking", + "name": "Trinity Large Thinking", + "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", + "family": "trinity", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-04-01", + "last_updated": "2026-05-28", "modalities": { "input": [ "text" @@ -70807,193 +74396,259 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 262144, + "output": 80000 }, "cost": { - "input": 0.19, - "output": 0.37, - "cache_read": 0.0028 + "input": 0.25, + "output": 0.8, + "cache_read": 0.06 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "~openai/gpt-terra-latest": { + "id": "~openai/gpt-terra-latest", + "name": "OpenAI: GPT Terra Latest", + "description": "This model always redirects to the latest model in the OpenAI GPT Terra family.", + "family": "gpt-terra", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0.56, - "output": 1.12, - "cache_read": 0.003625 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "deepseek/deepseek-reasoner": { - "id": "deepseek/deepseek-reasoner", - "name": "DeepSeek Reasoner", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "~openai/gpt-luna-latest": { + "id": "~openai/gpt-luna-latest", + "name": "OpenAI: GPT Luna Latest", + "description": "This model always redirects to the latest model in the OpenAI GPT Luna family.", + "family": "gpt-luna", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "structured_output": true, + "temperature": false, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.028 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125 } }, - "deepseek/deepseek-chat": { - "id": "deepseek/deepseek-chat", - "name": "DeepSeek Chat", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "~openai/gpt-sol-latest": { + "id": "~openai/gpt-sol-latest", + "name": "OpenAI: GPT Sol Latest", + "description": "This model always redirects to the latest model in the OpenAI GPT Sol family.", + "family": "gpt-sol", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "structured_output": true, + "temperature": false, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", - "attachment": false, + "~openai/gpt-astra-latest": { + "id": "~openai/gpt-astra-latest", + "name": "OpenAI: GPT Astra Latest ($$$$)", + "description": "This model always redirects to the latest model in the OpenAI GPT Astra family.", + "family": "gpt-astra", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": true, + "temperature": false, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, + "~openai/gpt-mini-latest": { + "id": "~openai/gpt-mini-latest", + "name": "OpenAI: GPT Mini Latest", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "structured_output": true, + "temperature": false, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "cognitivecomputations/dolphin-mistral-24b-venice-edition": { + "id": "cognitivecomputations/dolphin-mistral-24b-venice-edition", + "name": "Venice: Uncensored", + "description": "Venice Uncensored Dolphin Mistral 24B Venice Edition is a fine-tuned variant of Mistral-Small-24B-Instruct-2501, developed by dphn.ai in collaboration with Venice.ai. This model is designed as an “uncensored” instruct-tuned LLM, preserving...", + "family": "mistral", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text" @@ -71002,30 +74657,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.2, + "output": 0.9 } }, - "minimax/minimax-m2.5-highspeed": { - "id": "minimax/minimax-m2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "thedrummer/skyfall-36b-v2": { + "id": "thedrummer/skyfall-36b-v2", + "name": "TheDrummer: Skyfall 36B V2", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "release_date": "2025-03-10", + "last_updated": "2025-03-10", "modalities": { "input": [ "text" @@ -71034,46 +74686,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 32768, + "output": 29491 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.55, + "output": 0.8, + "cache_read": 0.25 } - } - } - }, - "helicone": { - "id": "helicone", - "env": [ - "HELICONE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://ai-gateway.helicone.ai/v1", - "name": "Helicone", - "doc": "https://helicone.ai/models", - "models": { - "chatgpt-4o-latest": { - "id": "chatgpt-4o-latest", - "name": "OpenAI ChatGPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + }, + "thedrummer/unslopnemo-12b": { + "id": "thedrummer/unslopnemo-12b", + "name": "TheDrummer: UnslopNemo 12B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2024-08-14", - "last_updated": "2024-08-14", + "release_date": "2024-11-08", + "last_updated": "2024-11-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -71081,31 +74718,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1024000, + "output": 819200 }, "cost": { - "input": 5, - "output": 20, - "cache_read": 2.5 + "input": 0.4, + "output": 0.4 } }, - "gpt-4.1-mini-2025-04-14": { - "id": "gpt-4.1-mini-2025-04-14", - "name": "OpenAI GPT-4.1 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "thedrummer/cydonia-24b-v4.1": { + "id": "thedrummer/cydonia-24b-v4.1", + "name": "TheDrummer: Cydonia 24B V4.1", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-09-27", + "last_updated": "2025-09-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -71113,30 +74747,39 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 131072, + "output": 117964 }, "cost": { - "input": 0.39999999999999997, - "output": 1.5999999999999999, - "cache_read": 0.09999999999999999 + "input": 0.3, + "output": 0.5, + "cache_read": 0.15 } }, - "deepseek-v3.1-terminus": { - "id": "deepseek-v3.1-terminus", - "name": "DeepSeek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "baidu/ernie-4.5-vl-424b-a47b": { + "id": "baidu/ernie-4.5-vl-424b-a47b", + "name": "Baidu: ERNIE 4.5 VL 424B A47B (retires Oct 8)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ernie", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ + "image", "text" ], "output": [ @@ -71145,31 +74788,42 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 123000, + "output": 16000 }, "cost": { - "input": 0.27, - "output": 1, - "cache_read": 0.21600000000000003 + "input": 0.42, + "output": 1.25 } }, - "claude-3.5-haiku": { - "id": "claude-3.5-haiku", - "name": "Anthropic: Claude 3.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": false, - "reasoning": false, - "tool_call": true, + "x-ai/grok-4.20-multi-agent": { + "id": "x-ai/grok-4.20-multi-agent", + "name": "SpaceXAI: Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -71177,31 +74831,44 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 2000000, + "output": 1800000 }, "cost": { - "input": 0.7999999999999999, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "llama-3.1-8b-instruct": { - "id": "llama-3.1-8b-instruct", - "name": "Meta Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "x-ai/grok-4.7": { + "id": "x-ai/grok-4.7", + "name": "Grok 4.7", + "description": "Grok 4.7 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -71209,30 +74876,43 @@ }, "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 500000, + "output": 450000 }, "cost": { - "input": 0.02, - "output": 0.049999999999999996 + "input": 1.6, + "output": 4.8, + "cache_read": 0.4 } }, - "o3": { - "id": "o3", - "name": "OpenAI o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": false, - "reasoning": false, + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "Grok 4.3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -71240,30 +74920,42 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 900000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "llama-prompt-guard-2-86m": { - "id": "llama-prompt-guard-2-86m", - "name": "Meta Llama Prompt Guard 2 86M", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + "x-ai/grok-4.5": { + "id": "x-ai/grok-4.5", + "name": "Grok 4.5", + "description": "Grok 4.5 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-01", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -71271,29 +74963,40 @@ }, "open_weights": false, "limit": { - "context": 512, - "output": 2 + "context": 500000, + "output": 450000 }, "cost": { - "input": 0.01, - "output": 0.01 + "input": 2, + "output": 6, + "cache_read": 0.3 } }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3 Coder 30B A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "x-ai/grok-build-0.1": { + "id": "x-ai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Grok coding model for agentic engineering, edits, and codebase workflows", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-31", - "last_updated": "2025-07-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -71301,29 +75004,41 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 230400 }, "cost": { - "input": 0.09999999999999999, - "output": 0.3 + "input": 1, + "output": 2, + "cache_read": 0.2 } }, - "hermes-2-pro-llama-3-8b": { - "id": "hermes-2-pro-llama-3-8b", - "name": "Hermes 2 Pro Llama 3 8B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "x-ai/grok-4.20": { + "id": "x-ai/grok-4.20", + "name": "SpaceXAI: Grok 4.20", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-05", - "release_date": "2024-05-27", - "last_updated": "2024-05-27", + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -71331,29 +75046,44 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 2000000, + "output": 1800000 }, "cost": { - "input": 0.14, - "output": 0.14 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "deepseek-v3": { - "id": "deepseek-v3", - "name": "DeepSeek V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "x-ai/grok-4.6": { + "id": "x-ai/grok-4.6", + "name": "Grok 4.6", + "description": "Grok 4.6 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-26", - "last_updated": "2024-12-26", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -71361,30 +75091,44 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 500000, + "output": 450000 }, "cost": { - "input": 0.56, - "output": 1.68, - "cache_read": 0.07 + "input": 2, + "output": 6, + "cache_read": 0.5 } }, - "grok-code-fast-1": { - "id": "grok-code-fast-1", - "name": "xAI Grok Code Fast 1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": false, - "reasoning": false, + "~anthropic/claude-opus-latest": { + "id": "~anthropic/claude-opus-latest", + "name": "Anthropic: Claude Opus Latest", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2024-08-25", - "last_updated": "2024-08-25", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -71392,30 +75136,42 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 10000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.19999999999999998, - "output": 1.5, - "cache_read": 0.02 + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 } }, - "o1-mini": { - "id": "o1-mini", - "name": "OpenAI: o1-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "~anthropic/claude-haiku-latest": { + "id": "~anthropic/claude-haiku-latest", + "name": "Anthropic: Claude Haiku Latest", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -71423,31 +75179,47 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "deepseek-r1-distill-llama-70b": { - "id": "deepseek-r1-distill-llama-70b", - "name": "DeepSeek R1 Distill Llama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "~anthropic/claude-sonnet-latest": { + "id": "~anthropic/claude-sonnet-latest", + "name": "Anthropic: Claude Sonnet Latest", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -71455,30 +75227,45 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.03, - "output": 0.13 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "~anthropic/claude-fable-latest": { + "id": "~anthropic/claude-fable-latest", + "name": "Anthropic: Claude Fable Latest ($$$$)", + "description": "This model always redirects to the latest model in the Claude Fable family.", + "family": "claude-fable", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "structured_output": true, + "temperature": false, + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -71486,40 +75273,43 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 40960 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.29, - "output": 0.59 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "claude-sonnet-4": { - "id": "claude-sonnet-4", - "name": "Anthropic: Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "upstage/solar-pro-3": { + "id": "upstage/solar-pro-3", + "name": "Upstage: Solar Pro 3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "solar-pro", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-14", - "last_updated": "2025-05-14", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -71527,33 +75317,44 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 117964 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.30000000000000004, - "cache_write": 3.75 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "upstage/solar-pro4": { + "id": "upstage/solar-pro4", + "name": "Upstage: Solar Pro 4", + "description": "Solar Pro 4 is a large language model from Upstage. It is suited for agentic workflows, office productivity, document-intensive work, and coding.", + "family": "solar", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -71561,30 +75362,44 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 16384 + "context": 524288, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 1.4 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "llama-4-maverick": { - "id": "llama-4-maverick", - "name": "Meta Llama 4 Maverick 17B 128E", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "upstage/solar-mini4": { + "id": "upstage/solar-mini4", + "name": "Upstage: Solar Mini 4", + "description": "Solar Mini 4 is Upstage's compact, cost-efficient language model, a 35B-parameter mixture-of-experts with 3B active parameters and a 524K context window. It is built for agentic use cases where response...", + "family": "solar", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -71592,26 +75407,38 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 524288, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 } }, - "mistral-nemo": { - "id": "mistral-nemo", - "name": "Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": false, + "~deepseek/deepseek-flash-latest": { + "id": "~deepseek/deepseek-flash-latest", + "name": "DeepSeek: DeepSeek Flash Latest", + "description": "This model always redirects to the latest model in the DeepSeek Flash family.", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-09-14", + "last_updated": "2026-09-14", "modalities": { "input": [ "text", @@ -71623,26 +75450,38 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16400 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 20, - "output": 40 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01 } }, - "llama-3.3-70b-versatile": { - "id": "llama-3.3-70b-versatile", - "name": "Meta Llama 3.3 70B Versatile", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "~deepseek/deepseek-pro-latest": { + "id": "~deepseek/deepseek-pro-latest", + "name": "DeepSeek: DeepSeek Pro Latest", + "description": "This model always redirects to the latest model in the DeepSeek Pro family.", + "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2026-09-14", + "last_updated": "2026-09-14", "modalities": { "input": [ "text" @@ -71653,26 +75492,38 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32678 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.59, - "output": 0.7899999999999999 + "input": 0.39996, + "output": 1.19988, + "cache_read": 0.012726 } }, - "gemma2-9b-it": { - "id": "gemma2-9b-it", - "name": "Google Gemma 2", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "~deepseek/deepseek-v4-flash-latest": { + "id": "~deepseek/deepseek-v4-flash-latest", + "name": "DeepSeek: DeepSeek V4 Flash Latest", + "description": "This model always redirects to the latest model in the DeepSeek V4 Flash family.", + "family": "deepseek", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-06-25", - "last_updated": "2024-06-25", + "release_date": "2026-08-01", + "last_updated": "2026-08-01", "modalities": { "input": [ "text" @@ -71683,37 +75534,43 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 0.01, - "output": 0.03 + "input": 0.038, + "output": 0.55, + "cache_read": 0.0228 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Google Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": false, + "~x-ai/grok-latest": { + "id": "~x-ai/grok-latest", + "name": "xAI: Grok Latest", + "description": "This model always redirects to the latest Grok model from xAI.", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -71721,32 +75578,30 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 500000, + "output": 450000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.3125, - "cache_write": 1.25 + "input": 1.6, + "output": 4.8, + "cache_read": 0.4 } }, - "gpt-5": { - "id": "gpt-5", - "name": "OpenAI GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, + "bytedance/ui-tars-1.5-7b": { + "id": "bytedance/ui-tars-1.5-7b", + "name": "ByteDance: UI-TARS 7B ", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" @@ -71754,31 +75609,38 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 128000, + "output": 2048 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 + "input": 0.1, + "output": 0.2, + "cache_read": 0.1 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Anthropic: Claude 4.5 Haiku (20251001)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "liquid/lfm-2.5-2.6b:free": { + "id": "liquid/lfm-2.5-2.6b:free", + "name": "LiquidAI: LFM2.5-2.6B (free)", + "description": "LFM2.5-2.6B is a compact reasoning model from Liquid AI. It is suited for agent workflows, data extraction, RAG, and long-context processing. Liquid advises against using it for agentic coding or...", + "family": "liquid", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-10", - "release_date": "2025-10-01", - "last_updated": "2025-10-01", + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -71786,134 +75648,145 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 65536, "output": 8192 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.09999999999999999, - "cache_write": 1.25 + "input": 0, + "output": 0 } }, - "claude-4.5-haiku": { - "id": "claude-4.5-haiku", - "name": "Anthropic: Claude 4.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": false, - "reasoning": false, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-10", - "release_date": "2025-10-01", - "last_updated": "2025-10-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ + "image", "text", - "image" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 + "context": 262144, + "output": 16384 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.09999999999999999, - "cache_write": 1.25 + "input": 0.09, + "output": 0.34, + "cache_read": 0.05 } }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "OpenAI: GPT-5 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", - "attachment": false, - "reasoning": false, + "google/gemini-3.1-flash-image-preview": { + "id": "google/gemini-3.1-flash-image-preview", + "name": "Nano Banana 2 Preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "high" + ] + } + ], "tool_call": false, - "temperature": false, + "structured_output": true, + "temperature": true, "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 65536, + "output": 58982 }, "cost": { - "input": 15, - "output": 120 + "input": 0.5, + "output": 3 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Google Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], - "tool_call": true, + "google/lyria-3-pro-preview": { + "id": "google/lyria-3-pro-preview", + "name": "Lyria 3 Pro Preview", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "lyria", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-03-25", + "last_updated": "2026-03-25", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { "context": 1048576, - "output": 65535 + "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075, - "cache_write": 0.3 + "input": 0, + "output": 0 } }, - "gpt-4o": { - "id": "gpt-4o", - "name": "OpenAI GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma 3 27B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-05", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text", @@ -71923,61 +75796,64 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 117964 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.08, + "output": 0.16 } }, - "o4-mini": { - "id": "o4-mini", - "name": "OpenAI o4 Mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, + "google/gemini-2.5-flash-image": { + "id": "google/gemini-2.5-flash-image", + "name": "Nano Banana", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": false, + "tool_call": false, + "structured_output": true, + "temperature": true, "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 32768, + "output": 8192 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333 } }, - "llama-3.1-8b-instant": { - "id": "llama-3.1-8b-instant", - "name": "Meta Llama 3.1 8B Instant", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", + "google/gemma-2-27b-it": { + "id": "google/gemma-2-27b-it", + "name": "Google: Gemma 2 27B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "release_date": "2024-07-13", + "last_updated": "2024-07-13", "modalities": { "input": [ "text" @@ -71988,29 +75864,45 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32678 + "context": 8192, + "output": 2048 }, "cost": { - "input": 0.049999999999999996, - "output": 0.08 + "input": 0.65, + "output": 0.65 } }, - "llama-prompt-guard-2-22m": { - "id": "llama-prompt-guard-2-22m", - "name": "Meta Llama Prompt Guard 2 22M", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + "google/gemini-3.6-flash": { + "id": "google/gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Gemini 3.6 Flash is a high-efficiency model from Google for coding, agentic workflows, and web and app development. It is designed to produce polished outputs with fewer unnecessary edits and...", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-01", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -72018,102 +75910,134 @@ }, "open_weights": false, "limit": { - "context": 512, - "output": 2 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.01, - "output": 0.01 + "input": 0.375, + "output": 1.875, + "reasoning": 1.875, + "cache_read": 0.0375, + "cache_write": 0.020833 } }, - "o3-pro": { - "id": "o3-pro", - "name": "OpenAI o3 Pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": false, - "reasoning": false, + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ + "image", "text", - "image" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 20, - "output": 80 + "input": 0.042, + "output": 0.22 } }, - "grok-3-mini": { - "id": "grok-3-mini", - "name": "xAI Grok 3 Mini", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": true, + "google/gemini-3-pro-image-preview": { + "id": "google/gemini-3-pro-image-preview", + "name": "Nano Banana Pro Preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "knowledge": "2025-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 65536, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 0.5, - "cache_read": 0.075 + "input": 1, + "output": 6, + "reasoning": 6, + "cache_read": 0.1, + "cache_write": 0.1875 } }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "Anthropic: Claude Opus 4.1 (20250805)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": false, + "google/gemini-3.5-flash-lite": { + "id": "google/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Gemini 3.5 Flash Lite is a high-efficiency model from Google with upgraded agentic capabilities. It is suited for subagents that execute focused tasks within complex, multi-agent workflows.", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -72121,102 +76045,123 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.15, + "output": 1.25, + "reasoning": 1.25, + "cache_read": 0.015, + "cache_write": 0.041667 } }, - "deepseek-tng-r1t2-chimera": { - "id": "deepseek-tng-r1t2-chimera", - "name": "DeepSeek TNG R1T2 Chimera", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": false, - "tool_call": true, + "google/gemini-3.1-flash-image": { + "id": "google/gemini-3.1-flash-image", + "name": "Nano Banana 2", + "description": "Gemini 3.1 Flash Image, a.k.a. \"Nano Banana 2,\" is Google’s latest state of the art image generation and editing model, delivering Pro-level visual quality at Flash speed. It combines advanced...", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-02", - "last_updated": "2025-07-02", + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 130000, - "output": 163840 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.5, + "output": 3 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Meta Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "google/lyria-3-clip-preview": { + "id": "google/lyria-3-clip-preview", + "name": "Lyria 3 Clip Preview", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "lyria", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2026-03-25", + "last_updated": "2026-03-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16400 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.13, - "output": 0.39 + "input": 0, + "output": 0 } }, - "sonar-reasoning-pro": { - "id": "sonar-reasoning-pro", - "name": "Perplexity Sonar Reasoning Pro", - "description": "Web-grounded reasoning model for multi-step research and cited answers", - "family": "sonar-reasoning", - "attachment": false, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" ] } ], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2025-01-27", - "last_updated": "2025-01-27", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -72224,61 +76169,80 @@ }, "open_weights": false, "limit": { - "context": 127000, - "output": 4096 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2, - "output": 8 + "input": 1, + "output": 6, + "reasoning": 6, + "cache_read": 0.1, + "cache_write": 0.1875 } }, - "grok-3": { - "id": "grok-3", - "name": "xAI Grok 3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": false, + "google/gemma-3-12b-it": { + "id": "google/gemma-3-12b-it", + "name": "Gemma 3 12B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, - "output": 131072 + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 + "input": 0.05, + "output": 0.15 } }, - "glm-4.6": { - "id": "glm-4.6", - "name": "Zai GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -72286,29 +76250,45 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.44999999999999996, - "output": 1.5 + "input": 0.75, + "output": 4.5, + "reasoning": 4.5, + "cache_read": 0.075, + "cache_write": 0.041667 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, - "reasoning": false, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -72316,30 +76296,43 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 262144 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.48, - "output": 2 + "input": 1.25, + "output": 10, + "reasoning": 10, + "cache_read": 0.125, + "cache_write": 0.375 } }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "xAI Grok 4.1 Fast Non-Reasoning", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", - "attachment": false, - "reasoning": false, + "google/gemini-3-pro-image": { + "id": "google/gemini-3-pro-image", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro is Google’s most advanced image-generation and editing model, built on Gemini 3 Pro. It extends the original Nano Banana with significantly improved multimodal reasoning, real-world grounding, and...", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-17", - "last_updated": "2025-11-17", + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ "text", @@ -72348,33 +76341,46 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 65536, + "output": 32768 }, "cost": { - "input": 0.19999999999999998, - "output": 0.5, - "cache_read": 0.049999999999999996 + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "qwen3-coder": { - "id": "qwen3-coder", - "name": "Qwen3 Coder 480B A35B Instruct Turbo", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", "audio", - "video" + "video", + "pdf" ], "output": [ "text" @@ -72382,94 +76388,139 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 0.22, - "output": 0.95 + "input": 0.3, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333 } }, - "gpt-5-chat-latest": { - "id": "gpt-5-chat-latest", - "name": "OpenAI GPT-5 Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2024-09", - "release_date": "2024-09-30", - "last_updated": "2024-09-30", + "google/gemini-3.1-flash-lite-image": { + "id": "google/gemini-3.1-flash-lite-image", + "name": "Nano Banana 2 Lite", + "description": "Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) is Google's fastest, most cost-efficient Gemini image model, built for high-velocity developer pipelines and rapid-fire visual exploration. It delivers text-to-image generation...", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 65536, + "output": 58982 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 + "input": 0.25, + "output": 1.5 } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "OpenAI: GPT-5.1 Codex", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-codex", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" + "google/gemini-3.7-flash": { + "id": "google/gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "Gemini 3.7 Flash is a multimodal model from Google for fast agentic workflows, coding, and complex multi-step reasoning. It is designed for tasks that require responsive performance and reliable multi-step...", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667 } }, - "o3-mini": { - "id": "o3-mini", - "name": "OpenAI o3 Mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, - "reasoning": false, + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2023-10", - "release_date": "2023-10-01", - "last_updated": "2023-10-01", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -72477,30 +76528,49 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "mistral-large-2411": { - "id": "mistral-large-2411", - "name": "Mistral-Large", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": false, - "reasoning": false, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-24", - "last_updated": "2024-07-24", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -72508,40 +76578,46 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2, - "output": 6 + "input": 0.25, + "output": 1.5, + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.041667 } }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Google Gemini 2.5 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": false, + "google/gemini-3.8-flash": { + "id": "google/gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Gemini 3.8 Flash is Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows.", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -72550,27 +76626,29 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65535 + "output": 65536 }, "cost": { - "input": 0.09999999999999999, - "output": 0.39999999999999997, - "cache_read": 0.024999999999999998, - "cache_write": 0.09999999999999999 + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667 } }, - "llama-guard-4": { - "id": "llama-guard-4", - "name": "Meta Llama Guard 4 12B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": false, + "google/gemma-3-4b-it": { + "id": "google/gemma-3-4b-it", + "name": "Gemma 3 4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": false, "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text", @@ -72580,42 +76658,45 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, - "output": 1024 + "output": 16384 }, "cost": { - "input": 0.21, - "output": 0.21 + "input": 0.05, + "output": 0.1 } }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Anthropic: Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": false, + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -72623,32 +76704,43 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.1, + "output": 0.4, + "reasoning": 0.4, + "cache_read": 0.01, + "cache_write": 0.083333 } }, - "claude-3.7-sonnet": { - "id": "claude-3.7-sonnet", - "name": "Anthropic: Claude 3.7 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": false, - "reasoning": false, + "google/gemini-2.5-pro-preview": { + "id": "google/gemini-2.5-pro-preview", + "name": "Google: Gemini 2.5 Pro Preview 06-05", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ + "pdf", + "image", "text", - "image" + "audio" ], "output": [ "text" @@ -72656,98 +76748,132 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.30000000000000004, - "cache_write": 3.75 + "input": 1.25, + "output": 10, + "reasoning": 10, + "cache_read": 0.125, + "cache_write": 0.375 } }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "OpenAI: GPT-5.1 Codex Mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-codex", - "attachment": false, - "reasoning": false, + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": false, + "structured_output": true, + "temperature": true, "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { "input": 0.25, - "output": 2, - "cache_read": 0.024999999999999998 + "output": 1.5, + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.083333 } }, - "gpt-5.1-chat-latest": { - "id": "gpt-5.1-chat-latest", - "name": "OpenAI GPT-5.1 Chat", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-codex", - "attachment": false, - "reasoning": false, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": false, + "structured_output": true, + "temperature": true, "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 + "input": 0.125, + "output": 0.75, + "reasoning": 0.75, + "cache_read": 0.0125, + "cache_write": 0.041667 } }, - "claude-3-haiku-20240307": { - "id": "claude-3-haiku-20240307", - "name": "Anthropic: Claude 3 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "writer/palmyra-x5": { + "id": "writer/palmyra-x5", + "name": "Writer: Palmyra X5", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "palmyra", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-03-07", - "last_updated": "2024-03-07", + "release_date": "2026-01-21", + "last_updated": "2026-01-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -72755,33 +76881,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 4096 + "context": 1040000, + "output": 8192 }, "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.03, - "cache_write": 0.3 + "input": 0.6, + "output": 6 } }, - "grok-4-fast-reasoning": { - "id": "grok-4-fast-reasoning", - "name": "xAI: Grok 4 Fast Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "ibm-granite/granite-4.0-h-micro": { + "id": "ibm-granite/granite-4.0-h-micro", + "name": "IBM: Granite 4.0 Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "granite", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-01", - "last_updated": "2025-09-01", + "release_date": "2025-10-20", + "last_updated": "2025-10-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -72789,31 +76911,39 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 131000, + "output": 117900 }, "cost": { - "input": 0.19999999999999998, - "output": 0.5, - "cache_read": 0.049999999999999996 + "input": 0.017, + "output": 0.112 } }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "OpenAI GPT-4.1 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "ibm-granite/granite-4.2-8b": { + "id": "ibm-granite/granite-4.2-8b", + "name": "IBM: Granite 4.2 8B", + "description": "Granite 4.2 8B is a dense reasoning model from IBM. It is suited for mathematics, code generation, multilingual dialogue, and agentic workflows that need multi-step reasoning. It supports full, low-effort,...", + "family": "granite", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-08-31", + "last_updated": "2026-08-31", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -72821,37 +76951,28 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 131072, + "output": 117964 }, "cost": { - "input": 0.09999999999999999, - "output": 0.39999999999999997, - "cache_read": 0.024999999999999998 + "input": 0.06, + "output": 0.25, + "cache_read": 0.015 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "OpenAI GPT-OSS 120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "mistralai/mistral-nemo": { + "id": "mistralai/mistral-nemo", + "name": "Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ "text" @@ -72860,31 +76981,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, - "output": 131072 + "output": 16384 }, "cost": { - "input": 0.04, - "output": 0.16 + "input": 0.019, + "output": 0.03 } }, - "sonar": { - "id": "sonar", - "name": "Perplexity Sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar", - "attachment": false, + "mistralai/ministral-8b-2512": { + "id": "mistralai/ministral-8b-2512", + "name": "Mistral: Ministral 3 8B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-27", - "last_updated": "2025-01-27", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -72892,26 +77014,27 @@ }, "open_weights": false, "limit": { - "context": 127000, - "output": 4096 + "context": 262144, + "output": 209715 }, "cost": { - "input": 1, - "output": 1 + "input": 0.15, + "output": 0.15, + "cache_read": 0.015 } }, - "qwen2.5-coder-7b-fast": { - "id": "qwen2.5-coder-7b-fast", - "name": "Qwen2.5 Coder 7B fast", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "mistralai/mistral-small-24b-instruct-2501": { + "id": "mistralai/mistral-small-24b-instruct-2501", + "name": "Mistral: Mistral Small 3", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-09-15", - "last_updated": "2024-09-15", + "release_date": "2025-01-30", + "last_updated": "2025-01-30", "modalities": { "input": [ "text" @@ -72922,29 +77045,30 @@ }, "open_weights": false, "limit": { - "context": 32000, - "output": 8192 + "context": 32768, + "output": 16384 }, "cost": { - "input": 0.03, - "output": 0.09 + "input": 0.05, + "output": 0.08 } }, - "o1": { - "id": "o1", - "name": "OpenAI: o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": false, + "mistralai/mistral-saba": { + "id": "mistralai/mistral-saba", + "name": "Mistral: Saba", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-02-17", + "last_updated": "2025-02-17", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -72952,31 +77076,41 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 32768, + "output": 26214 }, "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 0.2, + "output": 0.6, + "cache_read": 0.02 } }, - "ernie-4.5-21b-a3b-thinking": { - "id": "ernie-4.5-21b-a3b-thinking", - "name": "Baidu Ernie 4.5 21B A3B Thinking", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "ernie", - "attachment": false, + "mistralai/mistral-medium-3-5": { + "id": "mistralai/mistral-medium-3-5", + "name": "Mistral: Mistral Medium 3.5", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-03-16", - "last_updated": "2025-03-16", + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -72984,30 +77118,31 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8000 + "context": 262144, + "output": 209715 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 1.5, + "output": 7.5 } }, - "llama-4-scout": { - "id": "llama-4-scout", - "name": "Meta Llama 4 Scout 17B 16E", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": false, + "mistralai/mistral-medium-3.1": { + "id": "mistralai/mistral-medium-3.1", + "name": "Mistral: Mistral Medium 3.1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2025-08-13", + "last_updated": "2025-08-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -73016,27 +77151,29 @@ "open_weights": false, "limit": { "context": 131072, - "output": 8192 + "output": 104857 }, "cost": { - "input": 0.08, - "output": 0.3 + "input": 0.4, + "output": 2, + "cache_read": 0.04 } }, - "sonar-pro": { - "id": "sonar-pro", - "name": "Perplexity Sonar Pro", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", - "attachment": false, + "mistralai/mistral-small-3.2-24b-instruct": { + "id": "mistralai/mistral-small-3.2-24b-instruct", + "name": "Mistral: Mistral Small 3.2 24B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-27", - "last_updated": "2025-01-27", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", "modalities": { "input": [ + "image", "text" ], "output": [ @@ -73045,30 +77182,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 4096 + "context": 256000, + "output": 16384 }, "cost": { - "input": 3, - "output": 15 + "input": 0.09375, + "output": 0.25 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "OpenAI GPT-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, + "mistralai/mistral-large": { + "id": "mistralai/mistral-large", + "name": "Mistral Large", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2024-02-26", + "last_updated": "2024-02-26", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" @@ -73076,37 +77213,37 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 128000, + "output": 102400 }, "cost": { "input": 2, - "output": 8, - "cache_read": 0.5 + "output": 6, + "cache_read": 0.2 } }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "Anthropic: Claude Sonnet 4.5 (20250929)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": false, + "mistralai/mistral-small-2603": { + "id": "mistralai/mistral-small-2603", + "name": "Mistral Small 4", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text", @@ -73116,33 +77253,34 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 209715 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.30000000000000004, - "cache_write": 3.75 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "deepseek-reasoner": { - "id": "deepseek-reasoner", - "name": "DeepSeek Reasoner", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "mistralai/mistral-medium-3": { + "id": "mistralai/mistral-medium-3", + "name": "Mistral: Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -73150,75 +77288,64 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 64000 + "context": 131072, + "output": 104857 }, "cost": { - "input": 0.56, - "output": 1.68, - "cache_read": 0.07 + "input": 0.4, + "output": 2, + "cache_read": 0.04 } }, - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "xAI Grok 4.1 Fast Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "mistralai/voxtral-small-24b-2507": { + "id": "mistralai/voxtral-small-24b-2507", + "name": "Voxtral Small 24B 2507", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "voxtral", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-17", - "last_updated": "2025-11-17", + "release_date": "2025-07-15", + "last_updated": "2025-07-15", "modalities": { "input": [ "text", - "image" + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 32768, + "output": 26214 }, "cost": { - "input": 0.19999999999999998, - "output": 0.5, - "cache_read": 0.049999999999999996 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01 } }, - "gemini-3-pro-preview": { - "id": "gemini-3-pro-preview", - "name": "Google Gemini 3 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], + "mistralai/mistral-large-2407": { + "id": "mistralai/mistral-large-2407", + "name": "Mistral Large 2407", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2024-11-19", + "last_updated": "2024-11-19", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -73226,27 +77353,27 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 104857 }, "cost": { "input": 2, - "output": 12, - "cache_read": 0.19999999999999998 + "output": 6, + "cache_read": 0.2 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "OpenAI GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": false, + "mistralai/ministral-14b-2512": { + "id": "mistralai/ministral-14b-2512", + "name": "Mistral: Ministral 3 14B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "structured_output": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", @@ -73258,27 +77385,27 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 209715 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.024999999999999998 + "input": 0.2, + "output": 0.2, + "cache_read": 0.02 } }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "OpenAI GPT-4.1 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": false, + "mistralai/mistral-small-3.1-24b-instruct": { + "id": "mistralai/mistral-small-3.1-24b-instruct", + "name": "Mistral: Mistral Small 3.1 24B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-03-17", + "last_updated": "2025-03-17", "modalities": { "input": [ "text", @@ -73290,31 +77417,30 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 128000, + "output": 102400 }, "cost": { - "input": 0.39999999999999997, - "output": 1.5999999999999999, - "cache_read": 0.09999999999999999 + "input": 0.351, + "output": 0.555 } }, - "sonar-reasoning": { - "id": "sonar-reasoning", - "name": "Perplexity Sonar Reasoning", - "description": "Web-grounded reasoning model for multi-step research and cited answers", - "family": "sonar-reasoning", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "mistralai/ministral-3b-2512": { + "id": "mistralai/ministral-3b-2512", + "name": "Mistral: Ministral 3 3B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-27", - "last_updated": "2025-01-27", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -73322,40 +77448,31 @@ }, "open_weights": false, "limit": { - "context": 127000, - "output": 4096 + "context": 131072, + "output": 104857 }, "cost": { - "input": 1, - "output": 5 + "input": 0.1, + "output": 0.1, + "cache_read": 0.01 } }, - "sonar-deep-research": { - "id": "sonar-deep-research", - "name": "Perplexity Sonar Deep Research", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar-deep-research", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, + "mistralai/mixtral-8x22b-instruct": { + "id": "mistralai/mixtral-8x22b-instruct", + "name": "Mistral: Mixtral 8x22B Instruct", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-27", - "last_updated": "2025-01-27", + "release_date": "2024-04-17", + "last_updated": "2024-04-17", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -73363,29 +77480,31 @@ }, "open_weights": false, "limit": { - "context": 127000, - "output": 4096 + "context": 65536, + "output": 52428 }, "cost": { "input": 2, - "output": 8 + "output": 6, + "cache_read": 0.2 } }, - "kimi-k2-0905": { - "id": "kimi-k2-0905", - "name": "Kimi K2 (09/05)", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, + "mistralai/codestral-2508": { + "id": "mistralai/codestral-2508", + "name": "Mistral: Codestral 2508", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "codestral", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "release_date": "2025-08-01", + "last_updated": "2025-08-01", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -73393,31 +77512,42 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 256000, + "output": 204800 }, "cost": { - "input": 0.5, - "output": 2, - "cache_read": 0.39999999999999997 + "input": 0.3, + "output": 0.9, + "cache_read": 0.03 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "OpenAI GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": false, - "reasoning": false, + "sakana/fugu-max": { + "id": "sakana/fugu-max", + "name": "Sakana: Fugu Max", + "description": "Fugu Max is the cost-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to route...", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -73425,30 +77555,42 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.049999999999999996, - "output": 0.39999999999999997, - "cache_read": 0.005 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "grok-4": { - "id": "grok-4", - "name": "xAI Grok 4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": false, - "reasoning": false, + "sakana/fugu-ultra-v2": { + "id": "sakana/fugu-ultra-v2", + "name": "Sakana: Fugu Ultra v2", + "description": "Fugu Ultra v2 is the higher-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to...", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-09", - "last_updated": "2024-07-09", + "structured_output": true, + "temperature": false, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -73456,33 +77598,41 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "qwen3-235b-a22b-thinking": { - "id": "qwen3-235b-a22b-thinking", - "name": "Qwen3 235B A22B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "sakana/fugu-ultra": { + "id": "sakana/fugu-ultra", + "name": "Fugu Ultra", + "description": "Fugu Ultra is the higher-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to route...", + "family": "fugu", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -73490,40 +77640,41 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 81920 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 2.9000000000000004 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "claude-opus-4": { - "id": "claude-opus-4", - "name": "Anthropic: Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": false, + "sakana/sakana-namazu": { + "id": "sakana/sakana-namazu", + "name": "Sakana Namazu", + "description": "Sakana Namazu is a Japanese-specialized reasoning model from Sakana AI, based on Kimi K2.6 with additional training for Japanese language and business contexts. It is suited for Japanese instruction following,...", + "family": "sakana-namazu", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-14", - "last_updated": "2025-05-14", + "structured_output": true, + "temperature": false, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -73531,33 +77682,39 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.95, + "output": 4, + "cache_read": 0.15 } }, - "grok-4-fast-non-reasoning": { - "id": "grok-4-fast-non-reasoning", - "name": "xAI Grok 4 Fast Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "inclusionai/ling-3.0-flash-fin:free": { + "id": "inclusionai/ling-3.0-flash-fin:free", + "name": "inclusionAI: Ling 3.0 Flash Fin (free)", + "description": "Ling 3.0 Flash Fin is a finance-focused mixture-of-experts model from InclusionAI, built on Ling 3.0 Flash with 5.1B active parameters out of 124B total. It is designed for real-world investment...", + "family": "ling", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-19", - "last_updated": "2025-09-19", + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -73565,41 +77722,38 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.19999999999999998, - "output": 0.5, - "cache_read": 0.049999999999999996 + "input": 0, + "output": 0 } }, - "claude-4.5-opus": { - "id": "claude-4.5-opus", - "name": "Anthropic: Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "inclusionai/ling-3.0-flash-sante:free": { + "id": "inclusionai/ling-3.0-flash-sante:free", + "name": "inclusionAI: Ling 3.0 Flash Sante (free)", + "description": "Ling 3.0 Flash Sante is a health and medicine-focused mixture-of-experts model from InclusionAI, built on Ling 3.0 Flash with 5.1B active parameters out of 124B total. It is designed for...", + "family": "ling", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -73607,33 +77761,38 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0 } }, - "qwen3-vl-235b-a22b-instruct": { - "id": "qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "inclusionai/ling-3.0-flash": { + "id": "inclusionai/ling-3.0-flash", + "name": "inclusionAI: Ling 3.0 Flash", + "description": "*Ling-3.0-flash* is a *124B-parameter Mixture-of-Experts (MoE) model*, with approximately *5.1B parameters activated per token*. The model is designed with *token efficiency and production-scale agentic inference* as key priorities, enabling developers...", + "family": "ling", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-07-23", + "last_updated": "2026-07-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -73641,29 +77800,41 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 0.06, + "output": 0.18, + "cache_read": 0.012 } }, - "kimi-k2-0711": { - "id": "kimi-k2-0711", - "name": "Kimi K2 (07/11)", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "inclusionai/ling-3.0-flash-vl": { + "id": "inclusionai/ling-3.0-flash-vl", + "name": "inclusionAI: Ling 3.0 Flash VL", + "description": "Ling 3.0 Flash VL builds on Ling 3.0 Flash (124B total / 5.5B active MoE from InclusionAI), further strengthening its language capabilities while adding native visual perception and advanced visual...", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -73672,29 +77843,38 @@ "open_weights": false, "limit": { "context": 131072, - "output": 16384 + "output": 32768 }, "cost": { - "input": 0.5700000000000001, - "output": 2.3 + "input": 0.06, + "output": 0.18, + "cache_read": 0.012 } }, - "gemma-3-12b-it": { - "id": "gemma-3-12b-it", - "name": "Google Gemma 3 12B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "inclusionai/ling-3.0-flash-fin": { + "id": "inclusionai/ling-3.0-flash-fin", + "name": "inclusionAI: Ling 3.0 Flash Fin", + "description": "Ling 3.0 Flash Fin is a finance-focused mixture-of-experts model from InclusionAI, built on Ling 3.0 Flash with 5.1B active parameters out of 124B total. It is designed for real-world investment...", + "family": "ling", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -73702,131 +77882,162 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.049999999999999996, - "output": 0.09999999999999999 + "input": 0.06, + "output": 0.18, + "cache_read": 0.012 } }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "OpenAI GPT-4o-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": false, - "reasoning": false, + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "OpenAI GPT-OSS 20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.049999999999999996, - "output": 0.19999999999999998 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "claude-3.5-sonnet-v2": { - "id": "claude-3.5-sonnet-v2", - "name": "Anthropic: Claude 3.5 Sonnet v2", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 + "context": 262144, + "output": 98304 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.30000000000000004, - "cache_write": 3.75 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "qwen3-30b-a3b": { - "id": "qwen3-30b-a3b", - "name": "Qwen3 30B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-06-01", - "last_updated": "2025-06-01", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -73836,28 +78047,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 41000, - "output": 41000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.08, - "output": 0.29 + "input": 0.6, + "output": 3 } }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "OpenAI: GPT-5 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "MoonshotAI: Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "structured_output": true, + "temperature": true, + "release_date": "2025-09-04", + "last_updated": "2025-09-04", "modalities": { "input": [ "text" @@ -73868,27 +78079,26 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 98304 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 + "input": 0.6, + "output": 2.5 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "moonshotai/kimi-k2": { + "id": "moonshotai/kimi-k2", + "name": "MoonshotAI: Kimi K2 0711", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "release_date": "2025-07-11", + "last_updated": "2025-07-11", "modalities": { "input": [ "text" @@ -73899,36 +78109,35 @@ }, "open_weights": false, "limit": { - "context": 163840, - "output": 65536 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.27, - "output": 0.41 + "input": 0.57, + "output": 2.3 } }, - "claude-4.5-sonnet": { - "id": "claude-4.5-sonnet", - "name": "Anthropic: Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": false, + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "MoonshotAI: Kimi K2.7 Code is a coding-focused model in Moonshot AI's Kimi K2 family, built to complete end-to-end programming tasks reliably over long contexts. It uses a native multimodal mixture-of-experts...", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", @@ -73938,30 +78147,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.30000000000000004, - "cache_write": 3.75 + "input": 0.7062, + "output": 3.3, + "cache_read": 0.18 } }, - "mistral-small": { - "id": "mistral-small", - "name": "Mistral Small 3.2", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, - "reasoning": false, + "kilo-auto/small": { + "id": "kilo-auto/small", + "name": "Auto Small", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ "text", @@ -73971,31 +78180,35 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.075, - "output": 0.2 + "input": 0.05, + "output": 0.4, + "reasoning": 0, + "cache_read": 0.005 } }, - "llama-3.1-8b-instruct-turbo": { - "id": "llama-3.1-8b-instruct-turbo", - "name": "Meta Llama 3.1 8B Instruct Turbo", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", - "attachment": false, - "reasoning": false, + "kilo-auto/efficient": { + "id": "kilo-auto/efficient", + "name": "Auto Efficient", + "description": "Routes each request to the cheapest model that gets the job done, based on continuously benchmarked accuracy and cost.", + "family": "auto", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -74003,153 +78216,142 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.02, - "output": 0.03 + "input": 0.325, + "output": 1.95, + "reasoning": 0, + "cache_read": 0.0325, + "cache_write": 0.40625 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "OpenAI GPT-5.1", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "kilo-auto/free": { + "id": "kilo-auto/free", + "name": "Auto Free", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "structured_output": false, + "temperature": true, + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 256000, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 + "input": 0, + "output": 0, + "reasoning": 0, + "cache_read": 0, + "cache_write": 0 } - } - } - }, - "zai": { - "id": "zai", - "env": [ - "ZHIPU_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.z.ai/api/paas/v4", - "name": "Z.AI", - "doc": "https://docs.z.ai/guides/overview/pricing", - "models": { - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, + }, + "kilo-auto/frontier": { + "id": "kilo-auto/frontier", + "name": "Auto Frontier", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 5, + "output": 25, + "reasoning": 0, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "glm-4.5v": { - "id": "glm-4.5v", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "kilo-auto/balanced": { + "id": "kilo-auto/balanced", + "name": "Auto Balanced", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 64000, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 1.8 + "input": 0.325, + "output": 1.95, + "reasoning": 0, + "cache_read": 0.0325, + "cache_write": 0.40625 } }, - "glm-4.5": { - "id": "glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", + "rekaai/reka-flash-3": { + "id": "rekaai/reka-flash-3", + "name": "Reka Flash 3", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "reka", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "high" + ] } ], - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text" @@ -74158,75 +78360,69 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 65536, + "output": 58982 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 0.1, + "output": 0.2 } }, - "glm-4.7-flashx": { - "id": "glm-4.7-flashx", - "name": "GLM-4.7-FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "rekaai/reka-edge": { + "id": "rekaai/reka-edge", + "name": "Reka Edge", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "reka", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 16384, + "output": 14745 }, "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0 + "input": 0.1, + "output": 0.1 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "nvidia/nemotron-3-nano-30b-a3b": { + "id": "nvidia/nemotron-3-nano-30b-a3b", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ "text" @@ -74237,36 +78433,40 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 262144, + "output": 235929 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "input": 0.05, + "output": 0.2, + "cache_read": 0.03 } }, - "glm-4.6": { - "id": "glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, + "nvidia/nemotron-3.5-content-safety:free": { + "id": "nvidia/nemotron-3.5-content-safety:free", + "name": "NVIDIA: Nemotron 3.5 Content Safety (free)", + "description": "NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).", + "family": "nemotron", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -74274,40 +78474,35 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 0, + "output": 0 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "nvidia/nemotron-3.5-lightning": { + "id": "nvidia/nemotron-3.5-lightning", + "name": "Nemotron 3.5 Lightning 30B A3B", + "description": "NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that...", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", - "max" + "none", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ "text" @@ -74318,38 +78513,41 @@ }, "open_weights": true, "limit": { - "context": 1000000, + "context": 262144, "output": 131072 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "input": 0.065, + "output": 0.18 } }, - "glm-4.6v": { - "id": "glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free": { + "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free", + "name": "NVIDIA: Nemotron 3 Nano Omni (free)", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" @@ -74357,73 +78555,76 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0, + "output": 0 } }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 262144, + "output": 235929 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24, - "cache_write": 0 + "input": 0.08, + "output": 0.45 } }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "nvidia/nemotron-3-ultra-550b-a55b:free": { + "id": "nvidia/nemotron-3-ultra-550b-a55b:free", + "name": "NVIDIA: Nemotron 3 Ultra (free)", + "description": "NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -74434,33 +78635,36 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03, - "cache_write": 0 + "input": 0, + "output": 0 } }, - "glm-4.7-flash": { - "id": "glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", + "nvidia/nemotron-3-ultra-550b-a55b": { + "id": "nvidia/nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -74471,33 +78675,37 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 202800, + "output": 182520 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.5, + "output": 2.2, + "cache_read": 0.1 } }, - "glm-4.5-flash": { - "id": "glm-4.5-flash", - "name": "GLM-4.5-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "nvidia/nemotron-3-super-120b-a12b:free": { + "id": "nvidia/nemotron-3-super-120b-a12b:free", + "name": "NVIDIA: Nemotron 3 Super (free)", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text" @@ -74508,38 +78716,39 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 262144, + "output": 235929 }, "cost": { "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "output": 0 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "nvidia/nemotron-3.5-content-safety": { + "id": "nvidia/nemotron-3.5-content-safety", + "name": "Nemotron 3.5 Content Safety", + "description": "NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting...", + "family": "nemotron", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -74547,36 +78756,35 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 117964 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0 + "input": 0.2, + "output": 0.2 } }, - "glm-5-turbo": { - "id": "glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", + "nvidia/nemotron-3.5-lightning:free": { + "id": "nvidia/nemotron-3.5-lightning:free", + "name": "NVIDIA: Nemotron 3.5 Lightning (free)", + "description": "NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ "text" @@ -74585,88 +78793,90 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24, - "cache_write": 0 + "input": 0, + "output": 0 } - } - } - }, - "nearai": { - "id": "nearai", - "env": [ - "NEARAI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://cloud-api.near.ai/v1", - "name": "NEAR AI Cloud", - "doc": "https://docs.near.ai/", - "models": { - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + }, + "xiaomi/mimo-v2.6-pro": { + "id": "xiaomi/mimo-v2.6-pro", + "name": "MiMo-V2.6-Pro", + "description": "MiMo-V2.6-Pro is the flagship foundation model developed by Xiaomi. Built at a scale of over 1T parameters, it is designed to push the ceiling of capability for the most demanding...", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", "image", - "video", "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -74674,127 +78884,191 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.13, - "output": 0.4, - "cache_read": 0.026 + "input": 0.14, + "output": 0.28, + "cache_read": 0.003, + "tiers": [ + { + "input": 0.8, + "output": 4, + "cache_read": 0.16, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.8, + "output": 4, + "cache_read": 0.16 + } } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "xiaomi/mimo-v2.6-pro-ultraspeed": { + "id": "xiaomi/mimo-v2.6-pro-ultraspeed", + "name": "MiMo-V2.6-Pro-UltraSpeed", + "description": "MiMo-V2.6-Pro-UltraSpeed is the fast speed edition of Xiaomi's flagship foundation model, MiMo-V2.6-Pro. Built from the same 1T MiMo-V2.6-Pro checkpoint, it matches the original model in quality while delivering roughly 10x...", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ "text", "image", "audio", - "video", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, + "input": 4.35, + "output": 8.7, + "cache_read": 0.036 + } + }, + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.004, "tiers": [ { - "input": 2.5, - "output": 15, - "cache_read": 0.25, + "input": 2, + "output": 6, + "cache_read": 0.4, "tier": { "type": "context", - "size": 200000 + "size": 256000 } } ], "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 2, + "output": 6, + "cache_read": 0.4 } } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", + "xiaomi/mimo-v2.6-flash": { + "id": "xiaomi/mimo-v2.6-flash", + "name": "MiMo-V2.6-Flash", + "description": "MiMo-V2.6-Flash is an open-source foundation model developed by Xiaomi. Built on a Mixture-of-Experts architecture with 309B total parameters and 15B activated per token, it employs a hybrid attention mechanism for...", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", "image", "audio", - "video", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "undi95/remm-slerp-l2-13b": { + "id": "undi95/remm-slerp-l2-13b", + "name": "ReMM SLERP 13B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2023-07-22", + "last_updated": "2023-07-22", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -74802,37 +79076,28 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 6144, + "output": 5529 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 + "input": 0.35, + "output": 0.65 } }, - "google/gemini-3-pro": { - "id": "google/gemini-3-pro", - "name": "Gemini 3 Pro Preview", - "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "gryphe/mythomax-l2-13b": { + "id": "gryphe/mythomax-l2-13b", + "name": "MythoMax 13B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2023-07-02", + "last_updated": "2023-07-02", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -74840,36 +79105,69 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 4096, + "output": 3686 }, "cost": { - "input": 1.25, - "output": 15, - "cache_read": 0 + "input": 0.08, + "output": 0.11 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 + } + }, + "minimax/minimax-01": { + "id": "minimax/minimax-01", + "name": "MiniMax: MiniMax-01", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-01-15", + "last_updated": "2025-01-15", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -74877,27 +79175,26 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000192, + "output": 900172 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 + "input": 0.2, + "output": 1.1 } }, - "Qwen/Qwen3-Embedding-0.6B": { - "id": "Qwen/Qwen3-Embedding-0.6B", - "name": "Qwen3 Embedding 0.6B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "minimax/minimax-m2-her": { + "id": "minimax/minimax-m2-her", + "name": "MiniMax-M2 Her", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-06-03", - "last_updated": "2025-06-03", + "structured_output": false, + "temperature": true, + "release_date": "2026-01-23", + "last_updated": "2026-01-23", "modalities": { "input": [ "text" @@ -74906,27 +79203,37 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 1024 + "context": 65536, + "output": 2048 }, "cost": { - "input": 0.01, - "output": 0 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "Qwen/Qwen3-Reranker-0.6B": { - "id": "Qwen/Qwen3-Reranker-0.6B", - "name": "Qwen3 Reranker 0.6B", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "qwen", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-06-03", - "last_updated": "2025-06-03", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -74937,31 +79244,36 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 1024 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.01, - "output": 0.01 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "Qwen/Qwen3.6-35B-A3B-FP8": { - "id": "Qwen/Qwen3.6-35B-A3B-FP8", - "name": "Qwen 3.6 35B A3B FP8", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, + "minimax/minimax-m1": { + "id": "minimax/minimax-m1", + "name": "MiniMax: MiniMax M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text" @@ -74970,37 +79282,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 40000 }, "cost": { - "input": 0.17, - "output": 1.1, - "cache_read": 0.056 + "input": 0.4, + "output": 2.2 } }, - "Qwen/Qwen3.5-122B-A10B": { - "id": "Qwen/Qwen3.5-122B-A10B", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -75008,26 +79325,35 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 524288, + "output": 512000 }, "cost": { - "input": 0.4, - "output": 3.2 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "Qwen/Qwen3-30B-A3B-Instruct-2507": { - "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen3 30B-A3B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -75038,30 +79364,38 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.55 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "Qwen/Qwen3-VL-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", - "name": "Qwen3-VL 30B-A3B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -75069,19 +79403,48 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.55 + "input": 0.3, + "output": 1.2 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "mancer/weaver": { + "id": "mancer/weaver", + "name": "Mancer: Weaver (alpha)", + "description": "An attempt to recreate Claude-style verbosity, but don't expect the same level of coherence or memory. Meant for use in roleplay/narrative situations.", + "family": "alpha", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2023-08-02", + "last_updated": "2023-08-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8000, + "output": 6000 + }, + "cost": { + "input": 0.4, + "output": 0.75 + } + }, + "stepfun/step-3.7-flash:free": { + "id": "stepfun/step-3.7-flash:free", + "name": "StepFun: Step 3.7 Flash (free)", + "description": "Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters per token. The model supports a 256K context window and exposes selectable reasoning levels (high/medium/low), letting callers trade off speed, cost, and depth of reasoning. Designed for coding, agentic workflows, structured outputs, and long-context productivity tasks.", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -75095,82 +79458,74 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "structured_output": false, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0, + "output": 0, + "reasoning": 0, + "cache_read": 0 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "stepfun/step-3.5-flash": { + "id": "stepfun/step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.1, + "output": 0.3 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "stepfun/step-3.7-flash": { + "id": "stepfun/step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters...", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -75185,35 +79540,36 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 256000, + "input": 256000, + "output": 230400 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "dots-studio/dots-3-note-preview:free": { + "id": "dots-studio/dots-3-note-preview:free", + "name": "Dots Studio: Dots3-Note Preview (free)", + "description": "Dots3-Note Preview is an open-weight mixture-of-experts model from Dots Studio, with 16B active parameters out of 280B total. It is the lightest model in the Dots 3 family and is...", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -75221,19 +79577,15 @@ "type": "effort", "values": [ "none", - "low", - "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", @@ -75245,27 +79597,26 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 512000, + "output": 460800 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0, + "output": 0 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "inception/mercury-2.5": { + "id": "inception/mercury-2.5", + "name": "Inception: Mercury 2.5", + "description": "Mercury 2.5 is the fastest reasoning LLM, and the latest diffusion LLM (dLLM) from Inception. Instead of generating tokens sequentially, Mercury 2.5 produces and refines multiple tokens in parallel, achieving...", + "family": "mercury", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -75274,10 +79625,9 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ "text" @@ -75288,21 +79638,21 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 260000, + "output": 65536 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.2, + "output": 0.75, + "cache_read": 0.02 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, + "inception/mercury-2": { + "id": "inception/mercury-2", + "name": "Inception: Mercury 2", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -75311,21 +79661,18 @@ "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-03-04", + "last_updated": "2026-03-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -75333,29 +79680,27 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 50000 }, "cost": { - "input": 1.8, - "output": 15.5, - "cache_read": 0.18 + "input": 0.25, + "output": 0.75, + "cache_read": 0.025 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", + "amazon/nova-lite-v1": { + "id": "amazon/nova-lite-v1", + "name": "Amazon: Nova Lite 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text", @@ -75367,85 +79712,72 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 300000, + "output": 5120 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 0.06, + "output": 0.24 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT-OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "amazon/nova-2-lite-v1": { + "id": "amazon/nova-2-lite-v1", + "name": "Amazon: Nova 2 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131000, - "output": 32768 + "context": 1000000, + "output": 65535 }, "cost": { - "input": 0.15, - "output": 0.55 + "input": 0.3, + "output": 2.5 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "amazon/nova-pro-v1": { + "id": "amazon/nova-pro-v1", + "name": "Amazon: Nova Pro 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": false, + "temperature": true, + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -75453,57 +79785,26 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 300000, + "output": 5120 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0.8, + "output": 3.2 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "amazon/nova-premier-v1": { + "id": "amazon/nova-premier-v1", + "name": "Amazon: Nova Premier 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "structured_output": false, + "temperature": true, + "release_date": "2025-10-31", + "last_updated": "2025-10-31", "modalities": { "input": [ "text", @@ -75515,34 +79816,30 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 2.5, + "output": 12.5, + "cache_read": 0.625 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, + "amazon/nova-micro-v1": { + "id": "amazon/nova-micro-v1", + "name": "Amazon: Nova Micro 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75550,43 +79847,28 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 128000, + "output": 5120 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.035, + "output": 0.14 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "relace/relace-search": { + "id": "relace/relace-search", + "name": "Relace: Relace Search", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -75594,34 +79876,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 256000, "output": 128000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 1, + "output": 3 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, + "relace/relace-apply-3": { + "id": "relace/relace-apply-3", + "name": "Relace: Relace Apply 3", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "tool_call": false, + "structured_output": false, + "temperature": false, + "release_date": "2025-09-26", + "last_updated": "2025-09-26", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75629,43 +79905,36 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.85, + "output": 1.25 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, + "aion-labs/aion-2.0": { + "id": "aion-labs/aion-2.0", + "name": "AionLabs: Aion-2.0", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -75673,73 +79942,67 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 0.8, + "output": 1.6, + "cache_read": 0.2 } }, - "openai/whisper-large-v3": { - "id": "openai/whisper-large-v3", - "name": "Whisper Large v3", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", + "aion-labs/aion-rp-llama-3.1-8b": { + "id": "aion-labs/aion-rp-llama-3.1-8b", + "name": "AionLabs: Aion-RP 1.0 (8B)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2023-11-06", - "last_updated": "2023-11-06", + "structured_output": false, + "temperature": true, + "release_date": "2025-02-04", + "last_updated": "2025-02-04", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 448, - "output": 448 + "context": 32768, + "output": 29491 }, "cost": { - "input": 0.01, - "output": 0 + "input": 0.8, + "output": 1.6 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "aion-labs/aion-3.0": { + "id": "aion-labs/aion-3.0", + "name": "AionLabs: Aion-3.0", + "description": "Aion-3.0 is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It uses a collaborative generation process in which multiple specialized models each contribute...", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", - "medium", "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-07", + "last_updated": "2026-07-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -75747,46 +80010,39 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 3, + "output": 6, + "cache_read": 0.75 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "aion-labs/aion-3.5": { + "id": "aion-labs/aion-3.5", + "name": "AionLabs: Aion 3.5", + "description": "Aion 3.5 is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It uses a collaborative generation process in which multiple specialized models each...", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "structured_output": false, + "temperature": true, + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75794,58 +80050,39 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 3, + "output": 6, + "cache_read": 0.75 } }, - "anthropic/claude-sonnet-4-5": { - "id": "anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "aion-labs/aion-3.5-mini": { + "id": "aion-labs/aion-3.5-mini", + "name": "AionLabs: Aion 3.5 Mini", + "description": "Aion 3.5 Mini is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It is the smaller, lower-cost sibling of Aion 3.5 and uses...", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75853,38 +80090,37 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15.5, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.7, + "output": 1.4, + "cache_read": 0.18 } }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "aion-labs/aion-3.0-mini": { + "id": "aion-labs/aion-3.0-mini", + "name": "AionLabs: Aion-3.0-Mini", + "description": "Aion-3.0 Mini is a multi-model roleplaying and storytelling system from AionLabs, built on the DeepSeek family of models. It uses a collaborative generation process in which multiple specialized models each...", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "high" + ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-07", + "last_updated": "2026-07-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75892,42 +80128,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.7, + "output": 1.4, + "cache_read": 0.18 } }, - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "qwen/qwen3-235b-a22b-2507": { + "id": "qwen/qwen3-235b-a22b-2507", + "name": "Qwen: Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2025-07-21", + "last_updated": "2025-07-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75935,85 +80159,70 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.1495, + "output": 0.598 } }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "qwen/qwen3-vl-235b-a22b-instruct": { + "id": "qwen/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.26, + "output": 1.04 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -76022,38 +80231,43 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 1.475, + "output": 4.425, + "cache_read": 0.295, + "cache_write": 1.84375 } }, - "zai-org/GLM-5.1-FP8": { - "id": "zai-org/GLM-5.1-FP8", - "name": "GLM-5.1 FP8", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "qwen/qwen3.8-27b": { + "id": "qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Qwen3.8 27B is an open-weight dense vision-language model from Qwen. It is suited for coding, professional workflows, research, multimodal interaction, and long-running agent tasks, with flexible thinking that can be...", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -76061,68 +80275,78 @@ }, "open_weights": true, "limit": { - "context": 202752, + "context": 1000000, "output": 131072 }, "cost": { - "input": 0.85, - "output": 3.3 + "input": 0.425, + "output": 2.55, + "cache_read": 0.085, + "cache_write": 0.53125 } }, - "black-forest-labs/FLUX.2-klein-4B": { - "id": "black-forest-labs/FLUX.2-klein-4B", - "name": "FLUX.2 Klein 4B", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, + "qwen/qwen3-vl-235b-a22b-thinking": { + "id": "qwen/qwen3-vl-235b-a22b-thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", "image" ], "output": [ - "image" + "text" ] }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1, - "output": 1 + "input": 0.4, + "output": 4 } - } - } - }, - "llmgateway": { - "id": "llmgateway", - "env": [ - "LLMGATEWAY_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.llmgateway.io/v1", - "name": "LLM Gateway", - "doc": "https://llmgateway.io/docs", - "models": { - "qwen-coder-plus": { - "id": "qwen-coder-plus", - "name": "Qwen Coder Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + }, + "qwen/qwen3.8-2.4t-a95b": { + "id": "qwen/qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2024-09-18", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text" @@ -76131,94 +80355,44 @@ "text" ] }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.502, - "output": 1.004 - } - }, - "mistral-large-latest": { - "id": "mistral-large-latest", - "name": "Mistral Large (latest)", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, "open_weights": true, "limit": { - "context": 128000, - "output": 262144 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 4, - "output": 12 + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "qwen3-vl-235b-a22b-thinking": { - "id": "qwen3-vl-235b-a22b-thinking", - "name": "Qwen3 VL 235B A22B Thinking", + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5 27B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.98, - "output": 3.95 - } - }, - "devstral-small-2507": { - "id": "devstral-small-2507", - "name": "Devstral Small", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-07-10", - "last_updated": "2025-07-10", - "modalities": { - "input": [ - "text" + "image", + "video" ], "output": [ "text" @@ -76226,32 +80400,38 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 128000 + "context": 262144, + "output": 65536 }, - "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.195, + "output": 1.56 } }, - "qwen3-vl-30b-a3b-thinking": { - "id": "qwen3-vl-30b-a3b-thinking", - "name": "Qwen3 VL 30B A3B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-02", - "last_updated": "2025-10-02", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -76259,31 +80439,27 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.2, - "output": 1 + "input": 0.15, + "output": 1.2 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ "text" @@ -76294,25 +80470,24 @@ }, "open_weights": true, "limit": { - "context": 1050000, - "output": 384000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.3, + "output": 1.5 } }, - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "qwen/qwen3-coder": { + "id": "qwen/qwen3-coder", + "name": "Qwen: Qwen3 Coder 480B A35B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", "release_date": "2025-07-23", "last_updated": "2025-07-23", "modalities": { @@ -76325,28 +80500,36 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 262144, "output": 65536 }, "cost": { - "input": 6, - "output": 60, - "cache_read": 1.2, - "cache_write": 7.5 + "input": 0.975, + "output": 4.875 } }, - "minimax-m2.7-highspeed": { - "id": "minimax-m2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", + "qwen/qwen3-32b": { + "id": "qwen/qwen3-32b", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -76357,25 +80540,23 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 40960, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.08, + "output": 0.28 } }, - "qwen-plus": { - "id": "qwen-plus", + "qwen/qwen-plus": { + "id": "qwen/qwen-plus", "name": "Qwen Plus", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2024-04", "release_date": "2024-01-25", @@ -76390,36 +80571,42 @@ }, "open_weights": false, "limit": { - "context": 131072, + "context": 1000000, "output": 32768 }, "cost": { - "input": 0.4, - "output": 1.2, - "reasoning": 4, - "cache_read": 0.08, - "cache_write": 0.5 + "input": 0.26, + "output": 0.78, + "cache_read": 0.052, + "cache_write": 0.325 } }, - "o3": { - "id": "o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "qwen/qwen3.7-flash": { + "id": "qwen/qwen3.7-flash", + "name": "Qwen3.7 Flash", + "description": "Qwen3.7 Flash is a vision-language reasoning model from Alibaba. It is suited for multimodal agents, visual coding, search, and computer interaction, with strengths in object recognition, spatial understanding, and real-world...", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -76427,31 +80614,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "input": 991000, + "output": 65536 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.03, + "output": 0.13, + "cache_read": 0.006, + "cache_write": 0.038 } }, - "nemotron-3-ultra-550b": { - "id": "nemotron-3-ultra-550b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" @@ -76460,29 +80646,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 128000 + "output": 65536 }, "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.15 + "input": 0.78, + "output": 3.9, + "cache_read": 0.156, + "cache_write": 0.975 } }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "qwen/qwen3-coder-30b-a3b-instruct": { + "id": "qwen/qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -76493,33 +80681,41 @@ }, "open_weights": true, "limit": { - "context": 228700, - "output": 131072 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.2925, + "output": 1.4625 } }, - "grok-4-20-beta-0309-non-reasoning": { - "id": "grok-4-20-beta-0309-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "qwen/qwen3.8-omni-flash": { + "id": "qwen/qwen3.8-omni-flash", + "name": "Qwen3.8 Omni Flash", + "description": "Qwen3.8 Omni Flash is an omni-modal reasoning model from Alibaba, the first Qwen model built around agentic capabilities with native audio-video understanding. It is suited for audio-video analysis and summarization,...", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "release_date": "2026-09-17", + "last_updated": "2026-09-17", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -76527,102 +80723,69 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.15, + "output": 0.47, + "cache_read": 0.016 } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen/qwen3-vl-30b-a3b-instruct": { + "id": "qwen/qwen3-vl-30b-a3b-instruct", + "name": "Qwen: Qwen3 VL 30B A3B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.38, - "output": 1.98, - "cache_read": 0.19, - "cache_write": 0 + "input": 0.13, + "output": 0.52 } }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "qwen/qwen3-235b-a22b-thinking-2507": { + "id": "qwen/qwen3-235b-a22b-thinking-2507", + "name": "Qwen: Qwen3 235B A22B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", "high" ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -76630,28 +80793,36 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 117964 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 0.08333 + "input": 0.23, + "output": 2.3 } }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct (2507)", + "qwen/qwen3-235b-a22b": { + "id": "qwen/qwen3-235b-a22b", + "name": "Qwen3 235B-A22B", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-07-08", - "last_updated": "2025-07-08", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -76662,29 +80833,40 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 131072, "output": 8192 }, "cost": { - "input": 0.09, - "output": 0.58 + "input": 0.455, + "output": 1.82 } }, - "llama-3-70b-instruct": { - "id": "llama-3-70b-instruct", - "name": "Llama 3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-04-18", - "last_updated": "2024-04-18", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -76692,26 +80874,26 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.51, - "output": 0.74 + "input": 0.26, + "output": 2.08 } }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "qwen/qwen-plus-2025-07-28": { + "id": "qwen/qwen-plus-2025-07-28", + "name": "Qwen: Qwen Plus 0728", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-09-08", + "last_updated": "2025-09-08", "modalities": { "input": [ "text" @@ -76720,30 +80902,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 65536 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.07, - "output": 0.27 + "input": 0.26, + "output": 0.78 } }, - "seed-1-8-251228": { - "id": "seed-1-8-251228", - "name": "Seed 1.8 (251228)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "qwen/qwen3.6-35b-a3b": { + "id": "qwen/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", + "none", "high" ] } @@ -76751,12 +80931,13 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -76764,27 +80945,28 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 8192 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.25, - "output": 2, + "input": 0.15, + "output": 1, "cache_read": 0.05 } }, - "kimi-k2": { - "id": "kimi-k2", - "name": "Kimi K2", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-11", - "last_updated": "2025-07-11", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ "text" @@ -76795,69 +80977,34 @@ }, "open_weights": true, "limit": { - "context": 256000, + "context": 262144, "output": 16384 }, "cost": { - "input": 0.57, - "output": 2.3, - "cache_read": 0.5 - } - }, - "llama-3.1-70b-instruct": { - "id": "llama-3.1-70b-instruct", - "name": "Llama 3.1 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 2048 - }, - "status": "beta", - "cost": { - "input": 0.72, - "output": 0.72 + "input": 0.0975, + "output": 0.78 } }, - "gpt-5.2-pro": { - "id": "gpt-5.2-pro", - "name": "GPT-5.2 Pro", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", + "qwen/qwen3-vl-30b-a3b-thinking": { + "id": "qwen/qwen3-vl-30b-a3b-thinking", + "name": "Qwen: Qwen3 VL 30B A3B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "structured_output": true, + "temperature": true, + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", @@ -76869,58 +81016,69 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 21, - "output": 168 + "input": 0.2, + "output": 2.4 } }, - "minimax-m2.1": { - "id": "minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.27, - "output": 1.1 + "input": 0.1875, + "output": 1.125, + "cache_write": 0.234375 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "qwen/qwen3-coder-flash": { + "id": "qwen/qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -76929,63 +81087,74 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.3, - "reasoning": 8.4 + "input": 0.195, + "output": 0.975, + "cache_read": 0.039, + "cache_write": 0.24375 } }, - "pixtral-large-latest": { - "id": "pixtral-large-latest", - "name": "Pixtral Large (latest)", - "description": "Mistral's larger vision model for document-heavy image understanding and chat", - "family": "pixtral", - "attachment": true, + "qwen/qwen-2.5-coder-32b-instruct": { + "id": "qwen/qwen-2.5-coder-32b-instruct", + "name": "Qwen2.5 Coder 32B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2024-11-04", + "release_date": "2024-11-11", + "last_updated": "2024-11-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 32768, + "output": 29491 }, "cost": { - "input": 4, - "output": 12 + "input": 0.66, + "output": 1 } }, - "glm-4-32b-0414-128k": { - "id": "glm-4-32b-0414-128k", - "name": "GLM-4 32B (0414-128k)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": false, + "qwen/qwen3.5-flash-02-23": { + "id": "qwen/qwen3.5-flash-02-23", + "name": "Qwen: Qwen3.5-Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.5", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-02-25", + "last_updated": "2026-02-25", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -76993,28 +81162,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 0.065, + "output": 0.26 } }, - "seed-1-6-flash-250715": { - "id": "seed-1-6-flash-250715", - "name": "Seed 1.6 Flash (250715)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, + "qwen/qwen3-max-thinking": { + "id": "qwen/qwen3-max-thinking", + "name": "Qwen: Qwen3 Max Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", + "none", "high" ] } @@ -77022,40 +81189,39 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "release_date": "2026-02-09", + "last_updated": "2026-02-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.07, - "output": 0.3, - "cache_read": 0.015 + "input": 0.78, + "output": 3.9 } }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -77064,32 +81230,44 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 1.2 + "input": 0.65, + "output": 3.25, + "cache_read": 0.13, + "cache_write": 0.8125 } }, - "qwen3-vl-8b-instruct": { - "id": "qwen3-vl-8b-instruct", - "name": "Qwen3 VL 8B Instruct", + "qwen/qwen3.5-9b": { + "id": "qwen/qwen3.5-9b", + "name": "Qwen3.5 9B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-19", - "last_updated": "2025-08-19", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -77097,62 +81275,35 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.08, - "output": 0.5 - } - }, - "gpt-4o-mini-search-preview": { - "id": "gpt-4o-mini-search-preview", - "name": "GPT-4o Mini Search Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 + "context": 256000, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.1, + "output": 0.15 } }, - "glm-4.5v": { - "id": "glm-4.5v", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "qwen/qwen3.5-plus-20260420": { + "id": "qwen/qwen3.5-plus-20260420", + "name": "Qwen: Qwen3.5 Plus 2026-04-20", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.5", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", @@ -77163,34 +81314,43 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.6, + "input": 0.3, "output": 1.8, - "cache_read": 0.11 + "cache_write": 0.375 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "qwen/qwen3.8-flash": { + "id": "qwen/qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen3.8 Flash is a multimodal reasoning model from Alibaba. It is suited for coding assistance, agentic workflows, visual understanding, document and codebase analysis, desktop interaction, chart analysis, and long-video analysis.", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -77199,111 +81359,81 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 131072 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "cache_write": 0.5 + "input": 0.15, + "output": 0.47, + "cache_read": 0.016, + "cache_write": 0.2 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", + "none", "high" ] - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "audio", - "video", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.1625, + "output": 1.3 } }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "qwen/qwen3.6-max-preview": { + "id": "qwen/qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", + "none", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -77311,70 +81441,80 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.027, + "output": 6.162, + "cache_write": 1.28375 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "qwen/qwen3-30b-a3b": { + "id": "qwen/qwen3-30b-a3b", + "name": "Qwen3 30B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 40960, + "output": 16384 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.13, + "output": 0.52 } }, - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", "image", - "video", - "audio" + "video" ], "output": [ "text" @@ -77383,26 +81523,34 @@ "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "output": 235929 }, "cost": { - "input": 0.248, - "output": 1.485 + "input": 0.39, + "output": 2.34 } }, - "gpt-3.5-turbo": { - "id": "gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "qwen/qwen3-8b": { + "id": "qwen/qwen3-8b", + "name": "Qwen: Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, "structured_output": false, "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ "text" @@ -77413,38 +81561,26 @@ }, "open_weights": false, "limit": { - "context": 16385, - "output": 4096 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0 + "input": 0.117, + "output": 0.455 } }, - "seed-1-6-250615": { - "id": "seed-1-6-250615", - "name": "Seed 1.6 (250615)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "qwen/qwen3-vl-32b-instruct": { + "id": "qwen/qwen3-vl-32b-instruct", + "name": "Qwen: Qwen3 VL 32B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-06-25", - "last_updated": "2025-06-25", + "release_date": "2025-10-23", + "last_updated": "2025-10-23", "modalities": { "input": [ "text", @@ -77454,31 +81590,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05 + "input": 0.104, + "output": 0.416 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "qwen/qwen3-vl-8b-instruct": { + "id": "qwen/qwen3-vl-8b-instruct", + "name": "Qwen: Qwen3 VL 8B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2025-10-14", + "last_updated": "2025-10-14", "modalities": { "input": [ + "image", "text" ], "output": [ @@ -77487,33 +81623,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 3.75, - "cache_read": 0.125, - "cache_write": 3.125 + "input": 0.117, + "output": 0.455 } }, - "glm-4.5": { - "id": "glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", + "qwen/qwen3-30b-a3b-instruct-2507": { + "id": "qwen/qwen3-30b-a3b-instruct-2507", + "name": "Qwen: Qwen3 30B A3B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ "text" @@ -77522,39 +81651,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131000, - "output": 98304 + "context": 128000, + "output": 32000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 0.13, + "output": 0.52 } }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", + "qwen/qwen2.5-vl-72b-instruct": { + "id": "qwen/qwen2.5-vl-72b-instruct", + "name": "Qwen: Qwen2.5 VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "temperature": true, + "release_date": "2025-02-01", + "last_updated": "2025-02-01", "modalities": { "input": [ "text", @@ -77566,51 +81684,71 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 272000 + "context": 128000, + "output": 115200 }, "cost": { - "input": 15, - "output": 120 + "input": 0.8, + "output": 1, + "cache_read": 0.4 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", + "qwen/qwen-2.5-7b-instruct": { + "id": "qwen/qwen-2.5-7b-instruct", + "name": "Qwen: Qwen2.5 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-10-16", + "last_updated": "2024-10-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 29491 + }, + "cost": { + "input": 0.1, + "output": 0.2 + } + }, + "qwen/qwen3.5-plus-02-15": { + "id": "qwen/qwen3.5-plus-02-15", + "name": "Qwen: Qwen3.5 Plus 2026-02-15", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.5", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", + "none", "high" ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 24576 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", "image", - "audio", - "video", - "pdf" + "video" ], "output": [ "text" @@ -77618,33 +81756,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 0.26, + "output": 1.56 } }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, + "qwen/qwen-2.5-72b-instruct": { + "id": "qwen/qwen-2.5-72b-instruct", + "name": "Qwen2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -77652,59 +81786,34 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 32768, "output": 16384 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "ministral-8b-2512": { - "id": "ministral-8b-2512", - "name": "Ministral 8B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "mistral", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.15 + "input": 0.36, + "output": 0.4 } }, - "gpt-4": { - "id": "gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "qwen/qwen3-30b-a3b-thinking-2507": { + "id": "qwen/qwen3-30b-a3b-thinking-2507", + "name": "Qwen: Qwen3 30B A3B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, "structured_output": false, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ "text" @@ -77715,72 +81824,82 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 81920, + "output": 32768 }, "cost": { - "input": 30, - "output": 60 + "input": 0.2, + "output": 2.4 } }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "qwen/qwen3.6-27b": { + "id": "qwen/qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 262140 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 0.45, + "output": 2.7 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -77788,61 +81907,65 @@ }, "open_weights": false, "limit": { - "context": 262144, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.845, - "output": 3.38, - "cache_read": 0.6, - "cache_write": 3.75 + "input": 0.325, + "output": 1.95, + "cache_write": 0.40625 } }, - "mistral-small-2506": { - "id": "mistral-small-2506", - "name": "Mistral Small 3.2", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "qwen/qwen3-14b": { + "id": "qwen/qwen3-14b", + "name": "Qwen: Qwen3 14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 40960, "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.2275, + "output": 0.91 } }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "qwen/qwen3-vl-8b-thinking": { + "id": "qwen/qwen3-vl-8b-thinking", + "name": "Qwen: Qwen3 VL 8B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", "high" ] } @@ -77850,16 +81973,12 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2025-10-14", + "last_updated": "2025-10-14", "modalities": { "input": [ - "text", "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -77867,28 +81986,36 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 0.08333 + "input": 0.18, + "output": 2.1 } }, - "qwen-plus-latest": { - "id": "qwen-plus-latest", - "name": "Qwen Plus Latest", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Qwen3.7-Plus is a cost-effective model in Alibaba's Qwen3.7 series. It supports text and image input with text output, building on the series' text capabilities with a comprehensive upgrade to its...", "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-01-25", - "last_updated": "2025-01-25", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", @@ -77901,82 +82028,43 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 8192 + "output": 131072 }, "cost": { - "input": 0.4, - "output": 1.2, - "cache_read": 0.08, - "cache_write": 0.5 + "input": 0.32, + "output": 1.28, + "cache_read": 0.064, + "cache_write": 0.4 } }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen/qwen3.8-27b:free": { + "id": "qwen/qwen3.8-27b:free", + "name": "Qwen: Qwen3.8 27B (free)", + "description": "Qwen3.8 27B is an open-weight dense vision-language model from Qwen. It is suited for coding, professional workflows, research, multimodal interaction, and long-running agent tasks, with flexible thinking that can be...", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high", "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "gemma-4-31b-it": { - "id": "gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -77985,40 +82073,42 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 235929 }, "cost": { - "input": 0.13, - "output": 0.38 + "input": 0, + "output": 0 } }, - "grok-4-3": { - "id": "grok-4-3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "qwen/qwen3.8-max-0902": { + "id": "qwen/qwen3.8-max-0902", + "name": "Qwen3.8 Max 0902", + "description": "Qwen3.8 Max 0902 is an updated snapshot of Qwen3.8 Max from Alibaba's Qwen team. It is a 2.4-trillion-parameter mixture-of-experts model that accepts text, image, and video input and returns text,...", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -78027,73 +82117,57 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 30000 + "output": 131072 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.3125, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "llama-4-scout-17b-instruct": { - "id": "llama-4-scout-17b-instruct", - "name": "Llama 4 Scout 17B Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": true, + "morph/morph-v3-large": { + "id": "morph/morph-v3-large", + "name": "Morph: Morph V3 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "morph", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2025-07-07", + "last_updated": "2025-07-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 2048 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.18, - "output": 0.59 + "input": 0.9, + "output": 1.9 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": true, + "morph/morph-v3-fast": { + "id": "morph/morph-v3-fast", + "name": "Morph: Morph V3 Fast", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "morph", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2025-07-07", + "last_updated": "2025-07-07", "modalities": { "input": [ "text" @@ -78102,69 +82176,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 4096 + "context": 81920, + "output": 38000 }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 0.8, + "output": 1.2 } }, - "sonar-reasoning-pro": { - "id": "sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", - "family": "sonar-reasoning", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], - "tool_call": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 2, - "output": 8 - } - }, - "grok-4-20-non-reasoning": { - "id": "grok-4-20-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -78177,394 +82223,379 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "glm-4.7-flashx": { - "id": "glm-4.7-flashx", - "name": "GLM-4.7-FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, + "openai/gpt-5.4-image-2": { + "id": "openai/gpt-5.4-image-2", + "name": "OpenAI: GPT-5.4 Image 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0 - } - }, - "qwen3-30b-a3b-instruct-2507": { - "id": "qwen3-30b-a3b-instruct-2507", - "name": "Qwen3 30B A3B Instruct (2507)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": true, - "temperature": true, - "release_date": "2025-07-08", - "last_updated": "2025-07-08", + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ + "image", "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 8192 + "context": 272000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 8, + "output": 15, + "cache_read": 2 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 30, + "output": 180 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "openai/gpt-audio": { + "id": "openai/gpt-audio", + "name": "OpenAI: GPT Audio", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text" + "text", + "audio", + "pdf" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.931, - "output": 2.93, - "cache_read": 0.173, - "cache_write": 0 + "input": 2.5, + "output": 10 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1050000, - "output": 384000 + "context": 16385, + "output": 4096 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0.5, + "output": 1.5 } }, - "qwen3-next-80b-a3b-thinking": { - "id": "qwen3-next-80b-a3b-thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", - "attachment": false, + "openai/gpt-6-astra-pro": { + "id": "openai/gpt-6-astra-pro", + "name": "OpenAI: GPT-6 Astra Pro ($$$$)", + "description": "GPT-6 Astra Pro is the same underlying model as [GPT-6 Astra](https://openrouter.ai/openai/gpt-6-astra), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "structured_output": true, + "temperature": false, + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 1.2 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "glm-4.6": { - "id": "glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.55, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 30, + "output": 180 } }, - "qwen35-397b-a17b": { - "id": "qwen35-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text", + "pdf", "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "qwen3-235b-a22b-thinking-2507": { - "id": "qwen3-235b-a22b-thinking-2507", - "name": "Qwen3 235B A22B Thinking (2507)", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4o-2024-05-13": { + "id": "openai/gpt-4o-2024-05-13", + "name": "GPT-4o (2024-05-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-08", - "last_updated": "2025-07-08", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 8192 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 5, + "output": 15 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4o-mini-2024-07-18": { + "id": "openai/gpt-4o-mini-2024-07-18", + "name": "OpenAI: GPT-4o-mini (2024-07-18)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.06 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -78574,21 +82605,16 @@ "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -78601,21 +82627,21 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "gemini-pro-latest": { - "id": "gemini-pro-latest", - "name": "Gemini Pro Latest", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini", + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -78630,13 +82656,15 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-27", - "last_updated": "2026-02-27", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -78644,35 +82672,34 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "glm-4.5-x": { - "id": "glm-4.5-x", - "name": "GLM-4.5 X", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -78683,37 +82710,24 @@ "context": 128000, "output": 16384 }, - "status": "beta", "cost": { - "input": 2.2, - "output": 8.9, - "cache_read": 0.45 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "openai/gpt-chat-latest": { + "id": "openai/gpt-chat-latest", + "name": "OpenAI: GPT Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "release_date": "2026-05-05", + "last_updated": "2026-05-05", "modalities": { "input": [ "text", @@ -78726,27 +82740,60 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, "output": 128000 }, "cost": { "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "output": 30, + "cache_read": 0.5 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "openai/gpt-audio-mini": { + "id": "openai/gpt-audio-mini", + "name": "OpenAI: GPT Audio Mini", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text", + "audio", + "pdf" + ], + "output": [ + "text", + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.6, + "output": 2.4 + } + }, + "openai/gpt-5.6-sol-pro": { + "id": "openai/gpt-5.6-sol-pro", + "name": "GPT-5.6 Sol", + "description": "GPT-5.6 Sol Pro is the same underlying model as [GPT-5.6 Sol](https://openrouter.ai/openai/gpt-5.6-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -78755,16 +82802,17 @@ ] } ], - "tool_call": false, + "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -78772,32 +82820,46 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5 } }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "Grok 4.1 Fast Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -78805,28 +82867,27 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", "medium", "high", "xhigh" @@ -78837,12 +82898,13 @@ "structured_output": true, "temperature": false, "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ "text" @@ -78855,37 +82917,36 @@ "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 21, + "output": 168 } }, - "grok-4-20-beta-0309-reasoning": { - "id": "grok-4-20-beta-0309-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ @@ -78894,97 +82955,119 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "low", - "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 + } + }, + "openai/gpt-4": { + "id": "openai/gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 8191, + "output": 4096 }, "cost": { - "input": 1.26, - "output": 3.96, - "cache_read": 0.234, - "cache_write": 0 + "input": 30, + "output": 60 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -78992,30 +83075,34 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.9, - "cache_read": 0.015 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "glm-4.5-airx": { - "id": "glm-4.5-airx", - "name": "GLM-4.5 AirX", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ "text" @@ -79023,33 +83110,44 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 1.1, - "output": 4.5, - "cache_read": 0.22 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "gpt-5-chat-latest": { - "id": "gpt-5-chat-latest", - "name": "GPT-5 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", + "temperature": false, + "knowledge": "2024-05-30", "release_date": "2025-08-07", "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -79062,94 +83160,112 @@ "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "ministral-3b-2512": { - "id": "ministral-3b-2512", - "name": "Ministral 3B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "mistral", + "openai/o1": { + "id": "openai/o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "qwen2-5-vl-72b-instruct": { - "id": "qwen2-5-vl-72b-instruct", - "name": "Qwen2.5-VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 15, + "output": 120 } }, - "glm-4.6v-flashx": { - "id": "glm-4.6v-flashx", - "name": "GLM-4.6V FlashX", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "openai/gpt-3.5-turbo-instruct": { + "id": "openai/gpt-3.5-turbo-instruct", + "name": "OpenAI: GPT-3.5 Turbo Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2023-09-28", + "last_updated": "2023-09-28", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" @@ -79157,65 +83273,78 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16000 + "context": 4095, + "output": 3685 }, "cost": { - "input": 0.04, - "output": 0.4, - "cache_read": 0.004 + "input": 1.5, + "output": 2 } }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "openai/gpt-4o-2024-08-06": { + "id": "openai/gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "knowledge": "2023-09", + "release_date": "2024-08-06", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "qwen3-vl-plus": { - "id": "qwen3-vl-plus", - "name": "Qwen3-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "openai/gpt-6-astra": { + "id": "openai/gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's flagship model for demanding end-to-end work. It is suited for advanced analysis, software engineering, deep research, scientific work, and document creation, with particular strengths in long-horizon...", + "family": "gpt-astra", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -79223,45 +83352,43 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.6, - "reasoning": 4.8, - "cache_read": 0.04, - "cache_write": 0.25 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5.6-luna-pro": { + "id": "openai/gpt-5.6-luna-pro", + "name": "GPT-5.6 Luna", + "description": "GPT-5.6 Luna Pro is the same underlying model as [GPT-5.6 Luna](https://openrouter.ai/openai/gpt-5.6-luna), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -79274,27 +83401,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -79309,8 +83438,9 @@ "last_updated": "2025-11-13", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ "text" @@ -79328,66 +83458,109 @@ "cache_read": 0.125 } }, - "minimax-m2": { - "id": "minimax-m2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, + "openai/gpt-5-image-mini": { + "id": "openai/gpt-5-image-mini", + "name": "OpenAI: GPT-5 Image Mini", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "release_date": "2025-10-16", + "last_updated": "2025-10-16", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ + "image", "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, + "context": 400000, "output": 128000 }, "cost": { - "input": 0.2, - "output": 1, - "cache_read": 0.03 + "input": 2.5, + "output": 2, + "cache_read": 0.25 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 + } + }, + "openai/o3-pro": { + "id": "openai/o3-pro", + "name": "o3-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high", - "xhigh", - "max" + "none", + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2024-05", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", "modalities": { "input": [ "text", - "image", - "pdf" + "pdf", + "image" ], "output": [ "text" @@ -79395,107 +83568,117 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 20, + "output": 80 } }, - "gpt-5.3-chat-latest": { - "id": "gpt-5.3-chat-latest", - "name": "GPT-5.3 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": false, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.018, + "output": 0.09 } }, - "grok-4-5": { - "id": "grok-4-5", - "name": "Grok 4.5", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "openai/gpt-oss-safeguard-20b": { + "id": "openai/gpt-oss-safeguard-20b", + "name": "GPT OSS Safeguard 20B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", "high" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "release_date": "2025-10-29", + "last_updated": "2025-10-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 0.075, + "output": 0.3, + "cache_read": 0.0375 } }, - "claude-3-opus": { - "id": "claude-3-opus", - "name": "Claude 3 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude", + "openai/o3-mini-high": { + "id": "openai/o3-mini-high", + "name": "OpenAI: o3 Mini High", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2024-03-04", - "last_updated": "2024-03-04", + "structured_output": true, + "temperature": false, + "release_date": "2025-02-12", + "last_updated": "2025-02-12", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" @@ -79504,65 +83687,61 @@ "open_weights": false, "limit": { "context": 200000, - "output": 4096 + "output": 100000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "qwen-omni-turbo": { - "id": "qwen-omni-turbo", - "name": "Qwen-Omni Turbo", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, + "openai/gpt-3.5-turbo-16k": { + "id": "openai/gpt-3.5-turbo-16k", + "name": "OpenAI: GPT-3.5 Turbo 16k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01-19", - "last_updated": "2025-03-26", + "release_date": "2023-08-28", + "last_updated": "2023-08-28", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "pdf" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 2048 + "context": 16385, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 3, + "output": 4 } }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-5.2-chat": { + "id": "openai/gpt-5.2-chat", + "name": "OpenAI: GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "release_date": "2025-12-10", + "last_updated": "2025-12-10", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ @@ -79571,19 +83750,19 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 128000, + "output": 32000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "openai/gpt-6-sol-pro": { + "id": "openai/gpt-6-sol-pro", + "name": "OpenAI: GPT-6 Sol Pro", + "description": "GPT-6 Sol Pro is the same underlying model as [GPT-6 Sol](https://openrouter.ai/openai/gpt-6-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", "family": "gpt", "attachment": true, "reasoning": true, @@ -79595,20 +83774,21 @@ "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text", - "image" + "pdf", + "image", + "text" ], "output": [ "text" @@ -79616,231 +83796,110 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai/gpt-5-image": { + "id": "openai/gpt-5-image", + "name": "OpenAI: GPT-5 Image ($$$$)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high", - "xhigh" + "high" ] } ], - "tool_call": true, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "release_date": "2025-10-14", + "last_updated": "2025-10-14", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 10, + "output": 10, + "cache_read": 1.25 } }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "qwen-flash": { - "id": "qwen-flash", - "name": "Qwen Flash", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0.0625 - } - }, - "qwen3-4b-fp8": { - "id": "qwen3-4b-fp8", - "name": "Qwen3 4B FP8", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.03, - "output": 0.03 - } - }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text", + "pdf", "image", - "audio", - "video", - "pdf" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 - } - }, - "qwen2-5-vl-32b-instruct": { - "id": "qwen2-5-vl-32b-instruct", - "name": "Qwen2.5 VL 32B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-03-15", - "last_updated": "2025-03-15", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.2 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-nano", + "description": "GPT-5.6 Luna is a fast, cost-efficient model in OpenAI's GPT-5.6 series. It is suited for high-volume, latency-sensitive tasks such as chat, classification, and lightweight agentic workflows, providing capable reasoning for...", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -79879,14 +83938,15 @@ "output": 128000 }, "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 } }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "family": "gpt-codex", "attachment": true, @@ -79897,7 +83957,8 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh" ] } ], @@ -79910,7 +83971,8 @@ "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -79923,118 +83985,118 @@ "output": 128000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": false, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", + "pdf", "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.405, - "output": 1.98, - "cache_read": 0.225 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "seed-1-6-250915": { - "id": "seed-1-6-250915", - "name": "Seed 1.6 (250915)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-mini", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -80047,41 +84109,33 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "gpt-5.2-chat-latest": { - "id": "gpt-5.2-chat-latest", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "openai/gpt-4o-2024-11-20": { + "id": "openai/gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -80093,47 +84147,47 @@ "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "minimax-text-01": { - "id": "minimax-text-01", - "name": "MiniMax Text 01", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "openai/gpt-3.5-turbo-0613": { + "id": "openai/gpt-3.5-turbo-0613", + "name": "OpenAI: GPT-3.5 Turbo (older v0613)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-01-15", - "last_updated": "2025-01-15", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 4095, + "output": 3685 }, "cost": { - "input": 0.2, - "output": 1.1 + "input": 1, + "output": 2 } }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": false, "tool_call": true, @@ -80145,7 +84199,8 @@ "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -80153,40 +84208,95 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1047576, "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "openai/gpt-5.6-terra-pro": { + "id": "openai/gpt-5.6-terra-pro", + "name": "GPT-5.6 Terra", + "description": "GPT-5.6 Terra Pro is the same underlying model as [GPT-5.6 Terra](https://openrouter.ai/openai/gpt-5.6-terra), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", + "family": "gpt-terra", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ + "text", + "image", + "pdf" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "openai/gpt-6-luna": { + "id": "openai/gpt-6-luna", + "name": "GPT-6 Luna", + "description": "GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol. It is suited for high-volume and latency-sensitive workloads such as chat, classification, and lightweight agentic...", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" ], "output": [ "text" @@ -80194,30 +84304,42 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32766 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.25 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125 } }, - "qwen-vl-max": { - "id": "qwen-vl-max", - "name": "Qwen-VL Max", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/o4-mini-high": { + "id": "openai/o4-mini-high", + "name": "OpenAI: o4 Mini High", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-08", - "last_updated": "2025-08-13", + "structured_output": true, + "temperature": false, + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ "text" @@ -80225,55 +84347,74 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.8, - "output": 3.2 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "llama-3-8b-instruct": { - "id": "llama-3-8b-instruct", - "name": "Llama 3 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + "openai/gpt-6-luna-pro": { + "id": "openai/gpt-6-luna-pro", + "name": "OpenAI: GPT-6 Luna Pro", + "description": "GPT-6 Luna Pro is the same underlying model as [GPT-6 Luna](https://openrouter.ai/openai/gpt-6-luna), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "temperature": false, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0.04, - "output": 0.04 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "GPT-5.6 Terra is a balanced model in OpenAI's GPT-5.6 series, positioned between the flagship Sol tier and the cost-efficient Luna tier. It is suited for everyday coding, reasoning, and agentic...", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -80283,10 +84424,11 @@ } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -80299,31 +84441,35 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "sonar": { - "id": "sonar", - "name": "Sonar", - "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", - "family": "sonar", - "attachment": false, + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -80331,26 +84477,36 @@ }, "open_weights": false, "limit": { - "context": 130000, + "context": 128000, "output": 4096 }, "cost": { - "input": 1, - "output": 1 + "input": 10, + "output": 30 } }, - "qwen-max": { - "id": "qwen-max", - "name": "Qwen Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -80359,27 +84515,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 8192 + "context": 131072, + "output": 65536 }, "cost": { - "input": 1.6, - "output": 6.4 + "input": 0.03, + "output": 0.17, + "cache_read": 0.03 } }, - "o1": { - "id": "o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -80389,13 +84547,13 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ @@ -80404,52 +84562,64 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 0.25, + "output": 2, + "cache_read": 0.03 } }, - "minimax-m2.5-highspeed": { - "id": "minimax-m2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "deepseek-v3.1": { - "id": "deepseek-v3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -80465,76 +84635,112 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.56, - "output": 1.68, - "cache_read": 0.112 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "ministral-14b-2512": { - "id": "ministral-14b-2512", - "name": "Ministral 14B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "mistral", + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "GPT-5.6 Sol is the flagship model in OpenAI's GPT-5.6 series. It is suited for complex reasoning, coding, and agentic workflows, and is particularly strong at command-line and multi-step coding tasks...", + "family": "gpt-sol", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 8192 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5 } }, - "sonar-pro": { - "id": "sonar-pro", - "name": "Sonar Pro", - "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", - "family": "sonar-pro", + "openai/gpt-6-sol": { + "id": "openai/gpt-6-sol", + "name": "GPT-6 Sol", + "description": "GPT-6 Sol is the cost-efficient high-end model in OpenAI's GPT-6 series, positioned below the flagship GPT-6 Astra and above the fast GPT-6 Luna tier. It is suited for demanding professional...", + "family": "gpt-sol", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -80542,70 +84748,74 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "glm-4.6v": { - "id": "glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "openai/o1-pro": { + "id": "openai/o1-pro", + "name": "o1-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "tool_call": false, + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2025-03-19", + "last_updated": "2025-03-19", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.05 + "input": 150, + "output": 600 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "microsoft/phi-4": { + "id": "microsoft/phi-4", + "name": "Microsoft: Phi 4", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2025-01-10", + "last_updated": "2025-01-10", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -80613,33 +84823,28 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 16384, + "output": 14745 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.07, + "output": 0.14 } }, - "claude-haiku-4-5-free": { - "id": "claude-haiku-4-5-free", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, + "microsoft/wizardlm-2-8x22b": { + "id": "microsoft/wizardlm-2-8x22b", + "name": "WizardLM-2 8x22B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2024-04-16", + "last_updated": "2024-04-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -80647,27 +84852,38 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 200000 + "context": 65535, + "output": 8000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.62, + "output": 0.62 } - }, - "minimax-m2.1-lightning": { - "id": "minimax-m2.1-lightning", - "name": "MiniMax M2.1 Lightning", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + } + } + }, + "venice": { + "id": "venice", + "env": [ + "VENICE_API_KEY" + ], + "npm": "venice-ai-sdk-provider", + "name": "Venice AI", + "doc": "https://docs.venice.ai", + "models": { + "qwen-3-8-2-4t-a95b": { + "id": "qwen-3-8-2-4t-a95b", + "name": "Qwen 3.8 2.4T", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2026-08-12", + "last_updated": "2026-08-13", "modalities": { "input": [ "text" @@ -80678,19 +84894,20 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.12, - "output": 0.48 + "input": 2.5, + "output": 7.5, + "cache_read": 0.3125 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "openai-gpt-56-luna": { + "id": "openai-gpt-56-luna", + "name": "GPT-5.6 Luna", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -80701,21 +84918,21 @@ "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -80723,96 +84940,106 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 0.3125, + "tiers": [ + { + "input": 0.5, + "output": 2.25, + "cache_read": 0.05, + "cache_write": 0.625, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.5, + "output": 2.25, + "cache_read": 0.05, + "cache_write": 0.625 + } } }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "grok-4-5": { + "id": "grok-4-5", + "name": "Grok 4.5", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-07-07", + "last_updated": "2026-07-08", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 500000, + "output": 32000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028, + "input": 2.27, + "output": 6.8, + "cache_read": 0.34, "tiers": [ { - "input": 0.8, - "output": 4, - "cache_read": 0.16, + "input": 4.53, + "output": 13.6, + "cache_read": 0.68, "tier": { "type": "context", - "size": 256000 + "size": 200000 } } ], "context_over_200k": { - "input": 0.8, - "output": 4, - "cache_read": 0.16 + "input": 4.53, + "output": 13.6, + "cache_read": 0.68 } } }, - "mimo-v2-omni": { - "id": "mimo-v2-omni", - "name": "MiMo-V2-Omni", - "description": "MiMo omni model for text, image, video, audio, and agents", - "family": "mimo", - "attachment": true, - "reasoning": false, + "aion-labs-aion-3-5": { + "id": "aion-labs-aion-3-5", + "name": "Aion 3.5", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": true, + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -80820,20 +85047,20 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08 + "input": 3.75, + "output": 7.5, + "cache_read": 0.9375 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "kimi-k3-fast-api": { + "id": "kimi-k3-fast-api", + "name": "Kimi K3 Fast", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -80842,18 +85069,16 @@ "values": [ "none", "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", @@ -80863,35 +85088,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 4.5, + "output": 22.5, + "cache_read": 0.45 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "qwen3-6-27b": { + "id": "qwen3-6-27b", + "name": "Qwen 3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-04-24", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -80904,54 +85134,47 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 2.2, - "cache_read": 0.08 + "input": 0.325, + "output": 3.25 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, + "hermes-3-llama-3.1-405b": { + "id": "hermes-3-llama-3.1-405b", + "name": "Hermes 3 Llama 3.1 405b", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "hermes", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "tool_call": false, + "release_date": "2025-09-25", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 1.1, + "output": 3 } }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "kimi-k2-7-code": { + "id": "kimi-k2-7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -80966,110 +85189,79 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-06-13", + "last_updated": "2026-06-16", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 256000, "output": 65536 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.75, + "output": 3.5, + "cache_read": 0.16 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "google-gemma-3-27b-it": { + "id": "google-gemma-3-27b-it", + "name": "Google Gemma 3 27B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2024-08", + "release_date": "2025-11-04", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 198000, + "output": 16384 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.12, + "output": 0.2 } }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "qwen-3-7-max": { + "id": "qwen-3-7-max", + "name": "Qwen 3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-05-22", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -81077,171 +85269,104 @@ }, "open_weights": false, "limit": { - "context": 262144, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.108, - "output": 0.675, - "cache_read": 0.06 - } - }, - "llama-4-maverick-17b-instruct": { - "id": "llama-4-maverick-17b-instruct", - "name": "Llama 4 Maverick 17B Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 2048 - }, - "cost": { - "input": 0.27, - "output": 0.85 + "input": 2.7, + "output": 8.05, + "cache_read": 0.27, + "cache_write": 3.35 } }, - "qwen3-coder-480b-a35b-instruct": { - "id": "qwen3-coder-480b-a35b-instruct", - "name": "Qwen3-Coder 480B-A35B Instruct", - "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", + "qwen-3-8-max": { + "id": "qwen-3-8-max", + "name": "Qwen 3.8 Max", + "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", "family": "qwen", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-07-22", + "last_updated": "2026-07-19", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.3 + "input": 2.5, + "output": 7.5, + "cache_read": 0.3125, + "cache_write": 3.125 } }, - "devstral-2512": { - "id": "devstral-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2 - } - }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", "high", - "xhigh", "max" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-08-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.175, + "output": 0.35, + "cache_read": 0.035 } }, - "gemma-4-26b-a4b-it": { - "id": "gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "venice-uncensored-role-play": { + "id": "venice-uncensored-role-play", + "name": "Venice Role Play Uncensored", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "venice", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-02-20", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -81253,58 +85378,75 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.07, - "output": 0.34 + "input": 0.5, + "output": 2 } }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "grok-build-0-1": { + "id": "grok-build-0-1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-05-21", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 256000, "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.15 + "input": 1, + "output": 2, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 4, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 4, + "cache_read": 0.4 + } } }, - "qwq-plus": { - "id": "qwq-plus", - "name": "QwQ Plus", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "Qwen 3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-03-05", - "last_updated": "2025-03-05", + "release_date": "2025-04-29", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -81313,38 +85455,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.8, - "output": 2.4 + "input": 0.15, + "output": 0.75 } }, - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "Grok 4.1 Fast Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "openai-gpt-6-luna": { + "id": "openai-gpt-6-luna", + "name": "GPT-6 Luna", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", @@ -81356,21 +85502,41 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.125, + "output": 0.625, + "cache_read": 0.0125, + "cache_write": 0.15625, + "tiers": [ + { + "input": 0.25, + "output": 0.9375, + "cache_read": 0.025, + "cache_write": 0.3125, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.25, + "output": 0.9375, + "cache_read": 0.025, + "cache_write": 0.3125 + } } }, - "glm-5-2": { - "id": "glm-5-2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k3", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -81378,21 +85544,20 @@ "values": [ "none", "low", - "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -81400,27 +85565,40 @@ }, "open_weights": true, "limit": { - "context": 1024000, + "context": 1000000, "output": 131072 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 3.75, + "output": 18.75, + "cache_read": 0.375 } }, - "qwen3-vl-30b-a3b-instruct": { - "id": "qwen3-vl-30b-a3b-instruct", - "name": "Qwen3 VL 30B A3B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai-gpt-54-mini": { + "id": "openai-gpt-54-mini", + "name": "GPT-5.4 Mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-02", - "last_updated": "2025-10-02", + "knowledge": "2025-08-31", + "release_date": "2026-03-27", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -81430,32 +85608,35 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.7 + "input": 0.9375, + "output": 5.625, + "cache_read": 0.09375 } }, - "qwen3-vl-flash": { - "id": "qwen3-vl-flash", - "name": "Qwen3 VL Flash", + "qwen-3-8-flash": { + "id": "qwen-3-8-flash", + "name": "Qwen 3.8 Flash", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-10-09", - "last_updated": "2025-10-09", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -81463,27 +85644,27 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01 + "input": 0.14, + "output": 0.49, + "cache_read": 0.014 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, + "olafangensan-glm-4.7-flash-heretic": { + "id": "olafangensan-glm-4.7-flash-heretic", + "name": "GLM 4.7 Flash Heretic", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", "high" @@ -81492,77 +85673,51 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2026-02-04", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 200000, + "output": 24000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.07, + "output": 0.4, + "cache_read": 0.035 } }, - "llama-3.1-nemotron-ultra-253b": { - "id": "llama-3.1-nemotron-ultra-253b", - "name": "Llama 3.1 Nemotron Ultra 253B", - "description": "Flagship Nemotron model for high-throughput reasoning and complex agents", - "family": "nemotron", + "deepseek-v4-flash-0731-fast": { + "id": "deepseek-v4-flash-0731-fast", + "name": "DeepSeek V4 Flash 0731 Fast", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-07", - "last_updated": "2025-04-07", + "knowledge": "2025-05", + "release_date": "2026-08-09", + "last_updated": "2026-08-11", "modalities": { "input": [ "text" @@ -81573,27 +85728,29 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 1.8 + "input": 0.35, + "output": 0.7, + "cache_read": 0.0875 } }, - "gpt-4-turbo": { - "id": "gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "knowledge": "2025-05", + "release_date": "2025-12-06", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -81605,26 +85762,38 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 198000, + "output": 32768 }, "cost": { - "input": 10, - "output": 30 + "input": 6, + "output": 30, + "cache_read": 0.6, + "cache_write": 7.5 } }, - "qwen3-coder-flash": { - "id": "qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "zai-org-glm-5-2": { + "id": "zai-org-glm-5-2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ "text" @@ -81633,23 +85802,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 65536 + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.5, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "grok-build-0-1": { - "id": "grok-build-0-1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", + "claude-opus-4-8-fast": { + "id": "claude-opus-4-8-fast", + "name": "Claude Opus 4.8 Fast", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -81658,20 +85826,22 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -81679,43 +85849,28 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 4, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 4, - "cache_read": 0.4 - } + "input": 12, + "output": 60, + "cache_read": 1.2, + "cache_write": 15 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, + "z-ai-glm-5-turbo": { + "id": "z-ai-glm-5-turbo", + "name": "GLM 5 Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", "high" @@ -81724,75 +85879,54 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 - } - }, - "grok-4": { - "id": "grok-4", - "name": "Grok 4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "release_date": "2026-03-15", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 200000, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "llama-3.2-11b-instruct": { - "id": "llama-3.2-11b-instruct", - "name": "Llama 3.2 11B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "zai-org-glm-4.6": { + "id": "zai-org-glm-4.6", + "name": "GLM 4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2025-04", + "release_date": "2024-04-01", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -81803,25 +85937,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 198000, + "output": 16384 }, "cost": { - "input": 0.07, - "output": 0.33 + "input": 0.43, + "output": 1.75, + "cache_read": 0.08 } }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", + "openai-gpt-54-pro": { + "id": "openai-gpt-54-pro", "name": "GPT-5.4 Pro", "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "medium", "high", "xhigh" @@ -81829,11 +85965,11 @@ } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": false, "knowledge": "2025-08-31", "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -81845,129 +85981,77 @@ }, "open_weights": false, "limit": { - "context": 1050000, + "context": 1000000, "input": 922000, "output": 128000 }, "cost": { - "input": 30, - "output": 180 - } - }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" + "input": 37.5, + "output": 225, + "tiers": [ + { + "input": 75, + "output": 337.5, + "tier": { + "type": "context", + "size": 272000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131000, - "output": 98304 - }, - "cost": { - "input": 0.13, - "output": 0.85, - "cache_read": 0.025, - "cache_write": 0 + "context_over_200k": { + "input": 75, + "output": 337.5 + } } }, - "kimi-k2.7-code-highspeed": { - "id": "kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": true, + "aion-labs-aion-3-0": { + "id": "aion-labs-aion-3-0", + "name": "Aion 3.0", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 32768 }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 + "input": 3.75, + "output": 7.5, + "cache_read": 0.9375 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-07-31", + "release_date": "2025-01-15", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -81975,47 +86059,45 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 198000, "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 3.75, + "output": 18.75, + "cache_read": 0.375, + "cache_write": 4.69 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "claude-opus-5-5": { + "id": "claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-18", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -82023,60 +86105,33 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 4.8, + "output": 24, + "cache_read": 0.24, + "cache_write": 6 } }, - "glm-4.7-flash": { - "id": "glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, + "kimi-k2-5": { + "id": "kimi-k2-5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.06, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0 - } - }, - "qwen3-vl-235b-a22b-instruct": { - "id": "qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "temperature": false, + "knowledge": "2024-04", + "release_date": "2026-01-27", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -82088,37 +86143,44 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 0.56, + "output": 3.5, + "cache_read": 0.22 } }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", - "attachment": false, + "claude-fable-5-1": { + "id": "claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-08-29", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -82127,46 +86189,35 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 128000 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 12, + "output": 60, + "cache_read": 0.3, + "cache_write": 15 } }, - "gpt-4o-search-preview": { - "id": "gpt-4o-search-preview", - "name": "GPT-4o Search Preview", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "gemini-3-7-flash": { + "id": "gemini-3-7-flash", + "name": "Gemini 3.7 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", + "knowledge": "2026-03", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -82174,127 +86225,93 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10 - } - }, - "llama-3.2-3b-instruct": { - "id": "llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2024-09-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.03, - "output": 0.05 + "input": 0.9375, + "output": 4.6875, + "cache_read": 0.09375 } }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "grok-4-20": { + "id": "grok-4-20", + "name": "Grok 4.20", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "structured_output": true, + "release_date": "2026-03-12", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 2000000, + "output": 128000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036, + "input": 1.42, + "output": 2.83, + "cache_read": 0.23, "tiers": [ { - "input": 2, - "output": 6, - "cache_read": 0.4, + "input": 2.83, + "output": 5.67, + "cache_read": 0.45, "tier": { "type": "context", - "size": 256000 + "size": 200000 } } ], "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 + "input": 2.83, + "output": 5.67, + "cache_read": 0.45 } } }, - "gpt-5.5-pro": { - "id": "gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "openai-gpt-56-sol-pro": { + "id": "openai-gpt-56-sol-pro", + "name": "GPT-5.6 Sol Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -82302,17 +86319,20 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 30, - "output": 180, + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, "tiers": [ { - "input": 60, - "output": 270, + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, "tier": { "type": "context", "size": 272000 @@ -82320,29 +86340,31 @@ } ], "context_over_200k": { - "input": 60, - "output": 270 + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 } } }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", + "seed-2-1-turbo": { + "id": "seed-2-1-turbo", + "name": "Seed 2.1 Turbo", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-06-28", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -82350,26 +86372,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.625, + "output": 3.125, + "cache_read": 0.125 } }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "z-ai-glm-5v-turbo": { + "id": "z-ai-glm-5v-turbo", + "name": "GLM 5V Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -82379,11 +86402,12 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-01", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -82391,65 +86415,38 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32766 + "context": 200000, + "output": 32768 }, "cost": { - "input": 0.04, - "output": 0.15 + "input": 1.5, + "output": 5, + "cache_read": 0.3 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "openai-gpt-55-pro": { + "id": "openai-gpt-55-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 203000, - "output": 131072 - }, - "cost": { - "input": 0.72, - "output": 2.3, - "cache_read": 0.144, - "cache_write": 0 - } - }, - "qwen-max-latest": { - "id": "qwen-max-latest", - "name": "Qwen Max Latest", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-01-25", - "last_updated": "2025-01-25", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-24", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -82461,58 +86458,93 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 1000000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.6, - "output": 6.4 + "input": 37.5, + "output": 225 } }, - "mistral-large-2512": { - "id": "mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", + "qwen-3-7-plus": { + "id": "qwen-3-7-plus", + "name": "Qwen 3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 65536 }, "cost": { "input": 0.5, - "output": 1.5 + "output": 2, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 1.5, + "output": 6, + "cache_read": 0.15, + "cache_write": 1.875, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.5, + "output": 6, + "cache_read": 0.15, + "cache_write": 1.875 + } } }, - "qwen-turbo": { - "id": "qwen-turbo", - "name": "Qwen Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "qwen3-235b-a22b-thinking-2507": { + "id": "qwen3-235b-a22b-thinking-2507", + "name": "Qwen 3 235B A22B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11-01", - "last_updated": "2025-04-28", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "release_date": "2025-04-29", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -82521,60 +86553,70 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 128000, "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.2, - "reasoning": 0.5 + "input": 0.45, + "output": 3.5 } }, - "custom": { - "id": "custom", - "name": "Custom Model", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": true, - "reasoning": false, + "zai-org-glm-4.7-flash": { + "id": "zai-org-glm-4.7-flash", + "name": "GLM 4.7 Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "knowledge": "2025-04", + "release_date": "2026-01-29", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.06, + "output": 0.4, + "cache_read": 0.01 } }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt", + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -82586,14 +86628,13 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2026-05", + "release_date": "2026-07-23", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -82601,21 +86642,21 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 5, + "input": 6, "output": 30, - "cache_read": 0.5 + "cache_read": 0.6, + "cache_write": 7.5 } }, - "grok-4-20-reasoning": { - "id": "grok-4-20-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", + "openai-gpt-6-astra": { + "id": "openai-gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -82624,20 +86665,22 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-05", + "last_updated": "2026-09-04", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -82645,52 +86688,60 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, "tiers": [ { - "input": 2.5, - "output": 5, - "cache_read": 0.4, + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, "tier": { "type": "context", - "size": 200000 + "size": 272000 } } ], "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 } } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": true, + "minimax-m27": { + "id": "minimax-m27", + "name": "MiniMax M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-03-18", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -82698,28 +86749,37 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 16384 + "context": 198000, + "output": 32768 }, "cost": { - "input": 0.26, - "output": 0.38, - "cache_read": 0.13 + "input": 0.375, + "output": 1.5, + "cache_read": 0.06875 } }, - "qwen3.6-max-preview": { - "id": "qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "mercury-2-5": { + "id": "mercury-2-5", + "name": "Mercury 2.5", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "structured_output": true, + "release_date": "2026-09-08", + "last_updated": "2026-09-09", "modalities": { "input": [ "text" @@ -82730,59 +86790,68 @@ }, "open_weights": false, "limit": { - "context": 262144, + "context": 260000, "output": 65536 }, "cost": { - "input": 1.3, - "output": 7.8, - "cache_read": 0.13, - "cache_write": 1.625 + "input": 0.04999999999999999, + "output": 0.18749999999999994, + "cache_read": 0.004999999999999999 } }, - "auto": { - "id": "auto", - "name": "Auto Route", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": true, + "qwen3-next-80b": { + "id": "qwen3-next-80b", + "name": "Qwen 3 Next 80b", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "release_date": "2025-04-29", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 256000, "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.35, + "output": 1.9 } }, - "codestral-2508": { - "id": "codestral-2508", - "name": "Codestral", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "mistral", + "minimax-m25": { + "id": "minimax-m25", + "name": "MiniMax M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-07-30", - "last_updated": "2025-07-30", + "release_date": "2026-02-12", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -82793,30 +86862,32 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 198000, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0.27, + "output": 0.95, + "cache_read": 0.03 } }, - "qwen3-235b-a22b-fp8": { - "id": "qwen3-235b-a22b-fp8", - "name": "Qwen3 235B A22B FP8", + "qwen3-6-35b-a3b": { + "id": "qwen3-6-35b-a3b", + "name": "Qwen 3.6 35B A3B", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "release_date": "2026-07-20", + "last_updated": "2026-07-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -82824,26 +86895,26 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 8192 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0.1, + "output": 1 } }, - "fugu-ultra": { - "id": "fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "grok-4-20-multi-agent": { + "id": "grok-4-20-multi-agent", + "name": "Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-22", - "last_updated": "2026-06-22", + "tool_call": false, + "structured_output": true, + "release_date": "2026-03-12", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -82855,21 +86926,37 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 2000000, + "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 1.42, + "output": 2.83, + "cache_read": 0.23, + "tiers": [ + { + "input": 2.83, + "output": 5.67, + "cache_read": 0.45, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.83, + "output": 5.67, + "cache_read": 0.45 + } } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", - "attachment": true, + "zai-org-glm-4.7": { + "id": "zai-org-glm-4.7", + "name": "GLM 4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -82877,51 +86964,61 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-24", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 198000, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.55, + "output": 2.65, + "cache_read": 0.11 } }, - "qwen-vl-plus": { - "id": "qwen-vl-plus", - "name": "Qwen-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen3-5-9b": { + "id": "qwen3-5-9b", + "name": "Qwen 3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-08-15", + "release_date": "2026-03-05", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -82931,93 +87028,71 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 256000, + "output": 32768 }, "cost": { - "input": 0.21, - "output": 0.64 + "input": 0.1, + "output": 0.15 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "qwen3-vl-235b-a22b": { + "id": "qwen3-vl-235b-a22b", + "name": "Qwen3 VL 235B", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "structured_output": true, + "release_date": "2026-01-16", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 0.21, + "output": 1.9, + "cache_read": 0.1 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2026-01-31", + "release_date": "2026-06-10", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -83029,21 +87104,21 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 12, + "output": 60, + "cache_read": 1.2, + "cache_write": 15 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "qwen3-5-35b-a3b": { + "id": "qwen3-5-35b-a3b", + "name": "Qwen 3.5 35B A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -83053,82 +87128,20 @@ "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2026-02-25", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - } - } - }, - "alibaba-coding-plan-cn": { - "id": "alibaba-coding-plan-cn", - "env": [ - "ALIBABA_CODING_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://coding.dashscope.aliyuncs.com/v1", - "name": "Alibaba Coding Plan (China)", - "doc": "https://help.aliyun.com/zh/model-studio/coding-plan", - "models": { - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" + "video" ], "output": [ "text" @@ -83136,36 +87149,38 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.3125, + "output": 1.25, + "cache_read": 0.15625 } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -83176,33 +87191,41 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 16384 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.65, + "output": 4.95, + "cache_read": 0.165 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": false, + "openai-gpt-56-terra": { + "id": "openai-gpt-56-terra", + "name": "GPT-5.6 Terra", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-terra", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -83215,31 +87238,43 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25 + } } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "llama-3.3-70b": { + "id": "llama-3.3-70b", + "name": "Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2025-04-06", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -83248,40 +87283,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 4096 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 0.7, + "output": 2.8 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "openai-gpt-4o-mini-2024-07-18": { + "id": "openai-gpt-4o-mini-2024-07-18", + "name": "GPT-4o Mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2023-09", + "release_date": "2026-02-28", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -83289,64 +87317,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { "input": 0.1875, - "output": 1.125, - "cache_write": 0.234375 + "output": 0.75, + "cache_read": 0.09375 } }, - "qwen3-max-2026-01-23": { - "id": "qwen3-max-2026-01-23", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "mistral-small-3-2-24b-instruct": { + "id": "mistral-small-3-2-24b-instruct", + "name": "Mistral Small 3.2 24B Instruct", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-23", - "last_updated": "2026-01-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "structured_output": true, + "release_date": "2026-01-15", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -83356,42 +87346,44 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.09375, + "output": 0.25 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "inkling": { + "id": "inkling", + "name": "Inkling", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-07-16", + "last_updated": "2026-07-17", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ "text" @@ -83399,31 +87391,28 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 524288, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.25, + "output": 5.0625, + "cache_read": 0.2125 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "z-ai-glm-5-3": { + "id": "z-ai-glm-5-3", + "name": "GLM 5.3", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-08-18", + "last_updated": "2026-08-18", "modalities": { "input": [ "text" @@ -83434,31 +87423,43 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 24576 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.75, + "output": 5.5, + "cache_read": 0.325 } }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "google-gemma-4-26b-a4b-it": { + "id": "google-gemma-4-26b-a4b-it", + "name": "Google Gemma 4 26B A4B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-03", - "last_updated": "2026-02-03", + "release_date": "2026-04-02", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -83466,77 +87467,89 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 256000, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.13, + "output": 0.4, + "cache_read": 0.05 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "kimi-k2-6": { + "id": "kimi-k2-6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "knowledge": "2025-01", + "release_date": "2026-04-20", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202752, - "output": 16384 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.75, + "output": 3.5, + "cache_read": 0.16 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "openai-gpt-6-astra-pro": { + "id": "openai-gpt-6-astra-pro", + "name": "GPT-6 Astra Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "structured_output": true, + "release_date": "2026-09-05", + "last_updated": "2026-09-05", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -83544,41 +87557,59 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 12.5, + "output": 62.5, + "cache_read": 1.25, + "cache_write": 15.625, + "tiers": [ + { + "input": 25, + "output": 93.75, + "cache_read": 2.5, + "cache_write": 31.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 25, + "output": 93.75, + "cache_read": 2.5, + "cache_write": 31.25 + } } - } - } - }, - "abacus": { - "id": "abacus", - "env": [ - "ABACUS_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://routellm.abacus.ai/v1", - "name": "Abacus", - "doc": "https://abacus.ai/help/api", - "models": { - "o3": { - "id": "o3", - "name": "o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + }, + "openai-gpt-54": { + "id": "openai-gpt-54", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -83590,26 +87621,27 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "input": 922000, + "output": 131072 }, "cost": { - "input": 2, - "output": 8 + "input": 3.13, + "output": 18.8, + "cache_read": 0.313 } }, - "route-llm": { - "id": "route-llm", - "name": "Route LLM", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "gpt", + "gemma-4-uncensored": { + "id": "gemma-4-uncensored", + "name": "Gemma 4 Uncensored", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "structured_output": true, + "release_date": "2026-04-13", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -83619,27 +87651,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 256000, + "output": 8192 }, "cost": { - "input": 3, - "output": 15 + "input": 0.1625, + "output": 0.5 } }, - "grok-code-fast-1": { - "id": "grok-code-fast-1", - "name": "Grok Code Fast 1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-09-01", - "last_updated": "2025-09-01", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -83651,60 +87697,71 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.5 + "input": 6, + "output": 30, + "cache_read": 0.6, + "cache_write": 7.5 } }, - "gpt-5.3-codex-xhigh": { - "id": "gpt-5.3-codex-xhigh", - "name": "GPT-5.3 Codex XHigh", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", + "minimax-m3-preview": { + "id": "minimax-m3-preview", + "name": "MiniMax M3 Preview", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax-m3", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2026-06-12", + "last_updated": "2026-06-13", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 524288, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "llama-3.3-70b-versatile": { - "id": "llama-3.3-70b-versatile", - "name": "Llama 3.3 70B Versatile", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "openai-gpt-oss-120b": { + "id": "openai-gpt-oss-120b", + "name": "OpenAI GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2025-11-06", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -83716,33 +87773,42 @@ "open_weights": true, "limit": { "context": 128000, - "output": 32768 + "output": 16384 }, "cost": { - "input": 0.59, - "output": 0.79 + "input": 0.07, + "output": 0.3 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-29", + "last_updated": "2026-07-01", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -83750,27 +87816,40 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1.25, - "output": 10 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "deepseek-v4-1-flash": { + "id": "deepseek-v4-1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -83780,34 +87859,36 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.375, + "output": 1.5, + "cache_read": 0.0075 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "gemini-3-8-flash": { + "id": "gemini-3-8-flash", + "name": "Gemini 3.8 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -83815,34 +87896,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1, - "output": 5 + "input": 0.9375, + "output": 4.6875, + "cache_read": 0.09375 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "qwen-3-6-plus": { + "id": "qwen-3-6-plus", + "name": "Qwen 3.6 Plus Uncensored", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-05", + "knowledge": "2025-04", + "release_date": "2026-04-06", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", "image", - "audio", - "video", - "pdf" + "video" ], "output": [ "text" @@ -83850,27 +87931,58 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 0.625, + "output": 3.75, + "cache_read": 0.0625, + "cache_write": 0.78, + "tiers": [ + { + "input": 2.5, + "output": 7.5, + "cache_read": 0.0625, + "cache_write": 0.78, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 7.5, + "cache_read": 0.0625, + "cache_write": 0.78 + } } }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -83882,29 +87994,46 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 4.4 + "input": 6, + "output": 30, + "cache_read": 0.6, + "cache_write": 7.5 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "openai-gpt-6-sol": { + "id": "openai-gpt-6-sol", + "name": "GPT-6 Sol", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-sol", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -83912,63 +88041,107 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.2, - "output": 6 - } - }, - "o3-pro": { - "id": "o3-pro", - "name": "o3-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", + "input": 2.5, + "output": 12.5, + "cache_read": 0.25, + "cache_write": 3.125, + "tiers": [ + { + "input": 5, + "output": 18.75, + "cache_read": 0.5, + "cache_write": 6.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 18.75, + "cache_read": 0.5, + "cache_write": 6.25 + } + } + }, + "qwen-3-8-27b": { + "id": "qwen-3-8-27b", + "name": "Qwen 3.8 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "structured_output": true, + "temperature": true, + "release_date": "2026-08-17", + "last_updated": "2026-08-18", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 20, - "output": 40 + "input": 0.45, + "output": 3.2 } }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai-gpt-53-codex": { + "id": "openai-gpt-53-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-08-31", + "release_date": "2026-02-24", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -83976,29 +88149,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 15, - "output": 75 + "input": 2.19, + "output": 17.5, + "cache_read": 0.219 } }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "Grok 4.1 Fast (Non-Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": false, + "aion-labs-aion-3-0-mini": { + "id": "aion-labs-aion-3-0-mini", + "name": "Aion 3.0 Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2025-11-17", - "last_updated": "2025-11-17", + "structured_output": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -84006,32 +88180,42 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 16384 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 0.5 + "input": 0.875, + "output": 1.75, + "cache_read": 0.225 } }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "grok-4-3": { + "id": "grok-4-3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "release_date": "2026-04-18", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -84039,32 +88223,45 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 5, - "output": 25 + "input": 1.42, + "output": 2.83, + "cache_read": 0.23, + "tiers": [ + { + "input": 2.83, + "output": 5.67, + "cache_read": 0.45, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.83, + "output": 5.67, + "cache_read": 0.45 + } } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", - "attachment": true, + "aion-labs-aion-3-5-mini": { + "id": "aion-labs-aion-3-5-mini", + "name": "Aion 3.5 Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -84072,29 +88269,40 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.875, + "output": 1.75, + "cache_read": 0.225 } }, - "gpt-5.1-codex-max": { - "id": "gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", + "openai-gpt-56-terra-pro": { + "id": "openai-gpt-56-terra-pro", + "name": "GPT-5.6 Terra Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-terra", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -84106,31 +88314,67 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25 + } } }, - "claude-opus-4-20250514": { - "id": "claude-opus-4-20250514", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-05-14", - "last_updated": "2025-05-14", + "knowledge": "2025-01", + "release_date": "2025-12-19", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -84139,26 +88383,80 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 256000, + "output": 65536 }, "cost": { - "input": 15, - "output": 75 + "input": 0.7, + "output": 3.75, + "cache_read": 0.07 } }, - "gpt-5.3-chat-latest": { - "id": "gpt-5.3-chat-latest", - "name": "GPT-5.3 Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "mercury-2": { + "id": "mercury-2", + "name": "Mercury 2", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-02-20", + "last_updated": "2026-06-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 50000 + }, + "cost": { + "input": 0.3125, + "output": 0.9375, + "cache_read": 0.03125 + } + }, + "grok-4-6": { + "id": "grok-4-6", + "name": "Grok 4.6", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "knowledge": "2026-02-01", + "release_date": "2026-08-10", + "last_updated": "2026-08-12", "modalities": { "input": [ "text", @@ -84170,27 +88468,57 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 500000, + "output": 200000 }, "cost": { - "input": 1.75, - "output": 14 + "input": 2.27, + "output": 6.8, + "cache_read": 0.57, + "tiers": [ + { + "input": 4.53, + "output": 13.6, + "cache_read": 1.13, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4.53, + "output": 13.6, + "cache_read": 1.13 + } } }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "zai-org-glm-5-1": { + "id": "zai-org-glm-5-1", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -84199,58 +88527,76 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 200000, - "output": 100000 + "output": 80000 }, "cost": { - "input": 1.1, - "output": 4.4 + "input": 1.54, + "output": 4.84, + "cache_read": 0.286 } }, - "kimi-k2-turbo-preview": { - "id": "kimi-k2-turbo-preview", - "name": "Kimi K2 Turbo Preview", - "description": "Fast Kimi model for responsive chat, coding help, and agent loops", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "xiaomi-mimo-v2-5": { + "id": "xiaomi-mimo-v2-5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-07-08", - "last_updated": "2025-07-08", + "knowledge": "2024-12", + "release_date": "2026-06-11", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 8192 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 8 + "input": 0.4, + "output": 2, + "cache_read": 0.08 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", + "openai-gpt-4o-2024-11-20": { + "id": "openai-gpt-4o-2024-11-20", + "name": "GPT-4o", "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2026-02-28", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -84262,33 +88608,42 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14 + "input": 3.125, + "output": 12.5 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "openai-gpt-52": { + "id": "openai-gpt-52", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", "family": "gpt", - "attachment": true, + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2025-12-13", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -84296,31 +88651,36 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 256000, "input": 272000, - "output": 128000 + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14 + "input": 2.19, + "output": 17.5, + "cache_read": 0.219 } }, - "grok-4-0709": { - "id": "grok-4-0709", - "name": "Grok 4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "gemini-3-6-flash": { + "id": "gemini-3-6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "knowledge": "2026-03", + "release_date": "2026-07-09", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -84328,31 +88688,44 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 3, - "output": 15 + "input": 0.9375, + "output": 4.6875, + "cache_read": 0.09375 } }, - "claude-sonnet-4-20250514": { - "id": "claude-sonnet-4-20250514", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "gemini-3-1-pro-preview": { + "id": "gemini-3-1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-05-14", - "last_updated": "2025-05-14", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -84360,33 +88733,64 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 3, - "output": 15 + "input": 2.5, + "output": 15, + "cache_read": 0.5, + "cache_write": 0.5, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 0.5 + } } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -84394,27 +88798,39 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 1000000, "output": 32768 }, "cost": { - "input": 0.6, - "output": 3 + "input": 1.65, + "output": 3.301, + "cache_read": 0.33 } }, - "gpt-5.1-chat-latest": { - "id": "gpt-5.1-chat-latest", - "name": "GPT-5.1 Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -84426,27 +88842,41 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1.25, - "output": 10 + "input": 3.6, + "output": 18, + "cache_read": 0.36, + "cache_write": 4.5 } }, - "gpt-5.2-chat-latest": { - "id": "gpt-5.2-chat-latest", - "name": "GPT-5.2 Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "claude-opus-5-fast": { + "id": "claude-opus-5-fast", + "name": "Claude Opus 5 Fast", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2026-01-01", - "last_updated": "2026-01-01", + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-23", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", @@ -84458,26 +88888,41 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 1.75, - "output": 14 + "input": 12, + "output": 60, + "cache_read": 1.2, + "cache_write": 15 } }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai-gpt-55": { + "id": "openai-gpt-55", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -84489,31 +88934,62 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 1000000, + "input": 922000, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 6.25, + "output": 37.5, + "cache_read": 0.625, + "tiers": [ + { + "input": 12.5, + "output": 56.25, + "cache_read": 1.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 12.5, + "output": 56.25, + "cache_read": 1.25 + } } }, - "gpt-4o-2024-11-20": { - "id": "gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", + "openai-gpt-56-sol": { + "id": "openai-gpt-56-sol", + "name": "GPT-5.6 Sol", "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "family": "gpt-sol", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" @@ -84521,133 +88997,181 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "claude-3-7-sonnet-20250219": { - "id": "claude-3-7-sonnet-20250219", - "name": "Claude Sonnet 3.7", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen3-coder-480b-a35b-instruct-turbo": { + "id": "qwen3-coder-480b-a35b-instruct-turbo", + "name": "Qwen 3 Coder 480B Turbo", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "structured_output": true, + "release_date": "2026-01-27", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 256000, + "output": 65536 }, "cost": { - "input": 3, - "output": 15 + "input": 0.35, + "output": 1.5, + "cache_read": 0.04 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "z-ai-glm-5-3-flash": { + "id": "z-ai-glm-5-3-flash", + "name": "GLM 5.3 Flash", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 15 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "zai-org-glm-5": { + "id": "zai-org-glm-5", + "name": "GLM 5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-02-11", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 198000, + "output": 32000 }, "cost": { - "input": 2, - "output": 8 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "grok-4-7": { + "id": "grok-4-7", + "name": "Grok 4.7", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "knowledge": "2026-05", + "release_date": "2026-09-16", + "last_updated": "2026-09-21", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -84655,58 +89179,55 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 500000, + "output": 200000 }, "cost": { - "input": 2, - "output": 12 - } - }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" + "input": 2.27, + "output": 6.8, + "cache_read": 0.57, + "tiers": [ + { + "input": 4.53, + "output": 13.6, + "cache_read": 1.13, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25 + "context_over_200k": { + "input": 4.53, + "output": 13.6, + "cache_read": 1.13 + } } }, - "qwen-2.5-coder-32b": { - "id": "qwen-2.5-coder-32b", - "name": "Qwen 2.5 Coder 32B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-11-11", - "last_updated": "2024-11-11", + "knowledge": "2024-07", + "release_date": "2025-12-04", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -84717,32 +89238,35 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 160000, + "output": 32768 }, "cost": { - "input": 0.79, - "output": 0.79 + "input": 0.33, + "output": 0.48, + "cache_read": 0.16 } }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "gemini-3-5-flash-lite": { + "id": "gemini-3-5-flash-lite", + "name": "Gemini 3.5 Flash-Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2026-03", + "release_date": "2026-07-09", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -84750,31 +89274,44 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 3, - "output": 15 + "input": 0.375, + "output": 3.125, + "cache_read": 0.0375 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "gemini-3-5-flash": { + "id": "gemini-3-5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-22", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -84782,58 +89319,85 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 2 + "input": 1.55, + "output": 9.45, + "cache_read": 0.155, + "cache_write": 0.086 } }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "google-gemma-4-31b-it": { + "id": "google-gemma-4-31b-it", + "name": "Google Gemma 4 31B Instruct", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-04-03", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 256000, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 1.6 + "input": 0.12, + "output": 0.36, + "cache_read": 0.09 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -84845,58 +89409,27 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.4 - } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15 + "input": 6, + "output": 30, + "cache_read": 0.6, + "cache_write": 7.5 } }, - "grok-4-fast-non-reasoning": { - "id": "grok-4-fast-non-reasoning", - "name": "Grok 4 Fast (Non-Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "venice-uncensored-1-2": { + "id": "venice-uncensored-1-2", + "name": "Venice Uncensored 1.2", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "venice", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "structured_output": true, + "release_date": "2026-04-01", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -84906,96 +89439,114 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 16384 + "context": 128000, + "output": 8192 }, "cost": { "input": 0.2, - "output": 0.5 + "output": 0.9 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "qwen3-5-397b-a17b": { + "id": "qwen3-5-397b-a17b", + "name": "Qwen 3.5 397B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-02-16", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", "image", - "audio", - "video", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.75, + "output": 4.5 } }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "llama-3.2-3b": { + "id": "llama-3.2-3b", + "name": "Llama 3.2 3B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2023-12", + "release_date": "2024-10-03", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 16384 + "output": 4096 }, "cost": { "input": 0.15, "output": 0.6 } }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "GPT-5 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", - "attachment": false, + "openai-gpt-56-luna-pro": { + "id": "openai-gpt-56-luna-pro", + "name": "GPT-5.6 Luna Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-luna", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -85007,161 +89558,217 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 0.3125, + "tiers": [ + { + "input": 0.5, + "output": 2.25, + "cache_read": 0.05, + "cache_write": 0.625, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.5, + "output": 2.25, + "cache_read": 0.05, + "cache_write": 0.625 + } } }, - "gemini-3.1-flash-lite-preview": { - "id": "gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash", - "attachment": true, + "nvidia-nemotron-3-ultra-550b-a55b": { + "id": "nvidia-nemotron-3-ultra-550b-a55b", + "name": "NVIDIA Nemotron 3 Ultra", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "release_date": "2026-06-04", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 1 + "input": 0.625, + "output": 3.125, + "cache_read": 0.1875 } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", - "attachment": true, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash 0423", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.138, + "output": 0.275, + "cache_read": 0.028 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "nvidia-nemotron-3-nano-30b-a3b": { + "id": "nvidia-nemotron-3-nano-30b-a3b", + "name": "NVIDIA Nemotron 3 Nano 30B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": true, + "temperature": true, + "release_date": "2026-01-27", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.075, + "output": 0.3 } - }, - "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { - "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", - "name": "Llama 4 Maverick 17B 128E Instruct FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + } + } + }, + "alibaba-token-plan-cn": { + "id": "alibaba-token-plan-cn", + "env": [ + "ALIBABA_TOKEN_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://token-plan.cn-beijing.maas.aliyuncs.com/compatible-mode/v1", + "name": "Alibaba Token Plan (China)", + "doc": "https://www.alibabacloud.com/help/zh/model-studio/token-plan-overview", + "models": { + "happyhorse-1.1-r2v": { + "id": "happyhorse-1.1-r2v", + "name": "HappyHorse 1.1 Reference-to-Video", + "description": "Video model for reference-guided video generation", "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-07-17", + "last_updated": "2026-07-17", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 0, + "output": 0 }, "cost": { - "input": 0.14, - "output": 0.59 + "input": 0, + "output": 0 } }, - "meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo": { - "id": "meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo", - "name": "Llama 3.1 405B Instruct Turbo", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -85170,27 +89777,46 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 3.5, - "output": 3.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "meta-llama/Meta-Llama-3.1-8B-Instruct": { - "id": "meta-llama/Meta-Llama-3.1-8B-Instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -85201,116 +89827,149 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.02, - "output": 0.05 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "Qwen/qwen3-coder-480b-a35b-instruct": { - "id": "Qwen/qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.29, - "output": 1.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "Qwen/QwQ-32B": { - "id": "Qwen/QwQ-32B", - "name": "QwQ 32B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "happyhorse-1.1-t2v": { + "id": "happyhorse-1.1-t2v", + "name": "HappyHorse 1.1 Text-to-Video", + "description": "Video model for prompt-driven text-to-video generation", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2024-11-28", - "last_updated": "2024-11-28", + "release_date": "2026-07-17", + "last_updated": "2026-07-17", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 0, + "output": 0 }, "cost": { - "input": 0.4, - "output": 0.4 + "input": 0, + "output": 0 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen-image-2.0": { + "id": "qwen-image-2.0", + "name": "Qwen Image 2.0", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 8192 + "context": 8192, + "output": 0 }, "cost": { - "input": 0.13, - "output": 0.6 + "input": 0, + "output": 0 } }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -85319,30 +89978,51 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 202752, + "output": 16384 }, "cost": { - "input": 0.09, - "output": 0.29 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "Qwen/Qwen2.5-72B-Instruct": { - "id": "Qwen/Qwen2.5-72B-Instruct", - "name": "Qwen 2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -85350,19 +90030,21 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.11, - "output": 0.38 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT-OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "qwen3.8-max-preview": { + "id": "qwen3.8-max-preview", + "name": "Qwen3.8 Max Preview", + "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -85371,77 +90053,101 @@ "values": [ "low", "medium", - "high" + "xhigh" ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-07-19", + "last_updated": "2026-07-19", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 1000000, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.08, - "output": 0.44 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "zai-org/glm-4.7": { - "id": "zai-org/glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "qwen-image-2.0-pro": { + "id": "qwen-image-2.0-pro", + "name": "Qwen Image 2.0 Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2025-06-01", - "last_updated": "2025-06-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 8192, + "output": 0 }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 0, + "output": 0 } }, - "zai-org/glm-4.5": { - "id": "zai-org/glm-4.5", - "name": "GLM-4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -85449,88 +90155,98 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "zai-org/glm-4.6": { - "id": "zai-org/glm-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "wan2.7-image": { + "id": "wan2.7-image", + "name": "Wan2.7 Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2025-03-01", - "last_updated": "2025-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 8192, + "output": 0 }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 0, + "output": 0 } }, - "zai-org/glm-5": { - "id": "zai-org/glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "wan2.7-image-pro": { + "id": "wan2.7-image-pro", + "name": "Wan2.7 Image Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 8192, + "output": 0 }, "cost": { - "input": 1, - "output": 3.2 + "input": 0, + "output": 0 } }, - "deepseek-ai/DeepSeek-R1": { - "id": "deepseek-ai/DeepSeek-R1", - "name": "DeepSeek R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -85538,86 +90254,118 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 262144, + "output": 98304 }, "cost": { - "input": 3, - "output": 7 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "deepseek-ai/DeepSeek-V3.1-Terminus": { - "id": "deepseek-ai/DeepSeek-V3.1-Terminus", - "name": "DeepSeek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "happyhorse-1.1-i2v": { + "id": "happyhorse-1.1-i2v", + "name": "HappyHorse 1.1 Image-to-Video", + "description": "Video model for image-to-video generation", + "attachment": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-06-01", - "last_updated": "2025-06-01", + "release_date": "2026-07-17", + "last_updated": "2026-07-17", "modalities": { "input": [ + "image", "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 0, + "output": 0 }, "cost": { - "input": 0.27, - "output": 1 + "input": 0, + "output": 0 } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 131072 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-06-15", - "last_updated": "2025-06-15", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.27, - "output": 0.4 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "deepseek/deepseek-v3.1": { - "id": "deepseek/deepseek-v3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -85628,38 +90376,31 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.55, - "output": 1.66 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - } - } - }, - "cloudferro-sherlock": { - "id": "cloudferro-sherlock", - "env": [ - "CLOUDFERRO_SHERLOCK_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api-sherlock.cloudferro.com/openai/v1/", - "name": "CloudFerro Sherlock", - "doc": "https://docs.sherlock.cloudferro.com/", - "models": { - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + }, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-10-09", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -85670,96 +90411,94 @@ }, "open_weights": true, "limit": { - "context": 70000, - "output": 70000 + "context": 196608, + "input": 196601, + "output": 32768 }, "cost": { - "input": 2.92, - "output": 2.92 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "OpenAI GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "xhigh" ] + }, + { + "type": "budget_tokens", + "max": 262144 } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-28", - "last_updated": "2025-08-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131000, - "output": 131000 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 2.92, - "output": 2.92 - } - }, - "speakleash/Bielik-11B-v3.0-Instruct": { - "id": "speakleash/Bielik-11B-v3.0-Instruct", - "name": "Bielik 11B v3.0 Instruct", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 32000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.67, - "output": 0.67 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "speakleash/Bielik-11B-v2.6-Instruct": { - "id": "speakleash/Bielik-11B-v2.6-Instruct", - "name": "Bielik 11B v2.6 Instruct", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -85770,28 +90509,36 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 32000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.67, - "output": 0.67 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -85802,32 +90549,21 @@ }, "open_weights": true, "limit": { - "context": 196000, - "input": 180000, - "output": 16000 + "context": 202752, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - } - } - }, - "ollama-cloud": { - "id": "ollama-cloud", - "env": [ - "OLLAMA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://ollama.com/v1", - "name": "Ollama Cloud", - "doc": "https://docs.ollama.com/cloud", - "models": { - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "deepseek-v4-flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", + }, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -85843,6 +90579,12 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", "release_date": "2026-04-24", "last_updated": "2026-04-24", "modalities": { @@ -85855,95 +90597,68 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 - } - }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "minimax-m2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "knowledge": "2025-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 384000 }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "devstral-small-2:24b": { - "id": "devstral-small-2:24b", - "name": "devstral-small-2:24b", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": true, - "release_date": "2025-12-09", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - } - }, - "glm-4.7": { - "id": "glm-4.7", - "name": "glm-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 131072 } ], "tool_call": true, - "release_date": "2025-12-22", - "last_updated": "2026-01-19", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "cogito-2.1:671b": { - "id": "cogito-2.1:671b", - "name": "cogito-2.1:671b", - "description": "Legacy model retained for compatibility with older integrations", - "family": "cogito", + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -85952,8 +90667,11 @@ } ], "tool_call": true, - "release_date": "2025-11-19", - "last_updated": "2026-01-19", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-03", + "last_updated": "2025-12-05", "modalities": { "input": [ "text" @@ -85964,41 +90682,19 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 32000 - }, - "status": "deprecated" - }, - "minimax-m2.1": { - "id": "minimax-m2.1", - "name": "minimax-m2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "release_date": "2025-12-23", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 131072, + "output": 65536 }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 + "cost": { + "input": 0, + "output": 0 } }, - "gpt-oss:120b": { - "id": "gpt-oss:120b", - "name": "gpt-oss:120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -86006,44 +90702,19 @@ "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "release_date": "2025-08-05", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "interleaved": { + "field": "reasoning_content" }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - } - }, - "nemotron-3-nano:30b": { - "id": "nemotron-3-nano:30b", - "name": "nemotron-3-nano:30b", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2026-01-19", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -86054,62 +90725,19 @@ }, "open_weights": true, "limit": { - "context": 1048576, + "context": 1000000, "output": 131072 - } - }, - "ministral-3:8b": { - "id": "ministral-3:8b", - "name": "ministral-3:8b", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "release_date": "2024-12-01", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 128000 - } - }, - "rnj-1:8b": { - "id": "rnj-1:8b", - "name": "rnj-1:8b", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "rnj", - "attachment": false, - "reasoning": false, - "tool_call": true, - "release_date": "2025-12-06", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 4096 + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, "kimi-k2.7-code": { "id": "kimi-k2.7-code", - "name": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", "family": "kimi-k2", "attachment": true, @@ -86117,9 +90745,15 @@ "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, "knowledge": "2025-01", @@ -86128,7 +90762,8 @@ "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -86138,45 +90773,66 @@ "limit": { "context": 262144, "output": 262144 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "glm-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "release_date": "2026-03-27", - "last_updated": "2026-04-07", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "deepseek-v4-pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -86192,6 +90848,12 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", "release_date": "2026-04-24", "last_updated": "2026-04-24", "modalities": { @@ -86204,25 +90866,41 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "glm-4.6": { - "id": "glm-4.6", - "name": "glm-4.6", - "description": "Legacy model retained for compatibility with older integrations", - "family": "glm", + } + } + }, + "ai21": { + "id": "ai21", + "env": [ + "AI21_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.ai21.com/studio/v1", + "name": "AI21 Labs", + "doc": "https://docs.ai21.com/docs/jamba-foundation-models", + "models": { + "jamba-large": { + "id": "jamba-large", + "name": "Jamba Large", + "description": "AI21's hybrid SSM-Transformer long-context model for enterprise agents and grounded generation", + "family": "jamba", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "release_date": "2025-09-29", - "last_updated": "2026-01-19", + "structured_output": true, + "temperature": true, + "knowledge": "2024-08-22", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ "text" @@ -86233,23 +90911,27 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 256000, + "output": 4096 }, - "status": "deprecated" + "cost": { + "input": 2, + "output": 8 + } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "kimi-k2-thinking", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-thinking", + "jamba-mini": { + "id": "jamba-mini", + "name": "Jamba Mini", + "description": "AI21's efficient, lightweight hybrid SSM-Transformer model for a wide range of tasks", + "family": "jamba", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2026-01-19", + "structured_output": true, + "temperature": true, + "knowledge": "2024-08-22", + "release_date": "2026-01-01", + "last_updated": "2026-01-01", "modalities": { "input": [ "text" @@ -86260,27 +90942,38 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 4096 }, - "status": "deprecated" - }, - "nemotron-3-super": { - "id": "nemotron-3-super", - "name": "nemotron-3-super", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "cost": { + "input": 0.2, + "output": 0.4 + } + } + } + }, + "inference": { + "id": "inference", + "env": [ + "INFERENCE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.net/v1", + "name": "Inference", + "doc": "https://inference.net/models", + "models": { + "meta/llama-3.1-8b-instruct": { + "id": "meta/llama-3.1-8b-instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-12", + "knowledge": "2023-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -86291,20 +90984,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 16000, + "output": 4096 + }, + "cost": { + "input": 0.025, + "output": 0.025 } }, - "ministral-3:14b": { - "id": "ministral-3:14b", - "name": "ministral-3:14b", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "meta/llama-3.2-11b-vision-instruct": { + "id": "meta/llama-3.2-11b-vision-instruct", + "name": "Llama 3.2 11B Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", "attachment": true, "reasoning": false, "tool_call": true, - "release_date": "2024-12-01", - "last_updated": "2026-01-19", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", @@ -86316,34 +91015,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 128000 + "context": 16000, + "output": 4096 + }, + "cost": { + "input": 0.055, + "output": 0.055 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "meta/llama-3.2-1b-instruct": { + "id": "meta/llama-3.2-1b-instruct", + "name": "Llama 3.2 1B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2023-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -86354,41 +91045,29 @@ }, "open_weights": true, "limit": { - "context": 976000, - "output": 131072 + "context": 16000, + "output": 4096 + }, + "cost": { + "input": 0.01, + "output": 0.01 } }, - "minimax-m3": { - "id": "minimax-m3", - "name": "minimax-m3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax-m3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "meta/llama-3.2-3b-instruct": { + "id": "meta/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-31", - "last_updated": "2026-05-31", + "knowledge": "2023-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -86396,24 +91075,30 @@ }, "open_weights": true, "limit": { - "context": 512000, - "output": 131072 + "context": 16000, + "output": 4096 + }, + "cost": { + "input": 0.02, + "output": 0.02 } }, - "minimax-m2": { - "id": "minimax-m2", - "name": "minimax-m2", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "google/gemma-3": { + "id": "google/gemma-3", + "name": "Google Gemma 3", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, "tool_call": true, - "release_date": "2025-10-23", - "last_updated": "2026-01-19", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -86421,22 +91106,26 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 128000 + "context": 125000, + "output": 4096 }, - "status": "deprecated" + "cost": { + "input": 0.15, + "output": 0.3 + } }, - "qwen3-next:80b": { - "id": "qwen3-next:80b", - "name": "qwen3-next:80b", - "description": "Legacy model retained for compatibility with older integrations", - "family": "qwen", + "osmosis/osmosis-structure-0.6b": { + "id": "osmosis/osmosis-structure-0.6b", + "name": "Osmosis Structure 0.6B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "osmosis", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "release_date": "2025-09-15", - "last_updated": "2026-01-19", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -86447,21 +91136,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 4000, + "output": 2048 }, - "status": "deprecated" + "cost": { + "input": 0.1, + "output": 0.5 + } }, - "qwen3-coder:480b": { - "id": "qwen3-coder:480b", - "name": "qwen3-coder:480b", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "mistral/mistral-nemo-12b-instruct": { + "id": "mistral/mistral-nemo-12b-instruct", + "name": "Mistral Nemo 12B Instruct", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, "reasoning": false, "tool_call": true, - "release_date": "2025-07-22", - "last_updated": "2026-01-19", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -86472,21 +91166,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 16000, + "output": 4096 + }, + "cost": { + "input": 0.038, + "output": 0.1 } }, - "kimi-k2:1t": { - "id": "kimi-k2:1t", - "name": "kimi-k2:1t", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-k2", + "qwen/qwen3-embedding-4b": { + "id": "qwen/qwen3-embedding-4b", + "name": "Qwen 3 Embedding 4B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, - "knowledge": "2024-10", - "release_date": "2025-07-11", - "last_updated": "2026-01-19", + "tool_call": false, + "temperature": false, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -86497,29 +91196,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 32000, + "output": 2048 }, - "status": "deprecated" + "cost": { + "input": 0.01, + "output": 0 + } }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "minimax-m2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen/qwen-2.5-7b-vision-instruct": { + "id": "qwen/qwen-2.5-7b-vision-instruct", + "name": "Qwen 2.5 7B Vision Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, "tool_call": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -86527,54 +91227,38 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 196608 + "context": 125000, + "output": 4096 + }, + "cost": { + "input": 0.2, + "output": 0.2 } - }, - "deepseek-v3.1:671b": { - "id": "deepseek-v3.1:671b", - "name": "deepseek-v3.1:671b", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "release_date": "2025-08-21", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 163840 - } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "kimi-k2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + } + } + }, + "iflowcn": { + "id": "iflowcn", + "env": [ + "IFLOW_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://apis.iflow.cn/v1", + "name": "iFlow", + "doc": "https://platform.iflow.cn/en/docs", + "models": { + "qwen3-vl-plus": { + "id": "qwen3-vl-plus", + "name": "Qwen3-VL-Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", @@ -86584,57 +91268,62 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "gemma4:31b": { - "id": "gemma4:31b", - "name": "gemma4:31b", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "glm-4.6": { + "id": "glm-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-08", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "ministral-3:3b": { - "id": "ministral-3:3b", - "name": "ministral-3:3b", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3-32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "release_date": "2024-10-22", - "last_updated": "2026-01-19", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -86642,49 +91331,59 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 128000 + "context": 128000, + "output": 32000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "gemma3:12b": { - "id": "gemma3:12b", - "name": "gemma3:12b", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3-Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, - "release_date": "2024-12-01", - "last_updated": "2026-01-19", + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "gemma3:4b": { - "id": "gemma3:4b", - "name": "gemma3:4b", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "deepseek-v3": { + "id": "deepseek-v3", + "name": "DeepSeek-V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, - "tool_call": false, - "release_date": "2024-12-01", - "last_updated": "2026-01-19", + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-26", + "last_updated": "2024-12-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -86692,32 +91391,30 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 32000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "qwen3.5:397b": { - "id": "qwen3.5:397b", - "name": "qwen3.5:397b", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen3-235b-a22b-thinking-2507": { + "id": "qwen3-235b-a22b-thinking-2507", + "name": "Qwen3-235B-A22B-Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "release_date": "2026-02-15", - "last_updated": "2026-02-17", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -86725,50 +91422,56 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "kimi-k2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen3-max-preview": { + "id": "qwen3-max-preview", + "name": "Qwen3-Max-Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "qwen3-coder-next", + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3-Coder-Plus", "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "release_date": "2026-02-02", - "last_updated": "2026-02-08", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ "text" @@ -86779,24 +91482,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "qwen3-vl:235b-instruct": { - "id": "qwen3-vl:235b-instruct", - "name": "qwen3-vl:235b-instruct", - "description": "Legacy model retained for compatibility with older integrations", + "qwen3-235b-a22b-instruct": { + "id": "qwen3-235b-a22b-instruct", + "name": "Qwen3-235B-A22B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": false, "tool_call": true, - "release_date": "2025-09-22", - "last_updated": "2026-01-19", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -86804,76 +91512,90 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 256000, + "output": 64000 }, - "status": "deprecated" + "cost": { + "input": 0, + "output": 0 + } }, - "mistral-large-3:675b": { - "id": "mistral-large-3:675b", - "name": "mistral-large-3:675b", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, + "kimi-k2-0905": { + "id": "kimi-k2-0905", + "name": "Kimi-K2-0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, "tool_call": true, - "release_date": "2025-12-02", - "last_updated": "2026-01-19", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "gemma3:27b": { - "id": "gemma3:27b", - "name": "gemma3:27b", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "kimi-k2": { + "id": "kimi-k2", + "name": "Kimi-K2", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, - "tool_call": false, - "release_date": "2025-07-27", - "last_updated": "2026-01-19", + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 64000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "qwen3-vl:235b": { - "id": "qwen3-vl:235b", - "name": "qwen3-vl:235b", - "description": "Legacy model retained for compatibility with older integrations", + "qwen3-235b": { + "id": "qwen3-235b", + "name": "Qwen3-235B-A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "release_date": "2025-09-22", - "last_updated": "2026-01-19", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -86881,27 +91603,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "output": 32000 }, - "status": "deprecated" + "cost": { + "input": 0, + "output": 0 + } }, - "nemotron-3-ultra": { - "id": "nemotron-3-ultra", - "name": "nemotron-3-ultra", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2024-12", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -86912,30 +91634,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 128000 + "context": 128000, + "output": 32000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "gemini-3-flash-preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek-V3.2-Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2026-04-08", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -86943,33 +91664,47 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 64000 + }, + "cost": { + "input": 0, + "output": 0 } - }, - "gpt-oss:20b": { - "id": "gpt-oss:20b", - "name": "gpt-oss:20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + } + } + }, + "minimax-cn-coding-plan": { + "id": "minimax-cn-coding-plan", + "env": [ + "MINIMAX_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://api.minimax.cn/anthropic/v1", + "name": "MiniMax Token Plan (minimax.cn)", + "doc": "https://platform.minimaxi.com/docs/token-plan/intro", + "models": { + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "release_date": "2025-08-05", - "last_updated": "2026-01-19", + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-25", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -86977,28 +91712,28 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 512000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "glm-5": { - "id": "glm-5", - "name": "glm-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -87009,25 +91744,28 @@ }, "open_weights": true, "limit": { - "context": 202752, + "context": 204800, "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "deepseek-v3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "release_date": "2025-06-15", - "last_updated": "2026-01-19", + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -87038,20 +91776,28 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 65536 + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "devstral-2:123b": { - "id": "devstral-2:123b", - "name": "devstral-2:123b", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "release_date": "2025-12-09", - "last_updated": "2026-01-19", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -87062,34 +91808,28 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - } - } - }, - "cloudflare-ai-gateway": { - "id": "cloudflare-ai-gateway", - "env": [ - "CLOUDFLARE_API_TOKEN", - "CLOUDFLARE_ACCOUNT_ID", - "CLOUDFLARE_GATEWAY_ID" - ], - "npm": "ai-gateway-provider", - "name": "Cloudflare AI Gateway", - "doc": "https://developers.cloudflare.com/ai-gateway/", - "models": { - "workers-ai/@cf/baai/bge-m3": { - "id": "workers-ai/@cf/baai/bge-m3", - "name": "BGE M3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "bge", + }, + "MiniMax-M2.7-highspeed": { + "id": "MiniMax-M2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -87098,27 +91838,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.012, - "output": 0 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "workers-ai/@cf/baai/bge-small-en-v1.5": { - "id": "workers-ai/@cf/baai/bge-small-en-v1.5", - "name": "BGE Small EN v1.5", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "bge", + "MiniMax-M2": { + "id": "MiniMax-M2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ "text" @@ -87127,27 +91870,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.02, + "input": 0, "output": 0 } }, - "workers-ai/@cf/baai/bge-reranker-base": { - "id": "workers-ai/@cf/baai/bge-reranker-base", - "name": "BGE Reranker Base", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "bge", + "MiniMax-M2.5-highspeed": { + "id": "MiniMax-M2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-04-09", - "last_updated": "2025-04-09", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -87156,88 +91900,162 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.0031, - "output": 0 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "workers-ai/@cf/baai/bge-base-en-v1.5": { - "id": "workers-ai/@cf/baai/bge-base-en-v1.5", - "name": "BGE Base EN v1.5", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "bge", - "attachment": false, - "reasoning": false, - "tool_call": false, + } + } + }, + "thinkingmachines": { + "id": "thinkingmachines", + "env": [ + "TINKER_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://tinker.thinkingmachines.dev/services/tinker-prod/anthropic/api/v1", + "name": "Thinking Machines", + "doc": "https://tinker-docs.thinkingmachines.ai/tinker/compatible-apis/anthropic/", + "models": { + "thinkingmachines/Inkling:peft:262144": { + "id": "thinkingmachines/Inkling:peft:262144", + "name": "Inkling (256K)", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.067, - "output": 0 + "input": 3.74, + "output": 9.36, + "cache_read": 0.748 } }, - "workers-ai/@cf/baai/bge-large-en-v1.5": { - "id": "workers-ai/@cf/baai/bge-large-en-v1.5", - "name": "BGE Large EN v1.5", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "bge", - "attachment": false, - "reasoning": false, - "tool_call": false, + "thinkingmachines/Inkling": { + "id": "thinkingmachines/Inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 0 + "input": 1.87, + "output": 4.68, + "cache_read": 0.374 } - }, - "workers-ai/@cf/ai4bharat/indictrans2-en-indic-1B": { - "id": "workers-ai/@cf/ai4bharat/indictrans2-en-indic-1B", - "name": "IndicTrans2 EN-Indic 1B", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "indictrans", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + } + } + }, + "stepfun-step-plan": { + "id": "stepfun-step-plan", + "env": [ + "STEPFUN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.stepfun.com/step_plan/v1", + "name": "StepFun Step Plan (China)", + "doc": "https://platform.stepfun.com/docs/zh/step-plan/integrations/reasoning-api", + "models": { + "step-5-preview": { + "id": "step-5-preview", + "name": "Step 5 Preview", + "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "release_date": "2026-09-16", + "last_updated": "2026-09-20", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -87245,25 +92063,22 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.34, - "output": 0.34 + "context": 1000000, + "input": 1000000, + "output": 1000000 } }, - "workers-ai/@cf/ibm-granite/granite-4.0-h-micro": { - "id": "workers-ai/@cf/ibm-granite/granite-4.0-h-micro", - "name": "IBM Granite 4.0 H Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "granite", + "step-router-v1": { + "id": "step-router-v1", + "name": "Step Router v1", + "description": "StepFun routing model that dispatches requests to the appropriate Step model.", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text" @@ -87274,25 +92089,34 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.017, - "output": 0.11 + "context": 256000, + "input": 256000, + "output": 256000 } }, - "workers-ai/@cf/huggingface/distilbert-sst-2-int8": { - "id": "workers-ai/@cf/huggingface/distilbert-sst-2-int8", - "name": "DistilBERT SST-2 INT8", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "distilbert", + "step-3.5-flash-2603": { + "id": "step-3.5-flash-2603", + "name": "Step 3.5 Flash 2603", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "knowledge": "2025-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" @@ -87301,32 +92125,24 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.026, - "output": 0 + "context": 256000, + "input": 256000, + "output": 256000 } }, - "workers-ai/@cf/moonshotai/kimi-k2.5": { - "id": "workers-ai/@cf/moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, + "step-3.5-flash": { + "id": "step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "medium", "high" ] } @@ -87335,15 +92151,13 @@ "interleaved": { "field": "reasoning_content" }, - "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -87352,25 +92166,17 @@ "open_weights": true, "limit": { "context": 256000, + "input": 256000, "output": 256000 - }, - "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 } }, - "workers-ai/@cf/moonshotai/kimi-k2.6": { - "id": "workers-ai/@cf/moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "step-3.7-flash": { + "id": "step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -87384,15 +92190,15 @@ "interleaved": { "field": "reasoning_content" }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -87401,54 +92207,92 @@ "open_weights": true, "limit": { "context": 256000, + "input": 256000, "output": 256000 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 } - }, - "workers-ai/@cf/mistral/mistral-7b-instruct-v0.1": { - "id": "workers-ai/@cf/mistral/mistral-7b-instruct-v0.1", - "name": "Mistral 7B Instruct v0.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": false, - "reasoning": false, - "tool_call": false, + } + } + }, + "melious": { + "id": "melious", + "env": [ + "MELIOUS_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.melious.ai/v1", + "name": "Melious", + "doc": "https://melious.ai/docs/reference/models", + "models": { + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.11, - "output": 0.19 + "input": 0.11592, + "output": 0.46368, + "cache_read": 0.023184 } }, - "workers-ai/@cf/google/gemma-3-12b-it": { - "id": "workers-ai/@cf/google/gemma-3-12b-it", - "name": "Gemma 3 12B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-04-11", - "last_updated": "2025-04-11", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -87457,83 +92301,77 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.35, - "output": 0.56 + "input": 0.11592, + "output": 0.2898, + "cache_read": 0.023184 } }, - "workers-ai/@cf/myshell-ai/melotts": { - "id": "workers-ai/@cf/myshell-ai/melotts", - "name": "MyShell MeloTTS", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "melotts", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 3.1878, + "output": 15.939, + "cache_read": 0.788256 } }, - "workers-ai/@cf/openai/gpt-oss-120b": { - "id": "workers-ai/@cf/openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.35, - "output": 0.75 - } - }, - "workers-ai/@cf/openai/gpt-oss-20b": { - "id": "workers-ai/@cf/openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -87542,61 +92380,74 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 203000, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.3 + "input": 1.10124, + "output": 3.36168, + "cache_read": 0.266616 } }, - "workers-ai/@cf/mistralai/mistral-small-3.1-24b-instruct": { - "id": "workers-ai/@cf/mistralai/mistral-small-3.1-24b-instruct", - "name": "Mistral Small 3.1 24B Instruct", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, - "reasoning": false, - "tool_call": false, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-04-11", - "last_updated": "2025-04-11", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.35, - "output": 0.56 + "input": 0.23184, + "output": 1.1592, + "cache_read": 0.011592 } }, - "workers-ai/@cf/nvidia/nemotron-3-120b-a12b": { - "id": "workers-ai/@cf/nvidia/nemotron-3-120b-a12b", - "name": "Nemotron 3 Super 120B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } @@ -87605,12 +92456,16 @@ "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -87622,79 +92477,72 @@ "output": 256000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 0.81144, + "output": 4.0572, + "cache_read": 0.266616 } }, - "workers-ai/@cf/pfnet/plamo-embedding-1b": { - "id": "workers-ai/@cf/pfnet/plamo-embedding-1b", - "name": "PLaMo Embedding 1B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "plamo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.019, - "output": 0 - } - }, - "workers-ai/@cf/deepgram/aura-2-es": { - "id": "workers-ai/@cf/deepgram/aura-2-es", - "name": "Deepgram Aura 2 (ES)", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "aura", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.5796, + "output": 2.95596, + "cache_read": 0.139104 } }, - "workers-ai/@cf/deepgram/aura-2-en": { - "id": "workers-ai/@cf/deepgram/aura-2-en", - "name": "Deepgram Aura 2 (EN)", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "aura", + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -87703,27 +92551,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.1592, + "output": 3.4776, + "cache_read": 0.11592 } }, - "workers-ai/@cf/deepgram/nova-3": { - "id": "workers-ai/@cf/deepgram/nova-3", - "name": "Deepgram Nova 3", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "nova", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -87732,41 +92594,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.1592, + "output": 4.6368, + "cache_read": 0.2898 } }, - "workers-ai/@cf/zai-org/glm-4.7-flash": { - "id": "workers-ai/@cf/zai-org/glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -87777,54 +92639,40 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 203000, "output": 131072 }, "cost": { - "input": 0.06, - "output": 0.4 + "input": 1.50696, + "output": 4.6368, + "cache_read": 0.370944 } }, - "workers-ai/@cf/deepseek-ai/deepseek-r1-distill-qwen-32b": { - "id": "workers-ai/@cf/deepseek-ai/deepseek-r1-distill-qwen-32b", - "name": "DeepSeek R1 Distill Qwen 32B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.5, - "output": 4.88 - } - }, - "workers-ai/@cf/qwen/qwen3-embedding-0.6b": { - "id": "workers-ai/@cf/qwen/qwen3-embedding-0.6b", - "name": "Qwen3 Embedding 0.6B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -87833,27 +92681,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.012, - "output": 0 + "input": 1.85472, + "output": 3.70944, + "cache_read": 0.46368 } }, - "workers-ai/@cf/qwen/qwen3-30b-a3b-fp8": { - "id": "workers-ai/@cf/qwen/qwen3-30b-a3b-fp8", - "name": "Qwen3 30B A3B FP8", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "deepseek-r1-0528": { + "id": "deepseek-r1-0528", + "name": "DeepSeek-R1", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" @@ -87862,27 +92716,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 164000, + "output": 32768 }, "cost": { - "input": 0.051, - "output": 0.34 + "input": 0.69552, + "output": 2.78208, + "cache_read": 0.185472 } }, - "workers-ai/@cf/qwen/qwen2.5-coder-32b-instruct": { - "id": "workers-ai/@cf/qwen/qwen2.5-coder-32b-instruct", - "name": "Qwen 2.5 Coder 32B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-04-11", - "last_updated": "2025-04-11", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -87891,27 +92760,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 164000, + "output": 64000 }, "cost": { - "input": 0.66, - "output": 1 + "input": 0.34776, + "output": 0.5796, + "cache_read": 0.092736 } }, - "workers-ai/@cf/qwen/qwq-32b": { - "id": "workers-ai/@cf/qwen/qwq-32b", - "name": "QwQ 32B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-04-11", - "last_updated": "2025-04-11", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -87920,114 +92803,169 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.66, - "output": 1 + "input": 1.1592, + "output": 3.4776, + "cache_read": 0.23184 } }, - "workers-ai/@cf/pipecat-ai/smart-turn-v2": { - "id": "workers-ai/@cf/pipecat-ai/smart-turn-v2", - "name": "Pipecat Smart Turn v2", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "smart-turn", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.81144, + "output": 3.4776, + "cache_read": 0.220248 } - }, - "workers-ai/@cf/meta/llama-3.1-8b-instruct": { - "id": "workers-ai/@cf/meta/llama-3.1-8b-instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + } + } + }, + "berget": { + "id": "berget", + "env": [ + "BERGET_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.berget.ai/v1", + "name": "Berget.AI", + "doc": "https://api.berget.ai", + "models": { + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "knowledge": "2025-12", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "audio", + "image", + "text", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 16384 + "output": 8192 }, "cost": { - "input": 0.28, - "output": 0.8299999999999998 + "input": 0.275, + "output": 0.55 } }, - "workers-ai/@cf/meta/m2m100-1.2b": { - "id": "workers-ai/@cf/meta/m2m100-1.2b", - "name": "M2M100 1.2B", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "m2m", - "attachment": false, - "reasoning": false, - "tool_call": false, + "Qwen/Qwen3.8-27B-FP8": { + "id": "Qwen/Qwen3.8-27B-FP8", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "release_date": "2026-08-14", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.34, - "output": 0.34 + "input": 0.46, + "output": 3.48 } }, - "workers-ai/@cf/meta/llama-3.2-1b-instruct": { - "id": "workers-ai/@cf/meta/llama-3.2-1b-instruct", - "name": "Llama 3.2 1B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "mistralai/Mistral-Small-3.2-24B-Instruct-2506": { + "id": "mistralai/Mistral-Small-3.2-24B-Instruct-2506", + "name": "Mistral Small 3.2 24B Instruct 2506", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "knowledge": "2025-09", + "release_date": "2025-10-01", + "last_updated": "2025-10-01", "modalities": { "input": [ "text" @@ -88036,56 +92974,81 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 32000, + "output": 8192 }, "cost": { - "input": 0.027, - "output": 0.2 + "input": 0.33, + "output": 0.33 } }, - "workers-ai/@cf/meta/llama-3.2-11b-vision-instruct": { - "id": "workers-ai/@cf/meta/llama-3.2-11b-vision-instruct", - "name": "Llama 3.2 11B Vision Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "moonshotai/Kimi-K3": { + "id": "moonshotai/Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-27", + "last_updated": "2026-07-28", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 327680, + "output": 32768 }, "cost": { - "input": 0.049, - "output": 0.68 + "input": 3, + "output": 15 } }, - "workers-ai/@cf/meta/llama-4-scout-17b-16e-instruct": { - "id": "workers-ai/@cf/meta/llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout 17B 16E Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -88094,59 +93057,96 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 524288, + "output": 32768 }, "cost": { - "input": 0.27, - "output": 0.85 + "input": 1.54, + "output": 4.84 } }, - "workers-ai/@cf/meta/llama-guard-3-8b": { - "id": "workers-ai/@cf/meta/llama-guard-3-8b", - "name": "Llama Guard 3 8B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + "zai-org/GLM-5.3-Flash": { + "id": "zai-org/GLM-5.3-Flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "release_date": "2026-08-26", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 524288, "output": 16384 }, "cost": { - "input": 0.48, - "output": 0.03 + "input": 0.29, + "output": 0.58 } - }, - "workers-ai/@cf/meta/llama-3-8b-instruct-awq": { - "id": "workers-ai/@cf/meta/llama-3-8b-instruct-awq", - "name": "Llama 3 8B Instruct AWQ", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + } + } + }, + "snowflake-cortex": { + "id": "snowflake-cortex", + "env": [ + "SNOWFLAKE_ACCOUNT", + "SNOWFLAKE_CORTEX_PAT" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://${SNOWFLAKE_ACCOUNT}.snowflakecomputing.com/api/v2/cortex/v1", + "name": "Snowflake Cortex", + "doc": "https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-rest-api", + "models": { + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -88154,28 +93154,40 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 200000, "output": 16384 - }, - "cost": { - "input": 0.12, - "output": 0.27 } }, - "workers-ai/@cf/meta/llama-3.1-8b-instruct-awq": { - "id": "workers-ai/@cf/meta/llama-3.1-8b-instruct-awq", - "name": "Llama 3.1 8B Instruct AWQ", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "openai-gpt-5.5": { + "id": "openai-gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -88183,28 +93195,34 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "cost": { - "input": 0.12, - "output": 0.27 - } + "status": "beta" }, - "workers-ai/@cf/meta/llama-3.3-70b-instruct-fp8-fast": { - "id": "workers-ai/@cf/meta/llama-3.3-70b-instruct-fp8-fast", - "name": "Llama 3.3 70B Instruct FP8 Fast", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -88212,28 +93230,32 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.29, - "output": 2.25 + "context": 200000, + "output": 64000 } }, - "workers-ai/@cf/meta/llama-3-8b-instruct": { - "id": "workers-ai/@cf/meta/llama-3-8b-instruct", - "name": "Llama 3 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -88241,28 +93263,39 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 200000, "output": 16384 - }, - "cost": { - "input": 0.28, - "output": 0.83 } }, - "workers-ai/@cf/meta/llama-3.1-8b-instruct-fp8": { - "id": "workers-ai/@cf/meta/llama-3.1-8b-instruct-fp8", - "name": "Llama 3.1 8B Instruct FP8", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "openai-gpt-5-nano": { + "id": "openai-gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -88270,28 +93303,41 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, - "cost": { - "input": 0.15, - "output": 0.29 - } + "status": "beta" }, - "workers-ai/@cf/meta/llama-2-7b-chat-fp16": { - "id": "workers-ai/@cf/meta/llama-2-7b-chat-fp16", - "name": "Llama 2 7B Chat FP16", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "openai-gpt-5": { + "id": "openai-gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -88299,28 +93345,30 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, - "cost": { - "input": 0.56, - "output": 6.67 - } + "status": "beta" }, - "workers-ai/@cf/meta/llama-3.2-3b-instruct": { - "id": "workers-ai/@cf/meta/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "openai-gpt-4.1": { + "id": "openai-gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -88328,28 +93376,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.051, - "output": 0.34 + "context": 1047576, + "output": 32768 } }, - "workers-ai/@cf/facebook/bart-large-cnn": { - "id": "workers-ai/@cf/facebook/bart-large-cnn", - "name": "BART Large CNN", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "bart", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-09", - "last_updated": "2025-04-09", + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -88357,54 +93405,50 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 + "context": 1000000, + "output": 128000 } }, - "workers-ai/@cf/aisingapore/gemma-sea-lion-v4-27b-it": { - "id": "workers-ai/@cf/aisingapore/gemma-sea-lion-v4-27b-it", - "name": "Gemma SEA-LION v4 27B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, + "mistral-large2": { + "id": "mistral-large2", + "name": "Mistral Large (latest)", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.35, - "output": 0.56 + "context": 262144, + "output": 262144 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "openai-gpt-5.2": { + "id": "openai-gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high" @@ -88413,10 +93457,10 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -88428,31 +93472,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "context": 400000, + "input": 272000, + "output": 128000 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -88460,32 +93502,28 @@ }, "open_weights": false, "limit": { - "context": 16385, - "output": 4096 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 1.25 + "context": 1000000, + "output": 128000 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -88493,31 +93531,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "context": 1000000, + "output": 128000 } }, - "openai/gpt-4": { - "id": "openai/gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -88525,25 +93560,23 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 30, - "output": 60 + "context": 1000000, + "output": 128000 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "openai-gpt-5.6-luna": { + "id": "openai-gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high" @@ -88553,13 +93586,14 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -88567,42 +93601,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.28 - } + "status": "beta" }, - "openai/o3-pro": { - "id": "openai/o3-pro", - "name": "o3-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -88610,25 +93636,23 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 20, - "output": 80 + "context": 1000000, + "output": 128000 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai-gpt-5.6-sol": { + "id": "openai-gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high" @@ -88638,13 +93662,14 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -88652,26 +93677,25 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1050000, + "input": 922000, "output": 128000 }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } + "status": "beta" }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, + "openai-gpt-5-mini": { + "id": "openai-gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high" @@ -88681,12 +93705,13 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -88694,84 +93719,56 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 272000, + "input": 272000, + "output": 8192 }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 - } + "status": "beta" }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek-R1", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "context": 128000, + "output": 32768 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -88784,30 +93781,23 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "ai-gateway-provider" - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "context": 1000000, + "output": 16384 } }, - "openai/o1": { - "id": "openai/o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "openai-gpt-5.1": { + "id": "openai-gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high" @@ -88816,10 +93806,10 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -88831,20 +93821,42 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 400000, + "input": 272000, + "output": 128000 + } + }, + "snowflake-llama3.3-70b": { + "id": "snowflake-llama3.3-70b", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "openai-gpt-5.6-terra": { + "id": "openai-gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -88852,19 +93864,19 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -88881,32 +93893,26 @@ "input": 922000, "output": 128000 }, - "provider": { - "npm": "ai-gateway-provider" - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "status": "beta" }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -88914,31 +93920,52 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 }, - "cost": { - "input": 10, - "output": 30 + "status": "beta", + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "gemini-3.1-pro": { + "id": "gemini-3.1-pro", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -88946,39 +93973,35 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.08 + "context": 1048576, + "output": 65536 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai-gpt-5.4": { + "id": "openai-gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -88991,31 +94014,52 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, + "input": 922000, "output": 128000 }, - "provider": { - "npm": "ai-gateway-provider" - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "status": "beta", + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } } - }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + } + } + }, + "sarvam": { + "id": "sarvam", + "env": [ + "SARVAM_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.sarvam.ai/v1", + "name": "Sarvam AI", + "doc": "https://docs.sarvam.ai/api-reference-docs/getting-started/models", + "models": { + "sarvam-30b": { + "id": "sarvam-30b", + "name": "Sarvam-30B", + "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", + "family": "sarvam", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + null, "low", "medium", "high" @@ -89023,143 +94067,94 @@ } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-18", + "last_updated": "2026-03-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "context": 65536, + "output": 65536 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "sarvam-105b": { + "id": "sarvam-105b", + "name": "Sarvam-105B", + "description": "Flagship Indian-language reasoning model for enterprise multilingual applications", + "family": "sarvam", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + null, "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "anthropic/claude-3.5-haiku": { - "id": "anthropic/claude-3.5-haiku", - "name": "Claude Haiku 3.5 (latest)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": false, - "tool_call": true, "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "release_date": "2026-02-18", + "last_updated": "2026-03-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 - }, - "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "context": 131072, + "output": 131072 } - }, - "anthropic/claude-3.5-sonnet": { - "id": "anthropic/claude-3.5-sonnet", - "name": "Claude Sonnet 3.5 v2", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + } + } + }, + "nova": { + "id": "nova", + "env": [ + "NOVA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.nova.amazon.com/v1", + "name": "Nova", + "doc": "https://nova.amazon.com/dev/documentation", + "models": { + "nova-2-pro-v1": { + "id": "nova-2-pro-v1", + "name": "Nova 2 Pro", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "release_date": "2025-12-03", + "last_updated": "2026-01-03", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ @@ -89168,24 +94163,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0, + "reasoning": 0 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude Sonnet 4 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "nova-2-lite-v1": { + "id": "nova-2-lite-v1", + "name": "Nova 2 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova-lite", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -89197,13 +94194,13 @@ ], "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ @@ -89212,43 +94209,45 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 1000000, "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0, + "reasoning": 0 } - }, - "anthropic/claude-opus-4-5": { - "id": "anthropic/claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + } + } + }, + "abacus": { + "id": "abacus", + "env": [ + "ABACUS_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://routellm.abacus.ai/v1", + "name": "Abacus", + "doc": "https://abacus.ai/help/api", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -89256,43 +94255,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "anthropic/claude-sonnet-4-5": { - "id": "anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "Grok 4.1 Fast (Non-Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2025-11-17", + "last_updated": "2025-11-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -89300,45 +94286,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 2000000, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.2, + "output": 0.5 } }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -89347,82 +94317,64 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "output": 64000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2.5, + "output": 7.5 } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "gemini-3.1-flash-image-preview": { + "id": "gemini-3.1-flash-image-preview", + "name": "Nano Banana 2 Preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 32768 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.5, + "output": 3 } }, - "anthropic/claude-3-sonnet": { - "id": "anthropic/claude-3-sonnet", - "name": "Claude Sonnet 3", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-03-04", - "last_updated": "2024-03-04", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -89430,43 +94382,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 0.3 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "anthropic/claude-opus-4-8": { - "id": "anthropic/claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -89474,33 +94417,34 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "anthropic/claude-3-opus": { - "id": "anthropic/claude-3-opus", - "name": "Claude Opus 3", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-02-29", - "last_updated": "2024-02-29", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -89508,77 +94452,66 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "anthropic/claude-3-5-haiku": { - "id": "anthropic/claude-3-5-haiku", - "name": "Claude Haiku 3.5 (latest)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "gemini-2.5-flash-image": { + "id": "gemini-2.5-flash-image", + "name": "Nano Banana", + "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", + "family": "gemini-flash", "attachment": true, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "knowledge": "2024-06", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "input": 0.3, + "output": 30 } }, - "anthropic/claude-opus-4-1": { - "id": "anthropic/claude-opus-4-1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -89586,43 +94519,32 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -89631,42 +94553,31 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 131072 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 2, + "output": 6 } }, - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -89674,43 +94585,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -89718,46 +94620,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "anthropic/claude-3-haiku": { - "id": "anthropic/claude-3-haiku", - "name": "Claude Haiku 3", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "claude-3-7-sonnet-20250219": { + "id": "claude-3-7-sonnet-20250219", + "name": "Claude Sonnet 3.7", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-03-13", - "last_updated": "2024-03-13", + "knowledge": "2024-10-31", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ "text", @@ -89771,108 +94656,60 @@ "open_weights": false, "limit": { "context": 200000, - "output": 4096 + "output": 64000 }, "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.03, - "cache_write": 0.3 + "input": 3, + "output": 15 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "qwen-2.5-coder-32b": { + "id": "qwen-2.5-coder-32b", + "name": "Qwen 2.5 Coder 32B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "release_date": "2024-11-11", + "last_updated": "2024-11-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "ai-gateway-provider" + "context": 128000, + "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 0.79, + "output": 0.79 } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Claude Opus 4 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -89881,300 +94718,259 @@ "open_weights": false, "limit": { "context": 200000, - "output": 32000 + "output": 100000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 1.1, + "output": 4.4 } - } - } - }, - "moonshotai-cn": { - "id": "moonshotai-cn", - "env": [ - "MOONSHOT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.moonshot.cn/v1", - "name": "Moonshot AI (China)", - "doc": "https://platform.moonshot.cn/docs/api/chat", - "models": { - "kimi-k2.7-code-highspeed": { - "id": "kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code HighSpeed", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + }, + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 1.75, + "output": 14, + "cache_read": 0.18 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "kimi-k2-turbo-preview": { - "id": "kimi-k2-turbo-preview", - "name": "Kimi K2 Turbo", - "description": "Fast Kimi model for responsive chat, coding help, and agent loops", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "muse-spark-1.1": { + "id": "muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "release_date": "2026-04-08", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 2.4, - "output": 10, - "cache_read": 0.6 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "kimi-k2-0711-preview": { - "id": "kimi-k2-0711-preview", - "name": "Kimi K2 0711", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-07-14", - "last_updated": "2025-07-14", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", - "attachment": false, + "gpt-5.2-chat-latest": { + "id": "gpt-5.2-chat-latest", + "name": "GPT-5.2 Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "knowledge": "2024-09-30", + "release_date": "2026-01-01", + "last_updated": "2026-01-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 1.75, + "output": 14 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "gpt-5.1-chat-latest": { + "id": "gpt-5.1-chat-latest", + "name": "GPT-5.1 Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 1.25, + "output": 10 } }, - "kimi-k2-thinking-turbo": { - "id": "kimi-k2-thinking-turbo", - "name": "Kimi K2 Thinking Turbo", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -90183,74 +94979,67 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 16384 }, "cost": { - "input": 1.15, - "output": 8, - "cache_read": 0.15 + "input": 1.2, + "output": 6 } }, - "kimi-k2-0905-preview": { - "id": "kimi-k2-0905-preview", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "claude-opus-4-20250514": { + "id": "claude-opus-4-20250514", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "release_date": "2025-05-14", + "last_updated": "2025-05-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 15, + "output": 75 } - } - } - }, - "morph": { - "id": "morph", - "env": [ - "MORPH_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.morphllm.com/v1", - "name": "Morph", - "doc": "https://docs.morphllm.com/api-reference/introduction", - "models": { - "morph-v3-fast": { - "id": "morph-v3-fast", - "name": "Morph v3 Fast", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "morph", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", + }, + "gemini-3.6-flash": { + "id": "gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -90258,25 +95047,26 @@ }, "open_weights": false, "limit": { - "context": 16000, - "output": 16000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.8, - "output": 1.2 + "input": 1.5, + "output": 7.5, + "cache_read": 0.15 } }, - "morph-v3-large": { - "id": "morph-v3-large", - "name": "Morph v3 Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "morph", + "kimi-k2-turbo-preview": { + "id": "kimi-k2-turbo-preview", + "name": "Kimi K2 Turbo Preview", + "description": "Fast Kimi model for responsive chat, coding help, and agent loops", + "family": "kimi-k2", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", + "tool_call": true, + "temperature": true, + "release_date": "2025-07-08", + "last_updated": "2025-07-08", "modalities": { "input": [ "text" @@ -90287,28 +95077,31 @@ }, "open_weights": false, "limit": { - "context": 32000, - "output": 32000 + "context": 256000, + "output": 8192 }, "cost": { - "input": 0.9, - "output": 1.9 + "input": 0.15, + "output": 8 } }, - "auto": { - "id": "auto", - "name": "Auto", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "grok-4.3": { + "id": "grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -90316,47 +95109,26 @@ }, "open_weights": false, "limit": { - "context": 32000, - "output": 32000 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.85, - "output": 1.55 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } - } - } - }, - "sakana": { - "id": "sakana", - "env": [ - "SAKANA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.sakana.ai/v1", - "name": "Sakana AI", - "doc": "https://console.sakana.ai/models", - "models": { - "fugu-ultra-20260615": { - "id": "fugu-ultra-20260615", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", + }, + "grok-4-fast-non-reasoning": { + "id": "grok-4-fast-non-reasoning", + "name": "Grok 4 Fast (Non-Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "temperature": true, + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text", @@ -90368,59 +95140,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 - }, - "provider": { - "shape": "responses" + "context": 2000000, + "output": 16384 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.2, + "output": 0.5 } }, - "fugu": { - "id": "fugu", - "name": "Fugu", - "description": "Multi-agent model for routing expert agents across complex analytical tasks", - "family": "fugu", + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -90428,34 +95172,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 200000, + "output": 32000 }, - "provider": { - "shape": "responses" + "cost": { + "input": 15, + "output": 75 } }, - "fugu-ultra": { - "id": "fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", @@ -90467,239 +95204,197 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 - }, - "provider": { - "shape": "responses" + "context": 200000, + "output": 64000 }, "cost": { "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "output": 25 } - } - } - }, - "deepinfra": { - "id": "deepinfra", - "env": [ - "DEEPINFRA_API_KEY" - ], - "npm": "@ai-sdk/deepinfra", - "name": "Deep Infra", - "doc": "https://deepinfra.com/models", - "models": { - "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { - "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", - "name": "Llama 4 Maverick 17B FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + }, + "gemini-3-pro-image-preview": { + "id": "gemini-3-pro-image-preview", + "name": "Nano Banana Pro Preview", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "structured_output": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 16384 + "context": 65536, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "meta-llama/Llama-4-Scout-17B-16E-Instruct": { - "id": "meta-llama/Llama-4-Scout-17B-16E-Instruct", - "name": "Llama 4 Scout 17B", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 327680, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "meta-llama/Llama-3.3-70B-Instruct-Turbo": { - "id": "meta-llama/Llama-3.3-70B-Instruct-Turbo", - "name": "Llama 3.3 70B Turbo", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "gemini-3.1-flash-image": { + "id": "gemini-3.1-flash-image", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1048576, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.32 + "input": 0.5, + "output": 3 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", - "video" + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.75, - "output": 3.5, - "cache_read": 0.15 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 500000, "output": 32768 }, "cost": { - "input": 0.45, - "output": 2.25, - "cache_read": 0.07 + "input": 2, + "output": 6 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "family": "kimi-k2", - "attachment": true, + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -90713,72 +95408,62 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "output": 32768 }, "cost": { - "input": 0.74, - "output": 3.5, - "cache_read": 0.15 + "input": 0.6, + "output": 3 } }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.13, - "output": 0.38 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "google/gemma-4-26B-A4B-it": { - "id": "google/gemma-4-26B-A4B-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "llama-3.3-70b-versatile": { + "id": "llama-3.3-70b-versatile", + "name": "Llama 3.3 70B Versatile", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -90786,59 +95471,59 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 128000, "output": 32768 }, "cost": { - "input": 0.07, - "output": 0.34 + "input": 0.59, + "output": 0.79 } }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 81920 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.95 + "input": 5, + "output": 25 } }, - "Qwen/Qwen3.7-Max": { - "id": "Qwen/Qwen3.7-Max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -90849,116 +95534,99 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "tiers": [ - { - "input": 5, - "output": 15, - "cache_read": 1, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 6.25, - "output": 18.5, - "cache_read": 1.25, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 1, + "output": 3, + "cache_read": 0.2 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.32, - "output": 3.2 + "input": 0.15, + "output": 0.6 } }, - "Qwen/Qwen3-Next-80B-A3B-Instruct": { - "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "o3-pro": { + "id": "o3-pro", + "name": "o3-pro", + "description": "High-effort o3 tier for difficult technical reasoning and careful answers", + "family": "o-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.09, - "output": 1.1 + "input": 20, + "output": 40 } }, - "Qwen/Qwen3-Max": { - "id": "Qwen/Qwen3-Max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -90966,40 +95634,20 @@ }, "open_weights": false, "limit": { - "context": 256000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24, - "tiers": [ - { - "input": 2.4, - "output": 12, - "cache_read": 0.48, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 3, - "output": 15, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 1.5, + "output": 9, + "cache_read": 0.15 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen 3.5 397B A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [], @@ -91007,755 +95655,683 @@ "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-02-01", - "last_updated": "2026-04-20", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video" + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 81920 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.45, - "output": 3, - "cache_read": 0.22 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "Qwen/Qwen3.5-122B-A10B": { - "id": "Qwen/Qwen3.5-122B-A10B", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-3-pro-image": { + "id": "gemini-3-pro-image", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 65536, + "output": 32768 }, "cost": { - "input": 0.29, - "output": 2.4 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "Qwen/Qwen3.5-27B": { - "id": "Qwen/Qwen3.5-27B", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video", "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.26, - "output": 2.6 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "Qwen/Qwen3.5-9B": { - "id": "Qwen/Qwen3.5-9B", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.15 + "input": 10, + "output": 50 } }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, + "muse-spark-1.2": { + "id": "muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 16384 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.08, - "output": 0.28 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo", - "name": "Qwen3 Coder 480B A35B Instruct Turbo", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 66536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1, - "cache_read": 0.1 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "Qwen/Qwen3.5-35B-A3B": { - "id": "Qwen/Qwen3.5-35B-A3B", - "name": "Qwen 3.5 35B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-01", - "last_updated": "2026-04-20", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 81920 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 1, - "cache_read": 0.05 + "input": 1, + "output": 6, + "cache_read": 0.1 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "gpt-5.1-codex-max": { + "id": "gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.037, - "output": 0.17 + "input": 1.25, + "output": 10 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.03, - "output": 0.14 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "XiaomiMiMo/MiMo-V2.5": { - "id": "XiaomiMiMo/MiMo-V2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08 + "input": 5, + "output": 25 } }, - "XiaomiMiMo/MiMo-V2.5-Pro": { - "id": "XiaomiMiMo/MiMo-V2.5-Pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "grok-code-fast-1": { + "id": "grok-code-fast-1", + "name": "Grok Code Fast 1", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2025-09-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, + "context": 256000, "output": 16384 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2 + "input": 0.2, + "output": 1.5 } }, - "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning": { - "id": "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning", - "name": "Nemotron 3 Nano Omni 30B A3B Reasoning", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "nvidia/Nemotron-3-Nano-30B-A3B": { - "id": "nvidia/Nemotron-3-Nano-30B-A3B", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.05, - "output": 0.2 + "input": 3, + "output": 15 } }, - "nvidia/Llama-3.3-Nemotron-Super-49B-v1.5": { - "id": "nvidia/Llama-3.3-Nemotron-Super-49B-v1.5", - "name": "Llama 3.3 Nemotron Super 49B v1.5", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.4, - "output": 0.4 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "zai-org/GLM-4.7-Flash": { - "id": "zai-org/GLM-4.7-Flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 16384 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.06, - "output": 0.4, - "cache_read": 0.01 + "input": 3, + "output": 15 } }, - "zai-org/GLM-4.6": { - "id": "zai-org/GLM-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.43, - "output": 1.74, - "cache_read": 0.08 + "input": 5, + "output": 25 } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ - "text" - ], - "output": [ + "text", + "image", + "audio", + "video" + ], + "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.08, - "cache_read": 0.12 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 } }, - "zai-org/GLM-4.7": { - "id": "zai-org/GLM-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "gpt-5.3-codex-xhigh": { + "id": "gpt-5.3-codex-xhigh", + "name": "GPT-5.3 Codex XHigh", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 1.75, - "cache_read": 0.08 + "input": 1.75, + "output": 14 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 32768 + "output": 65536 }, "cost": { - "input": 0.93, + "input": 0.5, "output": 3, - "cache_read": 0.18 + "cache_read": 0.05 } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "gpt-4o-2024-11-20": { + "id": "gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2024-10", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, + "context": 128000, "output": 16384 }, "cost": { - "input": 1.05, - "output": 3.5, - "cache_read": 0.205 + "input": 2.5, + "output": 10 } }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", - "name": "DeepSeek-R1-0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "attachment": false, + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -91763,136 +96339,95 @@ }, "open_weights": false, "limit": { - "context": 163840, + "context": 200000, "output": 64000 }, "cost": { - "input": 0.5, - "output": 2.15, - "cache_read": 0.35 + "input": 1, + "output": 5 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.09, - "output": 0.18, - "cache_read": 0.018 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 16384 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1.3, - "output": 2.6, - "cache_read": 0.1 + "input": 3, + "output": 15 } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "gpt-5.3-chat-latest": { + "id": "gpt-5.3-chat-latest", + "name": "GPT-5.3 Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -91900,150 +96435,131 @@ }, "open_weights": false, "limit": { - "context": 163840, - "output": 64000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.26, - "output": 0.38, - "cache_read": 0.13 + "input": 1.75, + "output": 14 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "claude-sonnet-4-20250514": { + "id": "claude-sonnet-4-20250514", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2025-05-14", + "last_updated": "2025-05-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 1.15, - "cache_read": 0.03 + "input": 3, + "output": 15 } }, - "MiniMaxAI/MiniMax-M3": { - "id": "MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "MiniMaxAI/MiniMax-M2.7": { - "id": "MiniMaxAI/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, + "o3": { + "id": "o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 131072 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.25, - "output": 1, - "cache_read": 0.05 + "input": 2, + "output": 8, + "cache_read": 0.5 } - } - } - }, - "google-vertex-anthropic": { - "id": "google-vertex-anthropic", - "env": [ - "GOOGLE_VERTEX_PROJECT", - "GOOGLE_VERTEX_LOCATION", - "GOOGLE_APPLICATION_CREDENTIALS" - ], - "npm": "@ai-sdk/google-vertex/anthropic", - "name": "Vertex (Anthropic)", - "doc": "https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude", - "models": { - "claude-haiku-4-5@20251001": { - "id": "claude-haiku-4-5@20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + }, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -92051,39 +96567,32 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 5, + "output": 25 } }, - "claude-opus-4@20250514": { - "id": "claude-opus-4@20250514", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -92091,40 +96600,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 400000, + "input": 272000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "claude-opus-4-1@20250805": { - "id": "claude-opus-4-1@20250805", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "grok-4.6": { + "id": "grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -92132,47 +96635,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 500000, + "output": 32768 }, - "status": "deprecated", "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 2, + "output": 6, + "cache_read": 0.5 } }, - "claude-opus-4-5@20251101": { - "id": "claude-opus-4-5@20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gemini-3.1-flash-lite-preview": { + "id": "gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -92181,33 +96671,31 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 1 } }, - "claude-3-5-haiku@20241022": { - "id": "claude-3-5-haiku@20241022", - "name": "Claude Haiku 3.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "route-llm": { + "id": "route-llm", + "name": "RouteLLM", + "description": "RouteLLM routes prompts to an appropriate Abacus-backed text-generation model", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "release_date": "2024-01-01", + "last_updated": "2026-07-10", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -92215,40 +96703,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 128000, + "output": 64000 }, - "status": "deprecated", "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "input": 3, + "output": 15 } }, - "claude-sonnet-4@20250514": { - "id": "claude-sonnet-4@20250514", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "grok-4-0709": { + "id": "grok-4-0709", + "name": "Grok 4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -92256,46 +96734,32 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 256000, + "output": 16384 }, - "status": "deprecated", "cost": { "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "output": 15 } }, - "claude-opus-4-7@default": { - "id": "claude-opus-4-7@default", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -92308,52 +96772,29 @@ }, "cost": { "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "output": 30, + "cache_read": 0.5 } }, - "claude-sonnet-4-5@20250929": { - "id": "claude-sonnet-4-5@20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", "image", - "pdf" + "audio" ], "output": [ "text" @@ -92361,286 +96802,209 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025 } }, - "claude-sonnet-5@default": { - "id": "claude-sonnet-5@default", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "deepseek/deepseek-v3.1": { + "id": "deepseek/deepseek-v3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.55, + "output": 1.66 } }, - "claude-opus-4-6@default": { - "id": "claude-opus-4-6@default", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "meta-llama/Meta-Llama-3.3-70B-Instruct": { + "id": "meta-llama/Meta-Llama-3.3-70B-Instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } + "input": 0.59, + "output": 0.79 + } + }, + "meta-llama/Meta-Llama-3.1-8B-Instruct": { + "id": "meta-llama/Meta-Llama-3.1-8B-Instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.02, + "output": 0.05 } }, - "claude-opus-4-8@default": { - "id": "claude-opus-4-8@default", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo": { + "id": "meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo", + "name": "Llama 3.1 405B Instruct Turbo", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 3.5, + "output": 3.5 + } + }, + "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { + "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", + "name": "Llama 4 Maverick 17B Instruct", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 8192 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 0.14, + "output": 0.59 } }, - "claude-sonnet-4-6@default": { - "id": "claude-sonnet-4-6@default", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "thinkingmachines/Inkling": { + "id": "thinkingmachines/Inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 3.74, + "output": 9.36, + "cache_read": 0.748 } - } - } - }, - "v0": { - "id": "v0", - "env": [ - "V0_API_KEY" - ], - "npm": "@ai-sdk/vercel", - "name": "v0", - "doc": "https://sdk.vercel.ai/providers/ai-sdk-providers/vercel", - "models": { - "v0-1.0-md": { - "id": "v0-1.0-md", - "name": "v0-1.0-md", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "v0", + }, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -92650,112 +97014,118 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 3, - "output": 15 + "input": 0.14, + "output": 0.4 } }, - "v0-1.5-lg": { - "id": "v0-1.5-lg", - "name": "v0-1.5-lg", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "v0", - "attachment": true, + "Qwen/QwQ-32B": { + "id": "Qwen/QwQ-32B", + "name": "QwQ 32B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-06-09", - "last_updated": "2025-06-09", + "release_date": "2024-11-28", + "last_updated": "2024-11-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 512000, - "output": 32000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 15, - "output": 75 + "input": 0.4, + "output": 0.4 } }, - "v0-1.5-md": { - "id": "v0-1.5-md", - "name": "v0-1.5-md", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "v0", - "attachment": true, + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-06-09", - "last_updated": "2025-06-09", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, + "limit": { + "context": 262144, + "output": 8192 + }, + "cost": { + "input": 0.13, + "output": 0.6 + } + }, + "Qwen/Qwen2.5-72B-Instruct": { + "id": "Qwen/Qwen2.5-72B-Instruct", + "name": "Qwen 2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-09-19", + "last_updated": "2024-09-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, "limit": { "context": 128000, - "output": 32000 + "output": 8192 }, "cost": { - "input": 3, - "output": 15 + "input": 0.11, + "output": 0.38 } - } - } - }, - "azure": { - "id": "azure", - "env": [ - "AZURE_RESOURCE_NAME", - "AZURE_API_KEY" - ], - "npm": "@ai-sdk/azure", - "name": "Azure", - "doc": "https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models", - "models": { - "codex-mini": { - "id": "codex-mini", - "name": "Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex-mini", - "attachment": true, + }, + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2024-04", - "release_date": "2025-05-16", - "last_updated": "2025-05-16", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -92764,29 +97134,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 1.5, - "output": 6, - "cache_read": 0.375 + "input": 0.09, + "output": 0.29 } }, - "phi-3.5-moe-instruct": { - "id": "phi-3.5-moe-instruct", - "name": "Phi-3.5-MoE-instruct", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", + "Qwen/Qwen3-Coder-480B-A35B-Instruct": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "name": "Qwen3-Coder 480B-A35B Instruct", + "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-08-20", - "last_updated": "2024-08-20", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -92797,26 +97166,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.16, - "output": 0.64 + "input": 0.29, + "output": 1.2 } }, - "gpt-3.5-turbo-instruct": { - "id": "gpt-3.5-turbo-instruct", - "name": "GPT-3.5 Turbo Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-09-21", - "last_updated": "2023-09-21", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -92825,19 +97195,19 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4096, - "output": 4096 + "context": 262144, + "output": 8192 }, "cost": { - "input": 1.5, - "output": 2 + "input": 0.32, + "output": 3.2 } }, - "deepseek-r1-0528": { - "id": "deepseek-r1-0528", - "name": "DeepSeek-R1-0528", + "deepseek-ai/DeepSeek-R1": { + "id": "deepseek-ai/DeepSeek-R1", + "name": "DeepSeek R1", "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", "family": "deepseek-thinking", "attachment": false, @@ -92845,9 +97215,8 @@ "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -92858,23 +97227,23 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.35, - "output": 5.4 + "input": 3, + "output": 7 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek-V4-Flash", + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-05", @@ -92891,113 +97260,89 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 384000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" + "output": 32768 }, "cost": { - "input": 0.19, - "output": 0.51 + "input": 0.14, + "output": 0.28, + "cache_read": 0.03 } }, - "gpt-5.2-chat": { - "id": "gpt-5.2-chat", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 1.74, + "output": 3.48, + "cache_read": 0.15 } }, - "o3": { - "id": "o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", - "attachment": true, + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2025-06-15", + "last_updated": "2025-06-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.27, + "output": 0.4 } }, - "deepseek-v3-0324": { - "id": "deepseek-v3-0324", - "name": "DeepSeek-V3-0324", + "deepseek-ai/DeepSeek-V3.1-Terminus": { + "id": "deepseek-ai/DeepSeek-V3.1-Terminus", + "name": "DeepSeek V3.1 Terminus", "description": "DeepSeek chat model for instruction following, coding, and analysis", "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "release_date": "2025-06-01", + "last_updated": "2025-06-01", "modalities": { "input": [ "text" @@ -93008,29 +97353,31 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.14, - "output": 4.56 + "input": 0.27, + "output": 1 } }, - "phi-3-small-128k-instruct": { - "id": "phi-3-small-128k-instruct", - "name": "Phi-3-small-instruct (128k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, - "reasoning": false, - "tool_call": false, + "MiniMaxAI/MiniMax-M3": { + "id": "MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -93038,26 +97385,26 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.3, + "output": 1.2 } }, - "meta-llama-3-8b-instruct": { - "id": "meta-llama-3-8b-instruct", - "name": "Meta-Llama-3-8B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "MiniMaxAI/MiniMax-M2.7": { + "id": "MiniMaxAI/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-04-18", - "last_updated": "2024-04-18", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -93068,117 +97415,131 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 2048 + "context": 204800, + "output": 131072 }, "cost": { "input": 0.3, - "output": 0.61 + "output": 1.2 } }, - "mistral-small-2503": { - "id": "mistral-small-2503", - "name": "Mistral Small 3.1", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2025-03-01", - "last_updated": "2025-03-01", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "text-embedding-3-large": { - "id": "text-embedding-3-large", - "name": "text-embedding-3-large", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "moonshotai/Kimi-K3": { + "id": "moonshotai/Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8191, - "output": 3072 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.13, - "output": 0 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "o1-mini": { - "id": "o1-mini", - "name": "o1-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-09-12", - "last_updated": "2024-09-12", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "phi-3.5-mini-instruct": { - "id": "phi-3.5-mini-instruct", - "name": "Phi-3.5-mini-instruct", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-08-20", - "last_updated": "2024-08-20", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -93189,26 +97550,28 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.13, - "output": 0.52 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "mistral-nemo": { - "id": "mistral-nemo", - "name": "Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -93219,29 +97582,31 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "cohere-embed-v-4-0": { - "id": "cohere-embed-v-4-0", - "name": "Embed v4", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "zai-org/GLM-4.7": { + "id": "zai-org/GLM-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -93249,155 +97614,119 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 1536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.12, - "output": 0 + "input": 0.6, + "output": 2.2 } }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-08-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 204800, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1, + "output": 3.2 } }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "zai-org/GLM-4.6": { + "id": "zai-org/GLM-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 272000, - "output": 128000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.6, + "output": 2.2 } }, - "gpt-5-chat": { - "id": "gpt-5-chat", - "name": "GPT-5 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, + "zai-org/GLM-4.5": { + "id": "zai-org/GLM-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "temperature": false, - "knowledge": "2024-10-24", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 96000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.6, + "output": 2.2 } }, - "cohere-command-a": { - "id": "cohere-command-a", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -93408,30 +97737,40 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 8000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.08, + "output": 0.44 } - }, - "llama-3.2-11b-vision-instruct": { - "id": "llama-3.2-11b-vision-instruct", - "name": "Llama-3.2-11B-Vision-Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", - "attachment": true, + } + } + }, + "novita-ai": { + "id": "novita-ai", + "env": [ + "NOVITA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.novita.ai/openai", + "name": "NovitaAI", + "doc": "https://novita.ai/docs/guides/introduction", + "models": { + "kwaipilot/kat-coder-pro": { + "id": "kwaipilot/kat-coder-pro", + "name": "Kat Coder Pro", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "release_date": "2026-01-05", + "last_updated": "2026-01-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -93439,66 +97778,54 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.37, - "output": 0.37 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "deepseek/deepseek-prover-v2-671b": { + "id": "deepseek/deepseek-prover-v2-671b", + "name": "Deepseek Prover V2 671B", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-04-30", + "last_updated": "2025-04-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 272000 + "context": 160000, + "output": 160000 }, "cost": { - "input": 15, - "output": 120 + "input": 0.7, + "output": 2.5 } }, - "cohere-command-r-08-2024": { - "id": "cohere-command-r-08-2024", - "name": "Command R", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", + "deepseek/deepseek-r1-distill-qwen-14b": { + "id": "deepseek/deepseek-r1-distill-qwen-14b", + "name": "DeepSeek R1 Distill Qwen 14B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "deepseek-thinking", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -93509,58 +97836,59 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 32768, + "output": 16384 }, "cost": { "input": 0.15, - "output": 0.6 + "output": 0.15 } }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, + "deepseek/deepseek-v3-0324": { + "id": "deepseek/deepseek-v3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "knowledge": "2024-07", + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 163840, + "output": 163840 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.27, + "output": 1.12, + "cache_read": 0.135 } }, - "gpt-4": { - "id": "gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "deepseek/deepseek-r1-distill-llama-70b": { + "id": "deepseek/deepseek-r1-distill-llama-70b", + "name": "DeepSeek R1 Distill LLama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-03-14", - "last_updated": "2023-03-14", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", "modalities": { "input": [ "text" @@ -93569,70 +97897,55 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 8192, "output": 8192 }, "cost": { - "input": 60, - "output": 120 + "input": 0.8, + "output": 0.8 } }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "deepseek/deepseek-v3-turbo": { + "id": "deepseek/deepseek-v3-turbo", + "name": "DeepSeek V3 (Turbo)\t", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2025-03-05", + "last_updated": "2025-03-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 64000, + "output": 16000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 0.4, + "output": 1.3 } }, - "phi-3-medium-128k-instruct": { - "id": "phi-3-medium-128k-instruct", - "name": "Phi-3-medium-instruct (128k)", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", + "deepseek/deepseek-r1-0528-qwen3-8b": { + "id": "deepseek/deepseek-r1-0528-qwen3-8b", + "name": "DeepSeek R1 0528 Qwen3 8B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2025-05-29", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" @@ -93644,25 +97957,30 @@ "open_weights": true, "limit": { "context": 128000, - "output": 4096 + "output": 32000 }, "cost": { - "input": 0.17, - "output": 0.68 + "input": 0.06, + "output": 0.09 } }, - "gpt-4-32k": { - "id": "gpt-4-32k", - "name": "GPT-4 32K", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "deepseek/deepseek-v3.1-terminus": { + "id": "deepseek/deepseek-v3.1-terminus", + "name": "Deepseek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-03-14", - "last_updated": "2023-03-14", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ "text" @@ -93671,31 +97989,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, + "context": 131072, "output": 32768 }, "cost": { - "input": 60, - "output": 120 + "input": 0.27, + "output": 1, + "cache_read": 0.135 } }, - "meta-llama-3.1-405b-instruct": { - "id": "meta-llama-3.1-405b-instruct", - "name": "Meta-Llama-3.1-405B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "deepseek/deepseek-ocr-2": { + "id": "deepseek/deepseek-ocr-2", + "name": "deepseek/deepseek-ocr-2", + "description": "OCR model for extracting structured text from documents and screenshots", + "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "tool_call": false, + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -93703,29 +98020,29 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 8192, + "output": 8192 }, "cost": { - "input": 5.33, - "output": 16 + "input": 0.03, + "output": 0.03 } }, - "cohere-command-r-plus-08-2024": { - "id": "cohere-command-r-plus-08-2024", - "name": "Command R+", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", - "attachment": false, + "deepseek/deepseek-ocr": { + "id": "deepseek/deepseek-ocr", + "name": "DeepSeek-OCR", + "description": "OCR model for extracting structured text from documents and screenshots", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "release_date": "2025-10-24", + "last_updated": "2025-10-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -93733,26 +98050,25 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.03, + "output": 0.03 } }, - "phi-4-mini": { - "id": "phi-4-mini", - "name": "Phi-4-mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "deepseek/deepseek-r1-turbo": { + "id": "deepseek/deepseek-r1-turbo", + "name": "DeepSeek R1 (Turbo)\t", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2025-03-05", + "last_updated": "2025-03-05", "modalities": { "input": [ "text" @@ -93763,60 +98079,33 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 64000, + "output": 16000 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 0.7, + "output": 2.5 } }, - "gpt-3.5-turbo-1106": { - "id": "gpt-3.5-turbo-1106", - "name": "GPT-3.5 Turbo 1106", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "deepseek/deepseek-v3.2-exp": { + "id": "deepseek/deepseek-v3.2-exp", + "name": "Deepseek V3.2 Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-11-06", - "last_updated": "2023-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16384, - "output": 16384 - }, - "cost": { - "input": 1, - "output": 2 - } - }, - "llama-4-scout-17b-16e-instruct": { - "id": "llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout 17B 16E Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -93824,26 +98113,25 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 163840, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 0.78 + "input": 0.27, + "output": 0.41 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "deepseek/deepseek-r1-distill-qwen-32b": { + "id": "deepseek/deepseek-r1-distill-qwen-32b", + "name": "DeepSeek R1 Distill Qwen 32B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "deepseek-thinking", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -93854,26 +98142,31 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 64000, + "output": 32000 }, "cost": { - "input": 0.71, - "output": 0.71 + "input": 0.3, + "output": 0.3 } }, - "grok-4-20-non-reasoning": { - "id": "grok-4-20-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "deepseek/deepseek-v3.1": { + "id": "deepseek/deepseek-v3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ "text" @@ -93882,70 +98175,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 8192 + "context": 131072, + "output": 32768 }, - "status": "beta", "cost": { - "input": 2, - "output": 6 + "input": 0.27, + "output": 1, + "cache_read": 0.135 } }, - "gpt-5.1-chat": { - "id": "gpt-5.1-chat", - "name": "GPT-5.1 Chat", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt-codex", - "attachment": true, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "medium" + "low", + "medium", + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text", - "image", - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek-V4-Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, "structured_output": true, "temperature": true, "knowledge": "2025-05", @@ -93961,33 +98226,29 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" + "context": 1048576, + "output": 393216 }, "cost": { - "input": 1.74, - "output": 3.48 + "input": 1.6, + "output": 3.2, + "cache_read": 0.135 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "deepseek/deepseek-r1-0528": { + "id": "deepseek/deepseek-r1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-12-02", + "knowledge": "2024-07", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -93998,66 +98259,60 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 163840, + "output": 32768 }, "cost": { - "input": 0.6, + "input": 0.7, "output": 2.5, - "cache_read": 0.15 + "cache_read": 0.35 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "Deepseek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 163840, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.269, + "output": 0.4, + "cache_read": 0.1345 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -94066,63 +98321,56 @@ { "type": "effort", "values": [ + "minimal", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "status": "beta", - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 1048576, + "output": 393216 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "meta-llama/llama-4-maverick-17b-128e-instruct-fp8": { + "id": "meta-llama/llama-4-maverick-17b-128e-instruct-fp8", + "name": "Llama 4 Maverick Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "attachment": true, + "reasoning": false, "tool_call": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "release_date": "2025-04-06", + "last_updated": "2025-04-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -94130,106 +98378,85 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 1048576, + "output": 8192 }, "cost": { - "input": 1.35, - "output": 5.4 + "input": 0.27, + "output": 0.85 } }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "Grok 4.1 Fast (Non-Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, + "meta-llama/llama-3-70b-instruct": { + "id": "meta-llama/llama-3-70b-instruct", + "name": "Llama3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-06-27", - "last_updated": "2025-06-27", + "release_date": "2024-04-25", + "last_updated": "2024-04-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 8192, + "output": 8000 }, - "status": "beta", "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.51, + "output": 0.74 } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 Nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "meta-llama/llama-3.3-70b-instruct": { + "id": "meta-llama/llama-3.3-70b-instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-07", + "last_updated": "2024-12-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 120000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.135, + "output": 0.4 } }, - "phi-3-small-8k-instruct": { - "id": "phi-3-small-8k-instruct", - "name": "Phi-3-small-instruct (8k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "meta-llama/llama-3.1-8b-instruct": { + "id": "meta-llama/llama-3.1-8b-instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2024-07-24", + "last_updated": "2024-07-24", "modalities": { "input": [ "text" @@ -94240,35 +98467,24 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 2048 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.02, + "output": 0.05 } }, - "gpt-5.3-chat": { - "id": "gpt-5.3-chat", - "name": "GPT-5.3 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "meta-llama/llama-4-scout-17b-16e-instruct": { + "id": "meta-llama/llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-04-06", + "last_updated": "2025-04-06", "modalities": { "input": [ "text", @@ -94278,29 +98494,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.18, + "output": 0.59 } }, - "gpt-3.5-turbo-0125": { - "id": "gpt-3.5-turbo-0125", - "name": "GPT-3.5 Turbo 0125", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "meta-llama/llama-3.2-3b-instruct": { + "id": "meta-llama/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "knowledge": "2021-08", - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "release_date": "2024-09-18", + "last_updated": "2024-09-18", "modalities": { "input": [ "text" @@ -94309,27 +98523,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "output": 16384 + "context": 32768, + "output": 32000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 0.03, + "output": 0.05 } }, - "cohere-embed-v3-multilingual": { - "id": "cohere-embed-v3-multilingual", - "name": "Embed v3 Multilingual", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", + "meta-llama/llama-3-8b-instruct": { + "id": "meta-llama/llama-3-8b-instruct", + "name": "Llama 3 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2023-11-07", - "last_updated": "2023-11-07", + "temperature": true, + "release_date": "2024-04-25", + "last_updated": "2024-04-25", "modalities": { "input": [ "text" @@ -94340,26 +98554,25 @@ }, "open_weights": true, "limit": { - "context": 512, - "output": 1024 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0 + "input": 0.04, + "output": 0.04 } }, - "gpt-3.5-turbo-0613": { - "id": "gpt-3.5-turbo-0613", - "name": "GPT-3.5 Turbo 0613", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "nousresearch/hermes-2-pro-llama-3-8b": { + "id": "nousresearch/hermes-2-pro-llama-3-8b", + "name": "Hermes 2 Pro Llama 3 8B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-06-13", - "last_updated": "2023-06-13", + "release_date": "2024-06-27", + "last_updated": "2024-06-27", "modalities": { "input": [ "text" @@ -94368,90 +98581,69 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "output": 16384 + "context": 8192, + "output": 8192 }, "cost": { - "input": 3, - "output": 4 + "input": 0.14, + "output": 0.14 } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt-codex", + "xiaomimimo/mimo-v2-flash": { + "id": "xiaomimimo/mimo-v2-flash", + "name": "XiaomiMiMo/MiMo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-12-19", + "last_updated": "2025-12-19", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ - "text", - "image", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 32000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.1, + "output": 0.3, + "cache_read": 0.3 } }, - "gpt-5.1-codex-max": { - "id": "gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "xiaomimimo/mimo-v2-pro": { + "id": "xiaomimimo/mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-05-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -94459,28 +98651,48 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 2, + "output": 6, + "cache_read": 0.4, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "phi-4-reasoning": { - "id": "phi-4-reasoning", - "name": "Phi-4-reasoning", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", + "xiaomimimo/mimo-v2.5-pro": { + "id": "xiaomimimo/mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-27", "modalities": { "input": [ "text" @@ -94491,93 +98703,42 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 4096 - }, - "cost": { - "input": 0.125, - "output": 0.5 - } - }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-12-31", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 1048576, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, + "input": 0.522, + "output": 1.044, + "cache_read": 0.0043, "tiers": [ { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, + "input": 0.522, + "output": 1.044, + "cache_read": 0.0043, "tier": { "type": "context", - "size": 200000 + "size": 256000 } } ], "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.522, + "output": 1.044, + "cache_read": 0.0043 } } }, - "mistral-medium-2505": { - "id": "mistral-medium-2505", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "baidu/ernie-4.5-vl-28b-a3b": { + "id": "baidu/ernie-4.5-vl-28b-a3b", + "name": "ERNIE 4.5 VL 28B A3B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "release_date": "2025-06-30", + "last_updated": "2026-06-14", "modalities": { "input": [ "text", @@ -94587,31 +98748,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 30000, + "output": 8000 }, "cost": { - "input": 0.4, - "output": 2 + "input": 0.14, + "output": 0.56 } }, - "meta-llama-3.1-70b-instruct": { - "id": "meta-llama-3.1-70b-instruct", - "name": "Meta-Llama-3.1-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "baidu/ernie-4.5-vl-28b-a3b-thinking": { + "id": "baidu/ernie-4.5-vl-28b-a3b-thinking", + "name": "ERNIE-4.5-VL-28B-A3B-Thinking", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2025-11-26", + "last_updated": "2025-11-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -94619,24 +98782,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 131072, + "output": 65536 }, "cost": { - "input": 2.68, - "output": 3.54 + "input": 0.39, + "output": 0.39 } }, - "text-embedding-ada-002": { - "id": "text-embedding-ada-002", - "name": "text-embedding-ada-002", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "baidu/ernie-4.5-21B-a3b-thinking": { + "id": "baidu/ernie-4.5-21B-a3b-thinking", + "name": "ERNIE-4.5-21B-A3B-Thinking", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "ernie", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "release_date": "2022-12-15", - "last_updated": "2022-12-15", + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-09-19", + "last_updated": "2025-09-19", "modalities": { "input": [ "text" @@ -94645,38 +98811,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0 + "input": 0.07, + "output": 0.28 } }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "baidu/ernie-4.5-21B-a3b": { + "id": "baidu/ernie-4.5-21B-a3b", + "name": "ERNIE 4.5 21B A3B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "ernie", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ "text" @@ -94685,42 +98841,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 120000, + "output": 8000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.07, + "output": 0.28 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "baidu/ernie-4.5-vl-424b-a47b": { + "id": "baidu/ernie-4.5-vl-424b-a47b", + "name": "ERNIE 4.5 VL 424B A47B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ "text", @@ -94730,29 +98871,62 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 123000, + "output": 16000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.125 + "input": 0.42, + "output": 1.25 } }, - "gpt-4-turbo-vision": { - "id": "gpt-4-turbo-vision", - "name": "GPT-4 Turbo Vision", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "baidu/ernie-4.5-300b-a47b-paddle": { + "id": "baidu/ernie-4.5-300b-a47b-paddle", + "name": "ERNIE 4.5 300B A47B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-06-30", + "last_updated": "2025-06-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 123000, + "output": 12000 + }, + "cost": { + "input": 0.28, + "output": 1.1 + } + }, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -94762,71 +98936,63 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 131072 }, "cost": { - "input": 10, - "output": 30 + "input": 0.14, + "output": 0.4 } }, - "mistral-large-2411": { - "id": "mistral-large-2411", - "name": "Mistral Large 24.11", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": false, + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma 3 27B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-11-01", - "last_updated": "2024-11-01", + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 98304, + "output": 16384 }, "cost": { - "input": 2, - "output": 6 + "input": 0.119, + "output": 0.2 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -94836,58 +99002,58 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.13, + "output": 0.4 } }, - "text-embedding-3-small": { - "id": "text-embedding-3-small", - "name": "text-embedding-3-small", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, + "google/gemma-3-12b-it": { + "id": "google/gemma-3-12b-it", + "name": "Gemma 3 12B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": false, "tool_call": false, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "temperature": true, + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8191, - "output": 1536 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.02, - "output": 0 + "input": 0.05, + "output": 0.1 } }, - "deepseek-v3.2-speciale": { - "id": "deepseek-v3.2-speciale", - "name": "DeepSeek-V3.2-Speciale", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "minimaxai/minimax-m1-80k": { + "id": "minimaxai/minimax-m1-80k", + "name": "MiniMax M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text" @@ -94898,26 +99064,57 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 1000000, + "output": 40000 }, "cost": { - "input": 0.58, - "output": 1.68 + "input": 0.55, + "output": 2.2 } }, - "gpt-3.5-turbo-0301": { - "id": "gpt-3.5-turbo-0301", - "name": "GPT-3.5 Turbo 0301", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "mistralai/mistral-nemo": { + "id": "mistralai/mistral-nemo", + "name": "Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-03-01", - "last_updated": "2023-03-01", + "release_date": "2024-07-30", + "last_updated": "2024-07-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 60288, + "output": 16000 + }, + "cost": { + "input": 0.04, + "output": 0.17 + } + }, + "inclusionai/ring-2.6-1t": { + "id": "inclusionai/ring-2.6-1t", + "name": "Ring-2.6-1T", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "ring", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-08", + "last_updated": "2026-05-27", "modalities": { "input": [ "text" @@ -94928,26 +99125,27 @@ }, "open_weights": false, "limit": { - "context": 4096, - "output": 4096 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.5, - "output": 2 + "input": 0.3, + "output": 2.5, + "cache_read": 0.06 } }, - "meta-llama-3.1-8b-instruct": { - "id": "meta-llama-3.1-8b-instruct", - "name": "Meta-Llama-3.1-8B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "inclusionai/ling-2.6-1t": { + "id": "inclusionai/ling-2.6-1t", + "name": "Ling-2.6-1T", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "ling", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2026-04-23", + "last_updated": "2026-06-29", "modalities": { "input": [ "text" @@ -94958,108 +99156,98 @@ }, "open_weights": true, "limit": { - "context": 128000, + "context": 262144, "output": 32768 }, "cost": { "input": 0.3, - "output": 0.61 + "output": 2.5, + "cache_read": 0.06 } }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "inclusionai/ling-2.6-flash": { + "id": "inclusionai/ling-2.6-flash", + "name": "Ling-2.6-flash", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 262144, + "output": 32768 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", "family": "kimi-k2", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -95067,16 +99255,19 @@ } ], "tool_call": true, - "interleaved": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -95087,28 +99278,28 @@ "context": 262144, "output": 262144 }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" - }, "cost": { - "input": 0.6, - "output": 3 + "input": 0.8, + "output": 3.4, + "cache_read": 0.16 } }, - "phi-4": { - "id": "phi-4", - "name": "Phi-4", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2025-11-07", + "last_updated": "2026-06-29", "modalities": { "input": [ "text" @@ -95119,216 +99310,213 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.125, - "output": 0.5 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "grok-4-fast-reasoning": { - "id": "grok-4-fast-reasoning", - "name": "Grok 4 Fast (Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-09-19", - "last_updated": "2025-09-19", + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "gpt-image-1.5": { - "id": "gpt-image-1.5", - "name": "GPT-Image-1.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, + "moonshotai/kimi-k2-instruct": { + "id": "moonshotai/kimi-k2-instruct", + "name": "Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "tool_call": true, + "temperature": true, + "release_date": "2025-07-11", + "last_updated": "2025-07-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 32, - "cache_read": 1.25 + "input": 0.57, + "output": 2.3 } }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 0.6, + "output": 2.5 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 262144, + "output": 262144 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "o1": { - "id": "o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": false, + "zai-org/glm-4.6v": { + "id": "zai-org/glm-4.6v", + "name": "GLM 4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glmv", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", + "video", "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 0.3, + "output": 0.9, + "cache_read": 0.055 } }, - "deepseek-v3.1": { - "id": "deepseek-v3.1", - "name": "DeepSeek-V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "zai-org/glm-4.5": { + "id": "zai-org/glm-4.5", + "name": "GLM-4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -95340,29 +99528,76 @@ "open_weights": true, "limit": { "context": 131072, + "output": 98304 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 + } + }, + "zai-org/glm-4.6": { + "id": "zai-org/glm-4.6", + "name": "GLM 4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-09-30", + "last_updated": "2025-09-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, "output": 131072 }, "cost": { - "input": 0.56, - "output": 1.68 + "input": 0.55, + "output": 2.2, + "cache_read": 0.11 } }, - "llama-3.2-90b-vision-instruct": { - "id": "llama-3.2-90b-vision-instruct", - "name": "Llama-3.2-90B-Vision-Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", - "attachment": true, - "reasoning": false, + "zai-org/glm-5": { + "id": "zai-org/glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "release_date": "2026-02-11", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -95370,71 +99605,72 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 202800, + "output": 131072 }, "cost": { - "input": 2.04, - "output": 2.04 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "zai-org/glm-4.5v": { + "id": "zai-org/glm-4.5v", + "name": "GLM 4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glmv", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-02-31", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ "text", - "image", - "pdf" + "video", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 65536, + "output": 16384 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.6, + "output": 1.8, + "cache_read": 0.11 } }, - "ministral-3b": { - "id": "ministral-3b", - "name": "Ministral 3B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "zai-org/glm-4.7-flash": { + "id": "zai-org/glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -95445,129 +99681,103 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.04, - "output": 0.04 + "input": 0.07, + "output": 0.4, + "cache_read": 0.01 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, + "zai-org/glm-4.7": { + "id": "zai-org/glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 Mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, + "zai-org/glm-5.2": { + "id": "zai-org/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "zai-org/glm-5.1": { + "id": "zai-org/glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -95576,16 +99786,16 @@ } ], "tool_call": true, - "interleaved": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -95593,31 +99803,25 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.95, - "output": 4 + "input": 1.38, + "output": 4.4, + "cache_read": 0.26 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", + "zai-org/autoglm-phone-9b-multilingual": { + "id": "zai-org/autoglm-phone-9b-multilingual", + "name": "AutoGLM-Phone-9B-Multilingual", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-12-10", + "last_updated": "2025-12-10", "modalities": { "input": [ "text", @@ -95627,149 +99831,99 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 65536, + "output": 65536 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.035, + "output": 0.138 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "zai-org/glm-4.5-air": { + "id": "zai-org/glm-4.5-air", + "name": "GLM 4.5 Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "knowledge": "2025-04", + "release_date": "2025-10-13", + "last_updated": "2025-10-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 131072, + "output": 98304 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 0.13, + "output": 0.85, + "cache_read": 0.025 } }, - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "Grok 4.1 Fast (Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "gryphe/mythomax-l2-13b": { + "id": "gryphe/mythomax-l2-13b", + "name": "Mythomax L2 13B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-06-27", - "last_updated": "2025-06-27", + "release_date": "2024-04-25", + "last_updated": "2024-04-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 4096, + "output": 3200 }, - "status": "beta", "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.09, + "output": 0.09 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -95777,62 +99931,102 @@ }, "open_weights": false, "limit": { - "context": 272000, - "output": 128000 + "context": 204800, + "output": 131100 }, "cost": { - "input": 0.25, - "output": 2, + "input": 0.3, + "output": 1.2, "cache_read": 0.03 } }, - "gpt-image-1": { - "id": "gpt-image-1", - "name": "GPT-Image-1", - "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-04-24", - "last_updated": "2025-04-24", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax-m2.7", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 204800, + "output": 131072 }, "cost": { - "input": 5, - "output": 40, - "cache_read": 1.25 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-03-18", + "last_updated": "2026-05-27", "modalities": { "input": [ - "text", - "image" + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "minimax/minimax-m2.5-highspeed": { + "id": "minimax/minimax-m2.5-highspeed", + "name": "MiniMax M2.5 Highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax-m2.5", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -95840,104 +100034,97 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 204800, + "output": 131100 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.6, + "output": 2.4, + "cache_read": 0.03 } }, - "gpt-4-turbo": { - "id": "gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "Minimax M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": false, "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 10, - "output": 30 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 272000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "phi-3-medium-4k-instruct": { - "id": "phi-3-medium-4k-instruct", - "name": "Phi-3-medium-instruct (4k)", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", - "attachment": false, + "paddlepaddle/paddleocr-vl": { + "id": "paddlepaddle/paddleocr-vl", + "name": "PaddleOCR-VL", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, "reasoning": false, "tool_call": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2025-10-22", + "last_updated": "2025-10-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -95945,41 +100132,65 @@ }, "open_weights": true, "limit": { - "context": 4096, - "output": 1024 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0.17, - "output": 0.68 + "input": 0.02, + "output": 0.02 } }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "qwen/qwen3-vl-235b-a22b-instruct": { + "id": "qwen/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-09-24", + "last_updated": "2025-09-24", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 1.5 + } + }, + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7-Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -95987,43 +100198,32 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 1.25, + "output": 3.75, + "cache_read": 0.25, + "cache_write": 1.5625 } }, - "cohere-embed-v3-english": { - "id": "cohere-embed-v3-english", - "name": "Embed v3 English", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", - "attachment": false, - "reasoning": false, + "qwen/qwen3-vl-235b-a22b-thinking": { + "id": "qwen/qwen3-vl-235b-a22b-thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "temperature": false, - "release_date": "2023-11-07", - "last_updated": "2023-11-07", + "temperature": true, + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -96031,80 +100231,64 @@ }, "open_weights": true, "limit": { - "context": 512, - "output": 1024 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0 + "input": 0.98, + "output": 3.95 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5-27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 262144, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.3, + "output": 2.4 } }, - "codestral-2501": { - "id": "codestral-2501", - "name": "Codestral 25.01", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "codestral", + "qwen/qwen3-235b-a22b-instruct-2507": { + "id": "qwen/qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-03", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2025-04", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ "text" @@ -96113,29 +100297,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0.09, + "output": 0.58 } }, - "phi-4-reasoning-plus": { - "id": "phi-4-reasoning-plus", - "name": "Phi-4-reasoning-plus", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2025-09-10", + "last_updated": "2025-09-10", "modalities": { "input": [ "text" @@ -96146,31 +100329,29 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 4096 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.125, - "output": 0.5 + "input": 0.15, + "output": 1.5 } }, - "phi-4-multimodal": { - "id": "phi-4-multimodal", - "name": "Phi-4-multimodal", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "phi", - "attachment": true, + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -96178,89 +100359,99 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.08, - "output": 0.32, - "input_audio": 4 + "input": 0.2, + "output": 1.5 } }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", + "qwen/qwen3-omni-30b-a3b-instruct": { + "id": "qwen/qwen3-omni-30b-a3b-instruct", + "name": "Qwen3 Omni 30B A3B Instruct", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2024-04", + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ "text", + "video", + "audio", "image" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 65536, "output": 16384 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.25, + "output": 0.97, + "input_audio": 2.2, + "output_audio": 1.788 } }, - "model-router": { - "id": "model-router", - "name": "Model Router", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "model-router", + "qwen/qwen3-omni-30b-a3b-thinking": { + "id": "qwen/qwen3-omni-30b-a3b-thinking", + "name": "Qwen3 Omni 30B A3B Thinking", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "release_date": "2025-05-19", - "last_updated": "2025-11-18", + "structured_output": true, + "temperature": true, + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ "text", + "audio", + "video", "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 65536, "output": 16384 }, "cost": { - "input": 0.14, - "output": 0 + "input": 0.25, + "output": 0.97, + "input_audio": 2.2, + "output_audio": 1.788 } }, - "grok-4-20-reasoning": { - "id": "grok-4-20-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "knowledge": "2025-04", + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ "text" @@ -96271,70 +100462,85 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 8192 + "context": 262144, + "output": 65536 }, - "status": "beta", "cost": { - "input": 2, - "output": 6 + "input": 2.11, + "output": 8.45 } }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "qwen/qwen3-235b-a22b-fp8": { + "id": "qwen/qwen3-235b-a22b-fp8", + "name": "Qwen3 235B A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "release_date": "2025-04-29", + "last_updated": "2025-04-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 40960, + "output": 20000 + }, + "cost": { + "input": 0.2, + "output": 0.8 + } + }, + "qwen/qwen3-coder-480b-a35b-instruct": { + "id": "qwen/qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.38, + "output": 1.55 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "qwen/qwen3-coder-30b-a3b-instruct": { + "id": "qwen/qwen3-coder-30b-a3b-instruct", + "name": "Qwen3 Coder 30b A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2025-10-09", + "last_updated": "2025-10-09", "modalities": { "input": [ "text" @@ -96345,29 +100551,29 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 160000, + "output": 32768 }, "cost": { - "input": 0.58, - "output": 1.68 + "input": 0.07, + "output": 0.27 } }, - "llama-4-maverick-17b-128e-instruct-fp8": { - "id": "llama-4-maverick-17b-128e-instruct-fp8", - "name": "Llama 4 Maverick 17B 128E Instruct FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "qwen/qwen3-vl-30b-a3b-instruct": { + "id": "qwen/qwen3-vl-30b-a3b-instruct", + "name": "qwen/qwen3-vl-30b-a3b-instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2025-10-11", + "last_updated": "2025-10-11", "modalities": { "input": [ "text", + "video", "image" ], "output": [ @@ -96376,71 +100582,122 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.2, + "output": 0.7 } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "qwen/qwen3-235b-a22b-thinking-2507": { + "id": "qwen/qwen3-235b-a22b-thinking-2507", + "name": "Qwen3 235B A22b Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 3 + } + }, + "qwen/qwen2.5-7b-instruct": { + "id": "qwen/qwen2.5-7b-instruct", + "name": "Qwen2.5 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-04-16", + "last_updated": "2025-04-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.07, + "output": 0.07 + } + }, + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5-122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "temperature": true, + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.4, + "output": 3.2 } }, - "phi-4-mini-reasoning": { - "id": "phi-4-mini-reasoning", - "name": "Phi-4-mini-reasoning", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "qwen/qwen3-30b-a3b-fp8": { + "id": "qwen/qwen3-30b-a3b-fp8", + "name": "Qwen3 30B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -96451,26 +100708,25 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 40960, + "output": 20000 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 0.09, + "output": 0.45 } }, - "phi-3-mini-128k-instruct": { - "id": "phi-3-mini-128k-instruct", - "name": "Phi-3-mini-instruct (128k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2025-09-10", + "last_updated": "2025-09-10", "modalities": { "input": [ "text" @@ -96481,57 +100737,55 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.13, - "output": 0.52 + "input": 0.15, + "output": 1.5 } }, - "gpt-image-2": { - "id": "gpt-image-2", - "name": "GPT-Image-2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "qwen/qwen3-vl-30b-a3b-thinking": { + "id": "qwen/qwen3-vl-30b-a3b-thinking", + "name": "qwen/qwen3-vl-30b-a3b-thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-11", + "last_updated": "2025-10-11", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 1.25 + "input": 0.2, + "output": 1 } }, - "phi-3-mini-4k-instruct": { - "id": "phi-3-mini-4k-instruct", - "name": "Phi-3-mini-instruct (4k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "qwen/qwen-mt-plus": { + "id": "qwen/qwen-mt-plus", + "name": "Qwen MT Plus", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2025-09-03", + "last_updated": "2025-09-03", "modalities": { "input": [ "text" @@ -96542,26 +100796,25 @@ }, "open_weights": true, "limit": { - "context": 4096, - "output": 1024 + "context": 16384, + "output": 8192 }, "cost": { - "input": 0.13, - "output": 0.52 + "input": 0.25, + "output": 0.75 } }, - "meta-llama-3-70b-instruct": { - "id": "meta-llama-3-70b-instruct", - "name": "Meta-Llama-3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "qwen/qwen3-32b-fp8": { + "id": "qwen/qwen3-32b-fp8", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-04-18", - "last_updated": "2024-04-18", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -96572,162 +100825,133 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 2048 + "context": 40960, + "output": 20000 }, "cost": { - "input": 2.68, - "output": 3.54 + "input": 0.1, + "output": 0.45 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", + "qwen/qwen3-8b-fp8": { + "id": "qwen/qwen3-8b-fp8", + "name": "Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "release_date": "2025-04-29", + "last_updated": "2025-04-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 20000 + }, + "cost": { + "input": 0.035, + "output": 0.138 + } + }, + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5-35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "temperature": true, + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ "text", "image", - "audio" + "video" ], "output": [ - "text", - "image", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 272000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.25, + "output": 2 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5-397B-A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": true, + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 64000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.6, + "output": 3.6 } - } - } - }, - "cerebras": { - "id": "cerebras", - "env": [ - "CEREBRAS_API_KEY" - ], - "npm": "@ai-sdk/cerebras", - "name": "Cerebras", - "doc": "https://inference-docs.cerebras.ai/models/overview", - "models": { - "gemma-4-31b": { - "id": "gemma-4-31b", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + }, + "qwen/qwen3-vl-8b-instruct": { + "id": "qwen/qwen3-vl-8b-instruct", + "name": "qwen/qwen3-vl-8b-instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-07-01", + "release_date": "2025-10-17", + "last_updated": "2025-10-17", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -96736,39 +100960,29 @@ "open_weights": true, "limit": { "context": 131072, - "output": 40960 + "output": 32768 }, - "status": "beta", "cost": { - "input": 0.99, - "output": 1.49 + "input": 0.08, + "output": 0.5 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "qwen/qwen2.5-vl-72b-instruct": { + "id": "qwen/qwen2.5-vl-72b-instruct", + "name": "Qwen2.5 VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2026-06-10", + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -96776,33 +100990,27 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 40960 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.35, - "output": 0.75 + "input": 0.8, + "output": 0.8 } }, - "zai-glm-4.7": { - "id": "zai-glm-4.7", - "name": "Z.AI GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "qwen/qwen-2.5-72b-instruct": { + "id": "qwen/qwen-2.5-72b-instruct", + "name": "Qwen 2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01-07", - "last_updated": "2026-06-10", + "knowledge": "2024-04", + "release_date": "2024-10-15", + "last_updated": "2024-10-15", "modalities": { "input": [ "text" @@ -96813,49 +101021,25 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 40960 + "context": 32000, + "output": 8192 }, - "status": "beta", "cost": { - "input": 2.25, - "output": 2.75, - "cache_read": 0, - "cache_write": 0 + "input": 0.38, + "output": 0.4 } - } - } - }, - "zai-coding-plan": { - "id": "zai-coding-plan", - "env": [ - "ZHIPU_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.z.ai/api/coding/paas/v4", - "name": "Z.AI Coding Plan", - "doc": "https://docs.z.ai/devpack/overview", - "models": { - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + }, + "qwen/qwen3-4b-fp8": { + "id": "qwen/qwen3-4b-fp8", + "name": "Qwen3 4B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -96866,36 +101050,27 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 20000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.03, + "output": 0.03 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "baichuan/baichuan-m2-32b": { + "id": "baichuan/baichuan-m2-32b", + "name": "baichuan-m2-32b", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "baichuan", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "knowledge": "2024-12", + "release_date": "2025-08-13", + "last_updated": "2025-08-13", "modalities": { "input": [ "text" @@ -96904,45 +101079,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, + "context": 131072, "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.07, + "output": 0.07 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "OpenAI: GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", - "max" + "low", + "medium", + "high" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -96950,75 +101121,64 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.04, + "output": 0.15 } }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "OpenAI GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.05, + "output": 0.25 } }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", + "microsoft/wizardlm-2-8x22b": { + "id": "microsoft/wizardlm-2-8x22b", + "name": "Wizardlm 2 8x22B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2024-04-24", + "last_updated": "2024-04-24", "modalities": { "input": [ "text" @@ -97029,36 +101189,24 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 65535, + "output": 8000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.62, + "output": 0.62 } }, - "glm-5-turbo": { - "id": "glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", + "sao10K/l3-70b-euryale-v2.1": { + "id": "sao10K/l3-70b-euryale-v2.1", + "name": "L3 70B Euryale V2.1\t", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "release_date": "2024-06-18", + "last_updated": "2024-06-18", "modalities": { "input": [ "text" @@ -97067,41 +101215,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.48, + "output": 1.48 } - } - } - }, - "nvidia": { - "id": "nvidia", - "env": [ - "NVIDIA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://integrate.api.nvidia.com/v1", - "name": "Nvidia", - "doc": "https://docs.api.nvidia.com/nim/", - "models": { - "baai/bge-m3": { - "id": "baai/bge-m3", - "name": "BGE M3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "bge", + }, + "sao10K/L3-8B-stheno-v3.2": { + "id": "sao10K/L3-8B-stheno-v3.2", + "name": "L3 8B Stheno V3.2", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-01-30", - "last_updated": "2026-04-30", + "tool_call": true, + "temperature": true, + "release_date": "2024-11-29", + "last_updated": "2024-11-29", "modalities": { "input": [ "text" @@ -97113,25 +101247,24 @@ "open_weights": true, "limit": { "context": 8192, - "output": 1024 + "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.05, + "output": 0.05 } }, - "moonshotai/kimi-k2-instruct-0905": { - "id": "moonshotai/kimi-k2-instruct-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "sao10K/l3-8b-lunaris": { + "id": "sao10K/l3-8b-lunaris", + "name": "Sao10k L3 8B Lunaris\t", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "release_date": "2024-11-28", + "last_updated": "2024-11-28", "modalities": { "input": [ "text" @@ -97142,49 +101275,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.05, + "output": 0.05 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "sao10K/l31-70b-euryale-v2.2": { + "id": "sao10K/l31-70b-euryale-v2.2", + "name": "L31 70B Euryale V2.2", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -97192,65 +101303,79 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 1.48, + "output": 1.48 } - }, - "minimaxai/minimax-m3": { - "id": "minimaxai/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + } + } + }, + "302ai": { + "id": "302ai", + "env": [ + "302AI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.302.ai/v1", + "name": "302.AI", + "doc": "https://doc.302.ai", + "models": { + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "grok-4-1-fast-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "knowledge": "2025-06", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 16384 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 0.5 } }, - "minimaxai/minimax-m2.7": { - "id": "minimaxai/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "glm-4.6v": { + "id": "glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-04-11", + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -97258,76 +101383,93 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.145, + "output": 0.43 } }, - "stepfun-ai/step-3.7-flash": { - "id": "stepfun-ai/step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "gpt-5.4": { + "id": "gpt-5.4", + "name": "gpt-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 0, + "tiers": [ + { + "input": 5, + "output": 22.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5 + } } }, - "stepfun-ai/step-3.5-flash": { - "id": "stepfun-ai/step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "glm-4.5": { + "id": "glm-4.5", + "name": "GLM-4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "temperature": true, - "release_date": "2026-02-02", - "last_updated": "2026-02-02", + "knowledge": "2025-04", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ "text" @@ -97338,110 +101480,130 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0, - "output": 0 + "input": 0.286, + "output": 1.142 } }, - "google/gemma-3n-e4b-it": { - "id": "google/gemma-3n-e4b-it", - "name": "Gemma 3n E4b It", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "claude-haiku-4-5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-06-03", - "last_updated": "2025-06-03", + "knowledge": "2025-02-28", + "release_date": "2025-10-16", + "last_updated": "2025-10-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 5 } }, - "google/gemma-3n-e2b-it": { - "id": "google/gemma-3n-e2b-it", - "name": "Gemma 3n E2b It", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": true, - "reasoning": false, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-06-12", - "last_updated": "2025-06-12", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.8, + "output": 5.3 } }, - "google/google-paligemma": { - "id": "google/google-paligemma", - "name": "paligemma", - "description": "Gemini multimodal model for text, image, audio, video, and document tasks", + "gemini-3.1-flash-image-preview": { + "id": "gemini-3.1-flash-image-preview", + "name": "gemini-3.1-flash-image-preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2024-05-14", - "last_updated": "2024-08-26", + "knowledge": "2025-01", + "release_date": "2026-02-27", + "last_updated": "2026-02-27", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.5, + "output": 60 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma-4-31B-IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -97451,9 +101613,8 @@ ], "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", @@ -97466,86 +101627,110 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 1048576, + "output": 512000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.72, + "output": 2.88 } }, - "google/gemma-2-2b-it": { - "id": "google/gemma-2-2b-it", - "name": "Gemma 2 2b It", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": false, - "reasoning": false, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "gpt-5.4-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2024-07-16", - "last_updated": "2024-07-16", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-19", + "last_updated": "2026-03-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 1.25 } }, - "microsoft/phi-4-mini-instruct": { - "id": "microsoft/phi-4-mini-instruct", - "name": "Phi-4-Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, - "reasoning": false, + "gemini-3.5-flash-thinking": { + "id": "gemini-3.5-flash-thinking", + "name": "gemini-3.5-flash-thinking", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-01", - "last_updated": "2025-09-05", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.5, + "output": 9 } }, - "microsoft/phi-4-multimodal-instruct": { - "id": "microsoft/phi-4-multimodal-instruct", - "name": "Phi 4 Multimodal", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, + "gemini-3-pro-preview": { + "id": "gemini-3-pro-preview", + "name": "gemini-3-pro-preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -97553,21 +101738,20 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 12 } }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "family": "glm", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -97578,124 +101762,125 @@ "interleaved": { "field": "reasoning_content" }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, + "context": 200000, "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.72, + "output": 3.2 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT-OSS-120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "gemini-2.5-flash-image": { + "id": "gemini-2.5-flash-image", + "name": "gemini-2.5-flash-image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-04", - "last_updated": "2025-08-14", + "knowledge": "2025-01", + "release_date": "2025-10-08", + "last_updated": "2025-10-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 30 } }, - "openai/whisper-large-v3": { - "id": "openai/whisper-large-v3", - "name": "Whisper Large v3", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, + "gpt-4o": { + "id": "gpt-4o", + "name": "gpt-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, + "tool_call": true, + "structured_output": true, + "temperature": true, "knowledge": "2023-09", - "release_date": "2023-09-01", - "last_updated": "2025-09-05", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ - "audio" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, - "output": 4096 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 2.5, + "output": 10 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" @@ -97703,26 +101888,31 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.075, + "output": 0.25 } }, - "bytedance/seed-oss-36b-instruct": { - "id": "bytedance/seed-oss-36b-instruct", - "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "seed", + "glm-4.6": { + "id": "glm-4.6", + "name": "glm-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-09-04", - "last_updated": "2025-11-25", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -97731,27 +101921,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.286, + "output": 1.142 } }, - "mistralai/mistral-7b-instruct-v03": { - "id": "mistralai/mistral-7b-instruct-v03", - "name": "Mistral-7B-Instruct-v0.3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "qwen3-235b-a22b-instruct-2507", + "description": "Tool-capable chat model for instruction following and agentic application workflows", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-04-01", - "last_updated": "2025-04-01", + "knowledge": "2025-04", + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ "text" @@ -97760,29 +101950,52 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, + "context": 128000, "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.29, + "output": 1.143 } }, - "mistralai/magistral-small-2506": { - "id": "mistralai/magistral-small-2506", - "name": "Magistral Small 2506", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" @@ -97790,28 +102003,47 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 2.16, + "output": 6.36 } }, - "mistralai/mixtral-8x7b-instruct": { - "id": "mistralai/mixtral-8x7b-instruct", - "name": "Mistral: Mixtral 8x7B Instruct", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": false, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2023-12-10", - "last_updated": "2026-03-15", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -97819,29 +102051,45 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 16384 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15 } }, - "mistralai/mistral-medium-3-instruct": { - "id": "mistralai/mistral-medium-3-instruct", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "gpt-5.6-sol-pro": { + "id": "gpt-5.6-sol-pro", + "name": "gpt-5.6-sol-pro", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -97849,35 +102097,39 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 30 } }, - "mistralai/mistral-small-4-119b-2603": { - "id": "mistralai/mistral-small-4-119b-2603", - "name": "mistral-small-4-119b-2603", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "gpt-5-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", + "low", + "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ "text", @@ -97887,87 +102139,108 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.25, + "output": 2 } }, - "mistralai/mistral-nemotron": { - "id": "mistralai/mistral-nemotron", - "name": "mistral-nemotron", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "nemotron", + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "temperature": true, - "release_date": "2025-06-11", - "last_updated": "2025-06-12", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.12, + "output": 0.69 } }, - "mistralai/mistral-large-3-675b-instruct-2512": { - "id": "mistralai/mistral-large-3-675b-instruct-2512", - "name": "Mistral Large 3 675B Instruct 2512", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", + "claude-sonnet-4-6-thinking": { + "id": "claude-sonnet-4-6-thinking", + "name": "claude-sonnet-4-6-thinking", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "knowledge": "2025-08", + "release_date": "2026-02-18", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15 } }, - "mistralai/mixtral-8x22b-instruct": { - "id": "mistralai/mixtral-8x22b-instruct", - "name": "Mistral: Mixtral 8x22B Instruct", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "glm-5": { + "id": "glm-5", + "name": "glm-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2024-04-17", - "last_updated": "2024-04-17", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -97978,89 +102251,103 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 13108 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.6, + "output": 2.6 } }, - "nvidia/cosmos-transfer1-7b": { - "id": "nvidia/cosmos-transfer1-7b", - "name": "cosmos-transfer1-7b", - "description": "Video model for prompt-guided generation, editing, and motion workflows", + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "gpt-4.1-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-06-13", - "last_updated": "2025-06-30", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "video" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, - "output": 4096 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.4 } }, - "nvidia/cosmos-transfer2_5-2b": { - "id": "nvidia/cosmos-transfer2_5-2b", - "name": "cosmos-transfer2.5-2b", - "description": "Video model for prompt-guided generation, editing, and motion workflows", + "grok-4-fast-reasoning": { + "id": "grok-4-fast-reasoning", + "name": "grok-4-fast-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "video" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, - "output": 4096 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 0.5 } }, - "nvidia/llama-nemotron-embed-vl-1b-v2": { - "id": "nvidia/llama-nemotron-embed-vl-1b-v2", - "name": "llama-nemotron-embed-vl-1b-v2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "nemotron", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-02-10", - "last_updated": "2026-02-10", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -98068,103 +102355,106 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 2048 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.95, + "output": 4 } }, - "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning": { - "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning", - "name": "Nemotron 3 Nano Omni", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", + "gpt-5.2-chat-latest": { + "id": "gpt-5.2-chat-latest", + "name": "gpt-5.2-chat-latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": -1, - "max": 32768 + "type": "effort", + "values": [ + "medium" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-12", + "last_updated": "2025-12-12", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 1.75, + "output": 14 } }, - "nvidia/magpie-tts-zeroshot": { - "id": "nvidia/magpie-tts-zeroshot", - "name": "magpie-tts-zeroshot", - "description": "Speech generation model for controllable voice, narration, and audio delivery", + "gpt-5.1-chat-latest": { + "id": "gpt-5.1-chat-latest", + "name": "gpt-5.1-chat-latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-05-22", - "last_updated": "2025-06-12", + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text", - "audio" + "image" ], "output": [ - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, - "output": 4096 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 10 } }, - "nvidia/nvidia-nemotron-nano-9b-v2": { - "id": "nvidia/nvidia-nemotron-nano-9b-v2", - "name": "nvidia-nemotron-nano-9b-v2", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2025-08-18", - "last_updated": "2025-08-18", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -98175,57 +102465,66 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 204800, "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.2 } }, - "nvidia/synthetic-video-detector": { - "id": "nvidia/synthetic-video-detector", - "name": "synthetic-video-detector", - "description": "Video model for prompt-guided generation, editing, and motion workflows", + "gemini-2.5-flash-preview-09-2025": { + "id": "gemini-2.5-flash-preview-09-2025", + "name": "gemini-2.5-flash-preview-09-2025", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2025-01", + "release_date": "2025-09-26", + "last_updated": "2025-09-26", "modalities": { "input": [ - "video" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, - "output": 4096 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 2.5 } }, - "nvidia/nemotron-content-safety-reasoning-4b": { - "id": "nvidia/nemotron-content-safety-reasoning-4b", - "name": "nemotron-content-safety-reasoning-4b", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "nemotron", - "attachment": false, + "glm-4.5v": { + "id": "glm-4.5v", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": false, - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-12", + "last_updated": "2025-08-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -98233,55 +102532,89 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 64000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.29, + "output": 0.86 } }, - "nvidia/nv-embed-v1": { - "id": "nvidia/nv-embed-v1", - "name": "nv-embed-v1", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "attachment": false, - "reasoning": false, - "tool_call": false, + "claude-fable-5-1": { + "id": "claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "temperature": false, - "release_date": "2024-06-07", - "last_updated": "2025-07-22", + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 2048 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 10, + "output": 50 } }, - "nvidia/usdcode": { - "id": "nvidia/usdcode", - "name": "usdcode", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gemini-3.6-flash": { + "id": "gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-01-01", - "last_updated": "2026-01-01", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -98289,165 +102622,192 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.5, + "output": 7.5 } }, - "nvidia/riva-translate-4b-instruct-v1_1": { - "id": "nvidia/riva-translate-4b-instruct-v1_1", - "name": "riva-translate-4b-instruct-v1_1", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, + "claude-opus-5-thinking": { + "id": "claude-opus-5-thinking", + "name": "claude-opus-5-thinking", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": false, - "release_date": "2025-12-12", - "last_updated": "2025-12-12", + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25 } }, - "nvidia/sparsedrive": { - "id": "nvidia/sparsedrive", - "name": "sparsedrive", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "attachment": true, + "qwen3-coder-480b-a35b-instruct": { + "id": "qwen3-coder-480b-a35b-instruct", + "name": "qwen3-coder-480b-a35b-instruct", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-03-18", - "last_updated": "2025-07-20", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.86, + "output": 3.43 } }, - "nvidia/rerank-qa-mistral-4b": { - "id": "nvidia/rerank-qa-mistral-4b", - "name": "rerank-qa-mistral-4b", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "attachment": false, + "doubao-seed-1-6-vision-250815": { + "id": "doubao-seed-1-6-vision-250815", + "name": "doubao-seed-1-6-vision-250815", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-03-17", - "last_updated": "2025-01-17", + "tool_call": true, + "temperature": true, + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.114, + "output": 1.143 } }, - "nvidia/streampetr": { - "id": "nvidia/streampetr", - "name": "streampetr", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "claude-sonnet-4-5-20250929-thinking": { + "id": "claude-sonnet-4-5-20250929-thinking", + "name": "claude-sonnet-4-5-20250929-thinking", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-03", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "video" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15 } }, - "nvidia/active-speaker-detection": { - "id": "nvidia/active-speaker-detection", - "name": "Active Speaker Detection", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "grok-4.3": { + "id": "grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "video" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, - "output": 4096 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 2.5 } }, - "nvidia/llama-3_1-nemotron-safety-guard-8b-v3": { - "id": "nvidia/llama-3_1-nemotron-safety-guard-8b-v3", - "name": "llama-3.1-nemotron-safety-guard-8b-v3", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "nemotron", + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "kimi-k2-thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -98456,51 +102816,53 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.575, + "output": 2.3 } }, - "nvidia/llama-3_2-nemoretriever-300m-embed-v1": { - "id": "nvidia/llama-3_2-nemoretriever-300m-embed-v1", - "name": "llama-3_2-nemoretriever-300m-embed-v1", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "attachment": false, + "grok-4-fast-non-reasoning": { + "id": "grok-4-fast-non-reasoning", + "name": "grok-4-fast-non-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-07-24", - "last_updated": "2025-07-24", + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 2048 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 0.5 } }, - "nvidia/nemotron-voicechat": { - "id": "nvidia/nemotron-voicechat", - "name": "nemotron-voicechat", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", + "grok-4.20-beta-0309-reasoning": { + "id": "grok-4.20-beta-0309-reasoning", + "name": "grok-4.20-beta-0309-reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, "release_date": "2026-03-16", @@ -98508,217 +102870,274 @@ "modalities": { "input": [ "text", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 6 } }, - "nvidia/nv-embedcode-7b-v1": { - "id": "nvidia/nv-embedcode-7b-v1", - "name": "nv-embedcode-7b-v1", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-03-17", - "last_updated": "2025-05-29", + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "claude-opus-4-1-20250805", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 2048 - }, + "context": 200000, + "output": 32000 + }, "cost": { - "input": 0, - "output": 0 + "input": 15, + "output": 75 } }, - "nvidia/nemotron-3-ultra-550b-a55b": { - "id": "nvidia/nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "gpt-5-pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "high" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-08", + "last_updated": "2025-10-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 272000 }, "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.15 + "input": 15, + "output": 120 } }, - "nvidia/nemotron-3-nano-30b-a3b": { - "id": "nvidia/nemotron-3-nano-30b-a3b", - "name": "nemotron-3-nano-30b-a3b", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "claude-opus-4-5-20251101", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-12", - "last_updated": "2024-12", + "knowledge": "2025-03-31", + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25 } }, - "nvidia/cosmos-predict1-5b": { - "id": "nvidia/cosmos-predict1-5b", - "name": "cosmos-predict1-5b", - "description": "Video model for prompt-guided generation, editing, and motion workflows", + "gemini-3-pro-image-preview": { + "id": "gemini-3-pro-image-preview", + "name": "gemini-3-pro-image-preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-03-18", - "last_updated": "2025-03-18", + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "video" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, - "output": 4096 + "context": 32768, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 120 } }, - "nvidia/bevformer": { - "id": "nvidia/bevformer", - "name": "bevformer", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-03-18", - "last_updated": "2025-07-20", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "video" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 2.5 } }, - "nvidia/studiovoice": { - "id": "nvidia/studiovoice", - "name": "studiovoice", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "attachment": false, + "gemini-2.0-flash-lite": { + "id": "gemini-2.0-flash-lite", + "name": "gemini-2.0-flash-lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2024-10-03", - "last_updated": "2025-06-13", + "knowledge": "2024-11", + "release_date": "2025-06-16", + "last_updated": "2025-06-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 2000000, "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.075, + "output": 0.3 } }, - "nvidia/gliner-pii": { - "id": "nvidia/gliner-pii", - "name": "gliner-pii", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "qwen3-235b-a22b": { + "id": "qwen3-235b-a22b", + "name": "Qwen3-235B-A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2025-04", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -98727,152 +103146,221 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.29, + "output": 2.86 } }, - "nvidia/nemotron-mini-4b-instruct": { - "id": "nvidia/nemotron-mini-4b-instruct", - "name": "nemotron-mini-4b-instruct", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2024-08-21", - "last_updated": "2024-08-26", + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 10, + "output": 50 } }, - "nvidia/llama-nemotron-rerank-vl-1b-v2": { - "id": "nvidia/llama-nemotron-rerank-vl-1b-v2", - "name": "llama-nemotron-rerank-vl-1b-v2", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "nemotron", + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 12 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "Nemotron 3 Super", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 2, + "output": 6 } }, - "nvidia/usdvalidate": { - "id": "nvidia/usdvalidate", - "name": "usdvalidate", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gpt-5.6-luna-pro": { + "id": "gpt-5.6-luna-pro", + "name": "gpt-5.6-luna-pro", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2024-07-24", - "last_updated": "2025-01-08", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, - "output": 4096 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 1.2 } }, - "nvidia/nemotron-3-content-safety": { - "id": "nvidia/nemotron-3-content-safety", - "name": "nemotron-3-content-safety", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": false, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": false, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -98880,68 +103368,121 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.66, + "output": 3.3 } }, - "abacusai/dracarys-llama-3_1-70b-instruct": { - "id": "abacusai/dracarys-llama-3_1-70b-instruct", - "name": "dracarys-llama-3.1-70b-instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, + "gpt-5.1": { + "id": "gpt-5.1", + "name": "gpt-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2024-09-11", - "last_updated": "2025-05-22", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 10 } }, - "deepseek-ai/deepseek-v4-flash": { - "id": "deepseek-ai/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] }, - "structured_output": true, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25 + } + }, + "qwen3.7-max-2026-06-08": { + "id": "qwen3.7-max-2026-06-08", + "name": "qwen3.7-max-2026-06-08", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -98950,46 +103491,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 393216 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 1.8, + "output": 5.3 } }, - "deepseek-ai/deepseek-v4-pro": { - "id": "deepseek-ai/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -98997,26 +103534,24 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0.283, + "output": 1.705 } }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next-80B-A3B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "kimi-k2-0905-preview": { + "id": "kimi-k2-0905-preview", + "name": "kimi-k2-0905-preview", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-01", + "knowledge": "2025-06", + "release_date": "2025-09-05", "last_updated": "2025-09-05", "modalities": { "input": [ @@ -99029,90 +103564,124 @@ "open_weights": false, "limit": { "context": 262144, - "output": 16384 + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.632, + "output": 2.53 } }, - "qwen/qwen-image-edit": { - "id": "qwen/qwen-image-edit", - "name": "Qwen Image Edit", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", + "deepseek-flash": { + "id": "deepseek-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-08-19", - "last_updated": "2025-08-19", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", "image" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 } }, - "qwen/qwen-image": { - "id": "qwen/qwen-image", - "name": "Qwen Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.5, + "output": 9 } }, - "qwen/qwen3-coder-480b-a35b-instruct": { - "id": "qwen/qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "gemini-2.5-pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -99120,61 +103689,66 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 66536 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 10 } }, - "qwen/qwen2.5-coder-32b-instruct": { - "id": "qwen/qwen2.5-coder-32b-instruct", - "name": "Qwen2.5 Coder 32b Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "attachment": false, - "reasoning": false, + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-11-06", - "last_updated": "2024-11-06", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.188, + "output": 1.133 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5-397B-A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "gemini-2.5-flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", @@ -99184,123 +103758,132 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 8192 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 2.5 } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 10, + "output": 50 } }, - "sarvamai/sarvam-m": { - "id": "sarvamai/sarvam-m", - "name": "sarvam-m", - "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", - "attachment": false, - "reasoning": false, + "doubao-seed-1-6-thinking-250715": { + "id": "doubao-seed-1-6-thinking-250715", + "name": "doubao-seed-1-6-thinking-250715", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "release_date": "2025-07-15", + "last_updated": "2025-07-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 256000, + "output": 16000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.121, + "output": 1.21 } }, - "meta/llama-3.1-8b-instruct": { - "id": "meta/llama-3.1-8b-instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "gpt-5-thinking": { + "id": "gpt-5-thinking", + "name": "gpt-5-thinking", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2024-10", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16000, - "output": 4096 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 10 } }, - "meta/llama-3.1-70b-instruct": { - "id": "meta/llama-3.1-70b-instruct", - "name": "Llama 3.1 70b Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "MiniMax-M1": { + "id": "MiniMax-M1", + "name": "MiniMax-M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-07-16", - "last_updated": "2024-07-16", + "release_date": "2025-06-16", + "last_updated": "2025-06-16", "modalities": { "input": [ "text" @@ -99309,28 +103892,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.132, + "output": 1.254 } }, - "meta/llama-3.2-1b-instruct": { - "id": "meta/llama-3.2-1b-instruct", - "name": "Llama 3.2 1b Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-18", - "last_updated": "2024-09-18", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -99341,26 +103924,39 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.2 } }, - "meta/llama-3.2-11b-vision-instruct": { - "id": "meta/llama-3.2-11b-vision-instruct", - "name": "Llama 3.2 11b Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "gpt-5.4-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-18", - "last_updated": "2024-09-18", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-19", + "last_updated": "2026-03-19", "modalities": { "input": [ "text", @@ -99370,27 +103966,37 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.75, + "output": 4.5 } }, - "meta/llama-3.3-70b-instruct": { - "id": "meta/llama-3.3-70b-instruct", - "name": "Llama 3.3 70b Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "glm-4.7": { + "id": "glm-4.7", + "name": "glm-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2024-11-26", - "last_updated": "2024-11-26", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -99401,82 +104007,73 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.286, + "output": 1.142 } }, - "meta/llama-guard-4-12b": { - "id": "meta/llama-guard-4-12b", - "name": "Llama Guard 4 12B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-05", - "last_updated": "2026-04-30", + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/esmfold": { - "id": "meta/esmfold", - "name": "esmfold", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-03-15", - "last_updated": "2025-06-12", - "modalities": { - "input": [ - "text" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 1.2 } }, - "meta/esm2-650m": { - "id": "meta/esm2-650m", - "name": "esm2-650m", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2024-08-29", - "last_updated": "2025-03-10", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -99487,57 +104084,89 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.2 } }, - "meta/llama-3.2-90b-vision-instruct": { - "id": "meta/llama-3.2-90b-vision-instruct", - "name": "Llama-3.2-90B-Vision-Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.18, + "output": 0.564 } }, - "meta/llama-4-maverick-17b-128e-instruct": { - "id": "meta/llama-4-maverick-17b-128e-instruct", - "name": "Llama 4 Maverick 17b 128e Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", + "gpt-5.2": { + "id": "gpt-5.2", + "name": "gpt-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-02", - "release_date": "2025-04-01", - "last_updated": "2025-04-01", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-12", + "last_updated": "2025-12-12", "modalities": { "input": [ "text", @@ -99547,56 +104176,29 @@ "text" ] }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/llama-3.2-3b-instruct": { - "id": "meta/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2024-09-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.75, + "output": 14 } }, - "upstage/solar-10_7b-instruct": { - "id": "upstage/solar-10_7b-instruct", - "name": "solar-10.7b-instruct", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "deepseek-v3.2-thinking": { + "id": "deepseek-v3.2-thinking", + "name": "DeepSeek-V3.2-Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2024-06-05", - "last_updated": "2025-04-10", + "knowledge": "2024-12", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -99605,208 +104207,254 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 8192 + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.29, + "output": 0.43 } }, - "black-forest-labs/flux_1-schnell": { - "id": "black-forest-labs/flux_1-schnell", - "name": "FLUX.1-schnell", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "temperature": false, - "knowledge": "2024-07", - "release_date": "2024-08-01", - "last_updated": "2026-02-04", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "image" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 77, - "input": 77, - "output": 0 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25 } }, - "black-forest-labs/flux_1-kontext-dev": { - "id": "black-forest-labs/flux_1-kontext-dev", - "name": "FLUX.1-Kontext-dev", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-08-12", - "last_updated": "2025-08-12", + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "image" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 40960 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 30 } }, - "black-forest-labs/flux.1-dev": { - "id": "black-forest-labs/flux.1-dev", - "name": "FLUX.1-dev", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "claude-sonnet-4-5-20250929", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2024-08-01", - "last_updated": "2025-09-05", + "knowledge": "2025-07-31", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 4096, - "output": 0 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15 } }, - "black-forest-labs/flux_2-klein-4b": { - "id": "black-forest-labs/flux_2-klein-4b", - "name": "FLUX.2 Klein 4B", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, + "gpt-4.1": { + "id": "gpt-4.1", + "name": "gpt-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-01-14", - "last_updated": "2026-01-31", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ - "image" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 40960 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 8 } - } - } - }, - "evroc": { - "id": "evroc", - "env": [ - "EVROC_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://models.think.evroc.com/v1", - "name": "evroc", - "doc": "https://docs.evroc.com/products/think/overview.html", - "models": { - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.4375, - "output": 5.75 + "input": 2, + "output": 10 } }, - "google/gemma-4-26B-A4B-it": { - "id": "google/gemma-4-26B-A4B-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "qwen3.5-35b-a3b": { + "id": "qwen3.5-35b-a3b", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" @@ -99815,51 +104463,25 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.144, - "output": 0.575 - } - }, - "Qwen/Qwen3-Embedding-8B": { - "id": "Qwen/Qwen3-Embedding-8B", - "name": "Qwen3 Embedding 8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2025-07-30", - "last_updated": "2025-07-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "output": 4096 + "output": 65536 }, "cost": { - "input": 0.115, - "output": 0.115 + "input": 0.06, + "output": 0.46 } }, - "Qwen/Qwen3-Reranker-4B": { - "id": "Qwen/Qwen3-Reranker-4B", - "name": "Qwen3 Reranker 4B", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "qwen3-30b-a3b": { + "id": "qwen3-30b-a3b", + "name": "Qwen3-30B-A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "release_date": "2025-07-30", - "last_updated": "2025-07-30", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -99868,104 +104490,85 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 4096 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.0575, - "output": 0 + "input": 0.11, + "output": 1.08 } }, - "Qwen/Qwen3.6-35B-A3B-FP8": { - "id": "Qwen/Qwen3.6-35B-A3B-FP8", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ "text", "image", "video", - "audio" + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.345, - "output": 1.38 - } - }, - "Qwen/Qwen3-VL-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", - "name": "Qwen3 VL 30B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "release_date": "2025-07-30", - "last_updated": "2025-07-30", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 100000, - "output": 100000 - }, - "cost": { - "input": 0.23, - "output": 0.92 + "input": 0.75, + "output": 3.75 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -99976,93 +104579,86 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.23, - "output": 0.92 + "input": 1.4, + "output": 4.4 } }, - "openai/whisper-large-v3-turbo": { - "id": "openai/whisper-large-v3-turbo", - "name": "Whisper Large v3 Turbo", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, + "gemini-2.5-flash-nothink": { + "id": "gemini-2.5-flash-nothink", + "name": "gemini-2.5-flash-nothink", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": false, - "tool_call": false, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-24", + "last_updated": "2025-06-24", "modalities": { "input": [ - "audio" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 448, - "output": 448 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.0023, - "output": 0.0023, - "output_audio": 2.3 + "input": 0.3, + "output": 2.5 } }, - "openai/whisper-large-v3": { - "id": "openai/whisper-large-v3", - "name": "Whisper 3 Large", - "description": "Open Whisper checkpoint for robust multilingual transcription and captioning", - "family": "whisper", - "attachment": false, + "doubao-seed-1-8-251215": { + "id": "doubao-seed-1-8-251215", + "name": "doubao-seed-1-8-251215", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, "reasoning": false, - "tool_call": false, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", + "tool_call": true, + "temperature": true, + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ - "audio" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 448, - "output": 4096 + "context": 224000, + "output": 64000 }, "cost": { - "input": 0.0023, - "output": 0.0023, - "output_audio": 2.3 + "input": 0.114, + "output": 0.286 } }, - "mistralai/Mistral-Medium-3.5-128B": { - "id": "mistralai/Mistral-Medium-3.5-128B", - "name": "Mistral Medium 3.5", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", + "claude-opus-4-1-20250805-thinking": { + "id": "claude-opus-4-1-20250805-thinking", + "name": "claude-opus-4-1-20250805-thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "knowledge": "2025-03", + "release_date": "2025-05-27", + "last_updated": "2025-05-27", "modalities": { "input": [ "text", @@ -100072,29 +104668,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1.725, - "output": 6.9 + "input": 15, + "output": 75 } }, - "mistralai/Voxtral-Small-24B-2507": { - "id": "mistralai/Voxtral-Small-24B-2507", - "name": "Voxtral Small 24B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "voxtral", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2025-03-01", - "last_updated": "2025-03-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "audio", "text" ], "output": [ @@ -100103,70 +104708,71 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 32000 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.0023, - "output": 0.0023, - "output_audio": 2.3 + "input": 1.4, + "output": 4.4 } }, - "nvidia/Llama-3.3-70B-Instruct-FP8": { - "id": "nvidia/Llama-3.3-70B-Instruct-FP8", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "gemini-3-flash-preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-06", + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.15, - "output": 1.15 + "input": 0.5, + "output": 3 } }, - "evroc/roc": { - "id": "evroc/roc", - "name": "roc", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "claude-haiku-4-5-20251001", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-06-06", - "last_updated": "2026-06-06", + "knowledge": "2025-02-28", + "release_date": "2025-10-16", + "last_updated": "2025-10-16", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -100174,53 +104780,54 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 64000 }, "cost": { - "input": 2.875, - "output": 11.516 + "input": 1, + "output": 5 } }, - "KBLab/kb-whisper-large": { - "id": "KBLab/kb-whisper-large", - "name": "KB Whisper", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, + "mistral-large-2512": { + "id": "mistral-large-2512", + "name": "mistral-large-2512", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "attachment": true, "reasoning": false, - "tool_call": false, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ - "audio" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 448, - "output": 448 + "context": 128000, + "output": 262144 }, "cost": { - "input": 0.0023, - "output": 0.0023, - "output_audio": 2.3 + "input": 1.1, + "output": 3.3 } }, - "intfloat/multilingual-e5-large-instruct": { - "id": "intfloat/multilingual-e5-large-instruct", - "name": "E5 Multi-Lingual Large Embeddings 0.6B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "MiniMax-M2": { + "id": "MiniMax-M2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", "attachment": false, "reasoning": false, - "tool_call": false, - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "tool_call": true, + "temperature": true, + "release_date": "2025-10-26", + "last_updated": "2025-10-26", "modalities": { "input": [ "text" @@ -100229,136 +104836,109 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512, - "output": 512 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.114, - "output": 0.114 + "input": 0.33, + "output": 1.32 } - } - } - }, - "xiaomi": { - "id": "xiaomi", - "env": [ - "XIAOMI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.xiaomimimo.com/v1", - "name": "Xiaomi", - "doc": "https://platform.xiaomimimo.com/#/docs", - "models": { - "mimo-v2.5-pro-ultraspeed": { - "id": "mimo-v2.5-pro-ultraspeed", - "name": "MiMo-V2.5-Pro-UltraSpeed", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + }, + "gemini-2.5-flash-lite-preview-09-2025": { + "id": "gemini-2.5-flash-lite-preview-09-2025", + "name": "gemini-2.5-flash-lite-preview-09-2025", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-06-08", - "last_updated": "2026-06-09", + "knowledge": "2025-01", + "release_date": "2025-09-26", + "last_updated": "2025-09-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 65536 }, - "status": "beta", "cost": { - "input": 1.305, - "output": 2.61, - "cache_read": 0.0108 + "input": 0.1, + "output": 0.4 } }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "gpt-4.1-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-06-24", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.4, + "output": 1.6 } }, - "mimo-v2-omni": { - "id": "mimo-v2-omni", - "name": "MiMo-V2-Omni", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo", + "gpt-5.6-terra-pro": { + "id": "gpt-5.6-terra-pro", + "name": "gpt-5.6-terra-pro", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-06-24", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -100367,79 +104947,74 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 2, + "output": 12 } }, - "mimo-v2-flash": { - "id": "mimo-v2-flash", - "name": "MiMo-V2-Flash", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo", - "attachment": false, + "gemini-3.8-flash": { + "id": "gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12-01", - "release_date": "2025-12-16", - "last_updated": "2026-06-24", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1048576, "output": 65536 }, - "status": "deprecated", "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.75, + "output": 3.75 } }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "ministral-14b-2512": { + "id": "ministral-14b-2512", + "name": "ministral-14b-2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-06-24", + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -100447,21 +105022,19 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "input": 0.33, + "output": 0.33 } }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "glm-5-turbo": { + "id": "glm-5-turbo", + "name": "glm-5-turbo", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -100473,10 +105046,10 @@ "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-06-24", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text" @@ -100485,43 +105058,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, + "context": 200000, "output": 131072 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "input": 0.72, + "output": 3.2 } - } - } - }, - "inception": { - "id": "inception", - "env": [ - "INCEPTION_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.inceptionlabs.ai/v1/", - "name": "Inception", - "doc": "https://platform.inceptionlabs.ai/docs", - "models": { - "mercury-edit-2": { - "id": "mercury-edit-2", - "name": "Mercury Edit 2", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + }, + "grok-4.1": { + "id": "grok-4.1", + "name": "grok-4.1", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "attachment": true, + "reasoning": false, + "tool_call": true, "temperature": true, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", + "knowledge": "2025-06", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -100529,41 +105090,41 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.025 + "input": 2, + "output": 10 } }, - "mercury-2": { - "id": "mercury-2", - "name": "Mercury 2", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", - "attachment": false, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "claude-sonnet-4-6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "knowledge": "2025-08-31", + "release_date": "2026-02-18", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -100571,58 +105132,31 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 50000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.025 + "input": 3, + "output": 15 } - } - } - }, - "anthropic": { - "id": "anthropic", - "env": [ - "ANTHROPIC_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "name": "Anthropic", - "doc": "https://docs.anthropic.com/en/docs/about-claude/models", - "models": { - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + }, + "gpt-5.3-chat-latest": { + "id": "gpt-5.3-chat-latest", + "name": "GPT-5.3 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "knowledge": "2025-08-31", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -100630,40 +105164,39 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.75, + "output": 14 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -100671,35 +105204,40 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.3, + "output": 1.8 } }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -100712,41 +105250,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 2, + "output": 12 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "deepseek-v3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2024-12", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -100754,40 +105280,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.29, + "output": 0.43 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", + "claude-opus-4-7-thinking": { + "id": "claude-opus-4-7-thinking", + "name": "claude-opus-4-7-thinking", "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": false, "knowledge": "2026-01-31", - "release_date": "2026-04-14", + "release_date": "2026-04-16", "last_updated": "2026-04-16", "modalities": { "input": [ @@ -100804,61 +105316,34 @@ "context": 1000000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } - }, "cost": { "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "output": 25 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "o3": { + "id": "o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-29", - "last_updated": "2026-06-30", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", @@ -100871,73 +105356,68 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "output": 8 } }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-01", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.4, + "output": 4.4 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "claude-opus-4-7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -100950,11 +105430,10 @@ } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2026-01-31", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", @@ -100970,104 +105449,96 @@ "context": 1000000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } - }, "cost": { "input": 5, "output": 25, "cache_read": 0.5, - "cache_write": 6.25 + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 262144, + "output": 262144 }, - "status": "deprecated", "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.95, + "output": 4 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "gpt-5": { + "id": "gpt-5", + "name": "gpt-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "release_date": "2026-06-07", - "last_updated": "2026-06-09", + "knowledge": "2024-09-30", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -101075,40 +105546,43 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 1.25, + "output": 10 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "grok-4.6": { + "id": "grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -101116,48 +105590,44 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 500000, + "output": 500000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 2, + "output": 6 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "gemini-3.1-flash-lite-preview": { + "id": "gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-04", - "last_updated": "2026-03-13", + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -101166,60 +105636,68 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 65536 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "cost": { + "input": 0.25, + "output": 1.5 + } + }, + "qwen3-max-2025-09-23": { + "id": "qwen3-max-2025-09-23", + "name": "qwen3-max-2025-09-23", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-24", + "last_updated": "2025-09-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 258048, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.86, + "output": 3.43 } }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -101231,40 +105709,36 @@ "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.285, + "output": 1.15 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", + "xhigh", "max" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -101277,116 +105751,86 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - } - } - }, - "tencent-coding-plan": { - "id": "tencent-coding-plan", - "env": [ - "TENCENT_CODING_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.lkeap.cloud.tencent.com/coding/v3", - "name": "Tencent Coding Plan (China)", - "doc": "https://cloud.tencent.com/document/product/1772/128947", - "models": { - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 5, + "output": 30 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 1.5 } - }, - "hunyuan-turbos": { - "id": "hunyuan-turbos", - "name": "Hunyuan-TurboS", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", + } + } + }, + "openrouter": { + "id": "openrouter", + "env": [ + "OPENROUTER_API_KEY" + ], + "npm": "@openrouter/ai-sdk-provider", + "api": "https://openrouter.ai/api/v1", + "name": "OpenRouter", + "doc": "https://openrouter.ai/models", + "models": { + "sao10k/l3-lunaris-8b": { + "id": "sao10k/l3-lunaris-8b", + "name": "Llama 3 8B Lunaris", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-03-08", - "last_updated": "2026-03-08", + "knowledge": "2023-12-31", + "release_date": "2024-08-13", + "last_updated": "2024-08-13", "modalities": { "input": [ "text" @@ -101395,33 +105839,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 8192, + "output": 7372 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.04, + "output": 0.05 } }, - "hunyuan-t1": { - "id": "hunyuan-t1", - "name": "Hunyuan-T1", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", + "sao10k/l3.3-euryale-70b": { + "id": "sao10k/l3.3-euryale-70b", + "name": "Llama 3.3 Euryale 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-03-08", - "last_updated": "2026-03-08", + "knowledge": "2023-12-31", + "release_date": "2024-12-18", + "last_updated": "2024-12-18", "modalities": { "input": [ "text" @@ -101430,29 +105870,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.65, + "output": 0.75 } }, - "tc-code-latest": { - "id": "tc-code-latest", - "name": "Auto", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", + "sao10k/l3.1-euryale-70b": { + "id": "sao10k/l3.1-euryale-70b", + "name": "Llama 3.1 Euryale 70B v2.2", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-08", - "last_updated": "2026-03-08", + "knowledge": "2023-12-31", + "release_date": "2024-08-28", + "last_updated": "2024-08-28", "modalities": { "input": [ "text" @@ -101461,37 +105901,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.85, + "output": 0.85 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "kwaipilot/kat-coder-pro-v2.5": { + "id": "kwaipilot/kat-coder-pro-v2.5", + "name": "KAT-Coder-Pro V2.5", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "kat-coder", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "release_date": "2026-07-10", + "last_updated": "2026-07-10", "modalities": { "input": [ "text" @@ -101502,30 +105933,37 @@ }, "open_weights": false, "limit": { - "context": 202752, - "output": 16384 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.74, + "output": 2.96, + "cache_read": 0.15 } }, - "hunyuan-2.0-instruct": { - "id": "hunyuan-2.0-instruct", - "name": "Tencent HY 2.0 Instruct", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", - "attachment": false, - "reasoning": false, + "bytedance-seed/seed-1.6-flash": { + "id": "bytedance-seed/seed-1.6-flash", + "name": "Seed 1.6 Flash", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-08", - "last_updated": "2026-03-08", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" @@ -101533,34 +105971,55 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.075, + "output": 0.3, + "tiers": [ + { + "input": 0.1, + "output": 0.8, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "hunyuan-2.0-thinking": { - "id": "hunyuan-2.0-thinking", - "name": "Tencent HY 2.0 Think", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", - "attachment": false, + "bytedance-seed/seed-2.0-lite": { + "id": "bytedance-seed/seed-2.0-lite", + "name": "Seed 2.0 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-03-08", - "last_updated": "2026-03-08", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -101568,51 +106027,46 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 2, + "tiers": [ + { + "input": 0.5, + "output": 4, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } - } - } - }, - "freemodel": { - "id": "freemodel", - "env": [ - "FREEMODEL_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://cc.freemodel.dev/v1", - "name": "FreeModel", - "doc": "https://freemodel.dev", - "models": { - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + }, + "bytedance-seed/seed-2-1-turbo": { + "id": "bytedance-seed/seed-2-1-turbo", + "name": "Seed 2.1 Turbo", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -101620,45 +106074,45 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.5, + "output": 2.5 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "bytedance-seed/seed-2.0-mini": { + "id": "bytedance-seed/seed-2.0-mini", + "name": "Seed 2.0 Mini", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -101666,45 +106120,54 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.1, + "output": 0.4, + "tiers": [ + { + "input": 0.2, + "output": 0.8, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "bytedance-seed/seed-2.0-code": { + "id": "bytedance-seed/seed-2.0-code", + "name": "Seed 2.0 Code", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -101712,46 +106175,46 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.5, + "output": 3, + "tiers": [ + { + "input": 1, + "output": 6, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "bytedance-seed/seed-1.6": { + "id": "bytedance-seed/seed-1.6", + "name": "Seed 1.6", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", "image", - "pdf" + "text", + "video" ], "output": [ "text" @@ -101759,50 +106222,54 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.freemodel.dev/v1" + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175, - "cache_write": 1.75 + "input": 0.25, + "output": 2, + "tiers": [ + { + "input": 0.5, + "output": 4, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "~moonshotai/kimi-latest": { + "id": "~moonshotai/kimi-latest", + "name": "Kimi Latest", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -101810,242 +106277,217 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 1.4989, + "output": 10.758, + "cache_read": 0.3 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, + "poolside/laguna-s-2.1:free": { + "id": "poolside/laguna-s-2.1:free", + "name": "Laguna S 2.1 (free)", + "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", + "family": "laguna-s", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.freemodel.dev/v1" + "context": 262144, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 2.5 + "input": 0, + "output": 0 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, + "poolside/laguna-s-2.1": { + "id": "poolside/laguna-s-2.1", + "name": "Laguna S 2.1", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "laguna-s", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.freemodel.dev/v1" + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075, - "cache_write": 0.75 + "input": 0.09, + "output": 0.18, + "cache_read": 0.009 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "poolside/laguna-xs-2.1:free": { + "id": "poolside/laguna-xs-2.1:free", + "name": "Laguna XS 2.1 (free)", + "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", + "family": "laguna", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-02", + "last_updated": "2026-07-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "poolside/laguna-xs-2.1": { + "id": "poolside/laguna-xs-2.1", + "name": "Laguna XS 2.1", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", + "attachment": false, + "reasoning": true, + "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-07-02", + "last_updated": "2026-07-02", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.06, + "output": 0.12, + "cache_read": 0.03 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "prism-ml/ternary-bonsai-2-27b": { + "id": "prism-ml/ternary-bonsai-2-27b", + "name": "Ternary Bonsai 2 27B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", "medium", - "high", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-09-18", + "last_updated": "2026-09-18", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.075, + "output": 0.5 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "structured_output": false, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -102058,57 +106500,34 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.freemodel.dev/v1" + "context": 200000, + "output": 32000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 5 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } - } - } - }, - "sap-ai-core": { - "id": "sap-ai-core", - "env": [ - "AICORE_SERVICE_KEY" - ], - "npm": "@jerome-benoit/sap-ai-provider-v2", - "name": "SAP AI Core", - "doc": "https://help.sap.com/docs/sap-ai-core", - "models": { - "anthropic--claude-4.8-opus": { - "id": "anthropic--claude-4.8-opus", - "name": "anthropic--claude-4.8-opus", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + }, + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", @@ -102121,8 +106540,8 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { "input": 5, @@ -102131,36 +106550,40 @@ "cache_write": 6.25 } }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "gemini-3.1-flash-lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -102169,43 +106592,62 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "anthropic--claude-4.6-sonnet": { - "id": "anthropic--claude-4.6-sonnet", - "name": "anthropic--claude-4.6-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", "high", + "xhigh", "max" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-08", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -102219,32 +106661,50 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "anthropic--claude-3-sonnet": { - "id": "anthropic--claude-3-sonnet", - "name": "anthropic--claude-3-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "anthropic/claude-3-haiku": { + "id": "anthropic/claude-3-haiku", + "name": "Claude 3 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, "knowledge": "2023-08-31", - "release_date": "2024-03-04", - "last_updated": "2024-03-04", + "release_date": "2024-03-13", + "last_updated": "2024-03-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -102256,30 +106716,40 @@ "output": 4096 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.25, + "output": 1.25, + "cache_read": 0.03, + "cache_write": 0.3 } }, - "anthropic--claude-4-sonnet": { - "id": "anthropic--claude-4-sonnet", - "name": "anthropic--claude-4-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -102292,42 +106762,38 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "gemini-2.5-pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-25", - "last_updated": "2025-06-05", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -102336,43 +106802,49 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "gpt-5": { - "id": "gpt-5", - "name": "gpt-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -102380,45 +106852,45 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "gemini-2.5-flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "anthropic/claude-fable-5.1": { + "id": "anthropic/claude-fable-5.1", + "name": "Claude Fable 5.1", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-04-17", - "last_updated": "2025-06-05", + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -102427,46 +106899,38 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "gemini-3.5-flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -102475,34 +106939,59 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "anthropic--claude-4.5-haiku": { - "id": "anthropic--claude-4.5-haiku", - "name": "anthropic--claude-4.5-haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", @@ -102515,32 +107004,38 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "anthropic--claude-3-haiku": { - "id": "anthropic--claude-3-haiku", - "name": "anthropic--claude-3-haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-03-13", - "last_updated": "2024-03-13", + "knowledge": "2025-01-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ @@ -102550,33 +107045,61 @@ "open_weights": false, "limit": { "context": 200000, - "output": 4096 + "output": 64000 }, "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.03, - "cache_write": 0.3 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "anthropic--claude-4-opus": { - "id": "anthropic--claude-4-opus", - "name": "anthropic--claude-4-opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", @@ -102589,34 +107112,43 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "anthropic--claude-4.5-sonnet": { - "id": "anthropic--claude-4.5-sonnet", - "name": "anthropic--claude-4.5-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -102629,28 +107161,59 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { "input": 3, "output": 15, "cache_read": 0.3, - "cache_write": 3.75 + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "anthropic--claude-3.5-sonnet": { - "id": "anthropic--claude-3.5-sonnet", - "name": "anthropic--claude-3.5-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "anthropic/claude-opus-5.5": { + "id": "anthropic/claude-opus-5.5", + "name": "Claude Opus 5.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", @@ -102663,243 +107226,234 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 } }, - "anthropic--claude-4.6-opus": { - "id": "anthropic--claude-4.6-opus", - "name": "anthropic--claude-4.6-opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "nex-agi/nex-n2.5-mini": { + "id": "nex-agi/nex-n2.5-mini", + "name": "Nex-N2.5-Mini", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "agi", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", + "none", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.025, + "output": 0.1, + "cache_read": 0.0025 } }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "gemini-2.5-flash-lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "nex-agi/nex-n2.5-mini:free": { + "id": "nex-agi/nex-n2.5-mini:free", + "name": "Nex-N2.5-Mini (free)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "agi", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 + "type": "effort", + "values": [ + "none", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 + "input": 0, + "output": 0 } }, - "anthropic--claude-3.7-sonnet": { - "id": "anthropic--claude-3.7-sonnet", - "name": "anthropic--claude-3.7-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "nex-agi/nex-n2.5-pro:free": { + "id": "nex-agi/nex-n2.5-pro:free", + "name": "Nex-N2.5-Pro (free)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "agi", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "none", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2025-02-24", - "last_updated": "2025-02-24", + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0 } }, - "sonar": { - "id": "sonar", - "name": "sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar", - "attachment": false, - "reasoning": false, - "tool_call": false, + "nex-agi/nex-n2.5-pro": { + "id": "nex-agi/nex-n2.5-pro", + "name": "Nex-N2.5-Pro", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "agi", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 235929 }, "cost": { - "input": 1, - "output": 1 + "input": 0.075, + "output": 0.25, + "cache_read": 0.015 } }, - "sonar-pro": { - "id": "sonar-pro", - "name": "sonar-pro", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", - "attachment": true, + "cohere/command-r-08-2024": { + "id": "cohere/command-r-08-2024", + "name": "Command R", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 + "context": 128000, + "output": 4000 }, "cost": { - "input": 3, - "output": 15 + "input": 0.15, + "output": 0.6 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "gpt-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "cohere/command-a-plus": { + "id": "cohere/command-a-plus", + "name": "Command A+", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -102907,186 +107461,161 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 192000, + "output": 64000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 0.3, + "output": 1.5, + "cache_read": 0.15 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "gpt-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "cohere/command-a": { + "id": "cohere/command-a", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2024-08-31", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 256000, + "output": 8192 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 2.5, + "output": 10 } }, - "anthropic--claude-3-opus": { - "id": "anthropic--claude-3-opus", - "name": "anthropic--claude-3-opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": false, + "cohere/north-mini-code:free": { + "id": "cohere/north-mini-code:free", + "name": "North Mini Code (free)", + "description": "Cohere coding model for practical software engineering and agentic edits", + "family": "north", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-02-29", - "last_updated": "2024-02-29", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 4096 + "context": 256000, + "output": 64000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0, + "output": 0 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "gpt-5-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "cohere/command-r7b-12-2024": { + "id": "cohere/command-r7b-12-2024", + "name": "Command R7B", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2024-12-02", + "last_updated": "2024-12-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.0375, + "output": 0.15 } }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "gpt-4.1-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "cohere/command-r-plus-08-2024": { + "id": "cohere/command-r-plus-08-2024", + "name": "Command R+", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 2.5, + "output": 10 } }, - "sonar-deep-research": { - "id": "sonar-deep-research", - "name": "sonar-deep-research", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar-deep-research", + "deepseek/deepseek-chat-v3.1": { + "id": "deepseek/deepseek-chat-v3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-02-01", - "last_updated": "2025-09-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ "text" @@ -103095,231 +107624,207 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 163840, "output": 32768 }, "cost": { - "input": 2, - "output": 8, - "reasoning": 3 + "input": 0.25, + "output": 0.95, + "cache_read": 0.13 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "gpt-5-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, + "deepseek/deepseek-v4-flash-0731": { + "id": "deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1310720, + "output": 943718 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 0.04, + "output": 0.64, + "cache_read": 0.016 } }, - "anthropic--claude-4.5-opus": { - "id": "anthropic--claude-4.5-opus", - "name": "anthropic--claude-4.5-opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "deepseek/deepseek-v4.1-flash": { + "id": "deepseek/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01 } }, - "anthropic--claude-4.7-opus": { - "id": "anthropic--claude-4.7-opus", - "name": "anthropic--claude-4.7-opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "deepseek/deepseek-v4-flash-vision-exp": { + "id": "deepseek/deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "gpt-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, + "deepseek/deepseek-r1-distill-llama-70b": { + "id": "deepseek/deepseek-r1-distill-llama-70b", + "name": "R1 Distill Llama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-07-31", + "release_date": "2025-01-23", + "last_updated": "2025-01-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 8192, + "output": 7372 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.8, + "output": 0.8 } - } - } - }, - "opencode": { - "id": "opencode", - "env": [ - "OPENCODE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://opencode.ai/zen/v1", - "name": "OpenCode Zen", - "doc": "https://opencode.ai/docs/zen", - "models": { - "ring-2.6-1t-free": { - "id": "ring-2.6-1t-free", - "name": "Ring 2.6 1T Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "ring-1t-free", + }, + "deepseek/deepseek-v3.1-terminus": { + "id": "deepseek/deepseek-v3.1-terminus", + "name": "DeepSeek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-05-08", - "last_updated": "2026-05-08", + "knowledge": "2025-03-31", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ "text" @@ -103330,31 +107835,28 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 66000 + "context": 163840, + "output": 32768 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0 + "input": 0.27, + "output": 1, + "cache_read": 0.135 } }, - "mimo-v2-pro-free": { - "id": "mimo-v2-pro-free", - "name": "MiMo V2 Pro Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo-pro-free", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "deepseek/deepseek-chat": { + "id": "deepseek/deepseek-chat", + "name": "DeepSeek Chat", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ "text" @@ -103365,21 +107867,19 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 64000 + "context": 163840, + "output": 16384 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.32, + "output": 0.89 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "deepseek/deepseek-v4-pro-0813": { + "id": "deepseek/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -103389,20 +107889,17 @@ { "type": "effort", "values": [ + "low", "high", "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -103413,55 +107910,20 @@ }, "open_weights": true, "limit": { - "context": 1000000, + "context": 1048576, "output": 384000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 - } - }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.462, + "output": 1.386, + "cache_read": 0.0154 } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Legacy model retained for compatibility with older integrations", - "family": "glm", + "deepseek/deepseek-v3.2-exp": { + "id": "deepseek/deepseek-v3.2-exp", + "name": "DeepSeek V3.2 Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -103470,13 +107932,11 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text" @@ -103487,38 +107947,30 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 163840, + "output": 65536 }, - "status": "deprecated", "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.1 + "input": 0.27, + "output": 0.41 } }, - "mimo-v2.5-free": { - "id": "mimo-v2.5-free", - "name": "MiMo V2.5 Free", - "description": "MiMo omni model for text, image, video, audio, and agents", - "family": "mimo-v2.5-free", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "deepseek/deepseek-chat-v3-0324": { + "id": "deepseek/deepseek-chat-v3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-07-31", + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -103526,27 +107978,28 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 163840, + "output": 147456 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.25, + "output": 1 } }, - "kimi-k2": { - "id": "kimi-k2", - "name": "Kimi K2", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-k2", + "deepseek/deepseek-r1": { + "id": "deepseek/deepseek-r1", + "name": "DeepSeek-R1", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" @@ -103557,32 +108010,42 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 64000, + "output": 16000 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 2.5, - "cache_read": 0.4 + "input": 0.7, + "output": 2.5 } }, - "minimax-m2.1": { - "id": "minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -103593,32 +108056,29 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 1048576, + "output": 384000 }, - "status": "deprecated", "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.1 + "input": 0.951432, + "output": 1.902864, + "cache_read": 0.079286 } }, - "nemotron-3-ultra-free": { - "id": "nemotron-3-ultra-free", - "name": "Nemotron 3 Ultra Free", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron-free", + "deepseek/deepseek-r1-0528": { + "id": "deepseek/deepseek-r1-0528", + "name": "R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2025-03-31", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -103629,20 +108089,20 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 163840, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.5, + "output": 2.15, + "cache_read": 0.35 } }, - "glm-4.7-free": { - "id": "glm-4.7-free", - "name": "GLM-4.7 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "glm-free", + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -103651,13 +108111,11 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -103668,94 +108126,115 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 163840, + "output": 65536 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.269, + "output": 0.4, + "cache_read": 0.1345 } }, - "gemini-3-flash": { - "id": "gemini-3-flash", - "name": "Gemini 3 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", - "low", - "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 384000 }, - "provider": { - "npm": "@ai-sdk/google" + "cost": { + "input": 0.08246, + "output": 0.16492, + "cache_read": 0.016492 + } + }, + "unbiased/pareto": { + "id": "unbiased/pareto", + "name": "Pareto", + "description": "Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-09-17", + "last_updated": "2026-09-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 2.5, + "output": 7.5, + "cache_read": 0.25 } }, - "deepseek-v4-flash-free": { - "id": "deepseek-v4-flash-free", - "name": "DeepSeek V4 Flash Free", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash-free", + "tencent/hy3-preview": { + "id": "tencent/hy3-preview", + "name": "Hy3 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "max" + "none", + "low", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text" @@ -103766,189 +108245,196 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.18, + "output": 0.6, + "cache_read": 0.06 } }, - "claude-sonnet-4": { - "id": "claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "tencent/hunyuan-a13b-instruct": { + "id": "tencent/hunyuan-a13b-instruct", + "name": "Hunyuan A13B Instruct", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2025-07-08", + "last_updated": "2025-07-08", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 131072, + "output": 117964 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 0.14, + "output": 0.57 } }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, + "tencent/hy-mt2-30b-a3b": { + "id": "tencent/hy-mt2-30b-a3b", + "name": "Hy-MT2-30B-A3B", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2026-08-20", + "last_updated": "2026-08-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 8192, + "output": 4096 }, - "provider": { - "npm": "@ai-sdk/anthropic" + "cost": { + "input": 0.074, + "output": 0.295 + } + }, + "tencent/hy-mt2-1.8b": { + "id": "tencent/hy-mt2-1.8b", + "name": "Hy-MT2-1.8B", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-20", + "last_updated": "2026-08-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 4096 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.044, + "output": 0.177 } }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "tencent/hy3": { + "id": "tencent/hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", - "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, + "context": 262144, + "input": 192000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai" + "cost": { + "input": 0.0825, + "output": 0.33, + "cache_read": 0.020625 + } + }, + "tencent/hy-mt2-7b": { + "id": "tencent/hy-mt2-7b", + "name": "Hy-MT2-7B", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-19", + "last_updated": "2026-08-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 4096 }, "cost": { - "input": 1.07, - "output": 8.5, - "cache_read": 0.107 + "input": 0.074, + "output": 0.295 } }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "tencent/hy4-preview": { + "id": "tencent/hy4-preview", + "name": "Hy4 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", - "medium", "high" ] } @@ -103956,58 +108442,42 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-08-28", + "last_updated": "2026-08-28", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google" + "output": 64000 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 + "input": 0.834, + "output": 2.501, + "cache_read": 0.042 } }, - "minimax-m3-free": { - "id": "minimax-m3-free", - "name": "MiniMax-M3 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax-m3-free", + "anthracite-org/magnum-v4-72b": { + "id": "anthracite-org/magnum-v4-72b", + "name": "Magnum v4 72B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-31", - "last_updated": "2026-05-31", + "knowledge": "2024-06-30", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -104015,109 +108485,63 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 32768, + "output": 4096 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 2.5, + "output": 5 } }, - "gemini-3-pro": { - "id": "gemini-3-pro", - "name": "Gemini 3 Pro", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gemini-pro", + "meta-llama/llama-4-scout": { + "id": "meta-llama/llama-4-scout", + "name": "Llama 4 Scout", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "knowledge": "2024-08-31", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/google" + "context": 1310720, + "output": 16384 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.1, + "output": 0.3 } }, - "kimi-k2.5-free": { - "id": "kimi-k2.5-free", - "name": "Kimi K2.5 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-free", + "meta-llama/llama-guard-4-12b": { + "id": "meta-llama/llama-guard-4-12b", + "name": "Llama Guard 4 12B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "knowledge": "2024-08-31", + "release_date": "2025-04-30", + "last_updated": "2025-04-30", "modalities": { "input": [ - "text", "image", - "video" + "text" ], "output": [ "text" @@ -104125,38 +108549,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 163840, + "output": 16384 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.18, + "output": 0.18 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "meta-llama/llama-4-maverick": { + "id": "meta-llama/llama-4-maverick", + "name": "Llama 4 Maverick", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2024-08-31", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -104164,35 +108581,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 16384 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.1875, + "output": 0.6525 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "meta-llama/llama-3.3-70b-instruct": { + "id": "meta-llama/llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -104203,43 +108612,27 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 16384 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.1, + "output": 0.32 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "meta-llama/llama-3.1-8b-instruct": { + "id": "meta-llama/llama-3.1-8b-instruct", + "name": "Llama-3.1-8B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -104250,32 +108643,28 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 131072, + "output": 117964 }, "cost": { - "input": 1.74, - "output": 3.84, - "cache_read": 0.145 + "input": 0.05, + "output": 0.08, + "cache_read": 0.025 } }, - "glm-4.6": { - "id": "glm-4.6", - "name": "GLM-4.6", - "description": "Legacy model retained for compatibility with older integrations", - "family": "glm", + "meta-llama/llama-3.2-1b-instruct": { + "id": "meta-llama/llama-3.2-1b-instruct", + "name": "Llama 3.2 1B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2023-12-31", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text" @@ -104286,32 +108675,27 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 60000, + "output": 54000 }, - "status": "deprecated", "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.1 + "input": 0.027, + "output": 0.201 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-thinking", + "meta-llama/llama-3.2-3b-instruct": { + "id": "meta-llama/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "knowledge": "2023-12-31", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text" @@ -104322,83 +108706,50 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 117964 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 2.5, - "cache_read": 0.4 + "input": 0.05, + "output": 0.33 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "meta-llama/llama-3.1-70b-instruct": { + "id": "meta-llama/llama-3.1-70b-instruct", + "name": "Llama-3.1-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 131072, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 0.4, + "output": 0.4 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "~google/gemini-flash-latest": { + "id": "~google/gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -104407,22 +108758,23 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "pdf", + "audio" ], "output": [ "text" @@ -104430,24 +108782,22 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "~google/gemini-pro-latest": { + "id": "~google/gemini-pro-latest", + "name": "Gemini Pro Latest", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -104456,22 +108806,23 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text", + "audio", + "pdf", "image", - "pdf" + "text", + "video" ], "output": [ "text" @@ -104479,49 +108830,54 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 1048576, + "output": 65536 }, "cost": { "input": 2, - "output": 10, + "output": 12, + "reasoning": 12, "cache_read": 0.2, - "cache_write": 2.5 + "cache_write": 0.375, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "perceptron/perceptron-mk1": { + "id": "perceptron/perceptron-mk1", + "name": "Perceptron Mk1", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], - "tool_call": true, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -104529,34 +108885,37 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.15, + "output": 1.5 } }, - "qwen3-coder": { - "id": "qwen3-coder", - "name": "Qwen3 Coder", - "description": "Legacy model retained for compatibility with older integrations", - "family": "qwen", - "attachment": false, - "reasoning": false, + "z-ai/glm-4.6v": { + "id": "z-ai/glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -104564,39 +108923,33 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "output": 32768 }, - "status": "deprecated", "cost": { - "input": 0.45, - "output": 1.8 + "input": 0.3, + "output": 0.9, + "cache_read": 0.055 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "z-ai/glm-4.5": { + "id": "z-ai/glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -104607,63 +108960,75 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "output": 98304 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "hy3-preview-free": { - "id": "hy3-preview-free", - "name": "Hy3 preview Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "hy3-free", - "attachment": false, + "z-ai/glm-5v-turbo": { + "id": "z-ai/glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 202752, + "output": 131072 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "z-ai/glm-5.3-flash": { + "id": "z-ai/glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", @@ -104676,84 +109041,70 @@ }, "open_weights": true, "limit": { - "context": 512000, - "output": 128000 + "context": 1310720, + "output": 943718 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.15, + "output": 0.5, + "cache_read": 0.05 } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 204800, + "output": 16384 }, "cost": { - "input": 1.07, - "output": 8.5, - "cache_read": 0.107 + "input": 0.43, + "output": 1.75, + "cache_read": 0.08 } }, - "gpt-5.3-codex-spark": { - "id": "gpt-5.3-codex-spark", - "name": "GPT-5.3 Codex Spark", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex-spark", + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", + "temperature": true, "release_date": "2026-02-12", "last_updated": "2026-02-12", "modalities": { @@ -104764,45 +109115,35 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, + "context": 204800, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai" - }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.6, + "output": 1.92, + "cache_read": 0.12 } }, - "gpt-5.1-codex-max": { - "id": "gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "z-ai/glm-4.5v": { + "id": "z-ai/glm-4.5v", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ "text", @@ -104812,26 +109153,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 65536, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.6, + "output": 1.8, + "cache_read": 0.11 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "z-ai/glm-5.3-flashx": { + "id": "z-ai/glm-5.3-flashx", + "name": "GLM 5.3 FlashX", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -104839,23 +109176,21 @@ "type": "effort", "values": [ "low", - "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": false, + "temperature": true, + "release_date": "2026-09-18", + "last_updated": "2026-09-18", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -104863,249 +109198,160 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 1048576, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.37, + "output": 1.25, + "cache_read": 0.075 } }, - "ling-2.6-flash-free": { - "id": "ling-2.6-flash-free", - "name": "Ling 2.6 Flash Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "ling-flash-free", + "z-ai/glm-4.7-flash": { + "id": "z-ai/glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262100, - "output": 32800 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.5", - "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_details" + }, + "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "video" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 - } - }, - "claude-3-5-haiku": { - "id": "claude-3-5-haiku", - "name": "Claude Haiku 3.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-haiku", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 200000, - "output": 8192 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" + "output": 117964 }, "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "input": 0.0605, + "output": 0.4 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_details" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.4, + "output": 1.75, + "cache_read": 0.08 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", "high", "xhigh" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 1048576, + "output": 131072 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.6496, + "output": 2.0416, + "cache_read": 0.12064 } }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -105117,40 +109363,37 @@ "open_weights": true, "limit": { "context": 204800, - "output": 131072 + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.966, + "output": 3.036, + "cache_read": 0.1794 } }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's latest Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -105158,56 +109401,39 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "north-mini-code-free": { - "id": "north-mini-code-free", - "name": "North Mini Code Free", - "description": "Cohere coding model for practical software engineering and agentic edits", - "family": "north-free", + "z-ai/glm-5.2:free": { + "id": "z-ai/glm-5.2:free", + "name": "GLM 5.2 (free)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "high" + "high", + "xhigh" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-09-23", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -105218,26 +109444,32 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 32768, + "output": 29491 }, "cost": { "input": 0, "output": 0 } }, - "grok-code": { - "id": "grok-code", - "name": "Grok Code Fast 1", - "description": "Legacy model retained for compatibility with older integrations", - "family": "grok", - "attachment": true, + "z-ai/glm-4.5-air": { + "id": "z-ai/glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-08-20", - "last_updated": "2025-08-20", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -105246,134 +109478,102 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 131072, + "output": 98304 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.13, + "output": 0.85, + "cache_read": 0.025 } }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "z-ai/glm-5.3": { + "id": "z-ai/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 1310720, + "output": 131072 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.84, + "output": 2.64, + "cache_read": 0.156 } }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "inference-net/schematron-v2-small": { + "id": "inference-net/schematron-v2-small", + "name": "Schematron V2 Small", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2026-09-12", + "last_updated": "2026-09-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.05, + "output": 0.23, + "cache_read": 0.05 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "inference-net/schematron-v2-turbo": { + "id": "inference-net/schematron-v2-turbo", + "name": "Schematron V2 Turbo", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-09-12", + "last_updated": "2026-09-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -105381,79 +109581,45 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.08 + "input": 0.03, + "output": 0.15, + "cache_read": 0.03 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "thinkingmachines/inkling-small:free": { + "id": "thinkingmachines/inkling-small:free", + "name": "Inkling Small (free)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "hy3-free": { - "id": "hy3-free", - "name": "Hy3 Free", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hy3-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-26", - "last_updated": "2026-06-26", - "modalities": { - "input": [ - "text" + "audio" ], "output": [ "text" @@ -105461,63 +109627,65 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 1048576, + "output": 262144 }, "cost": { "input": 0, - "output": 0, - "cache_read": 0 + "output": 0 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "thinkingmachines/inkling-small": { + "id": "thinkingmachines/inkling-small", + "name": "Inkling Small", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 1048576, + "output": 262144 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.45, + "output": 1.2, + "cache_read": 0.1 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "thinkingmachines/inkling:free": { + "id": "thinkingmachines/inkling:free", + "name": "Inkling (free)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -105525,65 +109693,44 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 1048576, + "output": 262144 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0, + "output": 0 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "thinkingmachines/inkling": { + "id": "thinkingmachines/inkling", + "name": "Inkling", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -105591,57 +109738,60 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 1048576, + "output": 471859 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 1, + "output": 4.05, + "cache_read": 0.17 } }, - "minimax-m2.5-free": { - "id": "minimax-m2.5-free", - "name": "MiniMax-M2.5 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax-free", + "meituan/longcat-2.0": { + "id": "meituan/longcat-2.0", + "name": "LongCat 2.0", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "longcat", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-07-20", + "last_updated": "2026-07-20", "modalities": { "input": [ "text" @@ -105652,92 +109802,56 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 1048756, + "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.3, + "output": 1.2, + "cache_read": 0.006 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "openrouter/bodybuilder": { + "id": "openrouter/bodybuilder", + "name": "Body Builder (beta)", + "description": "Preview model for early access evaluation, prototyping, and compatibility testing", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": false, + "release_date": "2025-12-05", + "last_updated": "2025-12-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "context": 128000, + "output": 128000 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "openrouter/free": { + "id": "openrouter/free", + "name": "Free Models Router", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -105745,32 +109859,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 200000, + "input": 200000, + "output": 8000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0 } }, - "minimax-m2.1-free": { - "id": "minimax-m2.1-free", - "name": "MiniMax-M2.1 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax-free", + "openrouter/pareto-code": { + "id": "openrouter/pareto-code", + "name": "Pareto Code Router", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text" @@ -105779,49 +109887,26 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "context": 2000000, + "output": 200000 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "openrouter/fusion": { + "id": "openrouter/fusion", + "name": "Fusion", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -105829,100 +109914,58 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 } }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "openrouter/auto": { + "id": "openrouter/auto", + "name": "Auto Router", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "auto", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": true, + "temperature": true, + "release_date": "2023-11-08", + "last_updated": "2023-11-08", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "pdf", + "video" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 30, - "output": 180, - "cache_read": 30 + "context": 2000000, + "output": 2000000 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "perplexity/sonar-pro": { + "id": "perplexity/sonar-pro", + "name": "Sonar Pro", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "interleaved": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "release_date": "2025-03-07", + "last_updated": "2025-03-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -105930,35 +109973,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 200000, + "output": 8000 }, "cost": { "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "output": 15 } }, - "mimo-v2-flash-free": { - "id": "mimo-v2-flash-free", - "name": "MiMo V2 Flash Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo-flash-free", + "perplexity/sonar-deep-research": { + "id": "perplexity/sonar-deep-research", + "name": "Sonar Deep Research", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar-deep-research", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "release_date": "2025-03-07", + "last_updated": "2025-03-07", "modalities": { "input": [ "text" @@ -105967,48 +110006,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 115200 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 2, + "output": 8, + "reasoning": 3 } }, - "gemini-3.1-pro": { - "id": "gemini-3.1-pro", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "perplexity/sonar": { + "id": "perplexity/sonar", + "name": "Sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -106016,85 +110040,67 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google" + "context": 127072, + "output": 114364 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 1, + "output": 1 } }, - "trinity-large-preview-free": { - "id": "trinity-large-preview-free", - "name": "Trinity Large Preview", - "description": "Legacy model retained for compatibility with older integrations", - "family": "trinity", - "attachment": false, - "reasoning": false, - "tool_call": true, + "perplexity/sonar-reasoning-pro": { + "id": "perplexity/sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded reasoning model for multi-step research and cited answers", + "family": "sonar-reasoning", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-01-28", - "last_updated": "2026-01-28", + "release_date": "2025-03-07", + "last_updated": "2025-03-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 115200 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 8 } }, - "big-pickle": { - "id": "big-pickle", - "name": "Big Pickle", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "big-pickle", - "attachment": false, + "perplexity/sonar-pro-search": { + "id": "perplexity/sonar-pro-search", + "name": "Sonar Pro Search", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-10-17", - "last_updated": "2025-10-17", + "release_date": "2025-10-30", + "last_updated": "2025-10-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -106103,44 +110109,45 @@ "open_weights": false, "limit": { "context": 200000, - "input": 160000, - "output": 32000 + "output": 8000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 3, + "output": 15 } }, - "gpt-5.5-pro": { - "id": "gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "meta/muse-spark-1.3": { + "id": "meta/muse-spark-1.3", + "name": "Muse Spark 1.3", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", + "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "pdf", + "audio" ], "output": [ "text" @@ -106148,170 +110155,184 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 1048576, + "output": 943718 }, "cost": { - "input": 30, - "output": 180, - "cache_read": 30 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "qwen3.6-plus-free": { - "id": "qwen3.6-plus-free", - "name": "Qwen3.6 Plus Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "qwen-free", + "meta/muse-glimmer-30b": { + "id": "meta/muse-glimmer-30b", + "name": "Muse Glimmer 30B", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 65536 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 131072, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.3, + "output": 1.2, + "cache_read": 0.04 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "meta/muse-spark-1.1": { + "id": "meta/muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "release_date": "2026-04-08", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "mimo-v2-omni-free": { - "id": "mimo-v2-omni-free", - "name": "MiMo V2 Omni Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo-omni-free", + "meta/muse-spark-1.2": { + "id": "meta/muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ "text", "image", - "audio", - "pdf" + "video", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 64000 + "context": 1048576, + "output": 943718 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "GPT-5 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "meta/muse-spark-1.2-contributor": { + "id": "meta/muse-spark-1.2-contributor", + "name": "Muse Spark 1.2 Contributor", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-08-21", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf", + "audio" ], "output": [ "text" @@ -106319,48 +110340,47 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 1048576, + "output": 943718 }, "cost": { - "input": 1.07, - "output": 8.5, - "cache_read": 0.107 + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "meta/muse-spark-1.3-contributor": { + "id": "meta/muse-spark-1.3-contributor", + "name": "Muse Spark 1.3 Contributor", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "pdf", + "audio" ], "output": [ "text" @@ -106368,24 +110388,82 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 943718 }, - "provider": { - "npm": "@ai-sdk/openai" + "cost": { + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 + } + }, + "nousresearch/hermes-3-llama-3.1-70b": { + "id": "nousresearch/hermes-3-llama-3.1-70b", + "name": "Hermes 3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "nousresearch", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12-31", + "release_date": "2024-08-18", + "last_updated": "2024-08-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.7, + "output": 0.7 } }, - "glm-5-free": { - "id": "glm-5-free", - "name": "GLM-5 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "glm-free", + "nousresearch/hermes-3-llama-3.1-405b": { + "id": "nousresearch/hermes-3-llama-3.1-405b", + "name": "Hermes 3 405B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "nousresearch", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12-31", + "release_date": "2024-08-16", + "last_updated": "2024-08-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 1, + "output": 1 + } + }, + "nousresearch/hermes-4-405b": { + "id": "nousresearch/hermes-4-405b", + "name": "Hermes 4 405B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "hermes", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -106393,14 +110471,12 @@ "type": "toggle" } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "knowledge": "2024-08-31", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text" @@ -106411,37 +110487,36 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 117964 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1, + "output": 3 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.6", + "~z-ai/glm-flash-latest": { + "id": "~z-ai/glm-flash-latest", + "name": "GLM Flash Latest", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ "text", @@ -106454,35 +110529,69 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1310720, + "output": 131072 }, - "provider": { - "npm": "@ai-sdk/anthropic" + "cost": { + "input": 0.075, + "output": 0.25, + "cache_read": 0.015 + } + }, + "~z-ai/glm-latest": { + "id": "~z-ai/glm-latest", + "name": "GLM Latest", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-19", + "last_updated": "2026-08-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1310720, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625 + "input": 0.5614, + "output": 1.7644, + "cache_read": 0.10426 } }, - "nemotron-3-super-free": { - "id": "nemotron-3-super-free", - "name": "Nemotron 3 Super Free", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron-free", + "arcee-ai/trinity-large-thinking": { + "id": "arcee-ai/trinity-large-thinking", + "name": "Trinity Large Thinking", + "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", + "family": "trinity", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "release_date": "2026-04-01", + "last_updated": "2026-05-28", "modalities": { "input": [ "text" @@ -106493,33 +110602,46 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 128000 + "context": 262144, + "output": 80000 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.25, + "output": 0.8, + "cache_read": 0.06 } }, - "grok-build-0.1": { - "id": "grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Grok coding model for agentic engineering, edits, and codebase workflows", - "family": "grok-build", + "~openai/gpt-terra-latest": { + "id": "~openai/gpt-terra-latest", + "name": "GPT Terra Latest", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-terra", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-05-20", - "last_updated": "2026-05-20", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text", - "image" + "pdf", + "image", + "text" ], "output": [ "text" @@ -106527,20 +110649,39 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", + "~openai/gpt-luna-latest": { + "id": "~openai/gpt-luna-latest", + "name": "GPT Luna Latest", "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -106550,20 +110691,23 @@ "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2026-02-16", + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text", - "image" + "pdf", + "image", + "text" ], "output": [ "text" @@ -106571,24 +110715,39 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai" - }, "cost": { - "input": 1.07, - "output": 8.5, - "cache_read": 0.107 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "~openai/gpt-sol-latest": { + "id": "~openai/gpt-sol-latest", + "name": "GPT Sol Latest", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -106599,21 +110758,22 @@ "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "knowledge": "2026-02-16", + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text", + "pdf", "image", - "pdf" + "text" ], "output": [ "text" @@ -106622,21 +110782,19 @@ "open_weights": false, "limit": { "context": 1050000, - "input": 922000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai" - }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, "tiers": [ { - "input": 10, - "output": 45, - "cache_read": 1, + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, "tier": { "type": "context", "size": 272000 @@ -106644,97 +110802,136 @@ } ], "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 } } - } - } - }, - "inference": { - "id": "inference", - "env": [ - "INFERENCE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.net/v1", - "name": "Inference", - "doc": "https://inference.net/models", - "models": { - "mistral/mistral-nemo-12b-instruct": { - "id": "mistral/mistral-nemo-12b-instruct", - "name": "Mistral Nemo 12B Instruct", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, + }, + "~openai/gpt-astra-latest": { + "id": "~openai/gpt-astra-latest", + "name": "GPT Astra Latest", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "structured_output": true, + "temperature": false, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16000, - "output": 4096 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0.038, - "output": 0.1 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } } }, - "google/gemma-3": { - "id": "google/gemma-3", - "name": "Google Gemma 3", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "~openai/gpt-mini-latest": { + "id": "~openai/gpt-mini-latest", + "name": "GPT Mini Latest", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text", - "image" + "pdf", + "image", + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 125000, - "output": 4096 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.3 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "osmosis/osmosis-structure-0.6b": { - "id": "osmosis/osmosis-structure-0.6b", - "name": "Osmosis Structure 0.6B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "osmosis", + "cognitivecomputations/dolphin-mistral-24b-venice-edition": { + "id": "cognitivecomputations/dolphin-mistral-24b-venice-edition", + "name": "Uncensored", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2024-04-30", + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text" @@ -106745,26 +110942,26 @@ }, "open_weights": true, "limit": { - "context": 4000, - "output": 2048 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.5 + "input": 0.2, + "output": 0.9 } }, - "qwen/qwen3-embedding-4b": { - "id": "qwen/qwen3-embedding-4b", - "name": "Qwen 3 Embedding 4B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", + "thedrummer/skyfall-36b-v2": { + "id": "thedrummer/skyfall-36b-v2", + "name": "Skyfall 36B V2", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-30", + "release_date": "2025-03-10", + "last_updated": "2025-03-10", "modalities": { "input": [ "text" @@ -106775,30 +110972,30 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 2048 + "context": 32768, + "output": 29491 }, "cost": { - "input": 0.01, - "output": 0 + "input": 0.55, + "output": 0.8, + "cache_read": 0.25 } }, - "qwen/qwen-2.5-7b-vision-instruct": { - "id": "qwen/qwen-2.5-7b-vision-instruct", - "name": "Qwen 2.5 7B Vision Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "thedrummer/unslopnemo-12b": { + "id": "thedrummer/unslopnemo-12b", + "name": "UnslopNemo 12B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2024-04-30", + "release_date": "2024-11-08", + "last_updated": "2024-11-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -106806,26 +111003,26 @@ }, "open_weights": true, "limit": { - "context": 125000, - "output": 4096 + "context": 1024000, + "output": 819200 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 0.4, + "output": 0.4 } }, - "meta/llama-3.1-8b-instruct": { - "id": "meta/llama-3.1-8b-instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "thedrummer/cydonia-24b-v4.1": { + "id": "thedrummer/cydonia-24b-v4.1", + "name": "Cydonia 24B V4.1", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2024-04-30", + "release_date": "2025-09-27", + "last_updated": "2025-09-27", "modalities": { "input": [ "text" @@ -106836,28 +111033,36 @@ }, "open_weights": true, "limit": { - "context": 16000, - "output": 4096 + "context": 131072, + "output": 117964 }, "cost": { - "input": 0.025, - "output": 0.025 + "input": 0.3, + "output": 0.5, + "cache_read": 0.15 } }, - "meta/llama-3.2-1b-instruct": { - "id": "meta/llama-3.2-1b-instruct", - "name": "Llama 3.2 1B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, + "baidu/ernie-4.5-vl-424b-a47b": { + "id": "baidu/ernie-4.5-vl-424b-a47b", + "name": "ERNIE 4.5 VL 424B A47B ", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ernie", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2025-03-31", + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ + "image", "text" ], "output": [ @@ -106866,296 +111071,427 @@ }, "open_weights": true, "limit": { - "context": 16000, - "output": 4096 + "context": 123000, + "output": 16000 }, "cost": { - "input": 0.01, - "output": 0.01 + "input": 0.42, + "output": 1.25 } }, - "meta/llama-3.2-11b-vision-instruct": { - "id": "meta/llama-3.2-11b-vision-instruct", - "name": "Llama 3.2 11B Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", + "x-ai/grok-4.20-multi-agent": { + "id": "x-ai/grok-4.20-multi-agent", + "name": "Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2025-09-01", + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16000, - "output": 4096 + "context": 2000000, + "output": 1800000 }, "cost": { - "input": 0.055, - "output": 0.055 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "meta/llama-3.2-3b-instruct": { - "id": "meta/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "x-ai/grok-4.7": { + "id": "x-ai/grok-4.7", + "name": "Grok 4.7", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16000, - "output": 4096 + "context": 500000, + "output": 450000 }, "cost": { - "input": 0.02, - "output": 0.02 + "input": 1.6, + "output": 4.8, + "cache_read": 0.4, + "tiers": [ + { + "input": 3.2, + "output": 9.6, + "cache_read": 0.8, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 3.2, + "output": 9.6, + "cache_read": 0.8 + } } - } - } - }, - "inceptron": { - "id": "inceptron", - "env": [ - "INCEPTRON_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.inceptron.io/v1", - "name": "Inceptron", - "doc": "https://docs.inceptron.io", - "models": { - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + }, + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 900000 }, "cost": { - "input": 0.66, - "output": 3.5, + "input": 1.25, + "output": 2.5, "cache_read": 0.2, - "cache_write": 0 + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "x-ai/grok-4.5": { + "id": "x-ai/grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 500000, + "output": 450000 }, "cost": { - "input": 0.75, - "output": 3.5, - "cache_read": 0.2, - "cache_write": 0 + "input": 2, + "output": 6, + "cache_read": 0.3, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 0.6 + } } }, - "moonshotai/Kimi-K2.6-Fast": { - "id": "moonshotai/Kimi-K2.6-Fast", - "name": "Kimi K2.6 Fast", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "x-ai/grok-build-0.1": { + "id": "x-ai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 230400 }, - "status": "alpha", "cost": { - "input": 1.32, - "output": 7, - "cache_read": 0.4, - "cache_write": 0 + "input": 1, + "output": 2, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 4, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 4, + "cache_read": 0.4 + } } }, - "zai-org/GLM-5.1-FP8": { - "id": "zai-org/GLM-5.1-FP8", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "x-ai/grok-4.20": { + "id": "x-ai/grok-4.20", + "name": "Grok 4.20", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-09-01", + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 202752 + "context": 2000000, + "output": 1800000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "x-ai/grok-4.6": { + "id": "x-ai/grok-4.6", + "name": "Grok 4.6", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 500000, + "output": 450000 }, "cost": { - "input": 1.2, - "output": 4.2, - "cache_read": 0.26, - "cache_write": 0 + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, + "~anthropic/claude-opus-latest": { + "id": "~anthropic/claude-opus-latest", + "name": "Claude Opus Latest", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -107163,150 +111499,198 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 196608 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.9, - "cache_read": 0.05, - "cache_write": 0 + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 } - } - } - }, - "llama": { - "id": "llama", - "env": [ - "LLAMA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.llama.com/compat/v1/", - "name": "Llama", - "doc": "https://llama.developer.meta.com/docs/models", - "models": { - "llama-4-scout-17b-16e-instruct-fp8": { - "id": "llama-4-scout-17b-16e-instruct-fp8", - "name": "Llama-4-Scout-17B-16E-Instruct-FP8", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + }, + "~anthropic/claude-haiku-latest": { + "id": "~anthropic/claude-haiku-latest", + "name": "Claude Haiku Latest", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "cerebras-llama-4-maverick-17b-128e-instruct": { - "id": "cerebras-llama-4-maverick-17b-128e-instruct", - "name": "Cerebras-Llama-4-Maverick-17B-128E-Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "~anthropic/claude-sonnet-latest": { + "id": "~anthropic/claude-sonnet-latest", + "name": "Claude Sonnet Latest", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "~anthropic/claude-fable-latest": { + "id": "~anthropic/claude-fable-latest", + "name": "Claude Fable Latest", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "structured_output": true, + "temperature": false, + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "groq-llama-4-maverick-17b-128e-instruct": { - "id": "groq-llama-4-maverick-17b-128e-instruct", - "name": "Groq-Llama-4-Maverick-17B-128E-Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": true, - "reasoning": false, + "upstage/solar-pro-3": { + "id": "upstage/solar-pro-3", + "name": "Solar Pro 3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "solar-pro", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text" @@ -107315,28 +111699,43 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 131072, + "output": 117964 }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "cerebras-llama-4-scout-17b-16e-instruct": { - "id": "cerebras-llama-4-scout-17b-16e-instruct", - "name": "Cerebras-Llama-4-Scout-17B-16E-Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": true, - "reasoning": false, + "upstage/solar-pro4": { + "id": "upstage/solar-pro4", + "name": "Solar Pro 4", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "solar", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", "modalities": { "input": [ "text" @@ -107345,28 +111744,43 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 524288, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.09, + "output": 0.36, + "cache_read": 0.018 } }, - "llama-3.3-8b-instruct": { - "id": "llama-3.3-8b-instruct", - "name": "Llama-3.3-8B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": true, - "reasoning": false, + "upstage/solar-mini4": { + "id": "upstage/solar-mini4", + "name": "Solar Mini 4", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "solar", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ "text" @@ -107375,28 +111789,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 524288, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.05, + "output": 0.2, + "cache_read": 0.005 } }, - "llama-4-maverick-17b-128e-instruct-fp8": { - "id": "llama-4-maverick-17b-128e-instruct-fp8", - "name": "Llama-4-Maverick-17B-128E-Instruct-FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "~deepseek/deepseek-flash-latest": { + "id": "~deepseek/deepseek-flash-latest", + "name": "DeepSeek Flash Latest", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-09-14", + "last_updated": "2026-09-14", "modalities": { "input": [ "text", @@ -107406,38 +111834,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01 } - } - } - }, - "llmtr": { - "id": "llmtr", - "env": [ - "LLMTR_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://llmtr.com/v1", - "name": "LLMTR", - "doc": "https://llmtr.com/docs", - "models": { - "sincap": { - "id": "sincap", - "name": "Sincap", - "description": "General-purpose chat model for instruction following, writing, and analysis", + }, + "~deepseek/deepseek-pro-latest": { + "id": "~deepseek/deepseek-pro-latest", + "name": "DeepSeek Pro Latest", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-05", - "last_updated": "2026-05-05", + "release_date": "2026-09-14", + "last_updated": "2026-09-14", "modalities": { "input": [ "text" @@ -107448,24 +111880,40 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0, - "output": 0 + "input": 0.39996, + "output": 1.19988, + "cache_read": 0.012726 } }, - "magibu-11b-v8": { - "id": "magibu-11b-v8", - "name": "Magibu 11B v8", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "~deepseek/deepseek-v4-flash-latest": { + "id": "~deepseek/deepseek-v4-flash-latest", + "name": "DeepSeek V4 Flash Latest", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-05", - "last_updated": "2026-06-05", + "release_date": "2026-08-01", + "last_updated": "2026-08-01", "modalities": { "input": [ "text" @@ -107476,56 +111924,91 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 1310720, + "output": 943718 }, "cost": { - "input": 0, - "output": 0 + "input": 0.038, + "output": 0.55, + "cache_read": 0.0228 } }, - "gemma-4": { - "id": "gemma-4", - "name": "Gemma 4", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": false, - "reasoning": false, - "tool_call": false, + "~x-ai/grok-latest": { + "id": "~x-ai/grok-latest", + "name": "Grok Latest", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 500000, + "output": 450000 }, "cost": { - "input": 5, - "output": 10 + "input": 1.6, + "output": 4.8, + "cache_read": 0.4, + "tiers": [ + { + "input": 3.2, + "output": 9.6, + "cache_read": 0.8, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 3.2, + "output": 9.6, + "cache_read": 0.8 + } } }, - "medgemma-4b": { - "id": "medgemma-4b", - "name": "MedGemma 4B", + "bytedance/ui-tars-1.5-7b": { + "id": "bytedance/ui-tars-1.5-7b", + "name": "UI-TARS 7B ", "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, "reasoning": false, "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-04-26", - "last_updated": "2026-04-26", + "knowledge": "2025-01-31", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" @@ -107533,37 +112016,31 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 4096 + "context": 128000, + "output": 2048 }, "cost": { - "input": 3, - "output": 5 + "input": 0.1, + "output": 0.2, + "cache_read": 0.1 } }, - "qwen3-6-35b": { - "id": "qwen3-6-35b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, + "liquid/lfm-2.5-2.6b:free": { + "id": "liquid/lfm-2.5-2.6b:free", + "name": "LFM2.5-2.6B (free)", + "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", + "family": "liquid", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -107571,28 +112048,36 @@ }, "open_weights": true, "limit": { - "context": 16384, - "output": 65536 + "context": 65536, + "output": 8192 }, "cost": { - "input": 5, - "output": 10 + "input": 0, + "output": 0 } }, - "trendyol-7b": { - "id": "trendyol-7b", - "name": "Trendyol 7B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-06-06", - "last_updated": "2026-06-06", - "modalities": { + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" @@ -107600,89 +112085,105 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 8192 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.09, + "output": 0.34, + "cache_read": 0.05 } - } - } - }, - "cohere": { - "id": "cohere", - "env": [ - "COHERE_API_KEY" - ], - "npm": "@ai-sdk/cohere", - "name": "Cohere", - "doc": "https://docs.cohere.com/docs/models", - "models": { - "c4ai-aya-expanse-32b": { - "id": "c4ai-aya-expanse-32b", - "name": "Aya Expanse 32B", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "attachment": false, - "reasoning": false, + }, + "google/gemini-3.1-flash-image-preview": { + "id": "google/gemini-3.1-flash-image-preview", + "name": "Nano Banana 2 Preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2024-10-24", - "last_updated": "2024-10-24", + "knowledge": "2025-01", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ + "image", "text" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 + "context": 65536, + "output": 58982 + }, + "cost": { + "input": 0.5, + "output": 3 } }, - "command-a-03-2025": { - "id": "command-a-03-2025", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", - "attachment": false, + "google/lyria-3-pro-preview": { + "id": "google/lyria-3-pro-preview", + "name": "Lyria 3 Pro Preview", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "lyria", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "release_date": "2026-03-25", + "last_updated": "2026-03-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0, + "output": 0 } }, - "c4ai-aya-vision-32b": { - "id": "c4ai-aya-vision-32b", - "name": "Aya Vision 32B", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma 3 27B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-03-04", - "last_updated": "2025-05-14", + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text", @@ -107694,25 +112195,37 @@ }, "open_weights": true, "limit": { - "context": 16000, - "output": 4000 + "context": 131072, + "output": 117964 + }, + "cost": { + "input": 0.08, + "output": 0.45, + "cache_read": 0.04 } }, - "command-r7b-arabic-02-2025": { - "id": "command-r7b-arabic-02-2025", - "name": "Command R7B Arabic", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, + "google/gemma-4-31b-it:free": { + "id": "google/gemma-4-31b-it:free", + "name": "Gemma 4 31B (free)", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" @@ -107720,54 +112233,71 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.0375, - "output": 0.15 + "input": 0, + "output": 0 } }, - "c4ai-aya-vision-8b": { - "id": "c4ai-aya-vision-8b", - "name": "Aya Vision 8B", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "google/gemini-2.5-flash-image": { + "id": "google/gemini-2.5-flash-image", + "name": "Nano Banana", + "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", + "family": "gemini-flash", "attachment": true, "reasoning": false, "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-03-04", - "last_updated": "2025-05-14", + "knowledge": "2024-06", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16000, - "output": 4000 + "context": 32768, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333 } }, - "command-r-08-2024": { - "id": "command-r-08-2024", - "name": "Command R", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, + "google/gemma-4-26b-a4b-it:free": { + "id": "google/gemma-4-26b-a4b-it:free", + "name": "Gemma 4 26B A4B (free)", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" @@ -107775,26 +112305,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0, + "output": 0 } }, - "command-r7b-12-2024": { - "id": "command-r7b-12-2024", - "name": "Command R7B", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", + "google/gemma-2-27b-it": { + "id": "google/gemma-2-27b-it", + "name": "Gemma 2 27B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-12-02", - "last_updated": "2024-12-02", + "knowledge": "2024-06-30", + "release_date": "2024-07-13", + "last_updated": "2024-07-13", "modalities": { "input": [ "text" @@ -107805,71 +112336,85 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 8192, + "output": 2048 }, "cost": { - "input": 0.0375, - "output": 0.15 + "input": 0.65, + "output": 0.65 } }, - "command-a-vision-07-2025": { - "id": "command-a-vision-07-2025", - "name": "Command A Vision", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", - "attachment": false, - "reasoning": false, - "tool_call": false, + "google/gemini-3.6-flash": { + "id": "google/gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-07-31", - "last_updated": "2025-07-31", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667 } }, - "command-a-plus-05-2026": { - "id": "command-a-plus-05-2026", - "name": "Command A Plus", - "description": "Cohere's stronger command model for multilingual agents and enterprise workflows", - "family": "command-a", + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04-01", - "release_date": "2026-05-20", - "last_updated": "2026-06-09", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ + "image", "text", - "image" + "video" ], "output": [ "text" @@ -107877,173 +112422,262 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 64000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.09, + "output": 0.3, + "cache_read": 0.05 } }, - "command-a-translate-08-2025": { - "id": "command-a-translate-08-2025", - "name": "Command A Translate", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "family": "command-a", - "attachment": false, - "reasoning": false, - "tool_call": true, + "google/gemini-3-pro-image-preview": { + "id": "google/gemini-3-pro-image-preview", + "name": "Nano Banana Pro Preview", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "knowledge": "2025-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8000, - "output": 8000 + "context": 65536, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 10 + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "command-r-plus-08-2024": { - "id": "command-r-plus-08-2024", - "name": "Command R+", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", - "attachment": false, - "reasoning": false, + "google/gemini-3.5-flash-lite": { + "id": "google/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.3, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333 } }, - "command-a-reasoning-08-2025": { - "id": "command-a-reasoning-08-2025", - "name": "Command A Reasoning", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", - "attachment": false, + "google/gemini-3.1-flash-image": { + "id": "google/gemini-3.1-flash-image", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens", - "min": 1 + "type": "effort", + "values": [ + "minimal", + "high" + ] } ], - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ + "image", "text" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.5, + "output": 3 } }, - "north-mini-code-1-0": { - "id": "north-mini-code-1-0", - "name": "North Mini Code", - "description": "Cohere coding model for practical software engineering and agentic edits", - "family": "north", - "attachment": false, + "google/lyria-3-clip-preview": { + "id": "google/lyria-3-clip-preview", + "name": "Lyria 3 Clip Preview", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "lyria", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-03-25", + "last_updated": "2026-03-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], "tool_call": true, "interleaved": { - "field": "reasoning_content" + "field": "reasoning_details" }, "structured_output": true, "temperature": true, - "knowledge": "2025-09-23", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.cohere.ai/compatibility/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "c4ai-aya-expanse-8b": { - "id": "c4ai-aya-expanse-8b", - "name": "Aya Expanse 8B", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "attachment": false, + "google/gemma-3-12b-it": { + "id": "google/gemma-3-12b-it", + "name": "Gemma 3 12B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-10-24", - "last_updated": "2024-10-24", + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -108051,34 +112685,26 @@ }, "open_weights": true, "limit": { - "context": 8000, - "output": 4000 + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0.05, + "output": 0.15 } - } - } - }, - "sarvam": { - "id": "sarvam", - "env": [ - "SARVAM_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.sarvam.ai/v1", - "name": "Sarvam AI", - "doc": "https://docs.sarvam.ai/api-reference-docs/getting-started/models", - "models": { - "sarvam-105b": { - "id": "sarvam-105b", - "name": "Sarvam-105B", - "description": "Flagship Indian-language reasoning model for enterprise multilingual applications", - "family": "sarvam", - "attachment": false, + }, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - null, + "minimal", "low", "medium", "high" @@ -108086,236 +112712,303 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-18", - "last_updated": "2026-03-06", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "reasoning": 9, + "cache_read": 0.15, + "cache_write": 0.083333 } }, - "sarvam-30b": { - "id": "sarvam-30b", - "name": "Sarvam-30B", - "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", - "family": "sarvam", - "attachment": false, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - null, - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-18", - "last_updated": "2026-03-06", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, + "context": 1048576, "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 10, + "reasoning": 10, + "cache_read": 0.125, + "cache_write": 0.375, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } - } - } - }, - "stepfun": { - "id": "stepfun", - "env": [ - "STEPFUN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.stepfun.com/v1", - "name": "StepFun", - "doc": "https://platform.stepfun.com/docs/zh/overview/concept", - "models": { - "step-1-32k": { - "id": "step-1-32k", - "name": "Step 1 (32K)", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, + }, + "google/gemini-3-pro-image": { + "id": "google/gemini-3-pro-image", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-01-01", - "last_updated": "2026-02-13", + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, + "context": 131072, "output": 32768 }, "cost": { - "input": 2.05, - "output": 9.59, - "cache_read": 0.41 + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "step-3.7-flash": { - "id": "step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2026-01-01", - "release_date": "2026-05-29", - "last_updated": "2026-06-29", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "input": 0.3, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333 } }, - "step-3.5-flash-2603": { - "id": "step-3.5-flash-2603", - "name": "Step 3.5 Flash 2603", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, + "google/gemini-3.1-flash-lite-image": { + "id": "google/gemini-3.1-flash-lite-image", + "name": "Nano Banana 2 Lite", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", + "minimal", "high" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, + "structured_output": false, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 65536, + "output": 58982 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 0.25, + "output": 1.5 } }, - "stepaudio-2.5-tts": { - "id": "stepaudio-2.5-tts", - "name": "StepAudio 2.5 TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "step", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-16", - "last_updated": "2026-07-02", + "google/gemini-3.7-flash": { + "id": "google/gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667 } }, - "stepaudio-2.5-asr": { - "id": "stepaudio-2.5-asr", - "name": "StepAudio 2.5 ASR", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "step", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-24", - "last_updated": "2026-07-02", + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_details" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "audio" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -108323,85 +113016,236 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "step-3.5-flash": { - "id": "step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", - "attachment": false, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "interleaved": { - "field": "reasoning_content" + "field": "reasoning_details" }, + "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-06-15", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 - } - }, - "step-tts-2": { - "id": "step-tts-2", - "name": "Step TTS 2", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "step", - "attachment": false, + "input": 0.5, + "output": 3, + "reasoning": 3, + "cache_read": 0.05, + "cache_write": 0.083333 + } + }, + "google/gemini-3.8-flash": { + "id": "google/gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667 + } + }, + "google/gemma-3-4b-it": { + "id": "google/gemma-3-4b-it", + "name": "Gemma 3 4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-03-01", - "last_updated": "2026-07-02", + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0.05, + "output": 0.1 + } + }, + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "reasoning": 0.4, + "cache_read": 0.01, + "cache_write": 0.083333 } }, - "step-2-16k": { - "id": "step-2-16k", - "name": "Step 2 (16K)", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, + "google/gemini-2.5-pro-preview": { + "id": "google/gemini-2.5-pro-preview", + "name": "Gemini 2.5 Pro Preview 06-05", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-01-01", - "last_updated": "2026-02-13", + "knowledge": "2025-01-31", + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text" + "pdf", + "image", + "text", + "audio" ], "output": [ "text" @@ -108409,122 +113253,149 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5.21, - "output": 16.44, - "cache_read": 1.04 + "input": 1.25, + "output": 10, + "reasoning": 10, + "cache_read": 0.125, + "cache_write": 0.375, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } - } - } - }, - "hpc-ai": { - "id": "hpc-ai", - "env": [ - "HPC_AI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.hpc-ai.com/inference/v1", - "name": "HPC-AI", - "doc": "https://www.hpc-ai.com/doc/docs/quickstart/", - "models": { - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": false, + }, + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01-01", - "release_date": "2026-01-01", - "last_updated": "2026-06-01", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.3, + "input": 0.25, "output": 1.5, - "cache_read": 0.05 + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.083333 } }, - "zai-org/glm-5.1": { - "id": "zai-org/glm-5.1", - "name": "GLM 5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-06-01", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202000, - "output": 202000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.615, - "output": 2.46, - "cache_read": 0.133 + "input": 0.25, + "output": 1.5, + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.083333 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax-m2.5", + "writer/palmyra-x5": { + "id": "writer/palmyra-x5", + "name": "Palmyra X5", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "palmyra", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-06-01", + "release_date": "2026-01-21", + "last_updated": "2026-01-21", "modalities": { "input": [ "text" @@ -108533,41 +113404,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1040000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.6, + "output": 6 } - } - } - }, - "minimax-cn": { - "id": "minimax-cn", - "env": [ - "MINIMAX_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://api.minimaxi.com/anthropic/v1", - "name": "MiniMax (minimaxi.com)", - "doc": "https://platform.minimaxi.com/docs/guides/quickstart", - "models": { - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + }, + "ibm-granite/granite-4.0-h-micro": { + "id": "ibm-granite/granite-4.0-h-micro", + "name": "Granite 4.0 Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "granite", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2025-10-20", + "last_updated": "2025-10-20", "modalities": { "input": [ "text" @@ -108578,26 +113436,36 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 131000, + "output": 117900 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.017, + "output": 0.112 } }, - "MiniMax-M2.5-highspeed": { - "id": "MiniMax-M2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "ibm-granite/granite-4.2-8b": { + "id": "ibm-granite/granite-4.2-8b", + "name": "Granite 4.2 8B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "granite", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "release_date": "2026-08-31", + "last_updated": "2026-08-31", "modalities": { "input": [ "text" @@ -108608,28 +113476,28 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 117964 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.06, + "output": 0.25, + "cache_read": 0.015 } }, - "MiniMax-M2.7-highspeed": { - "id": "MiniMax-M2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "mistralai/mistral-nemo": { + "id": "mistralai/mistral-nemo", + "name": "Mistral Nemo", + "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", + "family": "mistral-nemo", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ "text" @@ -108640,31 +113508,30 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.019, + "output": 0.03 } }, - "MiniMax-M2": { - "id": "MiniMax-M2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "mistralai/ministral-8b-2512": { + "id": "mistralai/ministral-8b-2512", + "name": "Ministral 3 8B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -108672,26 +113539,28 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 128000 + "context": 262144, + "output": 209715 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.15, + "output": 0.15, + "cache_read": 0.015 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "mistralai/mistral-small-24b-instruct-2501": { + "id": "mistralai/mistral-small-24b-instruct-2501", + "name": "Mistral Small 3", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2023-10-31", + "release_date": "2025-01-30", + "last_updated": "2025-01-30", "modalities": { "input": [ "text" @@ -108702,167 +113571,138 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 32768, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.05, + "output": 0.08 } }, - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + "mistralai/mistral-saba": { + "id": "mistralai/mistral-saba", + "name": "Saba", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-25", + "knowledge": "2024-09-30", + "release_date": "2025-02-17", + "last_updated": "2025-02-17", "modalities": { "input": [ "text", - "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 32768, + "output": 26214 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "tiers": [ - { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12, - "tier": { - "type": "context", - "size": 512000 - } - } - ], - "context_over_200k": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 - } + "input": 0.2, + "output": 0.6, + "cache_read": 0.02 } }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "mistralai/mistral-medium-3-5": { + "id": "mistralai/mistral-medium-3-5", + "name": "Mistral Medium 3.5", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 209715 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 1.5, + "output": 7.5 } - } - } - }, - "alibaba-coding-plan": { - "id": "alibaba-coding-plan", - "env": [ - "ALIBABA_CODING_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://coding-intl.dashscope.aliyuncs.com/v1", - "name": "Alibaba Coding Plan", - "doc": "https://www.alibabacloud.com/help/en/model-studio/coding-plan", - "models": { - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + }, + "mistralai/mistral-medium-3.1": { + "id": "mistralai/mistral-medium-3.1", + "name": "Mistral Medium 3.1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "knowledge": "2025-06-30", + "release_date": "2025-08-13", + "last_updated": "2025-08-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 131072, + "output": 104857 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.4, + "output": 2, + "cache_read": 0.04 } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "mistralai/mistral-small-3.2-24b-instruct": { + "id": "mistralai/mistral-small-3.2-24b-instruct", + "name": "Mistral Small 3.2 24B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2023-10-31", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", "modalities": { "input": [ + "image", "text" ], "output": [ @@ -108871,37 +113711,31 @@ }, "open_weights": true, "limit": { - "context": 202752, + "context": 256000, "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.09375, + "output": 0.25 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "mistralai/mistral-large": { + "id": "mistralai/mistral-large", + "name": "Mistral Large", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2024-11-30", + "release_date": "2024-02-26", + "last_updated": "2024-02-26", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" @@ -108909,74 +113743,75 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "output": 102400 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "mistralai/mistral-small-2603": { + "id": "mistralai/mistral-small-2603", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "output": 209715 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "mistralai/mistral-medium-3": { + "id": "mistralai/mistral-medium-3", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2025-03-31", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -108984,69 +113819,65 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 131072, + "output": 104857 }, "cost": { - "input": 0.1875, - "output": 1.125, - "cache_write": 0.234375 + "input": 0.4, + "output": 2, + "cache_read": 0.04 } }, - "qwen3-max-2026-01-23": { - "id": "qwen3-max-2026-01-23", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "mistralai/voxtral-small-24b-2507": { + "id": "mistralai/voxtral-small-24b-2507", + "name": "Voxtral Small 24B 2507", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "voxtral", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-23", - "last_updated": "2026-01-23", + "release_date": "2025-07-15", + "last_updated": "2025-07-15", "modalities": { "input": [ - "text" + "text", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 32768, + "output": 26214 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01 } }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "mistralai/mistral-large-2407": { + "id": "mistralai/mistral-large-2407", + "name": "Mistral Large 2407", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "knowledge": "2024-03-31", + "release_date": "2024-11-19", + "last_updated": "2024-11-19", "modalities": { "input": [ "text", - "image", - "video" + "pdf" ], "output": [ "text" @@ -109054,41 +113885,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 131072, + "output": 104857 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "mistralai/ministral-14b-2512": { + "id": "mistralai/ministral-14b-2512", + "name": "Ministral 3 14B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -109097,33 +113918,31 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 209715 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.2, + "output": 0.2, + "cache_read": 0.02 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "mistralai/mistral-small-3.1-24b-instruct": { + "id": "mistralai/mistral-small-3.1-24b-instruct", + "name": "Mistral Small 3.1 24B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2023-10-31", + "release_date": "2025-03-17", + "last_updated": "2025-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -109131,32 +113950,30 @@ }, "open_weights": true, "limit": { - "context": 196608, - "input": 196601, - "output": 24576 + "context": 128000, + "output": 102400 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.351, + "output": 0.555 } }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "mistralai/ministral-3b-2512": { + "id": "mistralai/ministral-3b-2512", + "name": "Ministral 3 3B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-03", - "last_updated": "2026-02-03", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -109164,77 +113981,65 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "output": 104857 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.1, + "output": 0.1, + "cache_read": 0.01 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "mistralai/mixtral-8x22b-instruct": { + "id": "mistralai/mixtral-8x22b-instruct", + "name": "Mixtral 8x22B Instruct", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "knowledge": "2024-01-31", + "release_date": "2024-04-17", + "last_updated": "2024-04-17", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202752, - "output": 16384 + "context": 65536, + "output": 52428 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "mistralai/codestral-2508": { + "id": "mistralai/codestral-2508", + "name": "Codestral 2508", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "codestral", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-03-31", + "release_date": "2025-08-01", + "last_updated": "2025-08-01", "modalities": { "input": [ "text", - "image", - "video" + "pdf" ], "output": [ "text" @@ -109242,50 +114047,42 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 256000, + "output": 204800 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.3, + "output": 0.9, + "cache_read": 0.03 } - } - } - }, - "longcat": { - "id": "longcat", - "env": [ - "LONGCAT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.longcat.chat/openai", - "name": "LongCat", - "doc": "https://longcat.chat/platform/docs/", - "models": { - "LongCat-2.0": { - "id": "LongCat-2.0", - "name": "LongCat-2.0", - "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", - "family": "longcat", - "attachment": false, + }, + "sakana/fugu-max": { + "id": "sakana/fugu-max", + "name": "Fugu Max", + "description": "Multi-agent model for routing expert agents across complex analytical tasks", + "family": "fugu", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "structured_output": true, + "temperature": false, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -109294,66 +114091,42 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 128000 }, "cost": { - "input": 0.75, - "output": 2.95, - "cache_read": 0.015 - } - } - } - }, - "poe": { - "id": "poe", - "env": [ - "POE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.poe.com/v1", - "name": "Poe", - "doc": "https://creator.poe.com/docs/external-applications/openai-compatible-api", - "models": { - "trytako/tako": { - "id": "trytako/tako", - "name": "Tako", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "tako", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2048, - "output": 0 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "xai/grok-code-fast-1": { - "id": "xai/grok-code-fast-1", - "name": "Grok Code Fast 1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "sakana/fugu-ultra-v2": { + "id": "sakana/fugu-ultra-v2", + "name": "Fugu Ultra v2", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-08-22", - "last_updated": "2025-08-22", + "knowledge": "2026-08-28", + "release_date": "2026-09-11", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -109361,27 +114134,53 @@ }, "open_weights": false, "limit": { - "context": 256000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.5, - "cache_read": 0.02 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "xai/grok-4.1-fast-reasoning": { - "id": "xai/grok-4.1-fast-reasoning", - "name": "Grok-4.1-Fast-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "sakana/fugu-ultra": { + "id": "sakana/fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ "text", @@ -109393,34 +114192,57 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "xai/grok-3-mini": { - "id": "xai/grok-3-mini", - "name": "Grok 3 Mini", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "sakana/sakana-namazu": { + "id": "sakana/sakana-namazu", + "name": "Sakana Namazu", + "description": "Multi-agent model for routing expert agents across complex analytical tasks", + "family": "sakana-namazu", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-04-11", - "last_updated": "2025-04-11", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -109428,30 +114250,35 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 0.5, - "cache_read": 0.075 + "input": 0.95, + "output": 4, + "cache_read": 0.15 } }, - "xai/grok-4.1-fast-non-reasoning": { - "id": "xai/grok-4.1-fast-non-reasoning", - "name": "Grok-4.1-Fast-Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": false, + "inclusionai/ling-3.0-flash-fin:free": { + "id": "inclusionai/ling-3.0-flash-fin:free", + "name": "Ling 3.0 Flash Fin (free)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "structured_output": false, + "temperature": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -109459,21 +114286,31 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 } }, - "xai/grok-3": { - "id": "xai/grok-3", - "name": "Grok 3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, + "inclusionai/ling-3.0-flash-sante:free": { + "id": "inclusionai/ling-3.0-flash-sante:free", + "name": "Ling 3.0 Flash Sante (free)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-04-11", - "last_updated": "2025-04-11", + "structured_output": false, + "temperature": true, + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ "text" @@ -109484,94 +114321,108 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 + "input": 0, + "output": 0 } }, - "xai/grok-4-fast-reasoning": { - "id": "xai/grok-4-fast-reasoning", - "name": "Grok-4-Fast-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, + "inclusionai/ling-3.0-flash": { + "id": "inclusionai/ling-3.0-flash", + "name": "Ling 3.0 Flash", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-09-16", - "last_updated": "2025-09-16", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-23", + "last_updated": "2026-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.021, + "output": 0.063, + "cache_read": 0.0042 } }, - "xai/grok-4": { - "id": "xai/grok-4", - "name": "Grok-4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "inclusionai/ling-3.0-flash-vl": { + "id": "inclusionai/ling-3.0-flash-vl", + "name": "Ling 3.0 Flash VL", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ling", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-07-10", - "last_updated": "2025-07-10", + "structured_output": true, + "temperature": true, + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 + "input": 0.06, + "output": 0.18, + "cache_read": 0.012 } }, - "xai/grok-4-fast-non-reasoning": { - "id": "xai/grok-4-fast-non-reasoning", - "name": "Grok-4-Fast-Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": false, + "inclusionai/ling-3.0-flash-fin": { + "id": "inclusionai/ling-3.0-flash-fin", + "name": "Ling 3.0 Flash Fin", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-09-16", - "last_updated": "2025-09-16", + "structured_output": true, + "temperature": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -109579,80 +114430,82 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 128000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.06, + "output": 0.18, + "cache_read": 0.012 } }, - "xai/grok-4.20-multi-agent": { - "id": "xai/grok-4.20-multi-agent", - "name": "Grok-4.20-Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k3", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": false, - "release_date": "2026-03-13", - "last_updated": "2026-03-13", + "structured_output": true, + "temperature": true, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 0 + "context": 1048576, + "output": 943718 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "topazlabs-co/topazlabs": { - "id": "topazlabs-co/topazlabs", - "name": "TopazLabs", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "topazlabs", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] + "interleaved": { + "field": "reasoning_details" }, - "open_weights": false, - "limit": { - "context": 204, - "output": 0 - } - }, - "fireworks-ai/kimi-k2.5-fw": { - "id": "fireworks-ai/kimi-k2.5-fw", - "name": "Kimi-K2.5-FW", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -109662,404 +114515,382 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "input": 245760, - "output": 16384 + "output": 235929 }, "cost": { - "input": 0, - "output": 0 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "google/veo-3.1-fast": { - "id": "google/veo-3.1-fast", - "name": "Veo-3.1-Fast", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, - "reasoning": false, + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "video" - ] + "interleaved": { + "field": "reasoning_details" }, - "open_weights": false, - "limit": { - "context": 480, - "output": 0 - } - }, - "google/imagen-3": { - "id": "google/imagen-3", - "name": "Imagen-3", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2024-10-15", - "last_updated": "2024-10-15", + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 262144, + "output": 98304 + }, + "cost": { + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "google/nano-banana-pro": { - "id": "google/nano-banana-pro", - "name": "Nano-Banana-Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "nano-banana", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "interleaved": { + "field": "reasoning_details" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 0 + "context": 262144, + "output": 235929 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 0.45, + "output": 2.25, + "cache_read": 0.07 } }, - "google/lyria": { - "id": "google/lyria", - "name": "Lyria", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "lyria", - "attachment": true, + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-06-04", - "last_updated": "2025-06-04", + "structured_output": true, + "temperature": true, + "knowledge": "2024-12-31", + "release_date": "2025-09-04", + "last_updated": "2025-09-04", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "output": 98304 + }, + "cost": { + "input": 0.6, + "output": 2.5 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini-3.1-Flash-Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "high" - ] - } - ], + "moonshotai/kimi-k2": { + "id": "moonshotai/kimi-k2", + "name": "Kimi K2 0711", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2026-02-18", - "last_updated": "2026-02-18", + "structured_output": false, + "temperature": true, + "knowledge": "2024-12-31", + "release_date": "2025-07-11", + "last_updated": "2025-07-11", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.25, - "output": 1.5 + "input": 0.57, + "output": 2.3 } }, - "google/nano-banana": { - "id": "google/nano-banana", - "name": "Nano-Banana", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "nano-banana", + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 0 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.21, - "output": 1.8, - "cache_read": 0.021 + "input": 0.7062, + "output": 3.3, + "cache_read": 0.18 } }, - "google/gemini-deep-research": { - "id": "google/gemini-deep-research", - "name": "gemini-deep-research", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": true, + "rekaai/reka-flash-3": { + "id": "rekaai/reka-flash-3", + "name": "Reka Flash 3", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "reka", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "temperature": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01-31", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 0 + "context": 65536, + "output": 58982 }, - "status": "deprecated", "cost": { - "input": 1.6, - "output": 9.6 + "input": 0.1, + "output": 0.2 } }, - "google/veo-3": { - "id": "google/veo-3", - "name": "Veo-3", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", + "rekaai/reka-edge": { + "id": "rekaai/reka-edge", + "name": "Reka Edge", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "reka", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-05-21", - "last_updated": "2025-05-21", + "structured_output": true, + "temperature": true, + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 16384, + "output": 14745 + }, + "cost": { + "input": 0.1, + "output": 0.1 } }, - "google/gemini-3-flash": { - "id": "google/gemini-3-flash", - "name": "Gemini-3-Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "attachment": true, + "nvidia/nemotron-3-nano-30b-a3b": { + "id": "nvidia/nemotron-3-nano-30b-a3b", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-10-07", - "last_updated": "2025-10-07", + "structured_output": true, + "temperature": true, + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.04 - } - }, - "google/imagen-3-fast": { - "id": "google/imagen-3-fast", - "name": "Imagen-3-Fast", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2024-10-17", - "last_updated": "2024-10-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 480, - "output": 0 + "input": 0.05, + "output": 0.2, + "cache_read": 0.03 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini-2.5-Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "nvidia/nemotron-3.5-content-safety:free": { + "id": "nvidia/nemotron-3.5-content-safety:free", + "name": "Nemotron 3.5 Content Safety (free)", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "nemotron", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 0, - "max": 32768 + "type": "toggle" } ], - "tool_call": true, - "temperature": false, - "release_date": "2025-02-05", - "last_updated": "2025-02-05", + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1065535, - "output": 65535 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.87, - "output": 7, - "cache_read": 0.087 + "input": 0, + "output": 0 } }, - "google/veo-3.1": { - "id": "google/veo-3.1", - "name": "Veo-3.1", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, - "reasoning": false, + "nvidia/nemotron-3.5-lightning": { + "id": "nvidia/nemotron-3.5-lightning", + "name": "Nemotron 3.5 Lightning 30B A3B", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": true, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.08, + "output": 0.2, + "cache_read": 0.04 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini-2.5-Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free": { + "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free", + "name": "Nemotron 3 Nano Omni (free)", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 0, - "max": 24576 + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-04-26", - "last_updated": "2025-04-26", + "structured_output": false, + "temperature": true, + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ "text", @@ -110071,515 +114902,679 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1065535, - "output": 65535 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.21, - "output": 1.8, - "cache_read": 0.021 + "input": 0, + "output": 0 } }, - "google/imagen-4-fast": { - "id": "google/imagen-4-fast", - "name": "Imagen-4-Fast", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": true, - "reasoning": false, + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-06-25", - "last_updated": "2025-06-25", + "structured_output": true, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 262144, + "output": 235929 + }, + "cost": { + "input": 0.08, + "output": 0.45 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini-3.5-Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "nvidia/nemotron-3-ultra-550b-a55b:free": { + "id": "nvidia/nemotron-3-ultra-550b-a55b:free", + "name": "Nemotron 3 Ultra (free)", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", - "low", + "medium", "high" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "structured_output": false, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 1000000, "output": 65536 }, "cost": { - "input": 1.5152, - "output": 9.0909, - "cache_read": 0.1515 + "input": 0, + "output": 0 } }, - "google/gemma-4-31b": { - "id": "google/gemma-4-31b", - "name": "Gemma-4-31B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": true, - "reasoning": false, + "nvidia/nemotron-3-ultra-550b-a55b": { + "id": "nvidia/nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "medium", + "high" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "structured_output": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "output": 8192 + "output": 182520 }, "cost": { - "input": 0, - "output": 0 + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 } }, - "google/veo-3-fast": { - "id": "google/veo-3-fast", - "name": "Veo-3-Fast", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, - "reasoning": false, + "nvidia/nemotron-3-super-120b-a12b:free": { + "id": "nvidia/nemotron-3-super-120b-a12b:free", + "name": "Nemotron 3 Super (free)", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-10-13", - "last_updated": "2025-10-13", + "structured_output": true, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, + "context": 262144, + "output": 235929 + }, + "cost": { + "input": 0, "output": 0 } }, - "google/gemini-3-pro": { - "id": "google/gemini-3-pro", - "name": "Gemini-3-Pro", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gemini-pro", + "nvidia/nemotron-3.5-content-safety": { + "id": "nvidia/nemotron-3.5-content-safety", + "name": "Nemotron 3.5 Content Safety", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "nemotron", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "release_date": "2025-10-22", - "last_updated": "2025-10-22", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 131072, + "output": 117964 + }, + "cost": { + "input": 0.2, + "output": 0.2 + } + }, + "nvidia/nemotron-3.5-lightning:free": { + "id": "nvidia/nemotron-3.5-lightning:free", + "name": "Nemotron 3.5 Lightning (free)", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, "output": 65536 }, - "status": "deprecated", "cost": { - "input": 1.6, - "output": 9.6, - "cache_read": 0.16 + "input": 0, + "output": 0 } }, - "google/gemini-2.0-flash": { - "id": "google/gemini-2.0-flash", - "name": "Gemini-2.0-Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "xiaomi/mimo-v2.6-pro": { + "id": "xiaomi/mimo-v2.6-pro", + "name": "MiMo-V2.6-Pro", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "family": "mimo", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "structured_output": true, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", "image", - "video", - "audio" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 990000, - "output": 8192 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.42 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini-2.5-Flash-Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 0, - "max": 24576 + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-06-19", - "last_updated": "2025-06-19", + "interleaved": { + "field": "reasoning_details" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "video", - "audio" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1024000, - "output": 64000 + "context": 1050000, + "output": 131072 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "google/imagen-4": { - "id": "google/imagen-4", - "name": "Imagen-4", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", + "xiaomi/mimo-v2.6-pro-ultraspeed": { + "id": "xiaomi/mimo-v2.6-pro-ultraspeed", + "name": "MiMo-V2.6-Pro-UltraSpeed", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "family": "mimo", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "structured_output": true, + "temperature": true, + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 4.35, + "output": 8.7, + "cache_read": 0.036 } }, - "google/imagen-4-ultra": { - "id": "google/imagen-4-ultra", - "name": "Imagen-4-Ultra", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": true, - "reasoning": false, + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-05-24", - "last_updated": "2025-05-24", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 1050000, + "output": 131072 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } }, - "google/gemini-3.1-pro": { - "id": "google/gemini-3.1-pro", - "name": "Gemini-3.1-Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "xiaomi/mimo-v2.6-flash": { + "id": "xiaomi/mimo-v2.6-flash", + "name": "MiMo-V2.6-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "structured_output": true, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", "image", - "video", - "audio" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "google/veo-2": { - "id": "google/veo-2", - "name": "Veo-2", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, + "undi95/remm-slerp-l2-13b": { + "id": "undi95/remm-slerp-l2-13b", + "name": "ReMM SLERP 13B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2024-12-02", - "last_updated": "2024-12-02", + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2023-06-30", + "release_date": "2023-07-22", + "last_updated": "2023-07-22", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 6144, + "output": 5529 + }, + "cost": { + "input": 0.35, + "output": 0.65 } }, - "google/gemini-2.0-flash-lite": { - "id": "google/gemini-2.0-flash-lite", - "name": "Gemini-2.0-Flash-Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "gryphe/mythomax-l2-13b": { + "id": "gryphe/mythomax-l2-13b", + "name": "MythoMax 13B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2025-02-05", - "last_updated": "2025-02-05", + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2023-06-30", + "release_date": "2023-07-02", + "last_updated": "2023-07-02", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 990000, - "output": 8192 + "context": 8192, + "output": 3686 }, "cost": { - "input": 0.052, - "output": 0.21 + "input": 0.08, + "output": 0.11 } }, - "elevenlabs/elevenlabs-music": { - "id": "elevenlabs/elevenlabs-music", - "name": "ElevenLabs-Music", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "elevenlabs", - "attachment": true, - "reasoning": false, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-08-29", - "last_updated": "2025-08-29", + "interleaved": { + "field": "reasoning_details" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000, - "output": 0 + "context": 204800, + "output": 128000 + }, + "cost": { + "input": 0.27, + "output": 1.08, + "cache_read": 0.027 } }, - "elevenlabs/elevenlabs-v2.5-turbo": { - "id": "elevenlabs/elevenlabs-v2.5-turbo", - "name": "ElevenLabs-v2.5-Turbo", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "elevenlabs", + "minimax/minimax-01": { + "id": "minimax/minimax-01", + "name": "MiniMax-01", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2024-10-28", - "last_updated": "2024-10-28", + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-03-31", + "release_date": "2025-01-15", + "last_updated": "2025-01-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 0 + "context": 1000192, + "output": 900172 + }, + "cost": { + "input": 0.2, + "output": 1.1 } }, - "elevenlabs/elevenlabs-v3": { - "id": "elevenlabs/elevenlabs-v3", - "name": "ElevenLabs-v3", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "elevenlabs", - "attachment": true, + "minimax/minimax-m2-her": { + "id": "minimax/minimax-m2-her", + "name": "MiniMax-M2 Her", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-01-23", + "last_updated": "2026-01-23", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 0 + "context": 65536, + "output": 2048 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "openai/chatgpt-4o-latest": { - "id": "openai/chatgpt-4o-latest", - "name": "ChatGPT-4o-Latest", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gpt", - "attachment": true, - "reasoning": false, + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2024-08-14", - "last_updated": "2024-08-14", + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 204800, + "output": 131072 }, - "status": "deprecated", "cost": { - "input": 4.5, - "output": 14 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "openai/gpt-3.5-turbo-instruct": { - "id": "openai/gpt-3.5-turbo-instruct", - "name": "GPT-3.5-Turbo-Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, + "minimax/minimax-m1": { + "id": "minimax/minimax-m1", + "name": "MiniMax M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2023-09-20", - "last_updated": "2023-09-20", + "structured_output": false, + "temperature": true, + "knowledge": "2024-06-30", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -110587,96 +115582,134 @@ }, "open_weights": false, "limit": { - "context": 3500, - "output": 1024 + "context": 1000000, + "output": 40000 }, "cost": { - "input": 1.4, - "output": 1.8 + "input": 0.4, + "output": 2.2 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1048576, + "output": 512000 }, "cost": { - "input": 1.8, - "output": 7.2, - "cache_read": 0.45 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "openai/gpt-5.2-instant": { - "id": "openai/gpt-5.2-instant", - "name": "GPT-5.2-Instant", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "interleaved": { + "field": "reasoning_details" + }, + "structured_output": false, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.6, - "output": 13, - "cache_read": 0.16 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "openai/gpt-4o-search": { - "id": "openai/gpt-4o-search", - "name": "GPT-4o-Search", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-03-11", - "last_updated": "2025-03-11", + "interleaved": { + "field": "reasoning_details" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.255, + "output": 1.02 + } + }, + "mancer/weaver": { + "id": "mancer/weaver", + "name": "Weaver (alpha)", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "alpha", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2023-06-30", + "release_date": "2023-08-02", + "last_updated": "2023-08-02", "modalities": { "input": [ "text" @@ -110687,64 +115720,105 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 8000, + "output": 6000 }, "cost": { - "input": 2.2, - "output": 9 + "input": 0.4, + "output": 0.75 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2-Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "stepfun/step-3.5-flash": { + "id": "stepfun/step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "stepfun/step-3.7-flash": { + "id": "stepfun/step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "temperature": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "structured_output": true, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "input": 256000, + "output": 230400 }, "cost": { - "input": 19, - "output": 150 + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 } }, - "openai/gpt-4-classic-0314": { - "id": "openai/gpt-4-classic-0314", - "name": "GPT-4-Classic-0314", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gpt", + "dots-studio/dots-3-note-preview:free": { + "id": "dots-studio/dots-3-note-preview:free", + "name": "Dots3-Note Preview (free)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2024-08-26", - "last_updated": "2024-08-26", + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", @@ -110756,27 +115830,26 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 512000, + "output": 460800 }, - "status": "deprecated", "cost": { - "input": 27, - "output": 54 + "input": 0, + "output": 0 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "inception/mercury-2.5": { + "id": "inception/mercury-2.5", + "name": "Mercury 2.5", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", "high" @@ -110784,13 +115857,13 @@ } ], "tool_call": true, - "temperature": false, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": true, + "temperature": true, + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -110798,30 +115871,41 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 260000, + "output": 65536 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 0.04, + "output": 0.15, + "cache_read": 0.004 } }, - "openai/gpt-5-chat": { - "id": "openai/gpt-5-chat", - "name": "GPT-5-Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": false, + "inception/mercury-2": { + "id": "inception/mercury-2", + "name": "Mercury 2", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": true, + "release_date": "2026-03-04", + "last_updated": "2026-03-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -110830,25 +115914,27 @@ "open_weights": false, "limit": { "context": 128000, - "output": 16384 + "output": 50000 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 0.25, + "output": 0.75, + "cache_read": 0.025 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5-Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "amazon/nova-lite-v1": { + "id": "amazon/nova-lite-v1", + "name": "Nova Lite 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2023-09-13", - "last_updated": "2023-09-13", + "structured_output": false, + "temperature": true, + "knowledge": "2024-10-31", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text", @@ -110860,30 +115946,37 @@ }, "open_weights": false, "limit": { - "context": 16384, - "output": 2048 + "context": 300000, + "output": 5120 }, "cost": { - "input": 0.45, - "output": 1.4 + "input": 0.06, + "output": 0.24 } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5-Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "amazon/nova-2-lite-v1": { + "id": "amazon/nova-2-lite-v1", + "name": "Nova 2 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" @@ -110891,25 +115984,27 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 65535 }, "cost": { - "input": 14, - "output": 110 + "input": 0.3, + "output": 2.5 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "amazon/nova-pro-v1": { + "id": "amazon/nova-pro-v1", + "name": "Nova Pro 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "structured_output": false, + "temperature": true, + "knowledge": "2024-10-31", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text", @@ -110921,31 +116016,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 300000, + "output": 5120 + }, + "cost": { + "input": 0.8, + "output": 3.2 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "amazon/nova-premier-v1": { + "id": "amazon/nova-premier-v1", + "name": "Nova Premier 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "structured_output": false, + "temperature": true, + "release_date": "2025-10-31", + "last_updated": "2025-10-31", "modalities": { "input": [ "text", @@ -110957,66 +116047,60 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 0.99, - "output": 4, - "cache_read": 0.25 + "input": 2.5, + "output": 12.5, + "cache_read": 0.625 } }, - "openai/sora-2": { - "id": "openai/sora-2", - "name": "Sora-2", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "sora", - "attachment": true, + "amazon/nova-micro-v1": { + "id": "amazon/nova-micro-v1", + "name": "Nova Micro 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "structured_output": false, + "temperature": true, + "knowledge": "2024-10-31", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "output": 5120 + }, + "cost": { + "input": 0.035, + "output": 0.14 } }, - "openai/o3-pro": { - "id": "openai/o3-pro", - "name": "o3-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "relace/relace-search": { + "id": "relace/relace-search", + "name": "Relace Search", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -111024,29 +116108,28 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 256000, + "output": 128000 }, "cost": { - "input": 18, - "output": 72 + "input": 1, + "output": 3 } }, - "openai/gpt-4-classic": { - "id": "openai/gpt-4-classic", - "name": "GPT-4-Classic", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gpt", - "attachment": true, + "relace/relace-apply-3": { + "id": "relace/relace-apply-3", + "name": "Relace Apply 3", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": false, - "release_date": "2024-03-25", - "last_updated": "2024-03-25", + "release_date": "2025-09-26", + "last_updated": "2025-09-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -111054,30 +116137,29 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 256000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 27, - "output": 54 + "input": 0.85, + "output": 1.25 } }, - "openai/gpt-4o-aug": { - "id": "openai/gpt-4o-aug", - "name": "GPT-4o-Aug", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "aion-labs/aion-2.0": { + "id": "aion-labs/aion-2.0", + "name": "Aion-2.0", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2024-11-21", - "last_updated": "2024-11-21", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -111085,41 +116167,31 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 131072, + "output": 32768 }, "cost": { - "input": 2.2, - "output": 9, - "cache_read": 1.1 + "input": 0.8, + "output": 1.6, + "cache_read": 0.2 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4-Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "temperature": false, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "aion-labs/aion-rp-llama-3.1-8b": { + "id": "aion-labs/aion-rp-llama-3.1-8b", + "name": "Aion-RP 1.0 (8B)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12-31", + "release_date": "2025-02-04", + "last_updated": "2025-02-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -111127,67 +116199,69 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 32768, + "output": 29491 }, "cost": { - "input": 0.18, - "output": 1.1, - "cache_read": 0.018 + "input": 0.8, + "output": 1.6 } }, - "openai/sora-2-pro": { - "id": "openai/sora-2-pro", - "name": "Sora-2-Pro", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "sora", - "attachment": true, - "reasoning": false, + "aion-labs/aion-3.0": { + "id": "aion-labs/aion-3.0", + "name": "Aion-3.0", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-07", + "last_updated": "2026-07-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 3, + "output": 6, + "cache_read": 0.75 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "aion-labs/aion-3.5": { + "id": "aion-labs/aion-3.5", + "name": "Aion 3.5", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "temperature": false, - "release_date": "2025-11-12", - "last_updated": "2025-11-12", + "structured_output": false, + "temperature": true, + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -111195,29 +116269,39 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 3, + "output": 6, + "cache_read": 0.75 } }, - "openai/gpt-5.3-instant": { - "id": "openai/gpt-5.3-instant", - "name": "GPT-5.3-Instant", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, + "aion-labs/aion-3.5-mini": { + "id": "aion-labs/aion-3.5-mini", + "name": "Aion 3.5 Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": false, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "structured_output": false, + "temperature": true, + "release_date": "2026-09-23", + "last_updated": "2026-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -111225,27 +116309,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 111616, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.6, - "output": 13, - "cache_read": 0.16 + "input": 0.7, + "output": 1.4, + "cache_read": 0.18 } }, - "openai/gpt-5.3-codex-spark": { - "id": "openai/gpt-5.3-codex-spark", - "name": "GPT-5.3-Codex-Spark", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, + "aion-labs/aion-3.0-mini": { + "id": "aion-labs/aion-3.0-mini", + "name": "Aion-3.0-Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2026-03-04", - "last_updated": "2026-03-04", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-07", + "last_updated": "2026-07-07", "modalities": { "input": [ "text" @@ -111256,105 +116340,100 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.7, + "output": 1.4, + "cache_read": 0.18 } }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT-5.1-Codex-Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen/qwen3-235b-a22b-2507": { + "id": "qwen/qwen3-235b-a22b-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "structured_output": true, + "temperature": true, + "knowledge": "2025-06-30", + "release_date": "2025-07-21", + "last_updated": "2025-07-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 0.0875, + "output": 0.35, + "cache_read": 0.0175 } }, - "openai/dall-e-3": { - "id": "openai/dall-e-3", - "name": "DALL-E-3", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "dall-e", + "qwen/qwen3-vl-235b-a22b-instruct": { + "id": "qwen/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2023-11-06", - "last_updated": "2023-11-06", + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 800, - "output": 0 + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.21, + "output": 1.9, + "cache_read": 0.1 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": true, + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "structured_output": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -111362,77 +116441,76 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.99, - "output": 4 + "input": 1.475, + "output": 4.425, + "cache_read": 0.295, + "cache_write": 1.84375 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "qwen/qwen3.8-27b": { + "id": "qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", - "high", "xhigh" ] } ], "tool_call": true, - "temperature": false, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.6, - "output": 13, - "cache_read": 0.16 + "input": 0.42, + "output": 3, + "cache_read": 0.085 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "qwen/qwen3-vl-235b-a22b-thinking": { + "id": "qwen/qwen3-vl-235b-a22b-thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2026-02-10", - "last_updated": "2026-02-10", + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", @@ -111442,23 +116520,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.6, - "output": 13, - "cache_read": 0.16 + "input": 0.4, + "output": 4 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1-Codex-Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "qwen/qwen3.8-2.4t-a95b": { + "id": "qwen/qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -111466,14 +116543,15 @@ "values": [ "low", "medium", - "high" + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "release_date": "2025-11-12", - "last_updated": "2025-11-12", + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text" @@ -111482,284 +116560,262 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.22, - "output": 1.8, - "cache_read": 0.022 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "openai/o4-mini-deep-research": { - "id": "openai/o4-mini-deep-research", - "name": "o4-mini-deep-research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "family": "o-mini", + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-06-27", - "last_updated": "2025-06-27", + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.8, - "output": 7.2, - "cache_read": 0.45 - } - }, - "openai/gpt-image-1.5": { - "id": "openai/gpt-image-1.5", - "name": "gpt-image-1.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 0 + "input": 0.195, + "output": 1.56 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.09, - "output": 0.36, - "cache_read": 0.022 + "input": 0.15, + "output": 1.2 } }, - "openai/gpt-3.5-turbo-raw": { - "id": "openai/gpt-3.5-turbo-raw", - "name": "GPT-3.5-Turbo-Raw", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2023-09-27", - "last_updated": "2023-09-27", + "structured_output": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4524, - "output": 2048 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.45, - "output": 1.4 + "input": 0.12, + "output": 0.8, + "cache_read": 0.07 } }, - "openai/o1": { - "id": "openai/o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "qwen/qwen3-coder": { + "id": "qwen/qwen3-coder", + "name": "Qwen3 Coder 480B A35B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-12-18", - "last_updated": "2024-12-18", + "structured_output": true, + "temperature": true, + "knowledge": "2025-06-30", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 14, - "output": 54 + "input": 0.3, + "output": 1, + "cache_read": 0.1 } }, - "openai/o1-pro": { - "id": "openai/o1-pro", - "name": "o1-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": true, + "qwen/qwen3-32b": { + "id": "qwen/qwen3-32b", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-03-19", - "last_updated": "2025-03-19", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 140, - "output": 540 + "input": 0.08, + "output": 0.28 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen/qwen-plus": { + "id": "qwen/qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 2.2, - "output": 14, - "cache_read": 0.22 + "input": 0.26, + "output": 0.78, + "cache_read": 0.052, + "cache_write": 0.325, + "tiers": [ + { + "input": 0.78, + "output": 2.34, + "cache_read": 0.156, + "cache_write": 0.975, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.78, + "output": 2.34, + "cache_read": 0.156, + "cache_write": 0.975 + } } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4-Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "qwen/qwen3.7-flash": { + "id": "qwen/qwen3.7-flash", + "name": "Qwen3.7 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "temperature": false, - "release_date": "2026-03-12", - "last_updated": "2026-03-12", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -111767,31 +116823,55 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "input": 991000, + "output": 65536 }, "cost": { - "input": 0.68, - "output": 4, - "cache_read": 0.068 + "input": 0.03, + "output": 0.13, + "cache_read": 0.006, + "cache_write": 0.038, + "tiers": [ + { + "input": 0.1, + "output": 0.4, + "cache_read": 0.02, + "cache_write": 0.125, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.2, + "output": 0.8, + "cache_read": 0.04, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 256000 + } + } + ] } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -111799,27 +116879,51 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.8, - "output": 7.2, - "cache_read": 0.45 + "input": 0.78, + "output": 3.9, + "cache_read": 0.156, + "cache_write": 0.975, + "tiers": [ + { + "input": 1.56, + "output": 7.8, + "cache_read": 0.312, + "cache_write": 1.95, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 1.95, + "output": 9.75, + "cache_read": 0.39, + "cache_write": 2.4375, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "openai/o3-deep-research": { - "id": "openai/o3-deep-research", - "name": "o3-deep-research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen/qwen3-coder-30b-a3b-instruct": { + "id": "qwen/qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-06-27", - "last_updated": "2025-06-27", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -111828,43 +116932,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 9, - "output": 36, - "cache_read": 2.2 + "input": 0.07, + "output": 0.28 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "qwen/qwen3.8-omni-flash": { + "id": "qwen/qwen3.8-omni-flash", + "name": "Qwen3.8 Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-06-25", - "last_updated": "2025-06-25", + "structured_output": true, + "temperature": true, + "release_date": "2026-09-17", + "last_updated": "2026-09-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -111872,222 +116975,280 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.22, - "output": 1.8, - "cache_read": 0.022 + "input": 0.15, + "output": 0.47, + "cache_read": 0.016 } }, - "openai/gpt-image-1": { - "id": "openai/gpt-image-1", - "name": "GPT-Image-1", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "qwen/qwen3-vl-30b-a3b-instruct": { + "id": "qwen/qwen3-vl-30b-a3b-instruct", + "name": "Qwen3 VL 30B A3B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-03-31", - "last_updated": "2025-03-31", + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", "image" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 0 + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.13, + "output": 0.52 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + "qwen/qwen3-235b-a22b-thinking-2507": { + "id": "qwen/qwen3-235b-a22b-thinking-2507", + "name": "Qwen3 235B A22B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "structured_output": false, + "temperature": true, + "knowledge": "2025-06-30", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 131072, + "output": 117964 }, "cost": { - "input": 0.36, - "output": 1.4, - "cache_read": 0.09 + "input": 0.23, + "output": 2.3 } }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4-Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "qwen/qwen3-235b-a22b": { + "id": "qwen/qwen3-235b-a22b", + "name": "Qwen3 235B-A22B", + "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.455, + "output": 1.82 + } + }, + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2023-09-13", - "last_updated": "2023-09-13", + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 65536 }, "cost": { - "input": 9, - "output": 27 + "input": 0.26, + "output": 2.08 } }, - "openai/gpt-image-1-mini": { - "id": "openai/gpt-image-1-mini", - "name": "GPT-Image-1-Mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", - "attachment": true, + "qwen/qwen-plus-2025-07-28": { + "id": "qwen/qwen-plus-2025-07-28", + "name": "Qwen Plus 0728", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-08", + "last_updated": "2025-09-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 32768 + }, + "cost": { + "input": 0.26, + "output": 0.78, + "tiers": [ + { + "input": 0.78, + "output": 2.34, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.78, + "output": 2.34 + } } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "qwen/qwen3.6-35b-a3b": { + "id": "qwen/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 0.045, - "output": 0.36, - "cache_read": 0.0045 + "input": 0.15, + "output": 1, + "cache_read": 0.05 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4-Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 27, - "output": 160 + "input": 0.09, + "output": 1.1 } }, - "openai/o3-mini-high": { - "id": "openai/o3-mini-high", - "name": "o3-mini-high", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "qwen/qwen3-vl-30b-a3b-thinking": { + "id": "qwen/qwen3-vl-30b-a3b-thinking", + "name": "Qwen3 VL 30B A3B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", @@ -112097,74 +117258,86 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.99, - "output": 4 + "input": 0.2, + "output": 2.4 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5-Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 27.2727, - "output": 163.6364 + "input": 0.1875, + "output": 1.125, + "cache_write": 0.234375, + "tiers": [ + { + "input": 0.75, + "output": 3, + "cache_write": 0.9375, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.75, + "output": 3, + "cache_write": 0.9375 + } } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "qwen/qwen3-coder-flash": { + "id": "qwen/qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -112172,26 +117345,51 @@ }, "open_weights": false, "limit": { - "context": 124096, - "output": 4096 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.54, - "cache_read": 0.068 + "input": 0.195, + "output": 0.975, + "cache_read": 0.039, + "cache_write": 0.24375, + "tiers": [ + { + "input": 0.325, + "output": 1.625, + "cache_read": 0.065, + "cache_write": 0.40625, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.52, + "output": 2.6, + "cache_read": 0.104, + "cache_write": 0.65, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "openai/gpt-4o-mini-search": { - "id": "openai/gpt-4o-mini-search", - "name": "GPT-4o-mini-Search", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "qwen/qwen-2.5-coder-32b-instruct": { + "id": "qwen/qwen-2.5-coder-32b-instruct", + "name": "Qwen2.5 Coder 32B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2025-03-11", - "last_updated": "2025-03-11", + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-06-30", + "release_date": "2024-11-11", + "last_updated": "2024-11-11", "modalities": { "input": [ "text" @@ -112200,41 +117398,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 32768, + "output": 29491 }, "cost": { - "input": 0.14, - "output": 0.54 + "input": 0.66, + "output": 1 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "qwen/qwen3.5-flash-02-23": { + "id": "qwen/qwen3.5-flash-02-23", + "name": "Qwen3.5-Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "structured_output": true, + "temperature": true, + "release_date": "2026-02-25", + "last_updated": "2026-02-25", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -112242,39 +117437,34 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.1, - "output": 9 + "input": 0.065, + "output": 0.26 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, + "qwen/qwen3-max-thinking": { + "id": "qwen/qwen3-max-thinking", + "name": "Qwen3 Max Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "structured_output": true, + "temperature": true, + "release_date": "2026-02-09", + "last_updated": "2026-02-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -112282,31 +117472,48 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.6, - "output": 13, - "cache_read": 0.16 + "input": 0.78, + "output": 3.9, + "tiers": [ + { + "input": 1.56, + "output": 7.8, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 1.95, + "output": 9.75, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "openai/gpt-5.1-instant": { - "id": "openai/gpt-5.1-instant", - "name": "GPT-5.1-Instant", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-11-12", - "last_updated": "2025-11-12", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -112314,71 +117521,97 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 0.65, + "output": 3.25, + "cache_read": 0.13, + "cache_write": 0.8125, + "tiers": [ + { + "input": 1.17, + "output": 5.85, + "cache_read": 0.234, + "cache_write": 1.4625, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 1.95, + "output": 9.75, + "cache_read": 0.39, + "cache_write": 2.4375, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "openai/gpt-image-2": { - "id": "openai/gpt-image-2", - "name": "GPT-Image-2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "qwen/qwen3.5-9b": { + "id": "qwen/qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "output": 32768 }, "cost": { - "input": 5.0505, - "output": 32.3232, - "cache_read": 1.2626 + "input": 0.1, + "output": 0.15 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "qwen/qwen3.5-plus-20260420": { + "id": "qwen/qwen3.5-plus-20260420", + "name": "Qwen3.5 Plus 2026-04-20", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.5", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-11-12", - "last_updated": "2025-11-12", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -112386,123 +117619,176 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 0.3, + "output": 1.8, + "cache_write": 0.375, + "tiers": [ + { + "input": 0.375, + "output": 2.25, + "cache_write": 0.46875, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.375, + "output": 2.25, + "cache_write": 0.46875 + } } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "qwen/qwen3.8-flash": { + "id": "qwen/qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 4.5455, - "output": 27.2727, - "cache_read": 0.4545 + "input": 0.15, + "output": 0.47, + "cache_read": 0.016, + "cache_write": 0.2 } }, - "runwayml/runway": { - "id": "runwayml/runway", - "name": "Runway", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "runway", + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2024-10-11", - "last_updated": "2024-10-11", + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256, - "output": 0 + "context": 262144, + "output": 16384 + }, + "cost": { + "input": 0.3125, + "output": 1.25, + "cache_read": 0.15625 } }, - "runwayml/runway-gen-4-turbo": { - "id": "runwayml/runway-gen-4-turbo", - "name": "Runway-Gen-4-Turbo", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "runway", - "attachment": true, - "reasoning": false, + "qwen/qwen3.6-max-preview": { + "id": "qwen/qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-05-09", - "last_updated": "2025-05-09", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 256, - "output": 0 + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 1.027, + "output": 6.162, + "cache_write": 1.28375, + "tiers": [ + { + "input": 1.58, + "output": 9.48, + "cache_write": 1.975, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "cerebras/llama-3.1-8b-cs": { - "id": "cerebras/llama-3.1-8b-cs", - "name": "Llama-3.1-8B-CS", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": true, - "reasoning": false, + "qwen/qwen3-30b-a3b": { + "id": "qwen/qwen3-30b-a3b", + "name": "Qwen3 30B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-05-13", - "last_updated": "2025-05-13", + "structured_output": true, + "temperature": true, + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ "text" @@ -112511,62 +117797,72 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 0 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 0.12, + "output": 0.5 } }, - "cerebras/qwen3-235b-2507-cs": { - "id": "cerebras/qwen3-235b-2507-cs", - "name": "qwen3-235b-2507-cs", - "description": "Legacy model retained for compatibility with older integrations", + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "structured_output": true, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "output": 235929 }, - "status": "deprecated" + "cost": { + "input": 0.55, + "output": 3.5, + "cache_read": 0.225 + } }, - "cerebras/gpt-oss-120b-cs": { - "id": "cerebras/gpt-oss-120b-cs", - "name": "GPT-OSS-120B-CS", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "attachment": true, + "qwen/qwen3-8b": { + "id": "qwen/qwen3-8b", + "name": "Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "structured_output": false, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ "text" @@ -112575,192 +117871,196 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 0 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.35, - "output": 0.75 + "input": 0.117, + "output": 0.455 } }, - "cerebras/llama-3.3-70b-cs": { - "id": "cerebras/llama-3.3-70b-cs", - "name": "llama-3.3-70b-cs", - "description": "Legacy model retained for compatibility with older integrations", + "qwen/qwen3-vl-32b-instruct": { + "id": "qwen/qwen3-vl-32b-instruct", + "name": "Qwen3 VL 32B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-05-13", - "last_updated": "2025-05-13", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-23", + "last_updated": "2025-10-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "output": 32768 }, - "status": "deprecated" + "cost": { + "input": 0.104, + "output": 0.416 + } }, - "cerebras/qwen3-32b-cs": { - "id": "cerebras/qwen3-32b-cs", - "name": "qwen3-32b-cs", - "description": "Legacy model retained for compatibility with older integrations", + "qwen/qwen3-vl-8b-instruct": { + "id": "qwen/qwen3-vl-8b-instruct", + "name": "Qwen3 VL 8B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-05-15", - "last_updated": "2025-05-15", + "structured_output": true, + "temperature": true, + "release_date": "2025-10-14", + "last_updated": "2025-10-14", "modalities": { "input": [ + "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "output": 32768 }, - "status": "deprecated" + "cost": { + "input": 0.117, + "output": 0.455 + } }, - "anthropic/claude-sonnet-3.5": { - "id": "anthropic/claude-sonnet-3.5", - "name": "Claude-Sonnet-3.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-sonnet", - "attachment": true, + "qwen/qwen3-30b-a3b-instruct-2507": { + "id": "qwen/qwen3-30b-a3b-instruct-2507", + "name": "Qwen3 30B A3B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-06-05", - "last_updated": "2024-06-05", + "structured_output": true, + "temperature": true, + "knowledge": "2025-06-30", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 189096, - "output": 8192 + "context": 262144, + "output": 32000 }, - "status": "deprecated", "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "input": 0.04815, + "output": 0.19305 } }, - "anthropic/claude-sonnet-3.7": { - "id": "anthropic/claude-sonnet-3.7", - "name": "Claude-Sonnet-3.7", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "qwen/qwen2.5-vl-72b-instruct": { + "id": "qwen/qwen2.5-vl-72b-instruct", + "name": "Qwen2.5 VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-30", + "release_date": "2025-02-01", + "last_updated": "2025-02-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 196608, - "output": 128000 + "context": 128000, + "output": 115200 }, "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "input": 0.8, + "output": 1, + "cache_read": 0.4 } }, - "anthropic/claude-sonnet-3.5-june": { - "id": "anthropic/claude-sonnet-3.5-june", - "name": "Claude-Sonnet-3.5-June", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-sonnet", - "attachment": true, + "qwen/qwen-2.5-7b-instruct": { + "id": "qwen/qwen-2.5-7b-instruct", + "name": "Qwen2.5 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-11-18", - "last_updated": "2024-11-18", + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-30", + "release_date": "2024-10-16", + "last_updated": "2024-10-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 189096, - "output": 8192 + "context": 32768, + "output": 29491 }, - "status": "deprecated", "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "input": 0.1, + "output": 0.2 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude-Sonnet-4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "qwen/qwen3.5-plus-02-15": { + "id": "qwen/qwen3.5-plus-02-15", + "name": "Qwen3.5 Plus 2026-02-15", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 0, - "max": 31999 + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-09-26", - "last_updated": "2025-09-26", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -112768,106 +118068,152 @@ }, "open_weights": false, "limit": { - "context": 983040, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "input": 0.26, + "output": 1.56, + "tiers": [ + { + "input": 0.325, + "output": 1.95, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.325, + "output": 1.95 + } } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude-Sonnet-4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "qwen/qwen-2.5-72b-instruct": { + "id": "qwen/qwen-2.5-72b-instruct", + "name": "Qwen2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-30", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 0.36, + "output": 0.4 + } + }, + "qwen/qwen3-30b-a3b-thinking-2507": { + "id": "qwen/qwen3-30b-a3b-thinking-2507", + "name": "Qwen3 30B A3B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-05-21", - "last_updated": "2025-05-21", + "structured_output": false, + "temperature": true, + "knowledge": "2025-06-30", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 983040, - "output": 64000 + "context": 81920, + "output": 32768 }, "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "input": 0.2, + "output": 2.4 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude-Haiku-4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "qwen/qwen3.6-27b": { + "id": "qwen/qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 0, - "max": 63999 + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 192000, - "output": 64000 + "context": 262144, + "output": 262140 }, "cost": { - "input": 0.85, - "output": 4.3, - "cache_read": 0.085, - "cache_write": 1.1 + "input": 0.32, + "output": 2.7, + "cache_read": 0.15 } }, - "anthropic/claude-haiku-3": { - "id": "anthropic/claude-haiku-3", - "name": "Claude-Haiku-3", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2024-03-09", - "last_updated": "2024-03-09", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -112875,124 +118221,121 @@ }, "open_weights": false, "limit": { - "context": 189096, - "output": 8192 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.21, - "output": 1.1, - "cache_read": 0.021, - "cache_write": 0.26 + "input": 0.325, + "output": 1.95, + "cache_write": 0.40625, + "tiers": [ + { + "input": 1.3, + "output": 3.9, + "cache_write": 1.625, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.3, + "output": 3.9, + "cache_write": 1.625 + } } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude-Opus-4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "attachment": true, + "qwen/qwen3-14b": { + "id": "qwen/qwen3-14b", + "name": "Qwen3 14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2026-04-15", - "last_updated": "2026-04-15", + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 128000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 4.3, - "output": 21, - "cache_read": 0.43, - "cache_write": 5.4 + "input": 0.12, + "output": 0.24 } }, - "anthropic/claude-haiku-3.5": { - "id": "anthropic/claude-haiku-3.5", - "name": "Claude-Haiku-3.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "qwen/qwen3-vl-8b-thinking": { + "id": "qwen/qwen3-vl-8b-thinking", + "name": "Qwen3 VL 8B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", + "structured_output": true, + "temperature": true, + "release_date": "2025-10-14", + "last_updated": "2025-10-14", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 189096, - "output": 8192 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.68, - "output": 3.4, - "cache_read": 0.068, - "cache_write": 0.85 + "input": 0.18, + "output": 2.1 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude-Opus-4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -113000,85 +118343,108 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 4.2929, - "output": 21.4646 + "input": 0.32, + "output": 1.28, + "cache_read": 0.064, + "cache_write": 0.4, + "tiers": [ + { + "input": 0.96, + "output": 3.84, + "cache_read": 0.192, + "cache_write": 1.2, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.96, + "output": 3.84, + "cache_read": 0.192, + "cache_write": 1.2 + } } }, - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude-Opus-4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen/qwen3.8-27b:free": { + "id": "qwen/qwen3.8-27b:free", + "name": "Qwen3.8 27B (free)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 0, - "max": 31999 + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] } ], "tool_call": true, - "temperature": false, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 196608, - "output": 32000 + "context": 262144, + "output": 235929 }, "cost": { - "input": 13, - "output": 64, - "cache_read": 1.3, - "cache_write": 16 + "input": 0, + "output": 0 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude-Opus-4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen/qwen3.8-max-0902": { + "id": "qwen/qwen3.8-max-0902", + "name": "Qwen3.8 Max 0902", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high" + "high", + "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 63999 } ], "tool_call": true, - "temperature": false, - "release_date": "2025-11-21", - "last_updated": "2025-11-21", + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -113086,20 +118452,81 @@ }, "open_weights": false, "limit": { - "context": 196608, - "output": 64000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 4.3, - "output": 21, - "cache_read": 0.43, - "cache_write": 5.3 + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude-Sonnet-4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "morph/morph-v3-large": { + "id": "morph/morph-v3-large", + "name": "Morph V3 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "morph", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-07", + "last_updated": "2025-07-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.9, + "output": 1.9 + } + }, + "morph/morph-v3-fast": { + "id": "morph/morph-v3-fast", + "name": "Morph V3 Fast", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "morph", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-07-07", + "last_updated": "2025-07-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 81920, + "output": 38000 + }, + "cost": { + "input": 0.8, + "output": 1.2 + } + }, + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -113110,14 +118537,16 @@ "low", "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -113130,72 +118559,101 @@ }, "open_weights": false, "limit": { - "context": 983040, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Claude-Opus-4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5.4-image-2": { + "id": "openai/gpt-5.4-image-2", + "name": "GPT-5.4 Image 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": false, + "structured_output": true, "temperature": false, - "release_date": "2025-05-21", - "last_updated": "2025-05-21", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 192512, - "output": 28672 + "context": 272000, + "output": 128000 }, "cost": { - "input": 13, - "output": 64, - "cache_read": 1.3, - "cache_write": 16 + "input": 8, + "output": 15, + "cache_read": 2 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude-Opus-4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2026-02-04", - "last_updated": "2026-02-04", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -113208,53 +118666,74 @@ }, "open_weights": false, "limit": { - "context": 983040, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 4.3, - "output": 21, - "cache_read": 0.43, - "cache_write": 5.3 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "lumalabs/ray2": { - "id": "lumalabs/ray2", - "name": "Ray2", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ray", + "openai/gpt-audio": { + "id": "openai/gpt-audio", + "name": "GPT Audio", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-02-20", - "last_updated": "2025-02-20", + "structured_output": true, + "temperature": true, + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text", - "image" + "audio" ], "output": [ - "video" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 5000, - "output": 0 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10 } }, - "poetools/claude-code": { - "id": "poetools/claude-code", - "name": "claude-code", - "description": "Claude model for careful reasoning, writing, coding, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-11-27", - "last_updated": "2025-11-27", + "structured_output": true, + "temperature": true, + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" @@ -113265,210 +118744,357 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 16385, + "output": 4096 + }, + "cost": { + "input": 0.5, + "output": 1.5 } }, - "empiriolabs/deepseek-v4-pro-el": { - "id": "empiriolabs/deepseek-v4-pro-el", - "name": "DeepSeek-V4-Pro-EL", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "openai/gpt-6-astra-pro": { + "id": "openai/gpt-6-astra-pro", + "name": "GPT-6 Astra Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "release_date": "2026-04-24", - "last_updated": "2026-05-02", + "structured_output": true, + "temperature": false, + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 384000 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 1.67, - "output": 3.33 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } } }, - "empiriolabs/deepseek-v4-flash-el": { - "id": "empiriolabs/deepseek-v4-flash-el", - "name": "DeepSeek-V4-Flash-EL", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "release_date": "2026-04-24", - "last_updated": "2026-05-02", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 384000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.28 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "stabilityai/stablediffusionxl": { - "id": "stabilityai/stablediffusionxl", - "name": "StableDiffusionXL", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "stable-diffusion", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2023-07-09", - "last_updated": "2023-07-09", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text", - "image" + "pdf", + "image", + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 200, - "output": 0 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "ideogramai/ideogram": { - "id": "ideogramai/ideogram", - "name": "Ideogram", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ideogram", + "openai/gpt-4o-2024-05-13": { + "id": "openai/gpt-4o-2024-05-13", + "name": "GPT-4o (2024-05-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-04-03", - "last_updated": "2024-04-03", + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 150, - "output": 0 + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 5, + "output": 15 } }, - "ideogramai/ideogram-v2a": { - "id": "ideogramai/ideogram-v2a", - "name": "Ideogram-v2a", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ideogram", + "openai/gpt-4o-mini-2024-07-18": { + "id": "openai/gpt-4o-mini-2024-07-18", + "name": "GPT-4o-mini (2024-07-18)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "o-mini", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "structured_output": true, + "temperature": true, + "knowledge": "2023-10-31", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 150, - "output": 0 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "ideogramai/ideogram-v2": { - "id": "ideogramai/ideogram-v2", - "name": "Ideogram-v2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ideogram", + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2024-08-21", - "last_updated": "2024-08-21", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 150, - "output": 0 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "ideogramai/ideogram-v2a-turbo": { - "id": "ideogramai/ideogram-v2a-turbo", - "name": "Ideogram-v2a-Turbo", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ideogram", + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 150, - "output": 0 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "novita/glm-4.7": { - "id": "novita/glm-4.7", - "name": "glm-4.7", - "description": "Legacy model retained for compatibility with older integrations", + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -113476,29 +119102,32 @@ }, "open_weights": false, "limit": { - "context": 205000, - "output": 131072 + "context": 128000, + "output": 16384 }, - "status": "deprecated" + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } }, - "novita/minimax-m2.1": { - "id": "novita/minimax-m2.1", - "name": "minimax-m2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "openai/gpt-chat-latest": { + "id": "openai/gpt-chat-latest", + "name": "GPT Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-12-26", - "last_updated": "2025-12-26", + "release_date": "2026-05-05", + "last_updated": "2026-05-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -113506,85 +119135,78 @@ }, "open_weights": false, "limit": { - "context": 205000, - "output": 131072 + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "novita/glm-4.6": { - "id": "novita/glm-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "openai/gpt-audio-mini": { + "id": "openai/gpt-audio-mini", + "name": "GPT Audio Mini", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "o-mini", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "structured_output": true, + "temperature": true, + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 - } - }, - "novita/kimi-k2-thinking": { - "id": "novita/kimi-k2-thinking", - "name": "kimi-k2-thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": false, - "release_date": "2025-11-07", - "last_updated": "2025-11-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 128000, + "output": 16384 }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 0 + "cost": { + "input": 0.6, + "output": 2.4 } }, - "novita/kimi-k2.5": { - "id": "novita/kimi-k2.5", - "name": "Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "openai/gpt-5.6-sol-pro": { + "id": "openai/gpt-5.6-sol-pro", + "name": "GPT-5.6 Sol Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -113592,34 +119214,64 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 262144 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "novita/glm-4.6v": { - "id": "novita/glm-4.6v", - "name": "glm-4.6v", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -113627,52 +119279,65 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "novita/kimi-k2.6": { - "id": "novita/kimi-k2.6", - "name": "Kimi-K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2 Pro", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-05-02", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", "image", - "video" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.96, - "output": 4.04, - "cache_read": 0.16 + "input": 21, + "output": 168 } }, - "novita/glm-4.7-flash": { - "id": "novita/glm-4.7-flash", - "name": "glm-4.7-flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -113681,12 +119346,16 @@ } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ "text" @@ -113695,13 +119364,19 @@ "open_weights": false, "limit": { "context": 200000, - "output": 65500 + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "novita/glm-4.7-n": { - "id": "novita/glm-4.7-n", - "name": "glm-4.7-n", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -113710,12 +119385,15 @@ } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -113723,25 +119401,28 @@ }, "open_weights": false, "limit": { - "context": 205000, - "output": 131072 + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "novita/glm-5": { - "id": "novita/glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "openai/gpt-4": { + "id": "openai/gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ "text" @@ -113752,177 +119433,147 @@ }, "open_weights": false, "limit": { - "context": 205000, - "output": 131072 + "context": 8191, + "output": 4096 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 30, + "output": 60 } }, - "novita/deepseek-v3.2": { - "id": "novita/deepseek-v3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 0 - }, - "cost": { - "input": 0.27, - "output": 0.4, - "cache_read": 0.13 - } - } - } - }, - "kimi-for-coding": { - "id": "kimi-for-coding", - "env": [ - "KIMI_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://api.kimi.com/coding/v1", - "name": "Kimi For Coding", - "doc": "https://www.kimi.com/code/docs/en/third-party-tools/other-coding-agents.html", - "models": { - "k2p7": { - "id": "k2p7", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-11", - "last_updated": "2025-12", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1047576, "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "k2p5": { - "id": "k2p5", - "name": "Kimi K2.5", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "k2p6": { - "id": "k2p6", - "name": "Kimi K2.6", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, + "openai/o1": { + "id": "openai/o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -113931,152 +119582,86 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04", - "last_updated": "2026-04", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text", "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - } - } - }, - "dinference": { - "id": "dinference", - "env": [ - "DINFERENCE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.dinference.com/v1", - "name": "DInference", - "doc": "https://dinference.com", - "models": { - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 0.22, - "output": 0.88 - } - }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 200000, - "output": 128000 + "output": 100000 }, "cost": { - "input": 0.45, - "output": 1.65 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 3.89 + "input": 15, + "output": 120 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "openai/gpt-3.5-turbo-instruct": { + "id": "openai/gpt-3.5-turbo-instruct", + "name": "GPT-3.5 Turbo Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08", - "last_updated": "2025-08", + "knowledge": "2021-09-30", + "release_date": "2023-09-28", + "last_updated": "2023-09-28", "modalities": { "input": [ "text" @@ -114085,77 +119670,34 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 4095, + "output": 3685 }, "cost": { - "input": 0.0675, - "output": 0.27 + "input": 1.5, + "output": 2 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.75, - "output": 2.4 - } - } - } - }, - "perplexity-agent": { - "id": "perplexity-agent", - "env": [ - "PERPLEXITY_API_KEY" - ], - "npm": "@ai-sdk/openai", - "api": "https://api.perplexity.ai/v1", - "name": "Perplexity Agent", - "doc": "https://docs.perplexity.ai/docs/agent-api/models", - "models": { - "xai/grok-4-1-fast-non-reasoning": { - "id": "xai/grok-4-1-fast-non-reasoning", - "name": "Grok 4.1 Fast (Non-Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "openai/gpt-4o-2024-08-06": { + "id": "openai/gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "knowledge": "2023-09", + "release_date": "2024-08-06", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -114163,20 +119705,20 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "openai/gpt-6-astra": { + "id": "openai/gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-astra", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -114192,16 +119734,15 @@ } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-05", + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -114210,42 +119751,47 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, "tiers": [ { - "input": 2.5, - "output": 15, - "cache_read": 0.25, + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, "tier": { "type": "context", - "size": 200000 + "size": 272000 } } ], "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 } } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "openai/gpt-5.6-luna-pro": { + "id": "openai/gpt-5.6-luna-pro", + "name": "GPT-5.6 Luna Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -114255,16 +119801,15 @@ } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-05", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -114273,45 +119818,63 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", "image", - "video", - "audio", + "text", "pdf" ], "output": [ @@ -114320,122 +119883,69 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "openai/gpt-5-image-mini": { + "id": "openai/gpt-5-image-mini", + "name": "GPT-5 Image Mini", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2025-10-16", + "last_updated": "2025-10-16", "modalities": { "input": [ - "text", + "pdf", "image", - "video", - "audio", - "pdf" + "text" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "tiers": [ - { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 - } + "input": 2.5, + "output": 2, + "cache_read": 0.25 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -114443,43 +119953,37 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "openai/o3-pro": { + "id": "openai/o3-pro", + "name": "o3-pro", + "description": "High-effort o3 tier for difficult technical reasoning and careful answers", + "family": "o-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2024-05", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", "modalities": { "input": [ "text", + "pdf", "image" ], "output": [ @@ -114488,22 +119992,20 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 20, + "output": 80 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -114511,44 +120013,39 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.018, + "output": 0.09 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "openai/gpt-oss-safeguard-20b": { + "id": "openai/gpt-oss-safeguard-20b", + "name": "GPT OSS Safeguard 20B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -114556,66 +120053,58 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": true, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.075, + "output": 0.3, + "cache_read": 0.0375 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "openai/o3-mini-high": { + "id": "openai/o3-mini-high", + "name": "o3 Mini High", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "knowledge": "2023-10-31", + "release_date": "2025-02-12", + "last_updated": "2025-02-12", "modalities": { "input": [ "text", - "image", "pdf" ], "output": [ @@ -114624,40 +120113,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "Nemotron 3 Super 120B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "openai/gpt-3.5-turbo-16k": { + "id": "openai/gpt-3.5-turbo-16k", + "name": "GPT-3.5 Turbo 16k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "knowledge": "2021-09-30", + "release_date": "2023-08-28", + "last_updated": "2023-08-28", "modalities": { "input": [ "text" @@ -114666,27 +120143,62 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, + "context": 16385, + "output": 4096 + }, + "cost": { + "input": 3, + "output": 4 + } + }, + "openai/gpt-5.2-chat": { + "id": "openai/gpt-5.2-chat", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-10", + "last_updated": "2025-12-10", + "modalities": { + "input": [ + "pdf", + "image", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, "output": 32000 }, "cost": { - "input": 0.25, - "output": 2.5 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "anthropic/claude-opus-4-5": { - "id": "anthropic/claude-opus-4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-6-sol-pro": { + "id": "openai/gpt-6-sol-pro", + "name": "GPT-6 Sol Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -114696,60 +120208,116 @@ } ], "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "pdf", + "image", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } + } + }, + "openai/gpt-5-image": { + "id": "openai/gpt-5-image", + "name": "GPT-5 Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "knowledge": "2024-10-01", + "release_date": "2025-10-14", + "last_updated": "2025-10-14", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 10, + "output": 10, + "cache_read": 1.25 } }, - "anthropic/claude-sonnet-4-5": { - "id": "anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text", + "pdf", "image", - "pdf" + "text" ], "output": [ "text" @@ -114757,26 +120325,28 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -114786,10 +120356,11 @@ } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -114802,20 +120373,40 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } } }, - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -114825,21 +120416,20 @@ "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -114847,44 +120437,46 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", + "pdf", "image", - "pdf" + "text" ], "output": [ "text" @@ -114892,39 +120484,41 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -114937,30 +120531,50 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "perplexity/sonar": { - "id": "perplexity/sonar", - "name": "Sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar", - "attachment": false, + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -114968,125 +120582,98 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 2.5, - "cache_read": 0.0625 + "input": 2, + "output": 8, + "cache_read": 0.5 } - } - } - }, - "siliconflow": { - "id": "siliconflow", - "env": [ - "SILICONFLOW_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.siliconflow.com/v1", - "name": "SiliconFlow", - "doc": "https://cloud.siliconflow.com/models", - "models": { - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "moonshotai/Kimi-K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + }, + "openai/gpt-4o-2024-11-20": { + "id": "openai/gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-06-15", + "knowledge": "2023-09", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.77, - "output": 4, - "cache_read": 0.2 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "moonshotai/Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", + "openai/gpt-3.5-turbo-0613": { + "id": "openai/gpt-3.5-turbo-0613", + "name": "GPT-3.5 Turbo (older v0613)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "knowledge": "2021-09-30", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 4095, + "output": 3685 }, "cost": { - "input": 0.45, - "output": 2.25 + "input": 1, + "output": 2 } }, - "baidu/ERNIE-4.5-300B-A47B": { - "id": "baidu/ERNIE-4.5-300B-A47B", - "name": "baidu/ERNIE-4.5-300B-A47B", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "ernie", - "attachment": false, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-02", - "last_updated": "2025-11-25", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -115094,29 +120681,46 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.28, - "output": 1.1 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "ByteDance-Seed/Seed-OSS-36B-Instruct": { - "id": "ByteDance-Seed/Seed-OSS-36B-Instruct", - "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "seed", - "attachment": false, - "reasoning": false, + "openai/gpt-5.6-terra-pro": { + "id": "openai/gpt-5.6-terra-pro", + "name": "GPT-5.6 Terra Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-09-04", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -115124,30 +120728,66 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.21, - "output": 0.57 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "stepfun-ai/Step-3.5-Flash": { - "id": "stepfun-ai/Step-3.5-Flash", - "name": "stepfun-ai/Step-3.5-Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "family": "step", - "attachment": false, + "openai/gpt-6-luna": { + "id": "openai/gpt-6-luna", + "name": "GPT-6 Luna", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-luna", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -115155,29 +120795,61 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { "input": 0.1, - "output": 0.3 + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } } }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": false, - "reasoning": false, + "openai/o4-mini-high": { + "id": "openai/o4-mini-high", + "name": "o4 Mini High", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "knowledge": "2024-06-30", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ "text" @@ -115185,28 +120857,44 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "google/gemma-4-26B-A4B-it": { - "id": "google/gemma-4-26B-A4B-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, - "reasoning": false, + "openai/gpt-6-luna-pro": { + "id": "openai/gpt-6-luna-pro", + "name": "GPT-6 Luna Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ @@ -115215,29 +120903,65 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0.12, - "output": 0.4 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } } }, - "inclusionAI/Ling-flash-2.0": { - "id": "inclusionAI/Ling-flash-2.0", - "name": "inclusionAI/Ling-flash-2.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", - "attachment": false, - "reasoning": false, + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-09-18", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -115245,29 +120969,52 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": false, + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -115275,26 +121022,36 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 1.6 + "input": 10, + "output": 30 } }, - "Qwen/Qwen2.5-7B-Instruct": { - "id": "Qwen/Qwen2.5-7B-Instruct", - "name": "Qwen/Qwen2.5-7B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2025-11-25", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -115303,28 +121060,43 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 33000, - "output": 4000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.05, - "output": 0.05 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "Qwen/Qwen3-VL-235B-A22B-Instruct": { - "id": "Qwen/Qwen3-VL-235B-A22B-Instruct", - "name": "Qwen/Qwen3-VL-235B-A22B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-10-04", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -115336,29 +121108,39 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 0.25, + "output": 2, + "cache_read": 0.03 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -115366,29 +121148,44 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.3, - "output": 3.2 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -115396,36 +121193,47 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.39, - "output": 2.34 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-sol", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -115433,29 +121241,66 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.13, - "output": 0.6 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "name": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-6-sol": { + "id": "openai/gpt-6-sol", + "name": "GPT-6 Sol", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-07-31", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -115463,29 +121308,58 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "Qwen/Qwen3.5-122B-A10B": { - "id": "Qwen/Qwen3.5-122B-A10B", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, + "openai/o1-pro": { + "id": "openai/o1-pro", + "name": "o1-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2025-03-19", + "last_updated": "2025-03-19", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -115493,26 +121367,27 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.26, - "output": 2.08 + "input": 150, + "output": 600 } }, - "Qwen/Qwen3-Coder-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "name": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "microsoft/phi-4": { + "id": "microsoft/phi-4", + "name": "Phi 4", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2025-08-01", - "last_updated": "2025-11-25", + "knowledge": "2024-06-30", + "release_date": "2025-01-10", + "last_updated": "2025-01-10", "modalities": { "input": [ "text" @@ -115521,28 +121396,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 16384, + "output": 14745 }, "cost": { "input": 0.07, - "output": 0.28 + "output": 0.14 } }, - "Qwen/Qwen3.5-27B": { - "id": "Qwen/Qwen3.5-27B", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "microsoft/wizardlm-2-8x22b": { + "id": "microsoft/wizardlm-2-8x22b", + "name": "WizardLM-2 8x22B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2024-04-30", + "release_date": "2024-04-16", + "last_updated": "2024-04-16", "modalities": { "input": [ "text" @@ -115551,31 +121426,43 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 65535, + "output": 8000 }, "cost": { - "input": 0.25, - "output": 2 + "input": 0.62, + "output": 0.62 } - }, - "Qwen/Qwen3-30B-A3B-Instruct-2507": { - "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + } + } + }, + "perplexity": { + "id": "perplexity", + "env": [ + "PERPLEXITY_API_KEY" + ], + "npm": "@ai-sdk/perplexity", + "name": "Perplexity", + "doc": "https://docs.perplexity.ai", + "models": { + "sonar-pro": { + "id": "sonar-pro", + "name": "Sonar Pro", + "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", + "family": "sonar-pro", + "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2025-07-30", - "last_updated": "2025-11-25", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -115583,31 +121470,39 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.09, - "output": 0.3 + "input": 3, + "output": 15 } }, - "Qwen/Qwen3-VL-30B-A3B-Thinking": { - "id": "Qwen/Qwen3-VL-30B-A3B-Thinking", - "name": "Qwen/Qwen3-VL-30B-A3B-Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "sonar-deep-research": { + "id": "sonar-deep-research", + "name": "Perplexity Sonar Deep Research", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-11", - "last_updated": "2025-11-25", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-02-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -115615,30 +121510,30 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.29, - "output": 1 + "input": 2, + "output": 8, + "reasoning": 3 } }, - "Qwen/Qwen3-VL-8B-Instruct": { - "id": "Qwen/Qwen3-VL-8B-Instruct", - "name": "Qwen/Qwen3-VL-8B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "sonar": { + "id": "sonar", + "name": "Sonar", + "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", + "family": "sonar", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-11-25", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -115646,27 +121541,37 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.18, - "output": 0.68 + "input": 1, + "output": 1 } }, - "Qwen/Qwen3-VL-235B-A22B-Thinking": { - "id": "Qwen/Qwen3-VL-235B-A22B-Thinking", - "name": "Qwen/Qwen3-VL-235B-A22B-Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "sonar-reasoning-pro": { + "id": "sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", + "family": "sonar-reasoning", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, "temperature": true, - "release_date": "2025-10-04", - "last_updated": "2025-11-25", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text", @@ -115678,36 +121583,40 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.45, - "output": 3.5 + "input": 2, + "output": 8 } - }, - "Qwen/Qwen3-8B": { - "id": "Qwen/Qwen3-8B", - "name": "Qwen/Qwen3-8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + } + } + }, + "iteracompute": { + "id": "iteracompute", + "env": [ + "ITERACOMPUTE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.iteracompute.com/v1", + "name": "IteraCompute", + "doc": "https://iteracompute.com/docs.html", + "models": { + "deepseek/deepseek-v4-flash-0731": { + "id": "deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -115716,28 +121625,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 970000, + "output": 393216 }, "cost": { - "input": 0.06, - "output": 0.06 + "input": 0.34, + "output": 1.05, + "cache_read": 0.035 } }, - "Qwen/Qwen3.5-9B": { - "id": "Qwen/Qwen3.5-9B", - "name": "Qwen/Qwen3.5-9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "deepseek/deepseek-v4-pro-0813": { + "id": "deepseek/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-04-24", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -115746,68 +121657,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.1, - "output": 0.15 + "input": 1.1, + "output": 3.3, + "cache_read": 0.11 } }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen/Qwen3-32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "z-ai/glm-5.3-flash": { + "id": "z-ai/glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131000, - "output": 131000 - }, - "cost": { - "input": 0.14, - "output": 0.57 - } - }, - "Qwen/Qwen3-VL-32B-Instruct": { - "id": "Qwen/Qwen3-VL-32B-Instruct", - "name": "Qwen/Qwen3-VL-32B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-10-21", - "last_updated": "2025-11-25", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", @@ -115817,38 +121699,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 0.14, + "output": 0.49, + "cache_read": 0.03 } }, - "Qwen/Qwen3-14B": { - "id": "Qwen/Qwen3-14B", - "name": "Qwen/Qwen3-14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "z-ai/glm-5.3": { + "id": "z-ai/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -115857,58 +121731,62 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 1.2, + "output": 3.5, + "cache_read": 0.26 } }, - "Qwen/Qwen2.5-72B-Instruct": { - "id": "Qwen/Qwen2.5-72B-Instruct", - "name": "Qwen/Qwen2.5-72B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2025-11-25", + "structured_output": false, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 33000, - "output": 4000 + "context": 1048576, + "output": 999999 }, "cost": { - "input": 0.59, - "output": 0.59 + "input": 3, + "output": 14.9, + "cache_read": 0.29 } }, - "Qwen/Qwen3-VL-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", - "name": "Qwen/Qwen3-VL-30B-A3B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-10-05", - "last_updated": "2025-11-25", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", @@ -115918,29 +121796,37 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 1048576, + "output": 524288 }, "cost": { "input": 0.29, - "output": 1 + "output": 1.2, + "cache_read": 0.08 } }, - "Qwen/Qwen3-VL-32B-Thinking": { - "id": "Qwen/Qwen3-VL-32B-Thinking", - "name": "Qwen/Qwen3-VL-32B-Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "ornith-ai/ornith-1.5-35b-a3b": { + "id": "ornith-ai/ornith-1.5-35b-a3b", + "name": "Ornith 1.5 35B A3B", + "description": "Mixture-of-experts coding-reasoning model for agentic software tasks, tool use, and image understanding", + "family": "ornith", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-10-21", - "last_updated": "2025-11-25", + "release_date": "2026-08-18", + "last_updated": "2026-08-23", "modalities": { "input": [ "text", @@ -115950,95 +121836,135 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 327680, + "input": 262144, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 1.5 + "input": 0.3, + "output": 3, + "cache_read": 0.03 } }, - "Qwen/Qwen3.5-35B-A3B": { - "id": "Qwen/Qwen3.5-35B-A3B", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen/qwen3.8-27b": { + "id": "qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", "family": "qwen", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 327680, + "input": 262144, + "output": 65536 }, "cost": { - "input": 0.24, - "output": 1.8 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "openai/gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "qwen/qwen3.8-2.4t-a95b": { + "id": "qwen/qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-13", - "last_updated": "2025-11-25", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 8000 + "context": 970000, + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.45 + "input": 1.95, + "output": 5.95, + "cache_read": 0.2 } - }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "openai/gpt-oss-20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + } + } + }, + "the-grid-ai": { + "id": "the-grid-ai", + "env": [ + "THEGRID_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.thegrid.ai/v1", + "name": "The Grid AI", + "doc": "https://thegrid.ai/docs", + "models": { + "agent-prime": { + "id": "agent-prime", + "name": "Agent Prime", + "description": "Reliable models for dependable agentic applications, multi-step tool use, and reasoning workflows. Any model that meets the contract spec can serve your request.", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-08-13", - "last_updated": "2025-11-25", + "release_date": "2026-05-04", + "last_updated": "2026-07-15", "modalities": { "input": [ "text" @@ -116049,35 +121975,45 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 8000 + "context": 196608, + "input": 120000, + "output": 30000 }, - "cost": { - "input": 0.04, - "output": 0.18 - } + "status": "beta" }, - "tencent/Hy3-preview": { - "id": "tencent/Hy3-preview", - "name": "Hy3 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "code-max": { + "id": "code-max", + "name": "Code Max", + "description": "Frontier models for complex research, architectural decisions, debugging, and multi-file development. Any model that meets the contract spec can serve your request.", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-05-04", + "last_updated": "2026-07-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116085,40 +122021,45 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "input": 922000, + "output": 128000 }, - "cost": { - "input": 0.066, - "output": 0.26, - "cache_read": 0.029 - } + "status": "beta" }, - "tencent/Hunyuan-A13B-Instruct": { - "id": "tencent/Hunyuan-A13B-Instruct", - "name": "tencent/Hunyuan-A13B-Instruct", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", - "attachment": false, + "agent-max": { + "id": "agent-max", + "name": "Agent Max", + "description": "Frontier models for autonomous research, deep multi-step tool chains, and complex long-horizon tasks. Any model that meets the contract spec can serve your request.", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-11-25", + "temperature": false, + "release_date": "2026-05-04", + "last_updated": "2026-07-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116126,19 +122067,16 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "input": 922000, + "output": 128000 }, - "cost": { - "input": 0.14, - "output": 0.57 - } + "status": "beta" }, - "zai-org/GLM-5V-Turbo": { - "id": "zai-org/GLM-5V-Turbo", - "name": "zai-org/GLM-5V-Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "text-max": { + "id": "text-max", + "name": "Text Max", + "description": "Frontier models for deep reasoning, long context, and complex workflows. Any model that meets the contract spec can serve your request.", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -116146,18 +122084,24 @@ "type": "toggle" }, { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "structured_output": true, + "temperature": false, + "release_date": "2026-02-26", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", @@ -116169,27 +122113,25 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_write": 0 + "context": 1000000, + "input": 922000, + "output": 128000 } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "zai-org/GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "code-prime": { + "id": "code-prime", + "name": "Code Prime", + "description": "Reliable models for everyday software tasks, code completion, review, and standard debugging. Any model that meets the contract spec can serve your request.", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, @@ -116198,8 +122140,8 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-06-15", + "release_date": "2026-05-04", + "last_updated": "2026-07-15", "modalities": { "input": [ "text" @@ -116208,29 +122150,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 205000, - "output": 205000 + "context": 196608, + "input": 120000, + "output": 30000 }, - "cost": { - "input": 0.95, - "output": 2.55 - } + "status": "beta" }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "text-prime": { + "id": "text-prime", + "name": "Text Prime", + "description": "Reliable models for everyday text generation, editing, and analysis across diverse workflows. Any model that meets the contract spec can serve your request.", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", - "max" + "low", + "medium", + "high" ] } ], @@ -116240,39 +122180,8 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1049000, - "output": 262000 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_write": 0 - } - }, - "zai-org/GLM-4.5-Air": { - "id": "zai-org/GLM-4.5-Air", - "name": "zai-org/GLM-4.5-Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-11-25", + "release_date": "2026-02-26", + "last_updated": "2026-07-15", "modalities": { "input": [ "text" @@ -116283,26 +122192,25 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 - }, - "cost": { - "input": 0.14, - "output": 0.86 + "context": 196608, + "input": 120000, + "output": 30000 } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "zai-org/GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "code-standard": { + "id": "code-standard", + "name": "Code Standard", + "description": "Price-optimized models for rapid autocomplete, linting, high-frequency suggestions, and batch edits. Any model that meets the contract spec can serve your request.", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, @@ -116311,8 +122219,8 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "release_date": "2026-05-04", + "last_updated": "2026-07-15", "modalities": { "input": [ "text" @@ -116321,36 +122229,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 205000, - "output": 205000 + "context": 128000, + "input": 120000, + "output": 16000 }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_write": 0 - } + "status": "beta" }, - "deepseek-ai/DeepSeek-R1": { - "id": "deepseek-ai/DeepSeek-R1", - "name": "deepseek-ai/DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "text-standard": { + "id": "text-standard", + "name": "Text Standard", + "description": "Price-optimized models with low-latency, high-throughput and shorter maximum outputs. Any model that meets the contract spec can serve your request.", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-11-25", + "release_date": "2026-02-26", + "last_updated": "2026-07-15", "modalities": { "input": [ "text" @@ -116361,36 +122271,35 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 - }, - "cost": { - "input": 0.5, - "output": 2.18 + "context": 128000, + "input": 120000, + "output": 16000 } }, - "deepseek-ai/DeepSeek-V3.1-Terminus": { - "id": "deepseek-ai/DeepSeek-V3.1-Terminus", - "name": "deepseek-ai/DeepSeek-V3.1-Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "agent-standard": { + "id": "agent-standard", + "name": "Agent Standard", + "description": "Price-optimized models for fast tool calls, simple agent loops, high-throughput automation, and orchestration. Any model that meets the contract spec can serve your request.", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-11-25", + "release_date": "2026-05-04", + "last_updated": "2026-07-15", "modalities": { "input": [ "text" @@ -116401,39 +122310,56 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 128000, + "input": 120000, + "output": 16000 }, - "cost": { - "input": 0.27, - "output": 1 - } - }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "deepseek-ai/DeepSeek-V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "status": "beta" + } + } + }, + "meta": { + "id": "meta", + "env": [ + "META_MODEL_API_KEY" + ], + "npm": "@ai-sdk/openai", + "api": "https://api.meta.ai/v1", + "name": "Meta", + "doc": "https://dev.meta.ai/docs", + "models": { + "muse-spark-1.3": { + "id": "muse-spark-1.3", + "name": "Muse Spark 1.3", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-25", - "last_updated": "2025-11-25", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf", + "audio" ], "output": [ "text" @@ -116441,39 +122367,45 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.27, - "output": 1 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "deepseek-ai/DeepSeek-V3.2-Exp": { - "id": "deepseek-ai/DeepSeek-V3.2-Exp", - "name": "deepseek-ai/DeepSeek-V3.2-Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "muse-spark-1.1": { + "id": "muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-10", - "last_updated": "2025-11-25", + "release_date": "2026-04-08", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf", + "video" ], "output": [ "text" @@ -116481,123 +122413,140 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.27, - "output": 0.41 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "muse-spark-1.2": { + "id": "muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "muse-spark-1.2-contributor": { + "id": "muse-spark-1.2-contributor", + "name": "Muse Spark 1.2 Contributor", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.145 + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "deepseek-ai/DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "muse-spark-1.3-contributor": { + "id": "muse-spark-1.3-contributor", + "name": "Muse Spark 1.3 Contributor", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf", + "audio" ], "output": [ "text" @@ -116605,56 +122554,88 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.27, - "output": 0.42 + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 } - }, - "deepseek-ai/DeepSeek-V3": { - "id": "deepseek-ai/DeepSeek-V3", - "name": "deepseek-ai/DeepSeek-V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + } + } + }, + "cline-pass": { + "id": "cline-pass", + "env": [ + "CLINE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.cline.bot/api/v1", + "name": "ClinePass", + "doc": "https://docs.cline.bot/getting-started/clinepass", + "models": { + "cline-pass/mimo-v2.6-pro": { + "id": "cline-pass/mimo-v2.6-pro", + "name": "MiMo-V2.6-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2025-11-25", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 164000, - "output": 164000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMaxAI/MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "cline-pass/qwen3.7-max": { + "id": "cline-pass/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-06-15", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -116665,47 +122646,38 @@ }, "open_weights": false, "limit": { - "context": 197000, - "output": 131000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } - } - } - }, - "umans-ai-coding-plan": { - "id": "umans-ai-coding-plan", - "env": [ - "UMANS_AI_CODING_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.code.umans.ai/v1", - "name": "Umans AI Coding Plan", - "doc": "https://app.umans.ai/offers/code/docs", - "models": { - "umans-kimi-k2.7": { - "id": "umans-kimi-k2.7", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + }, + "cline-pass/mimo-v2.5": { + "id": "cline-pass/mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ @@ -116714,26 +122686,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "umans-glm-5.1": { - "id": "umans-glm-5.1", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "cline-pass/glm-5.3-flash": { + "id": "cline-pass/glm-5.3-flash", + "name": "cline-pass/glm-5.3-flash", + "description": "Latest natively multimodal model in the GLM-5 series", + "family": "glm-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, @@ -116742,99 +122718,104 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, + "context": 1000000, "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "umans-coder": { - "id": "umans-coder", - "name": "Umans Coder", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "cline-pass/qwen3.8-max": { + "id": "cline-pass/qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "umans-flash": { - "id": "umans-flash", - "name": "Umans Flash", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "cline-pass/kimi-k3": { + "id": "cline-pass/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -116842,43 +122823,39 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "umans-glm-5.2": { - "id": "umans-glm-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "cline-pass/deepseek-v4.1-flash": { + "id": "cline-pass/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "low", "high", "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -116890,48 +122867,38 @@ }, "open_weights": true, "limit": { - "context": 405504, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 } }, - "umans-qwen3.6-35b-a3b": { - "id": "umans-qwen3.6-35b-a3b", - "name": "Qwen3.6 35B A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "cline-pass/kimi-k2.6": { + "id": "cline-pass/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -116943,40 +122910,31 @@ "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } - } - } - }, - "io-net": { - "id": "io-net", - "env": [ - "IOINTELLIGENCE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.intelligence.io.solutions/api/v1", - "name": "IO.NET", - "doc": "https://io.net/docs/guides/intelligence/io-intelligence", - "models": { - "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { - "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", - "name": "Llama 4 Maverick 17B 128E Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + }, + "cline-pass/mimo-v2.5-pro": { + "id": "cline-pass/mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, "knowledge": "2024-12", - "release_date": "2025-01-15", - "last_updated": "2025-01-15", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -116984,31 +122942,44 @@ }, "open_weights": true, "limit": { - "context": 430000, - "output": 4096 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075, - "cache_write": 0.3 + "input": 1.74, + "output": 3.48, + "cache_read": 0.0145 } }, - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "cline-pass/mimo-v2.6-flash": { + "id": "cline-pass/mimo-v2.6-flash", + "name": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -117016,32 +122987,36 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.13, - "output": 0.38, - "cache_read": 0.065, - "cache_write": 0.26 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "meta-llama/Llama-3.2-90B-Vision-Instruct": { - "id": "meta-llama/Llama-3.2-90B-Vision-Instruct", - "name": "Llama 3.2 90B Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", - "attachment": false, - "reasoning": false, + "cline-pass/minimax-m3": { + "id": "cline-pass/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -117049,29 +123024,37 @@ }, "open_weights": true, "limit": { - "context": 16000, - "output": 4096 + "context": 1048576, + "output": 512000 }, "cost": { - "input": 0.35, - "output": 0.4, - "cache_read": 0.175, - "cache_write": 0.7 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "moonshotai/Kimi-K2-Thinking": { - "id": "moonshotai/Kimi-K2-Thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "cline-pass/glm-5.2": { + "id": "cline-pass/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2024-11-01", - "last_updated": "2024-11-01", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -117080,30 +123063,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 4096 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.55, - "output": 2.25, - "cache_read": 0.275, - "cache_write": 1.1 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "moonshotai/Kimi-K2-Instruct-0905": { - "id": "moonshotai/Kimi-K2-Instruct-0905", - "name": "Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "cline-pass/deepseek-v4-pro": { + "id": "cline-pass/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2024-09-05", - "last_updated": "2024-09-05", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -117112,63 +123105,84 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 4096 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.39, - "output": 1.9, - "cache_read": 0.195, - "cache_write": 0.78 + "input": 1.74, + "output": 3.48, + "cache_read": 0.0145 } }, - "Qwen/Qwen2.5-VL-32B-Instruct": { - "id": "Qwen/Qwen2.5-VL-32B-Instruct", - "name": "Qwen 2.5 VL 32B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "cline-pass/muse-spark-1.3-contributor": { + "id": "cline-pass/muse-spark-1.3-contributor", + "name": "Muse Spark 1.3 Contributor", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-11-01", - "last_updated": "2024-11-01", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 4096 - }, - "cost": { - "input": 0.05, - "output": 0.22, - "cache_read": 0.025, - "cache_write": 0.1 + "context": 1048576, + "output": 131072 } }, - "Qwen/Qwen3-Next-80B-A3B-Instruct": { - "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", - "name": "Qwen 3 Next 80B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "cline-pass/glm-5.3": { + "id": "cline-pass/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-10", - "last_updated": "2025-01-10", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -117179,32 +123193,34 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 4096 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.8, - "cache_read": 0.05, - "cache_write": 0.2 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen 3 235B Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "cline-pass/kimi-k2.7-code": { + "id": "cline-pass/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -117213,59 +123229,95 @@ "open_weights": true, "limit": { "context": 262144, - "output": 4096 + "output": 262144 }, "cost": { - "input": 0.11, - "output": 0.6, - "cache_read": 0.055, - "cache_write": 0.22 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT-OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": false, + "cline-pass/qwen3.7-plus": { + "id": "cline-pass/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 4096 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.04, - "output": 0.4, - "cache_read": 0.02, - "cache_write": 0.08 + "input": 0.4, + "output": 1.6, + "cache_read": 0.04, + "cache_write": 0.5, + "tiers": [ + { + "input": 1.2, + "output": 4.8, + "cache_read": 0.12, + "cache_write": 1.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.2, + "output": 4.8, + "cache_read": 0.12, + "cache_write": 1.5 + } } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT-OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "cline-pass/deepseek-v4-flash": { + "id": "cline-pass/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -117276,60 +123328,89 @@ }, "open_weights": true, "limit": { - "context": 64000, - "output": 4096 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.03, - "output": 0.14, - "cache_read": 0.015, - "cache_write": 0.06 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } - }, - "mistralai/Devstral-Small-2505": { - "id": "mistralai/Devstral-Small-2505", - "name": "Devstral Small 2505", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, + } + } + }, + "modal": { + "id": "modal", + "env": [ + "MODAL_PROXY_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.us-west.modal.direct/v1", + "name": "Modal", + "doc": "https://modal.com/docs/guide/endpoints", + "models": { + "thinkingmachines/Inkling-NVFP4": { + "id": "thinkingmachines/Inkling-NVFP4", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-05-01", - "last_updated": "2025-05-01", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 262144 }, "cost": { - "input": 0.05, - "output": 0.22, - "cache_read": 0.025, - "cache_write": 0.1 + "input": 1.2, + "output": 5, + "cache_read": 0.27 } }, - "mistralai/Magistral-Small-2506": { - "id": "mistralai/Magistral-Small-2506", - "name": "Magistral Small 2506", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-small", + "Qwen/Qwen3.8-2.4T-A95B": { + "id": "Qwen/Qwen3.8-2.4T-A95B", + "name": "Qwen3.8-Max", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-01", - "last_updated": "2025-06-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text" @@ -117338,30 +123419,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1010000, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.25, - "cache_write": 1 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "mistralai/Mistral-Large-Instruct-2411": { - "id": "mistralai/Mistral-Large-Instruct-2411", - "name": "Mistral Large Instruct 2411", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-11-01", - "last_updated": "2024-11-01", + "moonshotai/Kimi-K3": { + "id": "moonshotai/Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", @@ -117371,30 +123455,99 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 1, - "cache_write": 4 + "input": 3, + "output": 15, + "reasoning": 15, + "cache_read": 0.3 } }, - "mistralai/Mistral-Nemo-Instruct-2407": { - "id": "mistralai/Mistral-Nemo-Instruct-2407", - "name": "Mistral Nemo Instruct 2407", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + "zai-org/GLM-5.3-Flash": { + "id": "zai-org/GLM-5.3-Flash", + "name": "GLM 5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.45, + "output": 1.5, + "cache_read": 0.09 + } + } + } + }, + "coralbricks": { + "id": "coralbricks", + "env": [ + "CORAL_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.coralbricks.ai/v1", + "name": "CoralBricks", + "doc": "https://www.coralbricks.ai/docs", + "models": { + "glm-5.3-fp4": { + "id": "glm-5.3-fp4", + "name": "GLM 5.3 FP4", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-05", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -117405,31 +123558,44 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.02, - "output": 0.04, - "cache_read": 0.01, - "cache_write": 0.04 + "input": 1.12, + "output": 4.4, + "cache_read": 0, + "cache_write": 1.68 } }, - "Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar": { - "id": "Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar", - "name": "Qwen 3 Coder 480B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "glm-5.3-flash-fp4": { + "id": "glm-5.3-flash-fp4", + "name": "GLM 5.3 Flash FP4", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-15", - "last_updated": "2025-01-15", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -117437,28 +123603,43 @@ }, "open_weights": true, "limit": { - "context": 106000, - "output": 4096 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.22, - "output": 0.95, - "cache_read": 0.11, - "cache_write": 0.44 + "input": 0.15, + "output": 0.5, + "cache_read": 0, + "cache_write": 0.23 } }, - "zai-org/GLM-4.6": { - "id": "zai-org/GLM-4.6", - "name": "GLM 4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-11-15", - "last_updated": "2024-11-15", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -117467,31 +123648,43 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 4096 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.4, - "output": 1.75, - "cache_read": 0.2, - "cache_write": 0.8 + "input": 0.12, + "output": 0.6, + "cache_read": 0, + "cache_write": 0.18 } }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", - "name": "DeepSeek R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "deepseek-v4.1-flash-fast-fp4": { + "id": "deepseek-v4.1-flash-fast-fp4", + "name": "DeepSeek V4.1 Flash FP4", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-28", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text" @@ -117502,45 +123695,43 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2, - "output": 8.75, - "cache_read": 1, - "cache_write": 4 + "input": 0.3, + "output": 1.2, + "cache_read": 0, + "cache_write": 0.09 } } } }, - "gmicloud": { - "id": "gmicloud", + "routing-run": { + "id": "routing-run", "env": [ - "GMICLOUD_API_KEY" + "ROUTING_RUN_API_KEY" ], "npm": "@ai-sdk/openai-compatible", - "api": "https://api.gmi-serving.com/v1", - "name": "GMI Cloud", - "doc": "https://docs.gmicloud.ai/inference-engine/api-reference/llm-api-reference", + "api": "https://api.routing.run/v1", + "name": "routing.run", + "doc": "https://docs.routing.run/api-reference/models", "models": { - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "nemotron-3-ultra": { + "id": "nemotron-3-ultra", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -117551,19 +123742,18 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 131072, + "output": 32000 }, "cost": { - "input": 0.855, - "output": 3.6, - "cache_read": 0.144 + "input": 0.1, + "output": 0.1 } }, - "moonshotai/kimi-k2.7-code-highspeed": { - "id": "moonshotai/kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", "family": "kimi-k2", "attachment": true, "reasoning": true, @@ -117573,13 +123763,15 @@ "field": "reasoning_content" }, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -117587,72 +123779,67 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 + "input": 0.275, + "output": 1.1 } }, - "Qwen/Qwen3.7-Max": { - "id": "Qwen/Qwen3.7-Max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "output": 32000 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.25, - "cache_write": 3.125 + "input": 0.16, + "output": 0.48 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -117665,99 +123852,27 @@ }, "open_weights": false, "limit": { - "context": 1050000, + "context": 1000000, "input": 922000, "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - } - }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 409600, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } - }, - "cost": { - "input": 4.5, - "output": 22.5, - "cache_read": 0.45 + "input": 0.7, + "output": 4.2 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", + "claude-opus-4-8": { + "id": "claude-opus-4-8", "name": "Claude Opus 4.8", "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": false, "knowledge": "2026-01", "release_date": "2026-05-28", @@ -117775,71 +123890,29 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 32000 }, "cost": { "input": 5, - "output": 25, - "cache_read": 0.5 + "output": 25 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, + "glm-5.2-nitro": { + "id": "glm-5.2-nitro", + "name": "GLM 5.2 Nitro", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 409600, - "output": 64000 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 - } - ], - "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -117848,21 +123921,20 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 409600, - "output": 128000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.8, + "output": 2.4 } }, - "zai-org/GLM-5.1-FP8": { - "id": "zai-org/GLM-5.1-FP8", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, @@ -117873,8 +123945,8 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -117885,33 +123957,36 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.98, - "output": 3.08, - "cache_read": 0.182 + "input": 0.8, + "output": 2.4 } }, - "zai-org/GLM-5-FP8": { - "id": "zai-org/GLM-5-FP8", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "kimi-k2.6-nitro": { + "id": "kimi-k2.6-nitro", + "name": "Kimi K2.6 Nitro", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -117919,39 +123994,31 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.6, - "output": 1.92, - "cache_read": 0.12 + "input": 0.275, + "output": 1.1 } }, - "zai-org/GLM-5.2-FP8": { - "id": "zai-org/GLM-5.2-FP8", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -117963,56 +124030,55 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 131072 + "output": 64000 }, "cost": { - "input": 0.979, - "output": 3.08, - "cache_read": 0.182 + "input": 0.348, + "output": 0.696 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048575, - "output": 384000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.112, - "output": 0.224, - "cache_read": 0.022 + "input": 3, + "output": 15 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "kimi-k2.7-code-nitro": { + "id": "kimi-k2.7-code-nitro", + "name": "Kimi K2.7 Code Nitro", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, @@ -118020,13 +124086,15 @@ "field": "reasoning_content" }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -118034,173 +124102,196 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 384000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1.392, - "output": 2.784, - "cache_read": 0.116 + "input": 0.275, + "output": 1.1 } - } - } - }, - "xiaomi-token-plan-cn": { - "id": "xiaomi-token-plan-cn", - "env": [ - "XIAOMI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://token-plan-cn.xiaomimimo.com/v1", - "name": "Xiaomi Token Plan (China)", - "doc": "https://platform.xiaomimimo.com/#/docs", - "models": { - "mimo-v2.5-tts-voiceclone": { - "id": "mimo-v2.5-tts-voiceclone", - "name": "MiMo-V2.5-TTS-VoiceClone", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 1000000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.5, + "output": 9 } }, - "mimo-v2.5-tts-voicedesign": { - "id": "mimo-v2.5-tts-voicedesign", - "name": "MiMo-V2.5-TTS-VoiceDesign", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ - "audio" + "text" ] }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.275, + "output": 1.1 } }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 2.5, + "output": 15 } }, - "mimo-v2-tts": { - "id": "mimo-v2-tts", - "name": "MiMo-V2-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.112, + "output": 0.224 } - }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", + } + } + }, + "echo": { + "id": "echo", + "env": [ + "ECHO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://echo.tracerml.ai/v1", + "name": "Echo", + "doc": "https://echo.tracerml.ai/docs/api", + "models": { + "echo": { + "id": "echo", + "name": "Echo", + "description": "Adaptive model for coding, reasoning, and tool-driven agent workflows through one OpenAI-compatible endpoint", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-07-19", + "last_updated": "2026-08-16", "modalities": { "input": [ "text" @@ -118211,36 +124302,53 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 65536 }, - "status": "deprecated", + "status": "beta", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 10, + "output": 50 } - }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + } + } + }, + "neuralwatt": { + "id": "neuralwatt", + "env": [ + "NEURALWATT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.neuralwatt.com/v1", + "name": "Neuralwatt", + "doc": "https://portal.neuralwatt.com/docs", + "models": { + "glm-5.2-short-fast": { + "id": "glm-5.2-short-fast", + "name": "GLM 5.2 Short Fast", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "interleaved": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ "text" @@ -118251,79 +124359,87 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 199984, + "output": 32000 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.45, + "output": 4.5, + "cache_read": 0.145 } }, - "mimo-v2.5-tts": { - "id": "mimo-v2.5-tts", - "name": "MiMo-V2.5-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "kimi-k3-flex": { + "id": "kimi-k3-flex", + "name": "Kimi K3 Flex", + "description": "Kimi K3 on the flex tier: discounted, best-effort latency, requests may be held under load", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "audio" + "text" ] }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 1048560, + "output": 1048560 }, "cost": { - "input": 0, - "output": 0 + "input": 1.95, + "output": 9.75, + "cache_read": 0.195 } - } - } - }, - "zeldoc": { - "id": "zeldoc", - "env": [ - "ZELDOC_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.zeldoc.ai/v1", - "name": "Zeldoc", - "doc": "https://docs.zeldoc.ai", - "models": { - "z-code": { - "id": "z-code", - "name": "Z-Code", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + }, + "deepseek-v4-flash-flex": { + "id": "deepseek-v4-flash-flex", + "name": "DeepSeek V4 Flash Flex", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "high", "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "interleaved": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-15", - "last_updated": "2026-04-15", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -118332,44 +124448,47 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048560, + "output": 393216 }, "cost": { - "input": 0, - "output": 0 + "input": 0.091, + "output": 0.182, + "cache_read": 0.0182 } - } - } - }, - "scaleway": { - "id": "scaleway", - "env": [ - "SCALEWAY_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.scaleway.ai/v1", - "name": "Scaleway", - "doc": "https://www.scaleway.com/en/docs/generative-apis/", - "models": { - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", - "family": "qwen", - "attachment": false, + }, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, + "interleaved": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-01", - "last_updated": "2026-03-17", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -118377,27 +124496,41 @@ }, "open_weights": true, "limit": { - "context": 260000, - "output": 16384 + "context": 1048560, + "output": 1048560 }, "cost": { - "input": 0.75, - "output": 2.25, - "reasoning": 8.4 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", + "glm-5.2-short-flex": { + "id": "glm-5.2-short-flex", + "name": "GLM 5.2 Short Flex", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, + "interleaved": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2026-03-17", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ "text" @@ -118408,75 +124541,19 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.2, - "output": 0.8 - } - }, - "qwen3-embedding-8b": { - "id": "qwen3-embedding-8b", - "name": "Qwen3 Embedding 8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-25-11", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 4096 - }, - "cost": { - "input": 0.1, - "output": 0 - } - }, - "bge-multilingual-gemma2": { - "id": "bge-multilingual-gemma2", - "name": "BGE Multilingual Gemma2", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-07-26", - "last_updated": "2025-06-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8191, - "output": 3072 + "context": 199984, + "output": 32000 }, + "status": "deprecated", "cost": { - "input": 0.1, - "output": 0 + "input": 0.9425, + "output": 2.925, + "cache_read": 0.09425 } }, - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B A3B", + "qwen3.6-35b": { + "id": "qwen3.6-35b", + "name": "Qwen3.6 35B", "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", "family": "qwen", "attachment": true, @@ -118486,18 +124563,19 @@ "type": "effort", "values": [ "none", - "low", - "medium", "high" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-05-01", - "last_updated": "2026-05-22", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", @@ -118509,27 +124587,41 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131056, + "output": 131056 }, - "status": "beta", "cost": { - "input": 0.25, - "output": 1.5 + "input": 0.29, + "output": 1.15, + "cache_read": 0.029 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": true, - "reasoning": false, + "glm-5.2-fast": { + "id": "glm-5.2-fast", + "name": "GLM 5.2 Fast", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, + "interleaved": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2026-03-17", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ "text" @@ -118540,20 +124632,22 @@ }, "open_weights": true, "limit": { - "context": 100000, - "output": 16384 + "context": 1048560, + "output": 1048560 }, + "status": "deprecated", "cost": { - "input": 0.9, - "output": 0.9 + "input": 1.45, + "output": 4.5, + "cache_read": 0.145 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -118561,19 +124655,24 @@ "values": [ "none", "low", - "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "interleaved": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -118581,26 +124680,41 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 1048560, + "output": 1048560 }, "cost": { - "input": 1.8, - "output": 5.5 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "pixtral-12b-2409": { - "id": "pixtral-12b-2409", - "name": "Pixtral 12B 2409", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-09-25", - "last_updated": "2026-03-17", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -118612,26 +124726,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048560, + "output": 393216 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "mistral-small-3.2-24b-instruct-2506": { - "id": "mistral-small-3.2-24b-instruct-2506", - "name": "Mistral Small 3.2 24B Instruct (2506)", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, + "kimi-k3-fast": { + "id": "kimi-k3-fast", + "name": "Kimi K3 Fast", + "description": "Kimi K3 with thinking disabled for low-latency tool calling, vision, and JSON work", + "family": "kimi-k3", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2026-03-17", + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", @@ -118643,38 +124758,44 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 1048560, + "output": 1048560 }, "cost": { - "input": 0.15, - "output": 0.35 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT-OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "qwen3.6-35b-flex": { + "id": "qwen3.6-35b-flex", + "name": "Qwen3.6 35B Flex", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2026-03-17", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -118682,19 +124803,20 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 131056, + "output": 131056 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.1885, + "output": 0.7475, + "cache_read": 0.01885 } }, - "gemma-4-26b-a4b-it": { - "id": "gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "qwen-3.8-27b": { + "id": "qwen-3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -118704,16 +124826,19 @@ "none", "low", "medium", - "high" + "xhigh" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-01", - "last_updated": "2026-05-22", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", @@ -118725,40 +124850,44 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 262128, + "output": 131072 }, - "status": "beta", "cost": { - "input": 0.25, - "output": 0.5 + "input": 0.45, + "output": 3.2, + "cache_read": 0.25 } }, - "mistral-medium-3.5-128b": { - "id": "mistral-medium-3.5-128b", - "name": "Mistral Medium 3.5 128B", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, + "glm-5.2-short": { + "id": "glm-5.2-short", + "name": "GLM 5.2 Short", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, + "interleaved": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -118766,101 +124895,42 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 199984, + "output": 32000 }, + "status": "deprecated", "cost": { - "input": 1.5, - "output": 7.5 + "input": 1.45, + "output": 4.5, + "cache_read": 0.145 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "glm-5.3-flash-flex": { + "id": "glm-5.3-flash-flex", + "name": "GLM-5.3 Flash Flex", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 16384 - }, - "cost": { - "input": 0.6, - "output": 3.6 - } - }, - "whisper-large-v3": { - "id": "whisper-large-v3", - "name": "Whisper Large v3", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2023-09-01", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 0, - "output": 8192 - }, - "cost": { - "input": 0.003, - "output": 0 - } - }, - "gemma-3-27b-it": { - "id": "gemma-3-27b-it", - "name": "Gemma-3-27B-IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "interleaved": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-01", - "last_updated": "2026-03-17", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", @@ -118870,31 +124940,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 40000, - "output": 8192 + "context": 1048560, + "output": 1048560 }, "cost": { - "input": 0.25, - "output": 0.5 + "input": 0.0975, + "output": 0.325, + "cache_read": 0.0195 } }, - "voxtral-small-24b-2507": { - "id": "voxtral-small-24b-2507", - "name": "Voxtral Small 24B 2507", + "qwen3.6-35b-fast": { + "id": "qwen3.6-35b-fast", + "name": "Qwen3.6 35B Fast", "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "voxtral", + "family": "qwen3.6", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-07-01", - "last_updated": "2026-03-17", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", - "audio" + "image" ], "output": [ "text" @@ -118902,26 +124974,41 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 16384 + "context": 131056, + "output": 131056 }, "cost": { - "input": 0.15, - "output": 0.35 + "input": 0.29, + "output": 1.15, + "cache_read": 0.029 } }, - "devstral-2-123b-instruct-2512": { - "id": "devstral-2-123b-instruct-2512", - "name": "Devstral 2 123B Instruct (2512)", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", + "glm-5.2-flex": { + "id": "glm-5.2-flex", + "name": "GLM 5.2 Flex", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, + "interleaved": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-01-07", - "last_updated": "2026-03-17", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ "text" @@ -118932,38 +125019,40 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 1048560, + "output": 1048560 }, "status": "deprecated", "cost": { - "input": 0.4, - "output": 2 + "input": 0.9425, + "output": 2.925, + "cache_read": 0.09425 } - } - } - }, - "ovhcloud": { - "id": "ovhcloud", - "env": [ - "OVHCLOUD_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://oai.endpoints.kepler.ai.cloud.ovh.net/v1", - "name": "OVHcloud AI Endpoints", - "doc": "https://www.ovhcloud.com/en/public-cloud/ai-endpoints/catalog//", - "models": { - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder-30B-A3B-Instruct", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + }, + "deepseek-v4-flash-speed": { + "id": "deepseek-v4-flash-speed", + "name": "DeepSeek V4 Flash (Speed)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -118974,30 +125063,41 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048560, + "output": 393216 }, "cost": { - "input": 0.07, - "output": 0.26 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3-32B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM 5.2", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, + "interleaved": true, + "structured_output": false, "temperature": true, - "release_date": "2025-07-16", - "last_updated": "2025-07-16", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ "text" @@ -119008,24 +125108,42 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 1048560, + "output": 1048560 }, + "status": "deprecated", "cost": { - "input": 0.09, - "output": 0.25 + "input": 1.45, + "output": 4.5, + "cache_read": 0.145 } }, - "qwen3guard-gen-8b": { - "id": "qwen3guard-gen-8b", - "name": "Qwen3Guard-Gen-8B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "glm-5.2-short-fast-flex": { + "id": "glm-5.2-short-fast-flex", + "name": "GLM 5.2 Short Fast Flex", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": false, "temperature": true, - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ "text" @@ -119036,48 +125154,39 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 16384 - } - }, - "qwen3guard-gen-0.6b": { - "id": "qwen3guard-gen-0.6b", - "name": "Qwen3Guard-Gen-0.6B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-22", - "last_updated": "2026-01-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 199984, + "output": 32000 }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 16384 + "status": "deprecated", + "cost": { + "input": 0.9425, + "output": 2.925, + "cache_read": 0.09425 } }, - "meta-llama-3_3-70b-instruct": { - "id": "meta-llama-3_3-70b-instruct", - "name": "Meta-Llama-3_3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, + "kimi-k2.7-code-fast": { + "id": "kimi-k2.7-code-fast", + "name": "Kimi K2.7 Code Fast", + "description": "Kimi K2.7 Code with reasoning capped to a short budget for lower latency; reasoning cannot be disabled on this model", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], "tool_call": true, + "interleaved": true, "structured_output": true, - "temperature": true, - "release_date": "2025-04-01", - "last_updated": "2025-04-01", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -119085,25 +125194,37 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262128, + "output": 262128 }, "cost": { - "input": 0.74, - "output": 0.74 + "input": 0.95, + "output": 4, + "cache_read": 0.095 } }, - "mistral-small-3.2-24b-instruct-2506": { - "id": "mistral-small-3.2-24b-instruct-2506", - "name": "Mistral-Small-3.2-24B-Instruct-2506", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "gemma-4-31b": { + "id": "gemma-4-31b", + "name": "Gemma 4 31B", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "max" + ] + } + ], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-16", - "last_updated": "2025-07-16", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -119115,29 +125236,44 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262128, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.31 + "input": 0.144, + "output": 0.42, + "cache_read": 0.0144 } }, - "qwen2.5-vl-72b-instruct": { - "id": "qwen2.5-vl-72b-instruct", - "name": "Qwen2.5-VL-72B-Instruct", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "glm-5.3-flex": { + "id": "glm-5.3-flex", + "name": "GLM 5.3 Flex", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": false, "temperature": true, - "release_date": "2025-03-31", - "last_updated": "2025-03-31", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -119145,34 +125281,43 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 1048560, + "output": 1048560 }, "cost": { - "input": 1.01, - "output": 1.01 + "input": 0.9425, + "output": 2.925, + "cache_read": 0.09425 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "interleaved": true, "structured_output": true, - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -119183,28 +125328,47 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 1048560, + "output": 393216 }, + "status": "deprecated", "cost": { - "input": 0.09, - "output": 0.47 + "input": 1, + "output": 3, + "cache_read": 0.1 } }, - "mistral-7b-instruct-v0.3": { - "id": "mistral-7b-instruct-v0.3", - "name": "Mistral-7B-Instruct-v0.3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": false, + "qwen-3.8-27b-flex": { + "id": "qwen-3.8-27b-flex", + "name": "Qwen3.8 27B Flex", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-01", - "last_updated": "2025-04-01", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -119212,28 +125376,38 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 262128, + "output": 131072 }, "cost": { - "input": 0.11, - "output": 0.11 + "input": 0.2925, + "output": 2.08, + "cache_read": 0.1625 } }, - "mistral-nemo-instruct-2407": { - "id": "mistral-nemo-instruct-2407", - "name": "Mistral-Nemo-Instruct-2407", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": false, + "kimi-k2.7-code-flex": { + "id": "kimi-k2.7-code-flex", + "name": "Kimi K2.7 Code Flex", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], "tool_call": true, + "interleaved": true, "structured_output": true, - "temperature": true, - "release_date": "2024-11-20", - "last_updated": "2024-11-20", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -119241,41 +125415,44 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 262128, + "output": 262128 }, "cost": { - "input": 0.14, - "output": 0.14 + "input": 0.6175, + "output": 2.6, + "cache_read": 0.06175 } }, - "qwen3.6-27b": { - "id": "qwen3.6-27b", - "name": "Qwen3.6-27B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM 5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", - "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, + "interleaved": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -119283,36 +125460,34 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048560, + "output": 1048560 }, "cost": { - "input": 0.47, - "output": 3.19 + "input": 1.45, + "output": 4.5, + "cache_read": 0.145 } }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5-9B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "budget_tokens" } ], "tool_call": true, + "interleaved": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", @@ -119324,18 +125499,20 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 262128, + "output": 262128 }, "cost": { - "input": 0.12, - "output": 0.18 + "input": 0.95, + "output": 4, + "cache_read": 0.095 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5-397B-A17B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "deepseek-v4.1-flash-flex": { + "id": "deepseek-v4.1-flash-flex", + "name": "DeepSeek V4.1 Flash Flex", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -119344,16 +125521,19 @@ "values": [ "none", "low", - "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2026-05-18", - "last_updated": "2026-05-18", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -119365,34 +125545,39 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048560, + "output": 393216 }, "cost": { - "input": 0.71, - "output": 4.25 + "input": 0.0975, + "output": 0.39, + "cache_read": 0.00975 } }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "gpt-oss-20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "none", + "high", + "max" ] } ], "tool_call": true, + "interleaved": true, "structured_output": true, - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -119403,43 +125588,40 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 1048560, + "output": 393216 }, "cost": { - "input": 0.05, - "output": 0.18 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } } } }, - "friendli": { - "id": "friendli", + "gitlab": { + "id": "gitlab", "env": [ - "FRIENDLI_TOKEN" + "GITLAB_TOKEN" ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.friendli.ai/serverless/v1", - "name": "Friendli", - "doc": "https://friendli.ai/docs/guides/serverless_endpoints/introduction", + "npm": "gitlab-ai-provider", + "name": "GitLab Duo", + "doc": "https://docs.gitlab.com/user/duo_agent_platform/", "models": { - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "duo-chat-gpt-5-4-nano": { + "id": "duo-chat-gpt-5-4-nano", + "name": "Agentic Chat (GPT-5.4 Nano)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-nano", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -119449,248 +125631,256 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.4 + "input": 0, + "output": 0 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", - "family": "qwen", - "attachment": false, + "duo-chat-opus-5-5": { + "id": "duo-chat-opus-5-5", + "name": "Agentic Chat (Claude Opus 5.5)", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-29", - "last_updated": "2026-01-29", + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "duo-chat-gpt-6-astra": { + "id": "duo-chat-gpt-6-astra", + "name": "Agentic Chat (GPT-6 Astra)", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "duo-chat-opus-4-5": { + "id": "duo-chat-opus-4-5", + "name": "Agentic Chat (Claude Opus 4.5)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-03-31", + "release_date": "2026-01-08", + "last_updated": "2026-01-08", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 202752 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "duo-chat-gpt-5-mini": { + "id": "duo-chat-gpt-5-mini", + "name": "Agentic Chat (GPT-5 Mini)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.25 + "input": 0, + "output": 0 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, + "duo-chat-sonnet-4-5": { + "id": "duo-chat-sonnet-4-5", + "name": "Agentic Chat (Claude Sonnet 4.5)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-07-31", + "release_date": "2026-01-08", + "last_updated": "2026-01-08", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 196608 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - } - } - }, - "tencent-tokenhub": { - "id": "tencent-tokenhub", - "env": [ - "TENCENT_TOKENHUB_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://tokenhub.tencentmaas.com/v1", - "name": "Tencent TokenHub", - "doc": "https://cloud.tencent.com/document/product/1823/130050", - "models": { - "hy3": { - "id": "hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + }, + "duo-chat-gpt-5-6-sol": { + "id": "duo-chat-gpt-5-6-sol", + "name": "Agentic Chat (GPT-5.6 Sol)", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { "input": 0, @@ -119699,41 +125889,32 @@ "cache_write": 0 } }, - "hy3-preview": { - "id": "hy3-preview", - "name": "Hy3 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "duo-chat-sonnet-4-6": { + "id": "duo-chat-sonnet-4-6", + "name": "Agentic Chat (Claude Sonnet 4.6)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, + "context": 1000000, "output": 64000 }, "cost": { @@ -119742,227 +125923,252 @@ "cache_read": 0, "cache_write": 0 } - } - } - }, - "wandb": { - "id": "wandb", - "env": [ - "WANDB_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.inference.wandb.ai/v1", - "name": "Weights & Biases", - "doc": "https://docs.wandb.ai/guides/integrations/inference/", - "models": { - "ibm-granite/granite-4.1-8b": { - "id": "ibm-granite/granite-4.1-8b", - "name": "Granite 4.1 8B", - "description": "Granite 4.1 8B is a long-context instruct model capable of enhanced tool calling, instruction following, and chat capabilities.", - "family": "granite", - "attachment": false, - "reasoning": false, + }, + "duo-chat-gpt-5-4-mini": { + "id": "duo-chat-gpt-5-4-mini", + "name": "Agentic Chat (GPT-5.4 Mini)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.1, - "cache_read": 0.05 + "input": 0, + "output": 0 } }, - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama 3.3 70B", - "description": "Multilingual model excelling in conversational tasks, detailed instruction-following, and coding.", - "family": "llama", - "attachment": false, - "reasoning": false, + "duo-chat-haiku-4-5": { + "id": "duo-chat-haiku-4-5", + "name": "Agentic Chat (Claude Haiku 4.5)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "knowledge": "2025-02-28", + "release_date": "2026-01-08", + "last_updated": "2026-01-08", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.71, - "output": 0.71, - "cache_read": 0.71 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "meta-llama/Llama-3.1-70B-Instruct": { - "id": "meta-llama/Llama-3.1-70B-Instruct", - "name": "Llama 3.1 70B", - "description": "Efficient conversational model optimized for responsive multilingual chatbot interactions.", - "family": "llama", - "attachment": false, - "reasoning": false, + "duo-chat-fable-5-1": { + "id": "duo-chat-fable-5-1", + "name": "Agentic Chat (Claude Fable 5.1)", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.8, - "output": 0.8, - "cache_read": 0.8 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "meta-llama/Llama-3.1-8B-Instruct": { - "id": "meta-llama/Llama-3.1-8B-Instruct", - "name": "Llama 3.1 8B", - "description": "Efficient conversational model optimized for responsive multilingual chatbot interactions.", - "family": "llama", - "attachment": false, - "reasoning": false, + "duo-chat-gpt-5-5": { + "id": "duo-chat-gpt-5-5", + "name": "Agentic Chat (GPT-5.5)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 0.22, - "output": 0.22, - "cache_read": 0.22 + "input": 0, + "output": 0 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Kimi K2.6 is a multimodal Mixture-of-Experts language model featuring 32 billion activated parameters and a total of 1 trillion parameters.", - "family": "kimi-k2", + "duo-chat-fable-5": { + "id": "duo-chat-fable-5", + "name": "Agentic Chat (Claude Fable 5)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Kimi K2.5 is a multimodal Mixture-of-Experts language model featuring 32 billion activated parameters and a total of 1 trillion parameters.", - "family": "kimi-k2", + "duo-chat-gpt-5-3-codex": { + "id": "duo-chat-gpt-5-3-codex", + "name": "Agentic Chat (GPT-5.3 Codex)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-02", - "last_updated": "2026-02-02", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 0, + "output": 0 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Kimi K2.7 Code is a 1T-parameter MoE model with 32B active parameters purpose-built for long-horizon agentic coding and software engineering.", - "family": "kimi-k2", + "duo-chat-gpt-5-1": { + "id": "duo-chat-gpt-5-1", + "name": "Agentic Chat (GPT-5.1)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ "text", @@ -119972,513 +126178,698 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.94, - "output": 4, - "cache_read": 0.19 + "input": 0, + "output": 0 } }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B", - "description": "Gemma 4 31B Dense is designed for advanced reasoning, agentic workflows, and longer context and is natively trained on 140+ languages.", - "family": "gemma", + "duo-chat-opus-4-7": { + "id": "duo-chat-opus-4-7", + "name": "Agentic Chat (Claude Opus 4.7)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.12, - "output": 0.35, - "cache_read": 0.09 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "microsoft/Phi-4-mini-instruct": { - "id": "microsoft/Phi-4-mini-instruct", - "name": "Phi 4 Mini 3.8B", - "description": "Compact, efficient model ideal for fast responses in resource-constrained environments.", - "family": "phi", - "attachment": false, - "reasoning": false, + "duo-chat-gpt-5-2": { + "id": "duo-chat-gpt-5-2", + "name": "Agentic Chat (GPT-5.2)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-10", - "release_date": "2025-02-01", - "last_updated": "2025-02-01", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-01-23", + "last_updated": "2026-01-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 400000, + "input": 272000, "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.08, - "output": 0.35, - "cache_read": 0.08 + "input": 0, + "output": 0 } }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B A3B", - "description": "Qwen3.6-35B-A3B is an MoE multimodal model with 262K context optimized for agentic coding workflows.", - "family": "qwen3.6", + "duo-chat-opus-5": { + "id": "duo-chat-opus-5", + "name": "Agentic Chat (Claude Opus 5)", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "duo-chat-opus-4-6": { + "id": "duo-chat-opus-4-6", + "name": "Agentic Chat (Claude Opus 4.6)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2026-04-15", - "last_updated": "2026-04-15", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.25 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen3.6-27B is a 27B dense multimodal model with 262K context built for flagship-level agentic coding.", - "family": "qwen3.6", + "duo-chat-gpt-5-6-luna": { + "id": "duo-chat-gpt-5-6-luna", + "name": "Agentic Chat (GPT-5.6 Luna)", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3.6, - "cache_read": 0.12 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen3 235B A22B Thinking-2507", - "description": "High-performance Mixture-of-Experts model optimized for structured reasoning, math, and long-form generation.", - "family": "qwen", + "duo-chat-gpt-5-codex": { + "id": "duo-chat-gpt-5-codex", + "name": "Agentic Chat (GPT-5 Codex)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.1 + "input": 0, + "output": 0 } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "name": "Qwen3 Coder 480B A35B", - "description": "Mixture-of-Experts model optimized for agentic coding tasks such as function calling, tool use, and long-context reasoning.", - "family": "qwen", - "attachment": false, - "reasoning": false, + "duo-chat-sonnet-5": { + "id": "duo-chat-sonnet-5", + "name": "Agentic Chat (Claude Sonnet 5)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1, - "output": 1.5, - "cache_read": 1 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "Qwen/Qwen3.5-27B": { - "id": "Qwen/Qwen3.5-27B", - "name": "Qwen3.5-27B", - "description": "Qwen3.5-27B is a dense model from the Qwen3.5 family built for high performance across a large range of benchmarks.", - "family": "qwen3.5", + "duo-chat-opus-4-8": { + "id": "duo-chat-opus-4-8", + "name": "Agentic Chat (Claude Opus 4.8)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.39, - "output": 3.12, - "cache_read": 0.08 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "Qwen/Qwen3-30B-A3B-Instruct-2507": { - "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen3 30B A3B Instruct 2507", - "description": "Qwen3-30B-A3B-Instruct-2507 is a 30.5B MoE instruction-tuned model with enhanced reasoning, coding, and long-context understanding.", - "family": "qwen", - "attachment": false, - "reasoning": false, + "duo-chat-gpt-5-2-codex": { + "id": "duo-chat-gpt-5-2-codex", + "name": "Agentic Chat (GPT-5.2 Codex)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.1 + "input": 0, + "output": 0 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B-2507", - "description": "Efficient multilingual, Mixture-of-Experts, instruction-tuned model, optimized for logical reasoning.", - "family": "qwen", - "attachment": false, - "reasoning": false, + "duo-chat-gpt-5-4": { + "id": "duo-chat-gpt-5-4", + "name": "Agentic Chat (GPT-5.4)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.1 + "input": 0, + "output": 0 } }, - "Qwen/Qwen3.5-35B-A3B": { - "id": "Qwen/Qwen3.5-35B-A3B", - "name": "Qwen3.5-35B-A3B", - "description": "Qwen3.5-35B-A3B is an open-weights multimodal MoE model built for efficient, high-throughput inference across chat, reasoning, and agentic tasks.", - "family": "qwen3.5", + "duo-chat-gpt-5-6-terra": { + "id": "duo-chat-gpt-5-6-terra", + "name": "Agentic Chat (GPT-5.6 Terra)", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, + "context": 1050000, + "input": 922000, + "output": 128000 + }, "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.25 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Efficient Mixture-of-Experts model designed for high-reasoning, agentic and general-purpose use cases.", - "family": "gpt-oss", - "attachment": false, + } + } + }, + "infer": { + "id": "infer", + "env": [ + "INFER_API_KEY" + ], + "npm": "@ai-sdk/openai", + "api": "https://infer.flow7.org/v1", + "name": "Infer by Flow7", + "doc": "https://infer.flow7.org/opencode", + "models": { + "infer/gpt-6-astra:official": { + "id": "infer/gpt-6-astra:official", + "name": "GPT-6 Astra (Official API)", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 271999, + "output": 128000 + }, + "provider": { + "shape": "responses" }, "cost": { - "input": 0.04, - "output": 0.14, - "cache_read": 0.04 + "input": 12.5, + "output": 62.5, + "cache_read": 1.25, + "cache_write": 15.625 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "gpt-oss-20b", - "description": "Lower latency Mixture-of-Experts model trained on OpenAI's Harmony response format with reasoning capabilities.", - "family": "gpt-oss", - "attachment": false, + "infer/gpt-5.6-sol:official": { + "id": "infer/gpt-5.6-sol:official", + "name": "GPT-5.6 Sol (Official API)", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 271999, + "output": 128000 + }, + "provider": { + "shape": "responses" }, "cost": { - "input": 0.03, - "output": 0.13, - "cache_read": 0.03 + "input": 2.5, + "output": 12.5, + "cache_read": 0.25, + "cache_write": 3.125 } - }, - "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B": { - "id": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B", - "name": "Nemotron 3 Ultra", - "description": "Nemotron 3 Ultra is a powerful MoE model designed for long-running agents across coding, deep research, and enterprise automation.", - "family": "nemotron", - "attachment": false, + } + } + }, + "azure": { + "id": "azure", + "env": [ + "AZURE_RESOURCE_NAME", + "AZURE_API_KEY" + ], + "npm": "@ai-sdk/azure", + "name": "Azure", + "doc": "https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models", + "models": { + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "Grok 4.1 Fast (Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "release_date": "2025-06-27", + "last_updated": "2025-06-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 0.75, - "output": 2.75, - "cache_read": 0.15 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8": { - "id": "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8", - "name": "Nemotron 3 Super", - "description": "Nemotron 3 is a LatentMoE model designed to deliver strong agentic, reasoning, and conversational capabilities.", - "family": "nemotron", - "attachment": false, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.8, - "cache_read": 0.2 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "JetBrains/Mellum2-12B-A2.5B-Instruct": { - "id": "JetBrains/Mellum2-12B-A2.5B-Instruct", - "name": "Mellum2 12B A2.5B", - "description": "Mellum2-12B-A2.5B-Instruct is a fast MoE model with 131K context built for coding, tool use, and low-latency AI workflows.", + "deepseek-v3.2-speciale": { + "id": "deepseek-v3.2-speciale", + "name": "DeepSeek-V3.2-Speciale", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -120489,130 +126880,161 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.1, - "cache_read": 0.05 + "input": 0.58, + "output": 1.68 } }, - "OpenPipe/Qwen3-14B-Instruct": { - "id": "OpenPipe/Qwen3-14B-Instruct", - "name": "Qwen3 14B Instruct", - "description": "An efficient multilingual, dense, instruction-tuned model, optimized by OpenPipe for building agents with finetuning.", - "family": "qwen", - "attachment": false, - "reasoning": false, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "knowledge": "2025-02-28", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.05, - "output": 0.22, - "cache_read": 0.05 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM 5.2", - "description": "GLM-5.2 is a Mixture-of-Experts language model featuring 40 billion activated parameters and a total of 744 billion parameters.", - "family": "glm", - "attachment": false, + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.39, - "output": 4.4, - "cache_read": 0.26 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM 5.1", - "description": "Powerful MoE model for long-horizon agentic engineering and advanced reasoning.", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "Grok 4.1 Fast (Non-Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-06-27", + "last_updated": "2025-06-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 202752 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek V3.1", - "description": "A large hybrid model that supports both thinking and non-thinking modes via prompt templates.", - "family": "deepseek", + "phi-4-reasoning": { + "id": "phi-4-reasoning", + "name": "Phi-4-reasoning", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -120623,33 +127045,24 @@ }, "open_weights": true, "limit": { - "context": 161000, - "output": 161000 + "context": 32000, + "output": 4096 }, "cost": { - "input": 0.55, - "output": 1.65, - "cache_read": 0.55 + "input": 0.125, + "output": 0.5 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "DeepSeek V4-Flash is an MoE model with 1M context length great for coding, reasoning, and agentic workloads.", - "family": "deepseek", + "text-embedding-3-small": { + "id": "text-embedding-3-small", + "name": "text-embedding-3-small", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning": false, + "tool_call": false, + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -120658,167 +127071,171 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 8191, + "output": 1536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.07 + "input": 0.02, + "output": 0 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "DeepSeek V4-Pro is a 1.6T-parameter MoE model with 49B active parameters excelling at advanced reasoning, coding, and complex agentic workloads.", - "family": "deepseek", - "attachment": false, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 Nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.14 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax M2.5", - "description": "MoE model with a highly sparse architecture designed for high-throughput and low latency with strong coding capabilities.", - "family": "minimax-m2.5", + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 196608 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.3 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } - } - } - }, - "kuae-cloud-coding-plan": { - "id": "kuae-cloud-coding-plan", - "env": [ - "KUAE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://coding-plan-endpoint.kuaecloud.net/v1", - "name": "KUAE Cloud Coding Plan", - "doc": "https://docs.mthreads.com/kuaecloud/kuaecloud-doc-online/coding_plan/", - "models": { - "GLM-4.7": { - "id": "GLM-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + }, + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt-codex", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ - "text" + "text", + "image", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } - } - } - }, - "gitlab": { - "id": "gitlab", - "env": [ - "GITLAB_TOKEN" - ], - "npm": "gitlab-ai-provider", - "name": "GitLab Duo", - "doc": "https://docs.gitlab.com/user/duo_agent_platform/", - "models": { - "duo-chat-opus-4-5": { - "id": "duo-chat-opus-4-5", - "name": "Agentic Chat (Claude Opus 4.5)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + }, + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2026-01-08", - "last_updated": "2026-01-08", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -120826,64 +127243,68 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "duo-chat-opus-4-8": { - "id": "duo-chat-opus-4-8", - "name": "Agentic Chat (Claude Opus 4.8)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "ministral-3b": { + "id": "ministral-3b", + "name": "Ministral 3B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "knowledge": "2024-03", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.04, + "output": 0.04 } }, - "duo-chat-opus-4-7": { - "id": "duo-chat-opus-4-7", - "name": "Agentic Chat (Claude Opus 4.7)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-chat-latest": { + "id": "gpt-chat-latest", + "name": "GPT Chat Latest", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2025-12-01", + "release_date": "2026-05-05", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -120896,35 +127317,43 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "input": 111616, + "output": 16384 }, + "status": "beta", "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "duo-chat-gpt-5-2-codex": { - "id": "duo-chat-gpt-5-2-codex", - "name": "Agentic Chat (GPT-5.2 Codex)", + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "GPT-5-Codex", "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "family": "gpt-codex", - "attachment": true, + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -120937,39 +127366,38 @@ "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "duo-chat-fable-5": { - "id": "duo-chat-fable-5", - "name": "Agentic Chat (Claude Fable 5)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -120977,105 +127405,73 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 2, + "cache_read": 0.03 } }, - "duo-chat-gpt-5-5": { - "id": "duo-chat-gpt-5-5", - "name": "Agentic Chat (GPT-5.5)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "gpt-image-2": { + "id": "gpt-image-2", + "name": "GPT-Image-2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0, + "context": 0, "output": 0 - } - }, - "duo-chat-opus-4-6": { - "id": "duo-chat-opus-4-6", - "name": "Agentic Chat (Claude Opus 4.6)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 5, + "output": 30, + "cache_read": 1.25 } }, - "duo-chat-gpt-5-4": { - "id": "duo-chat-gpt-5-4", - "name": "Agentic Chat (GPT-5.4)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -121083,29 +127479,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 100000 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "duo-chat-gpt-5-codex": { - "id": "duo-chat-gpt-5-codex", - "name": "Agentic Chat (GPT-5 Codex)", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "llama-4-maverick-17b-128e-instruct-fp8": { + "id": "llama-4-maverick-17b-128e-instruct-fp8", + "name": "Llama 4 Maverick 17B 128E Instruct FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -121115,35 +127510,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.25, + "output": 1 } }, - "duo-chat-gpt-5-4-nano": { - "id": "duo-chat-gpt-5-4-nano", - "name": "Agentic Chat (GPT-5.4 Nano)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-nano", - "attachment": true, + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -121151,28 +127552,42 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 100000 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "duo-chat-sonnet-4-6": { - "id": "duo-chat-sonnet-4-6", - "name": "Agentic Chat (Claude Sonnet 4.6)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-08-01", "modalities": { "input": [ "text", @@ -121185,34 +127600,33 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 200000, "output": 64000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "duo-chat-gpt-5-mini": { - "id": "duo-chat-gpt-5-mini", - "name": "Agentic Chat (GPT-5 Mini)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "text-embedding-ada-002": { + "id": "text-embedding-ada-002", + "name": "text-embedding-ada-002", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2022-12-15", + "last_updated": "2022-12-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -121220,47 +127634,43 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 8192, + "output": 1536 }, "cost": { - "input": 0, + "input": 0.1, "output": 0 } }, - "duo-chat-sonnet-5": { - "id": "duo-chat-sonnet-5", - "name": "Agentic Chat (Claude Sonnet 5)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "knowledge": "2025-08-31", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -121268,30 +127678,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "duo-chat-gpt-5-4-mini": { - "id": "duo-chat-gpt-5-4-mini", - "name": "Agentic Chat (GPT-5.4 Mini)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-mini", + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -121303,34 +127711,43 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1047576, + "output": 32768 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "duo-chat-gpt-5-3-codex": { - "id": "duo-chat-gpt-5-3-codex", - "name": "Agentic Chat (GPT-5.3 Codex)", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -121343,59 +127760,63 @@ "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.05, + "output": 0.4, + "cache_read": 0.01 } }, - "duo-chat-haiku-4-5": { - "id": "duo-chat-haiku-4-5", - "name": "Agentic Chat (Claude Haiku 4.5)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "phi-4-multimodal": { + "id": "phi-4-multimodal", + "name": "Phi-4-multimodal", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "phi", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2026-01-08", - "last_updated": "2026-01-08", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text", "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.08, + "output": 0.32, + "input_audio": 4 } }, - "duo-chat-gpt-5-2": { - "id": "duo-chat-gpt-5-2", - "name": "Agentic Chat (GPT-5.2)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-23", - "last_updated": "2026-01-23", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", @@ -121405,30 +127826,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 0, - "output": 0 + "input": 0.95, + "output": 4 } }, - "duo-chat-sonnet-4-5": { - "id": "duo-chat-sonnet-4-5", - "name": "Agentic Chat (Claude Sonnet 4.5)", + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5", "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2025-07-31", - "release_date": "2026-01-08", - "last_updated": "2026-01-08", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", @@ -121444,31 +127875,47 @@ "context": 200000, "output": 64000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "duo-chat-gpt-5-1": { - "id": "duo-chat-gpt-5-1", - "name": "Agentic Chat (GPT-5.1)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "claude-opus-5-5": { + "id": "claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -121476,70 +127923,81 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + }, "cost": { - "input": 0, - "output": 0 + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 } - } - } - }, - "kilo": { - "id": "kilo", - "env": [ - "KILO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.kilo.ai/api/gateway", - "name": "Kilo Gateway", - "doc": "https://kilo.ai", - "models": { - "inclusionai/ling-2.6-1t": { - "id": "inclusionai/ling-2.6-1t", - "name": "inclusionAI: Ling-2.6-1T", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, + }, + "gpt-image-2.5-flare": { + "id": "gpt-image-2.5-flare", + "name": "GPT Image 2.5 Flare", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-04-23", - "last_updated": "2026-05-16", + "tool_call": false, + "temperature": false, + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 0, + "output": 0 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.06 + "input": 5, + "output": 30, + "cache_read": 1.25 } }, - "inclusionai/ring-2.6-1t": { - "id": "inclusionai/ring-2.6-1t", - "name": "inclusionAI: Ring-2.6-1T", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "claude-fable-5-1": { + "id": "claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-05-08", - "last_updated": "2026-05-16", + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -121547,28 +128005,36 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.075, - "output": 0.625, - "cache_read": 0.015 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "inclusionai/ling-2.6-flash": { - "id": "inclusionai/ling-2.6-flash", - "name": "inclusionAI: Ling-2.6 Flash", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, + "mistral-medium-2505": { + "id": "mistral-medium-2505", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-05-01", + "knowledge": "2025-05", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -121576,25 +128042,25 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.08, - "output": 0.24, - "cache_read": 0.016 + "input": 0.4, + "output": 2 } }, - "ibm-granite/granite-4.0-h-micro": { - "id": "ibm-granite/granite-4.0-h-micro", - "name": "IBM: Granite 4.0 Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "cohere-embed-v3-multilingual": { + "id": "cohere-embed-v3-multilingual", + "name": "Embed v3 Multilingual", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-10-20", - "last_updated": "2026-03-15", + "temperature": false, + "release_date": "2023-11-07", + "last_updated": "2023-11-07", "modalities": { "input": [ "text" @@ -121605,27 +128071,40 @@ }, "open_weights": true, "limit": { - "context": 131000, - "output": 32768 + "context": 512, + "output": 1024 }, "cost": { - "input": 0.017, - "output": 0.11 + "input": 0.1, + "output": 0 } }, - "ibm-granite/granite-4.1-8b": { - "id": "ibm-granite/granite-4.1-8b", - "name": "IBM: Granite 4.1 8B", - "description": "Tool-capable chat model for instruction following and agentic application workflows", + "o1": { + "id": "o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-05-01", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -121633,25 +128112,38 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "output": 100000 }, + "status": "deprecated", "cost": { - "input": 0.05, - "output": 0.1, - "cache_read": 0.05 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "meta-llama/llama-3.1-8b-instruct": { - "id": "meta-llama/llama-3.1-8b-instruct", - "name": "Meta: Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, + "codex-mini": { + "id": "codex-mini", + "name": "Codex Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2025-12-23", + "temperature": false, + "knowledge": "2024-04", + "release_date": "2025-05-16", + "last_updated": "2025-05-16", "modalities": { "input": [ "text" @@ -121660,26 +128152,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 200000, + "output": 100000 }, + "status": "deprecated", "cost": { - "input": 0.02, - "output": 0.05 + "input": 1.5, + "output": 6, + "cache_read": 0.375 } }, - "meta-llama/llama-3-70b-instruct": { - "id": "meta-llama/llama-3-70b-instruct", - "name": "Meta: Llama 3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "phi-4-reasoning-plus": { + "id": "phi-4-reasoning-plus", + "name": "Phi-4-reasoning-plus", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -121690,80 +128187,87 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8000 + "context": 32000, + "output": 4096 }, "cost": { - "input": 0.51, - "output": 0.74 + "input": 0.125, + "output": 0.5 } }, - "meta-llama/llama-3.1-70b-instruct": { - "id": "meta-llama/llama-3.1-70b-instruct", - "name": "Meta: Llama 3.1 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, + "claude-mythos-5": { + "id": "claude-mythos-5", + "name": "Claude Mythos 5", + "description": "Restricted Claude model for advanced cybersecurity and biology research workflows", + "family": "claude-mythos", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2024-07-16", - "last_updated": "2024-07-23", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 26215 - }, - "cost": { - "input": 0.4, - "output": 0.4 - } - }, - "meta-llama/llama-3.2-1b-instruct": { - "id": "meta-llama/llama-3.2-1b-instruct", - "name": "Meta: Llama 3.2 1B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2026-01-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 128000 }, - "open_weights": true, - "limit": { - "context": 60000, - "output": 12000 + "status": "beta", + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.027, - "output": 0.2 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "meta-llama/llama-4-maverick": { - "id": "meta-llama/llama-4-maverick", - "name": "Meta: Llama 4 Maverick", - "description": "Open multimodal Llama model for strong reasoning and fast responses", + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-04-05", - "last_updated": "2025-12-24", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", @@ -121773,55 +128277,59 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 16384 + "context": 400000, + "output": 272000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 15, + "output": 120 } }, - "meta-llama/llama-3.2-11b-vision-instruct": { - "id": "meta-llama/llama-3.2-11b-vision-instruct", - "name": "Meta: Llama 3.2 11B Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "attachment": true, + "gpt-3.5-turbo-instruct": { + "id": "gpt-3.5-turbo-instruct", + "name": "GPT-3.5 Turbo Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2021-08", + "release_date": "2023-09-21", + "last_updated": "2023-09-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 4096, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 0.049, - "output": 0.049 + "input": 1.5, + "output": 2 } }, - "meta-llama/llama-3.3-70b-instruct": { - "id": "meta-llama/llama-3.3-70b-instruct", - "name": "Meta: Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "gpt-3.5-turbo-0125": { + "id": "gpt-3.5-turbo-0125", + "name": "GPT-3.5 Turbo 0125", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2024-08-01", - "last_updated": "2026-02-04", + "knowledge": "2021-08", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -121830,26 +128338,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 16384, "output": 16384 }, + "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.32 + "input": 0.5, + "output": 1.5 } }, - "meta-llama/llama-guard-3-8b": { - "id": "meta-llama/llama-guard-3-8b", - "name": "Llama Guard 3 8B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "text-embedding-3-large": { + "id": "text-embedding-3-large", + "name": "text-embedding-3-large", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-04-18", - "last_updated": "2026-02-04", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -121858,58 +128367,76 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 26215 + "context": 8191, + "output": 3072 }, "cost": { - "input": 0.02, - "output": 0.06 + "input": 0.13, + "output": 0 } }, - "meta-llama/llama-guard-4-12b": { - "id": "meta-llama/llama-guard-4-12b", - "name": "Meta: Llama Guard 4 12B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-03-31", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 32768 + "context": 200000, + "output": 32000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.18, - "output": 0.18 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "meta-llama/llama-3-8b-instruct": { - "id": "meta-llama/llama-3-8b-instruct", - "name": "Meta: Llama 3 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, + "cohere-embed-v-4-0": { + "id": "cohere-embed-v-4-0", + "name": "Embed v4", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", + "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-04-25", - "last_updated": "2025-04-03", + "tool_call": false, + "temperature": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -121917,53 +128444,93 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 16384 + "context": 128000, + "output": 1536 }, "cost": { - "input": 0.03, - "output": 0.04 + "input": 0.12, + "output": 0 } }, - "meta-llama/llama-4-scout": { - "id": "meta-llama/llama-4-scout", - "name": "Meta: Llama 4 Scout", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 327680, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.08, - "output": 0.3 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } } }, - "meta-llama/llama-3.2-3b-instruct": { - "id": "meta-llama/llama-3.2-3b-instruct", - "name": "Meta: Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "codestral-2501": { + "id": "codestral-2501", + "name": "Codestral 25.01", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "codestral", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2026-03-15", + "knowledge": "2024-03", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -121972,118 +128539,115 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 80000, - "output": 16384 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.051, - "output": 0.34 + "input": 0.3, + "output": 0.9 } }, - "~anthropic/claude-haiku-latest": { - "id": "~anthropic/claude-haiku-latest", - "name": "Anthropic: Claude Haiku Latest", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "attachment": true, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "knowledge": "2025-01", + "release_date": "2026-02-06", + "last_updated": "2026-02-06", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.6, + "output": 3 } }, - "~anthropic/claude-sonnet-latest": { - "id": "~anthropic/claude-sonnet-latest", - "name": "Anthropic: Claude Sonnet Latest", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text", "image", - "pdf" + "audio" ], "output": [ - "text" + "text", + "image", + "audio" ] }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "~anthropic/claude-opus-latest": { - "id": "~anthropic/claude-opus-latest", - "name": "Anthropic: Claude Opus Latest", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -122096,9 +128660,11 @@ } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2026-04-16", - "last_updated": "2026-05-01", + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", @@ -122114,6 +128680,10 @@ "context": 1000000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + }, "cost": { "input": 5, "output": 25, @@ -122121,17 +128691,18 @@ "cache_write": 6.25 } }, - "kilo-auto/balanced": { - "id": "kilo-auto/balanced", - "name": "Kilo Auto Balanced", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-03-15", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -122140,31 +128711,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 3 + "input": 0.71, + "output": 0.71 } }, - "kilo-auto/small": { - "id": "kilo-auto/small", - "name": "Kilo Auto Small", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-03-15", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -122172,83 +128744,58 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 128000, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 0.05, - "output": 0.4 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "kilo-auto/free": { - "id": "kilo-auto/free", - "name": "Kilo Auto Free", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "cohere-command-a": { + "id": "cohere-command-a", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "kilo-auto/frontier": { - "id": "kilo-auto/frontier", - "name": "Kilo Auto Frontier", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-03-15", + "knowledge": "2024-06-01", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 5, - "output": 25 + "input": 2.5, + "output": 10 } }, - "moonshotai/kimi-k2": { - "id": "moonshotai/kimi-k2", - "name": "MoonshotAI: Kimi K2 0711", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "cohere-embed-v3-english": { + "id": "cohere-embed-v3-english", + "name": "Embed v3 English", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-07-11", - "last_updated": "2026-03-15", + "tool_call": false, + "temperature": false, + "release_date": "2023-11-07", + "last_updated": "2023-11-07", "modalities": { "input": [ "text" @@ -122259,282 +128806,298 @@ }, "open_weights": true, "limit": { - "context": 131000, - "output": 26215 + "context": 512, + "output": 1024 }, "cost": { - "input": 0.55, - "output": 2.2 + "input": 0.1, + "output": 0 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "MoonshotAI: Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "model-router": { + "id": "model-router", + "name": "Model Router", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "model-router", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-11-06", - "last_updated": "2026-03-15", + "release_date": "2025-05-19", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65535 + "context": 200000, + "output": 16384 }, "cost": { - "input": 0.47, - "output": 2, - "cache_read": 0.2 + "input": 0.14, + "output": 0 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "MoonshotAI: Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "gpt-image-2.5-sunburst": { + "id": "gpt-image-2.5-sunburst", + "name": "GPT Image 2.5 Sunburst", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-03-15", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65535 + "context": 0, + "output": 0 }, "cost": { - "input": 0.45, - "output": 2.2 + "input": 5, + "output": 30, + "cache_read": 1.25 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "MoonshotAI: Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-05-12", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65535 - }, - "cost": { - "input": 0.75, - "output": 3.5, - "cache_read": 0.375 - } - }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "MoonshotAI: Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-09-05", - "last_updated": "2025-09-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 128000 }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 26215 + "status": "beta", + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.15 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "baidu/ernie-4.5-300b-a47b": { - "id": "baidu/ernie-4.5-300b-a47b", - "name": "Baidu: ERNIE 4.5 300B A47B ", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, + "gpt-image-1": { + "id": "gpt-image-1", + "name": "GPT-Image-1", + "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", + "family": "gpt-image", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2026-01", + "temperature": false, + "release_date": "2025-04-24", + "last_updated": "2025-04-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 123000, - "output": 12000 + "context": 0, + "output": 0 }, "cost": { - "input": 0.28, - "output": 1.1 + "input": 5, + "output": 40, + "cache_read": 1.25 } }, - "baidu/ernie-4.5-vl-28b-a3b": { - "id": "baidu/ernie-4.5-vl-28b-a3b", - "name": "Baidu: ERNIE 4.5 VL 28B A3B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 Mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 30000, - "output": 8000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.56 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "baidu/ernie-4.5-vl-424b-a47b": { - "id": "baidu/ernie-4.5-vl-424b-a47b", - "name": "Baidu: ERNIE 4.5 VL 424B A47B ", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], - "tool_call": false, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2026-01", + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 123000, - "output": 16000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.42, - "output": 1.25 - } - }, - "baidu/ernie-4.5-21b-a3b-thinking": { - "id": "baidu/ernie-4.5-21b-a3b-thinking", - "name": "Baidu: ERNIE 4.5 21B A3B Thinking", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2025-09-19", - "last_updated": "2025-09-19", - "modalities": { - "input": [ - "text" + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 65536 - }, - "cost": { - "input": 0.07, - "output": 0.28 + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } } }, - "baidu/cobuddy:free": { - "id": "baidu/cobuddy:free", - "name": "Baidu: CoBuddy (free)", - "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", - "attachment": false, + "gpt-5.1-codex-max": { + "id": "gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2026-05-06", - "last_updated": "2026-05-07", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -122542,56 +129105,44 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "baidu/ernie-4.5-21b-a3b": { - "id": "baidu/ernie-4.5-21b-a3b", - "name": "Baidu: ERNIE 4.5 21B A3B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 120000, - "output": 8000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "baidu/qianfan-ocr-fast": { - "id": "baidu/qianfan-ocr-fast", - "name": "Baidu: Qianfan-OCR-Fast", - "description": "OCR model for extracting structured text from documents and screenshots", + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-05-16", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -122599,30 +129150,45 @@ }, "open_weights": false, "limit": { - "context": 65536, - "output": 28672 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.68, - "output": 2.81 + "input": 1.75, + "output": 14, + "cache_read": 0.125 } }, - "perceptron/perceptron-mk1": { - "id": "perceptron/perceptron-mk1", - "name": "Perceptron: Perceptron Mk1", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-16", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2025-12-31", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "image", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -122630,24 +129196,50 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "alfredpros/codellama-7b-instruct-solidity": { - "id": "alfredpros/codellama-7b-instruct-solidity", - "name": "AlfredPros: CodeLLaMa 7B Instruct Solidity", + "phi-4": { + "id": "phi-4", + "name": "Phi-4", "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2025-04-14", - "last_updated": "2026-03-15", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -122658,24 +129250,26 @@ }, "open_weights": true, "limit": { - "context": 4096, + "context": 128000, "output": 4096 }, "cost": { - "input": 0.8, - "output": 1.2 + "input": 0.125, + "output": 0.5 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Google: Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -122684,16 +129278,16 @@ } ], "tool_call": true, - "temperature": true, - "release_date": "2026-05-07", - "last_updated": "2026-05-16", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -122701,114 +129295,110 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.08333 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "google/gemma-3n-e4b-it": { - "id": "google/gemma-3n-e4b-it", - "name": "Google: Gemma 3n 4B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": false, + "mistral-small-2503": { + "id": "mistral-small-2503", + "name": "Mistral Small 3.1", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "knowledge": "2024-09", + "release_date": "2025-03-01", + "last_updated": "2025-03-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 6554 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.02, - "output": 0.04 + "input": 0.1, + "output": 0.3 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Google: Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "attachment": true, + "phi-4-mini-reasoning": { + "id": "phi-4-mini-reasoning", + "name": "Phi-4-mini-reasoning", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-03-20", - "last_updated": "2026-03-15", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ - "audio", - "image", - "pdf", - "text", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.25, - "output": 10, - "reasoning": 10, - "cache_read": 0.125, - "cache_write": 0.375 + "input": 0.075, + "output": 0.3 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Google: Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-07-17", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image" ], "output": [ "text" @@ -122816,45 +129406,49 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 1047576, + "output": 32768 }, + "status": "deprecated", "cost": { - "input": 0.3, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Google: Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-05-19", - "last_updated": "2026-05-27", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -122862,34 +129456,52 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 1.5, - "output": 9, - "reasoning": 9, - "cache_read": 0.15, - "cache_write": 0.08333 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "google/gemini-2.0-flash-lite-001": { - "id": "google/gemini-2.0-flash-lite-001", - "name": "Google: Gemini 2.0 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "temperature": true, - "release_date": "2024-12-11", - "last_updated": "2026-03-15", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-07-17", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -122897,36 +129509,54 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 8192 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "google/gemini-2.5-flash-lite-preview-09-2025": { - "id": "google/gemini-2.5-flash-lite-preview-09-2025", - "name": "Google: Gemini 2.5 Flash Lite Preview 09-2025", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "gpt-4-turbo-vision": { + "id": "gpt-4-turbo-vision", + "name": "GPT-4 Turbo Vision", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-09-25", - "last_updated": "2026-03-15", + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image" ], "output": [ "text" @@ -122934,34 +129564,30 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.4, - "reasoning": 0.4, - "cache_read": 0.01, - "cache_write": 0.083333 + "input": 10, + "output": 30 } }, - "google/gemini-2.0-flash-001": { - "id": "google/gemini-2.0-flash-001", - "name": "Google: Gemini 2.0 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "attachment": true, + "grok-4-20-non-reasoning": { + "id": "grok-4-20-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-12-11", - "last_updated": "2026-03-15", + "knowledge": "2025-09", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ - "audio", - "image", - "pdf", - "text", - "video" + "text" ], "output": [ "text" @@ -122969,109 +129595,110 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 262000, "output": 8192 }, + "status": "beta", "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025, - "cache_write": 0.083333 + "input": 2, + "output": 6 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Google: Gemma 4 31B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "gpt-3.5-turbo-1106": { + "id": "gpt-3.5-turbo-1106", + "name": "GPT-3.5 Turbo 1106", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-11", + "knowledge": "2021-08", + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 16384, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 0.14, - "output": 0.4 + "input": 1, + "output": 2 } }, - "google/lyria-3-clip-preview": { - "id": "google/lyria-3-clip-preview", - "name": "Google: Lyria 3 Clip Preview", - "description": "Speech generation model for controllable voice, narration, and audio delivery", + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-03-30", - "last_updated": "2026-04-11", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ - "audio", "text" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1047576, + "output": 32768 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Google: Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "gpt-6-luna": { + "id": "gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -123079,114 +129706,137 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } } }, - "google/gemini-3-pro-image-preview": { - "id": "google/gemini-3-pro-image-preview", - "name": "Google: Nano Banana Pro (Gemini 3 Pro Image Preview)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": false, "temperature": true, - "release_date": "2025-11-20", - "last_updated": "2026-03-15", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "image", "text" ], "output": [ - "image", "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 32768 + "context": 163840, + "output": 163840 }, + "status": "deprecated", "cost": { - "input": 2, - "output": 12, - "reasoning": 12 + "input": 1.35, + "output": 5.4 } }, - "google/gemini-2.5-flash-image": { - "id": "google/gemini-2.5-flash-image", - "name": "Google: Nano Banana (Gemini 2.5 Flash Image)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": false, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek-V4-Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-10-08", - "last_updated": "2026-03-15", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "image", "text" ], "output": [ - "image", "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 1000000, + "output": 384000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 1.74, + "output": 3.48 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Google: Gemini 2.5 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] }, { "type": "budget_tokens", - "min": 0, - "max": 24576 + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-06-17", - "last_updated": "2026-03-15", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -123194,80 +129844,99 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.1, - "output": 0.4, - "reasoning": 0.4, - "cache_read": 0.01, - "cache_write": 0.083333 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "google/gemini-3.1-flash-image-preview": { - "id": "google/gemini-3.1-flash-image-preview", - "name": "Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "llama-4-scout-17b-16e-instruct": { + "id": "llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout 17B 16E Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-03-15", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ - "image", "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.2, + "output": 0.78 } }, - "google/gemini-2.5-pro-preview-05-06": { - "id": "google/gemini-2.5-pro-preview-05-06", - "name": "Google: Gemini 2.5 Pro Preview 05-06", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "gpt-image-1.5": { + "id": "gpt-image-1.5", + "name": "GPT-Image-1.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-11-25", + "last_updated": "2025-11-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + }, + "cost": { + "input": 5, + "output": 32, + "cache_read": 1.25 + } + }, + "grok-4-20-reasoning": { + "id": "grok-4-20-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-05-06", - "last_updated": "2026-03-15", + "knowledge": "2025-09", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ - "audio", - "image", - "pdf", - "text", - "video" + "text" ], "output": [ "text" @@ -123275,45 +129944,46 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 262000, + "output": 8192 }, + "status": "beta", "cost": { - "input": 1.25, - "output": 10, - "reasoning": 10, - "cache_read": 0.125, - "cache_write": 0.375 + "input": 2, + "output": 6 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Google: Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-19", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -123321,35 +129991,51 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { "input": 2, "output": 12, - "reasoning": 12 + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Google: Gemma 4 26B A4B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": true, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-04-03", - "last_updated": "2026-04-11", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" @@ -123357,35 +130043,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.12, - "output": 0.4 + "input": 0.58, + "output": 1.68 } }, - "google/gemini-2.5-pro-preview": { - "id": "google/gemini-2.5-pro-preview", - "name": "Google: Gemini 2.5 Pro Preview 06-05", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "gpt-4-turbo": { + "id": "gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-06-05", - "last_updated": "2026-03-15", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "audio", - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -123393,22 +130074,21 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 1.25, - "output": 10, - "reasoning": 10, - "cache_read": 0.125, - "cache_write": 0.375 + "input": 10, + "output": 30 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Google: Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "attachment": true, + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -123416,22 +130096,20 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2025-12-17", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image" ], "output": [ "text" @@ -123439,90 +130117,129 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 3, - "reasoning": 3, - "cache_read": 0.05, - "cache_write": 0.083333 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "google/gemma-3-12b-it": { - "id": "google/gemma-3-12b-it", - "name": "Google: Gemma 3 12B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "o3": { + "id": "o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-03-13", - "last_updated": "2026-03-15", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.04, - "output": 0.13, - "cache_read": 0.015 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "google/gemma-3-4b-it": { - "id": "google/gemma-3-4b-it", - "name": "Google: Gemma 3 4B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-03-13", - "last_updated": "2026-03-15", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 19200 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.04, - "output": 0.08 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Google: Gemma 3 27B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2025-03-12", - "last_updated": "2026-03-15", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -123530,101 +130247,135 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 65536 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 0.03, - "output": 0.11, - "cache_read": 0.02 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "google/lyria-3-pro-preview": { - "id": "google/lyria-3-pro-preview", - "name": "Google: Lyria 3 Pro Preview", - "description": "Speech generation model for controllable voice, narration, and audio delivery", + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-03-30", - "last_updated": "2026-04-11", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ - "audio", "text" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "google/gemma-2-27b-it": { - "id": "google/gemma-2-27b-it", - "name": "Google: Gemma 2 27B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": false, - "reasoning": false, - "tool_call": false, + "grok-4.6": { + "id": "grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-06-24", - "last_updated": "2024-06-24", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 2048 + "context": 200000, + "output": 128000 }, - "cost": { - "input": 0.65, - "output": 0.65 - } + "status": "beta" }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Google: Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -123632,25 +130383,79 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5 + "input": 4, + "output": 20, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "liquid/lfm-2-24b-a2b": { - "id": "liquid/lfm-2-24b-a2b", - "name": "LiquidAI: LFM2-24B-A2B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "phi-4-mini": { + "id": "phi-4-mini", + "name": "Phi-4-mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.075, + "output": 0.3 + } + }, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek-V4-Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-03-15", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -123661,34 +130466,50 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 1000000, + "output": 384000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 0.03, - "output": 0.12 + "input": 0.19, + "output": 0.51 } }, - "x-ai/grok-4.20": { - "id": "x-ai/grok-4.20", - "name": "xAI: Grok 4.20", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "gpt-6-sol": { + "id": "gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-31", - "last_updated": "2026-04-11", + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -123696,19 +130517,52 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { "input": 2, - "output": 6, - "cache_read": 0.2 + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } } - }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "xAI: Grok 4.3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + } + } + }, + "freemodel": { + "id": "freemodel", + "env": [ + "FREEMODEL_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://cc.freemodel.dev/v1", + "name": "FreeModel", + "doc": "https://freemodel.dev", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -123718,14 +130572,17 @@ "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -123738,28 +130595,33 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 4096 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.freemodel.dev/v1" }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "x-ai/grok-4.20-multi-agent": { - "id": "x-ai/grok-4.20-multi-agent", - "name": "xAI: Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -123767,15 +130629,17 @@ ] } ], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-31", - "last_updated": "2026-04-11", + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -123783,30 +130647,50 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.freemodel.dev/v1" }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 1.75, + "output": 14, + "cache_read": 0.175, + "cache_write": 1.75 } }, - "x-ai/grok-build-0.1": { - "id": "x-ai/grok-build-0.1", - "name": "xAI: Grok Build 0.1", - "description": "Grok coding model for agentic engineering, edits, and codebase workflows", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-05-20", - "last_updated": "2026-05-27", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -123814,42 +130698,45 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "~google/gemini-pro-latest": { - "id": "~google/gemini-pro-latest", - "name": "Google: Gemini Pro Latest", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -123857,43 +130744,49 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.freemodel.dev/v1" }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075, + "cache_write": 0.75 } }, - "~google/gemini-flash-latest": { - "id": "~google/gemini-flash-latest", - "name": "Google: Gemini Flash Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -123902,177 +130795,140 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.08333333333333334 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "microsoft/phi-4-mini-instruct": { - "id": "microsoft/phi-4-mini-instruct", - "name": "Microsoft: Phi 4 Mini Instruct", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-10-17", - "last_updated": "2026-05-07", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 1050000, + "input": 922000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.freemodel.dev/v1" + }, "cost": { - "input": 0.08, - "output": 0.35, - "cache_read": 0.08 - } - }, - "microsoft/phi-4": { - "id": "microsoft/phi-4", - "name": "Microsoft: Phi 4", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-12-11", - "last_updated": "2024-12-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 16384, - "output": 16384 - }, - "cost": { - "input": 0.06, - "output": 0.14 - } - }, - "microsoft/wizardlm-2-8x22b": { - "id": "microsoft/wizardlm-2-8x22b", - "name": "WizardLM-2 8x22B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-04-24", - "last_updated": "2024-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 65535, - "output": 8000 - }, - "cost": { - "input": 0.62, - "output": 0.62 + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 5 } }, - "poolside/laguna-xs.2:free": { - "id": "poolside/laguna-xs.2:free", - "name": "Poolside: Laguna XS.2 (free)", - "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", - "family": "laguna", - "attachment": false, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-06-13", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "poolside/laguna-m.1:free": { - "id": "poolside/laguna-m.1:free", - "name": "Poolside: Laguna M.1 (free)", - "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", - "family": "laguna", - "attachment": false, + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "writer/palmyra-x5": { - "id": "writer/palmyra-x5", - "name": "Writer: Palmyra X5", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -124080,398 +130936,455 @@ }, "open_weights": false, "limit": { - "context": 1040000, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 6 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "Z.ai: GLM 4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", + "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "temperature": true, - "release_date": "2025-12-22", - "last_updated": "2026-03-15", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 65535 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.38, - "output": 1.98, - "cache_read": 0.2 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "z-ai/glm-4.5v": { - "id": "z-ai/glm-4.5v", - "name": "Z.ai: GLM 4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", + "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } - }, - "z-ai/glm-4.5": { - "id": "z-ai/glm-4.5", - "name": "Z.ai: GLM 4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, + } + } + }, + "azure-cognitive-services": { + "id": "azure-cognitive-services", + "env": [ + "AZURE_COGNITIVE_SERVICES_RESOURCE_NAME", + "AZURE_COGNITIVE_SERVICES_API_KEY" + ], + "npm": "@ai-sdk/azure", + "name": "Azure Cognitive Services", + "doc": "https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "low", + "medium", "high", "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2026-03-15", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.175 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "Z.ai: GLM 5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "knowledge": "2025-02-28", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 1.26, - "output": 3.96 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "Z.ai: GLM 4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "medium", "high", "xhigh" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2025-09-30", - "last_updated": "2026-03-15", + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 204800 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.39, - "output": 1.9, - "cache_read": 0.175 - } - }, - "z-ai/glm-4-32b": { - "id": "z-ai/glm-4-32b", - "name": "Z.ai: GLM 4 32B ", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-07-25", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.1 + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "z-ai/glm-4.6v": { - "id": "z-ai/glm-4.6v", - "name": "Z.ai: GLM 4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 Nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "low", + "medium", "high", "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-09-30", - "last_updated": "2026-01-10", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "image", "text", - "video" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "z-ai/glm-5v-turbo": { - "id": "z-ai/glm-5v-turbo", - "name": "Z.ai: GLM 5V Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "gpt-chat-latest": { + "id": "gpt-chat-latest", + "name": "GPT Chat Latest", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "xhigh" + "low", + "medium", + "high" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-11", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-05-05", + "last_updated": "2026-05-28", "modalities": { "input": [ - "image", "text", - "video" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 128000, + "input": 111616, + "output": 16384 }, + "status": "beta", "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "z-ai/glm-4.5-air": { - "id": "z-ai/glm-4.5-air", - "name": "Z.ai: GLM 4.5 Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "attachment": false, + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "xhigh" + "low", + "medium", + "high" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-08-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.13, - "output": 0.85, - "cache_read": 0.025 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "z-ai/glm-4.7-flash": { - "id": "z-ai/glm-4.7-flash", - "name": "Z.ai: GLM 4.7 Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -124479,111 +131392,145 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 40551 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 0.06, - "output": 0.4, - "cache_read": 0.01 + "input": 0.95, + "output": 4 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "Z.ai: GLM 5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-03-15", + "knowledge": "2025-07-31", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.72, - "output": 2.3 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "Z.ai: GLM 5 Turbo", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "attachment": false, + "claude-opus-5-5": { + "id": "claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", + "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-04-11", + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 } }, - "openai/gpt-4o-mini-2024-07-18": { - "id": "openai/gpt-4o-mini-2024-07-18", - "name": "OpenAI: GPT-4o-mini (2024-07-18)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "claude-fable-5-1": { + "id": "claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2024-07-18", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -124591,19 +131538,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "openai/gpt-oss-safeguard-20b": { - "id": "openai/gpt-oss-safeguard-20b", - "name": "OpenAI: gpt-oss-safeguard-20b", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "attachment": false, + "claude-mythos-5": { + "id": "claude-mythos-5", + "name": "Claude Mythos 5", + "description": "Restricted Claude model for advanced cybersecurity and biology research workflows", + "family": "claude-mythos", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -124611,17 +131565,23 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -124629,28 +131589,45 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 1000000, + "output": 128000 + }, + "status": "beta", + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.075, - "output": 0.3, - "cache_read": 0.037 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "openai/gpt-3.5-turbo-instruct": { - "id": "openai/gpt-3.5-turbo-instruct", - "name": "OpenAI: GPT-3.5 Turbo Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2023-03-01", - "last_updated": "2023-09-21", + "knowledge": "2025-03-31", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -124658,49 +131635,68 @@ }, "open_weights": false, "limit": { - "context": 4095, - "output": 4096 + "context": 200000, + "output": 32000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 1.5, - "output": 2 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "openai/gpt-5.2-chat": { - "id": "openai/gpt-5.2-chat", - "name": "OpenAI: GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-12-11", - "last_updated": "2026-03-15", + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.6, + "output": 3 } }, - "openai/o3": { - "id": "openai/o3", - "name": "OpenAI: o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -124709,19 +131705,22 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "temperature": false, - "release_date": "2025-04-16", - "last_updated": "2026-03-15", + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -124729,43 +131728,49 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/o4-mini-high": { - "id": "openai/o4-mini-high", - "name": "OpenAI: o4 Mini High", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "release_date": "2025-04-17", - "last_updated": "2026-03-15", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -124773,54 +131778,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4 - } - }, - "openai/gpt-audio": { - "id": "openai/gpt-audio", - "name": "OpenAI: GPT Audio", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-20", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "audio", - "text" - ], - "output": [ - "audio", - "text" - ] + "context": 1000000, + "output": 128000 }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 + "status": "beta", + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 2.5, - "output": 10 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "OpenAI: GPT-5.2 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 Mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", "medium", "high", "xhigh" @@ -124828,14 +131813,16 @@ } ], "tool_call": true, - "temperature": false, - "release_date": "2025-12-11", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -124844,67 +131831,46 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 21, - "output": 168 - } - }, - "openai/gpt-4o-mini-search-preview": { - "id": "openai/gpt-4o-mini-search-preview", - "name": "OpenAI: GPT-4o-mini Search Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-01", - "last_updated": "2025-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "OpenAI: GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-08-07", - "last_updated": "2026-03-15", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -124912,30 +131878,64 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1, + "output": 6, + "cache_read": 0.1, + "cache_write": 1.25, + "tiers": [ + { + "input": 2, + "output": 9, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 9, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "openai/gpt-5-chat": { - "id": "openai/gpt-5-chat", - "name": "OpenAI: GPT-5 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "temperature": false, - "release_date": "2025-08-07", - "last_updated": "2026-03-15", + "knowledge": "2025-12-31", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -124943,66 +131943,71 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "OpenAI: GPT-3.5 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2023-03-01", - "last_updated": "2023-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 128000 }, - "open_weights": false, - "limit": { - "context": 16385, - "output": 4096 + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "OpenAI: GPT-5 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "high" + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-10-06", - "last_updated": "2026-03-15", + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -125010,29 +132015,52 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + }, "cost": { - "input": 15, - "output": 120 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "OpenAI: GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "temperature": true, - "release_date": "2024-05-13", - "last_updated": "2026-03-15", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-07-17", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -125040,72 +132068,71 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "openai/gpt-4": { - "id": "openai/gpt-4", - "name": "OpenAI: GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2023-03-14", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 128000 }, - "open_weights": false, - "limit": { - "context": 8191, - "output": 4096 + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 30, - "output": 60 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "OpenAI: o4 Mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "temperature": false, - "release_date": "2025-04-16", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -125113,68 +132140,51 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 - } - }, - "openai/gpt-3.5-turbo-16k": { - "id": "openai/gpt-3.5-turbo-16k", - "name": "OpenAI: GPT-3.5 Turbo 16k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2023-08-28", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 64000 }, - "open_weights": false, - "limit": { - "context": 16385, - "output": 4096 + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { "input": 3, - "output": 4 + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openai/o3-pro": { - "id": "openai/o3-pro", - "name": "OpenAI: o3 Pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-04-16", - "last_updated": "2026-03-15", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -125182,29 +132192,64 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 20, - "output": 80 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25 + } } }, - "openai/gpt-5.1-chat": { - "id": "openai/gpt-5.1-chat", - "name": "OpenAI: GPT-5.1 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "temperature": false, - "release_date": "2025-11-13", - "last_updated": "2026-03-15", + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -125212,77 +132257,25 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-4o-2024-05-13": { - "id": "openai/gpt-4o-2024-05-13", - "name": "OpenAI: GPT-4o (2024-05-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-05-13", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "image", - "pdf", - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 128000 }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { "input": 5, - "output": 15 - } - }, - "openai/gpt-4-0314": { - "id": "openai/gpt-4-0314", - "name": "OpenAI: GPT-4 (older v0314)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2023-05-28", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8191, - "output": 4096 - }, - "cost": { - "input": 30, - "output": 60 + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "OpenAI: GPT-5.4 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -125293,19 +132286,22 @@ "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2026-03-17", - "last_updated": "2026-04-11", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -125313,54 +132309,44 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "openai/gpt-5.3-chat": { - "id": "openai/gpt-5.3-chat", - "name": "OpenAI: GPT-5.3 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "attachment": true, - "reasoning": false, - "tool_call": true, - "release_date": "2026-03-04", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "image", - "pdf", - "text" + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.75, - "output": 14 + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "openai/gpt-3.5-turbo-0613": { - "id": "openai/gpt-3.5-turbo-0613", - "name": "OpenAI: GPT-3.5 Turbo (older v0613)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "phi-4-mini": { + "id": "phi-4-mini", + "name": "Phi-4-mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2023-06-13", - "last_updated": "2023-06-13", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -125369,52 +132355,65 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4095, + "context": 128000, "output": 4096 }, "cost": { - "input": 1, - "output": 2 + "input": 0.075, + "output": 0.3 } }, - "openai/gpt-5-image-mini": { - "id": "openai/gpt-5-image-mini", - "name": "OpenAI: GPT-5 Image Mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-10-16", - "last_updated": "2026-03-15", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 2.5, - "output": 2 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "OpenAI: GPT-5.1-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "o3": { + "id": "o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -125429,8 +132428,9 @@ ], "tool_call": true, "temperature": false, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", @@ -125442,20 +132442,21 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "OpenAI: GPT-5.1-Codex-Max", + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex Mini", "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -125463,15 +132464,16 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text", @@ -125484,29 +132486,31 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "openai/gpt-4o-2024-08-06": { - "id": "openai/gpt-4o-2024-08-06", - "name": "OpenAI: GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "gpt-4-turbo": { + "id": "gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-08-06", - "last_updated": "2026-03-15", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -125515,179 +132519,155 @@ "open_weights": false, "limit": { "context": 128000, - "output": 16384 + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 10, + "output": 30 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "OpenAI: o3 Mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "attachment": true, - "reasoning": false, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2024-12-20", - "last_updated": "2026-03-15", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "pdf", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 128000, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.58, + "output": 1.68 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "OpenAI: GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "llama-4-scout-17b-16e-instruct": { + "id": "llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout 17B 16E Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-12-11", - "last_updated": "2026-03-15", + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.2, + "output": 0.78 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "OpenAI: GPT-5.3-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "release_date": "2026-02-25", - "last_updated": "2026-03-15", + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 163840, + "output": 163840 }, + "status": "deprecated", "cost": { - "input": 1.75, - "output": 14 + "input": 1.35, + "output": 5.4 } }, - "openai/gpt-audio-mini": { - "id": "openai/gpt-audio-mini", - "name": "OpenAI: GPT Audio Mini", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "attachment": false, + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-01-20", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "audio", - "text" + "text", + "image" ], "output": [ - "audio", "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1047576, + "output": 32768 }, + "status": "deprecated", "cost": { - "input": 0.6, - "output": 2.4 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "OpenAI: GPT-5.1-Codex-Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": false, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "gpt-3.5-turbo-1106": { + "id": "gpt-3.5-turbo-1106", + "name": "GPT-3.5 Turbo 1106", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2021-08", + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -125696,44 +132676,31 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 100000 + "context": 16384, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 1, + "output": 2 } }, - "openai/o4-mini-deep-research": { - "id": "openai/o4-mini-deep-research", - "name": "OpenAI: o4 Mini Deep Research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "gpt-4-turbo-vision": { + "id": "gpt-4-turbo-vision", + "name": "GPT-4 Turbo Vision", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-06-26", - "last_updated": "2026-03-15", + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -125741,30 +132708,31 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 128000, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 10, + "output": 30 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "OpenAI: GPT-4.1 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, + "knowledge": "2024-04", "release_date": "2025-04-14", - "last_updated": "2026-03-15", + "last_updated": "2025-04-14", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -125775,32 +132743,26 @@ "context": 1047576, "output": 32768 }, + "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "OpenAI: gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "phi-4-mini-reasoning": { + "id": "phi-4-mini-reasoning", + "name": "Phi-4-mini-reasoning", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -125811,29 +132773,30 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 26215 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.039, - "output": 0.19 + "input": 0.075, + "output": 0.3 } }, - "openai/gpt-4o-2024-11-20": { - "id": "openai/gpt-4o-2024-11-20", - "name": "OpenAI: GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "mistral-small-2503": { + "id": "mistral-small-2503", + "name": "Mistral Small 3.1", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-11-20", - "last_updated": "2026-03-15", + "knowledge": "2024-09", + "release_date": "2025-03-01", + "last_updated": "2025-03-01", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -125842,29 +132805,43 @@ "open_weights": false, "limit": { "context": 128000, - "output": 16384 + "output": 32768 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.1, + "output": 0.3 } }, - "openai/o1": { - "id": "openai/o1", - "name": "OpenAI: o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2024-12-05", - "last_updated": "2026-03-15", + "knowledge": "2025-12-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -125872,40 +132849,91 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "openai/o1-pro": { - "id": "openai/o1-pro", - "name": "OpenAI: o1-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "phi-4": { + "id": "phi-4", + "name": "Phi-4", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.125, + "output": 0.5 + } + }, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-03-19", - "last_updated": "2026-03-15", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -125913,105 +132941,121 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 150, - "output": 600 + "input": 1.75, + "output": 14, + "cache_read": 0.125 } }, - "openai/gpt-chat-latest": { - "id": "openai/gpt-chat-latest", - "name": "OpenAI: GPT Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "model-router": { + "id": "model-router", + "name": "Model Router", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "model-router", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, + "release_date": "2025-05-19", + "last_updated": "2025-11-18", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 16384 + }, + "cost": { + "input": 0.14, + "output": 0 + } + }, + "cohere-embed-v3-english": { + "id": "cohere-embed-v3-english", + "name": "Embed v3 English", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "release_date": "2026-05-05", - "last_updated": "2026-05-07", + "release_date": "2023-11-07", + "last_updated": "2023-11-07", "modalities": { "input": [ - "image", - "pdf", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 512, + "output": 1024 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.1, + "output": 0 } }, - "openai/gpt-5-image": { - "id": "openai/gpt-5-image", - "name": "OpenAI: GPT-5 Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, + "cohere-command-a": { + "id": "cohere-command-a", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-10-14", - "last_updated": "2026-03-15", + "knowledge": "2024-06-01", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ - "image", - "pdf", "text" ], "output": [ - "image", "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 10, + "input": 2.5, "output": 10 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "OpenAI: GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "release_date": "2026-03-06", - "last_updated": "2026-03-15", + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -126019,18 +133063,51 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "output": 128000 + "context": 128000, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 2.5, - "output": 15 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "OpenAI: GPT-5.4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.71, + "output": 0.71 + } + }, + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -126040,50 +133117,54 @@ "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2026-03-17", - "last_updated": "2026-04-11", + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "audio" ], "output": [ - "text" + "text", + "image", + "audio" ] }, "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "OpenAI: GPT-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": true, + "codestral-2501": { + "id": "codestral-2501", + "name": "Codestral 25.01", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "codestral", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-04-14", - "last_updated": "2026-03-15", + "knowledge": "2024-03", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "image", - "pdf", "text" ], "output": [ @@ -126092,67 +133173,56 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 256000, + "output": 256000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.3, + "output": 0.9 } }, - "openai/gpt-4o-audio-preview": { - "id": "openai/gpt-4o-audio-preview", - "name": "OpenAI: GPT-4o Audio", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "attachment": false, + "cohere-embed-v-4-0": { + "id": "cohere-embed-v-4-0", + "name": "Embed v4", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", + "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-08-15", - "last_updated": "2026-03-15", + "tool_call": false, + "temperature": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "audio", - "text" + "text", + "image" ], "output": [ - "audio", "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 16384 + "output": 1536 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.12, + "output": 0 } }, - "openai/o3-deep-research": { - "id": "openai/o3-deep-research", - "name": "OpenAI: o3 Deep Research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2024-06-26", - "last_updated": "2026-03-15", + "text-embedding-3-large": { + "id": "text-embedding-3-large", + "name": "text-embedding-3-large", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ - "image", - "pdf", "text" ], "output": [ @@ -126161,25 +133231,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 8191, + "output": 3072 }, "cost": { - "input": 10, - "output": 40, - "cache_read": 2.5 + "input": 0.13, + "output": 0 } }, - "openai/gpt-4-turbo-preview": { - "id": "openai/gpt-4-turbo-preview", - "name": "OpenAI: GPT-4 Turbo Preview", + "gpt-3.5-turbo-0125": { + "id": "gpt-3.5-turbo-0125", + "name": "GPT-3.5 Turbo 0125", "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, + "knowledge": "2021-08", "release_date": "2024-01-25", - "last_updated": "2026-03-15", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -126190,39 +133261,29 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 16384, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 10, - "output": 30 + "input": 0.5, + "output": 1.5 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "OpenAI: GPT-5 Mini", + "gpt-3.5-turbo-instruct": { + "id": "gpt-3.5-turbo-instruct", + "name": "GPT-3.5 Turbo Instruct", "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": false, - "release_date": "2025-08-07", - "last_updated": "2026-03-15", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2021-08", + "release_date": "2023-09-21", + "last_updated": "2023-09-21", "modalities": { "input": [ - "image", - "pdf", "text" ], "output": [ @@ -126231,30 +133292,40 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 4096, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 1.5, + "output": 2 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "OpenAI: GPT-4.1 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-04-14", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -126262,55 +133333,56 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 400000, + "output": 272000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 15, + "output": 120 } }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "OpenAI: GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, - "tool_call": true, + "phi-4-reasoning-plus": { + "id": "phi-4-reasoning-plus", + "name": "Phi-4-reasoning-plus", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2023-09-13", - "last_updated": "2024-04-09", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 32000, "output": 4096 }, "cost": { - "input": 10, - "output": 30 + "input": 0.125, + "output": 0.5 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "OpenAI: GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "codex-mini": { + "id": "codex-mini", + "name": "Codex Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -126319,12 +133391,11 @@ ], "tool_call": true, "temperature": false, - "release_date": "2025-08-07", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-05-16", + "last_updated": "2025-05-16", "modalities": { "input": [ - "image", - "pdf", "text" ], "output": [ @@ -126333,39 +133404,42 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 200000, + "output": 100000 }, + "status": "deprecated", "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 1.5, + "output": 6, + "cache_read": 0.375 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "OpenAI: GPT-5.4 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "attachment": true, + "o1": { + "id": "o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "release_date": "2026-03-06", - "last_updated": "2026-03-15", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -126373,129 +133447,136 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "output": 128000 + "context": 200000, + "output": 100000 }, + "status": "deprecated", "cost": { - "input": 30, - "output": 180 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "openai/o3-mini-high": { - "id": "openai/o3-mini-high", - "name": "OpenAI: o3 Mini High", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "attachment": true, + "cohere-embed-v3-multilingual": { + "id": "cohere-embed-v3-multilingual", + "name": "Embed v3 Multilingual", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": false, - "release_date": "2025-01-31", - "last_updated": "2026-03-15", + "release_date": "2023-11-07", + "last_updated": "2023-11-07", "modalities": { "input": [ - "pdf", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 512, + "output": 1024 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.1, + "output": 0 } }, - "openai/gpt-5.4-image-2": { - "id": "openai/gpt-5.4-image-2", - "name": "OpenAI: GPT-5.4 Image 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "mistral-medium-2505": { + "id": "mistral-medium-2505", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-05-01", + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { - "context": 272000, + "context": 128000, "output": 128000 }, "cost": { - "input": 8, - "output": 15, - "cache_read": 2 + "input": 0.4, + "output": 2 } }, - "openai/gpt-4o-search-preview": { - "id": "openai/gpt-4o-search-preview", - "name": "OpenAI: GPT-4o Search Preview", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": false, + "phi-4-multimodal": { + "id": "phi-4-multimodal", + "name": "Phi-4-multimodal", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "phi", + "attachment": true, "reasoning": false, "tool_call": false, - "release_date": "2025-03-13", - "last_updated": "2026-03-15", + "temperature": true, + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 16384 + "output": 4096 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.08, + "output": 0.32, + "input_audio": 4 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "OpenAI: GPT-5.5 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", + "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "temperature": false, - "release_date": "2026-04-24", - "last_updated": "2026-05-01", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -126503,29 +133584,32 @@ }, "open_weights": false, "limit": { - "context": 1050000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 30, - "output": 180 + "input": 0.05, + "output": 0.4, + "cache_read": 0.01 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "OpenAI: GPT-4o-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-07-18", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -126533,63 +133617,72 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1047576, + "output": 32768 }, + "status": "deprecated", "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "OpenAI: gpt-oss-20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 26215 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.03, - "output": 0.14 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai/gpt-4-1106-preview": { - "id": "openai/gpt-4-1106-preview", - "name": "OpenAI: GPT-4 Turbo (older v1106)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "text-embedding-ada-002": { + "id": "text-embedding-ada-002", + "name": "text-embedding-ada-002", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2023-11-06", - "last_updated": "2026-03-15", + "tool_call": false, + "release_date": "2022-12-15", + "last_updated": "2022-12-15", "modalities": { "input": [ "text" @@ -126600,19 +133693,20 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 8192, + "output": 1536 }, "cost": { - "input": 10, - "output": 30 + "input": 0.1, + "output": 0 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "OpenAI: GPT-5 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -126626,8 +133720,41 @@ ], "tool_call": true, "temperature": false, - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "status": "deprecated", + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 + } + }, + "llama-4-maverick-17b-128e-instruct-fp8": { + "id": "llama-4-maverick-17b-128e-instruct-fp8", + "name": "Llama 4 Maverick 17B 128E Instruct FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -126637,21 +133764,21 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.25, + "output": 1 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "OpenAI: GPT-5.2-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -126660,15 +133787,15 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "temperature": false, - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", @@ -126680,26 +133807,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 200000, + "output": 100000 }, + "status": "deprecated", "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "OpenAI: GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", "high" @@ -126708,13 +133837,13 @@ ], "tool_call": true, "temperature": false, - "release_date": "2025-11-13", - "last_updated": "2026-03-15", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -126723,41 +133852,41 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.25, + "output": 2, + "cache_read": 0.03 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "OpenAI: GPT-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "attachment": true, + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "temperature": false, - "release_date": "2026-04-24", - "last_updated": "2026-05-01", + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -126765,25 +133894,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "thedrummer/cydonia-24b-v4.1": { - "id": "thedrummer/cydonia-24b-v4.1", - "name": "TheDrummer: Cydonia 24B V4.1", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "ministral-3b": { + "id": "ministral-3b", + "name": "Ministral 3B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-09-27", - "last_updated": "2026-03-15", + "knowledge": "2024-03", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ "text" @@ -126794,111 +133926,151 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 0.5 + "input": 0.04, + "output": 0.04 } }, - "thedrummer/skyfall-36b-v2": { - "id": "thedrummer/skyfall-36b-v2", - "name": "TheDrummer: Skyfall 36B V2", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-03-11", - "last_updated": "2026-03-15", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 128000, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 0.55, - "output": 0.8 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "thedrummer/unslopnemo-12b": { - "id": "thedrummer/unslopnemo-12b", - "name": "TheDrummer: UnslopNemo 12B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt-codex", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2024-11-09", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ - "text" + "text", + "image", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 0.4 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "thedrummer/rocinante-12b": { - "id": "thedrummer/rocinante-12b", - "name": "TheDrummer: Rocinante 12B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-09-30", - "last_updated": "2026-03-15", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.17, - "output": 0.43 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "bytedance/ui-tars-1.5-7b": { - "id": "bytedance/ui-tars-1.5-7b", - "name": "ByteDance: UI-TARS 7B ", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "text-embedding-3-small": { + "id": "text-embedding-3-small", + "name": "text-embedding-3-small", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-07-23", - "last_updated": "2026-03-15", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -126907,25 +134079,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 2048 + "context": 8191, + "output": 1536 }, "cost": { - "input": 0.1, - "output": 0.2 + "input": 0.02, + "output": 0 } }, - "rekaai/reka-flash-3": { - "id": "rekaai/reka-flash-3", - "name": "Reka Flash 3", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "phi-4-reasoning": { + "id": "phi-4-reasoning", + "name": "Phi-4-reasoning", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": false, "temperature": true, - "release_date": "2025-03-12", - "last_updated": "2026-04-11", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -126936,54 +134110,27 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.2 - } - }, - "rekaai/reka-edge": { - "id": "rekaai/reka-edge", - "name": "Reka Edge", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-03-20", - "last_updated": "2026-04-11", - "modalities": { - "input": [ - "image", - "text", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 16384, - "output": 16384 + "context": 32000, + "output": 4096 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 0.125, + "output": 0.5 } }, - "mistralai/mistral-large-2407": { - "id": "mistralai/mistral-large-2407", - "name": "Mistral Large 2407", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "deepseek-v3.2-speciale": { + "id": "deepseek-v3.2-speciale", + "name": "DeepSeek-V3.2-Speciale", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2024-11-19", - "last_updated": "2026-03-15", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -126994,27 +134141,40 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 128000 }, "cost": { - "input": 2, - "output": 6 + "input": 0.58, + "output": 1.68 } - }, - "mistralai/mistral-small-3.2-24b-instruct": { - "id": "mistralai/mistral-small-3.2-24b-instruct", - "name": "Mistral: Mistral Small 3.2 24B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "attachment": true, + } + } + }, + "pendra": { + "id": "pendra", + "env": [ + "PENDRA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.pendra.ai/api/v1", + "name": "Pendra", + "doc": "https://pendra.ai/docs/integrations/opencode", + "models": { + "llama3.3:70b": { + "id": "llama3.3:70b", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-06-20", - "last_updated": "2025-06-20", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -127023,25 +134183,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.03 + "input": 0, + "output": 0 } }, - "mistralai/mistral-nemo": { - "id": "mistralai/mistral-nemo", - "name": "Mistral: Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "qwen3-coder:30b": { + "id": "qwen3-coder:30b", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-07-01", - "last_updated": "2024-07-30", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -127052,63 +134213,71 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.02, - "output": 0.04 + "input": 0, + "output": 0 } }, - "mistralai/mistral-medium-3-5": { - "id": "mistralai/mistral-medium-3-5", - "name": "Mistral: Mistral Medium 3.5", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": true, + "gpt-oss:120b": { + "id": "gpt-oss:120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-05-07", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 7.5 + "input": 0, + "output": 0 } }, - "mistralai/ministral-8b-2512": { - "id": "mistralai/ministral-8b-2512", - "name": "Mistral: Ministral 3 8B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "qwen3.6:27b": { + "id": "qwen3.6:27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2026-03-15", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -127117,23 +134286,26 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 0, + "output": 0 } }, - "mistralai/devstral-small": { - "id": "mistralai/devstral-small", - "name": "Mistral: Devstral Small 1.1", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "glm-4.7-flash": { + "id": "glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-05-07", - "last_updated": "2025-07-10", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -127144,27 +134316,39 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 26215 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0, + "output": 0 } }, - "mistralai/mistral-small-3.1-24b-instruct": { - "id": "mistralai/mistral-small-3.1-24b-instruct", - "name": "Mistral: Mistral Small 3.1 24B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "attachment": true, - "reasoning": false, - "tool_call": false, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-03-17", - "last_updated": "2026-03-15", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -127173,25 +134357,42 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.35, - "output": 0.56, - "cache_read": 0.015 + "input": 0, + "output": 0 } - }, - "mistralai/mistral-saba": { - "id": "mistralai/mistral-saba", - "name": "Mistral: Saba", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + } + } + }, + "moark": { + "id": "moark", + "env": [ + "MOARK_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://moark.com/v1", + "name": "Moark", + "doc": "https://moark.com/docs/openapi/v1#tag/%E6%96%87%E6%9C%AC%E7%94%9F%E6%88%90", + "models": { + "GLM-4.7": { + "id": "GLM-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-02-17", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -127202,24 +134403,26 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 3.5, + "output": 14 } }, - "mistralai/mistral-large": { - "id": "mistralai/mistral-large", - "name": "Mistral Large", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2024-07-24", - "last_updated": "2025-12-02", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -127230,56 +134433,70 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 25600 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2, - "output": 6 + "input": 2.1, + "output": 8.4, + "cache_read": 2.1, + "cache_write": 8.4 } - }, - "mistralai/mistral-medium-3.1": { - "id": "mistralai/mistral-medium-3.1", - "name": "Mistral: Mistral Medium 3.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": true, + } + } + }, + "atomic-chat": { + "id": "atomic-chat", + "env": [ + "ATOMIC_CHAT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "http://127.0.0.1:1337/v1", + "name": "Atomic Chat", + "doc": "https://atomic.chat", + "models": { + "gemma-4-E4B-it-MLX-4bit": { + "id": "gemma-4-E4B-it-MLX-4bit", + "name": "Gemma 4 E4B Instruct (MLX 4-bit)", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2025-08-12", - "last_updated": "2025-08-12", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 26215 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 2 + "input": 0, + "output": 0 } }, - "mistralai/pixtral-large-2411": { - "id": "mistralai/pixtral-large-2411", - "name": "Mistral: Pixtral Large 2411", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "attachment": true, + "Meta-Llama-3_1-8B-Instruct-GGUF": { + "id": "Meta-Llama-3_1-8B-Instruct-GGUF", + "name": "Meta Llama 3.1 8B Instruct (GGUF)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-11-19", - "last_updated": "2026-03-15", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -127289,26 +134506,28 @@ "open_weights": true, "limit": { "context": 131072, - "output": 32768 + "output": 4096 }, "cost": { - "input": 2, - "output": 6 + "input": 0, + "output": 0 } }, - "mistralai/devstral-medium": { - "id": "mistralai/devstral-medium", - "name": "Mistral: Devstral Medium", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "attachment": false, + "Qwen3_5-9B-Q4_K_M": { + "id": "Qwen3_5-9B-Q4_K_M", + "name": "Qwen 3.5 9B (Q4_K_M)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-07-10", - "last_updated": "2025-07-10", + "release_date": "2026-03-05", + "last_updated": "2026-04-04", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -127316,24 +134535,25 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 26215 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 2 + "input": 0, + "output": 0 } }, - "mistralai/mistral-small-24b-instruct-2501": { - "id": "mistralai/mistral-small-24b-instruct-2501", - "name": "Mistral: Mistral Small 3", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "gemma-4-E4B-it-IQ4_XS": { + "id": "gemma-4-E4B-it-IQ4_XS", + "name": "Gemma 4 E4B Instruct (IQ4_XS)", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2025-12-29", - "last_updated": "2026-01-10", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" @@ -127345,27 +134565,28 @@ "open_weights": true, "limit": { "context": 32768, - "output": 16384 + "output": 8192 }, "cost": { - "input": 0.05, - "output": 0.08 + "input": 0, + "output": 0 } }, - "mistralai/ministral-3b-2512": { - "id": "mistralai/ministral-3b-2512", - "name": "Mistral: Ministral 3 3B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "Qwen3_5-9B-MLX-4bit": { + "id": "Qwen3_5-9B-MLX-4bit", + "name": "Qwen 3.5 9B (MLX 4-bit)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2026-03-15", + "release_date": "2026-03-05", + "last_updated": "2026-04-04", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -127373,86 +134594,111 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 0, + "output": 0 } - }, - "mistralai/mistral-small-2603": { - "id": "mistralai/mistral-small-2603", - "name": "Mistral: Mistral Small 4", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + } + } + }, + "qihang-ai": { + "id": "qihang-ai", + "env": [ + "QIHANG_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.qhaigc.net/v1", + "name": "QiHang", + "doc": "https://www.qhaigc.net/docs", + "models": { + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-04-11", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 0.14, + "output": 1.14 } }, - "mistralai/mistral-large-2411": { - "id": "mistralai/mistral-large-2411", - "name": "Mistral Large 2411", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "attachment": false, - "reasoning": false, + "gemini-3-pro-preview": { + "id": "gemini-3-pro-preview", + "name": "Gemini 3 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2024-07-24", - "last_updated": "2024-11-04", + "knowledge": "2025-11", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 26215 + "context": 1000000, + "output": 65000 }, "cost": { - "input": 2, - "output": 6 + "input": 0.57, + "output": 3.43 } }, - "mistralai/mistral-7b-instruct-v0.1": { - "id": "mistralai/mistral-7b-instruct-v0.1", - "name": "Mistral: Mistral 7B Instruct v0.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5-Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -127460,24 +134706,27 @@ }, "open_weights": false, "limit": { - "context": 2824, - "output": 565 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.11, - "output": 0.19 + "input": 0.04, + "output": 0.29 } }, - "mistralai/ministral-14b-2512": { - "id": "mistralai/ministral-14b-2512", - "name": "Mistral: Ministral 3 14B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "knowledge": "2025-03", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", @@ -127489,81 +134738,78 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 52429 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 0.71, + "output": 3.57 } }, - "mistralai/devstral-2512": { - "id": "mistralai/devstral-2512", - "name": "Mistral: Devstral 2 2512", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "attachment": false, - "reasoning": false, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-09-12", - "last_updated": "2026-03-15", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.025 - } - }, - "mistralai/mixtral-8x22b-instruct": { - "id": "mistralai/mixtral-8x22b-instruct", - "name": "Mistral: Mixtral 8x22B Instruct", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-04-17", - "last_updated": "2024-04-17", - "modalities": { - "input": [ - "text" + "input": 0.09, + "output": 0.71, + "tiers": [ + { + "input": 0.09, + "output": 0.71, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 65536, - "output": 13108 - }, - "cost": { - "input": 2, - "output": 6 + "context_over_200k": { + "input": 0.09, + "output": 0.71 + } } }, - "mistralai/mistral-medium-3": { - "id": "mistralai/mistral-medium-3", - "name": "Mistral: Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -127575,113 +134821,116 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 26215 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.4, + "input": 0.25, "output": 2 } }, - "mistralai/voxtral-small-24b-2507": { - "id": "mistralai/voxtral-small-24b-2507", - "name": "Mistral: Voxtral Small 24B 2507", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "attachment": false, - "reasoning": false, + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "audio" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 6400 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.43, + "output": 2.14 } }, - "mistralai/mistral-large-2512": { - "id": "mistralai/mistral-large-2512", - "name": "Mistral: Mistral Large 3 2512", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-11-01", - "last_updated": "2025-12-16", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 52429 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "mistralai/codestral-2508": { - "id": "mistralai/codestral-2508", - "name": "Mistral: Codestral 2508", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-08-01", - "last_updated": "2025-08-01", - "modalities": { - "input": [ - "text" + "input": 0.07, + "output": 0.43, + "tiers": [ + { + "input": 0.07, + "output": 0.43, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 51200 - }, - "cost": { - "input": 0.3, - "output": 0.9 + "context_over_200k": { + "input": 0.07, + "output": 0.43 + } } }, - "morph/morph-v3-fast": { - "id": "morph/morph-v3-fast", - "name": "Morph: Morph V3 Fast", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, - "reasoning": false, - "tool_call": false, + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", + "knowledge": "2025-07-31", + "release_date": "2025-10-01", + "last_updated": "2025-10-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -127689,27 +134938,56 @@ }, "open_weights": false, "limit": { - "context": 81920, - "output": 38000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.8, - "output": 1.2 + "input": 0.14, + "output": 0.71 } - }, - "morph/morph-v3-large": { - "id": "morph/morph-v3-large", - "name": "Morph: Morph V3 Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, + } + } + }, + "ai-router": { + "id": "ai-router", + "env": [ + "AI_ROUTER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.ai-router.dev/v1", + "name": "AI-ROUTER", + "doc": "https://ai-router.dev/openai-compatible-api-gateway/", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -127717,87 +134995,95 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.9, - "output": 1.9 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "bytedance-seed/seed-1.6-flash": { - "id": "bytedance-seed/seed-1.6-flash", - "name": "ByteDance Seed: Seed 1.6 Flash", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "image", "text", - "video" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 1, + "output": 6, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "bytedance-seed/seed-1.6": { - "id": "bytedance-seed/seed-1.6", - "name": "ByteDance Seed: Seed 1.6", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2025-09", - "last_updated": "2025-09", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "image", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -127805,87 +135091,144 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 2 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "bytedance-seed/seed-2.0-mini": { - "id": "bytedance-seed/seed-2.0-mini", - "name": "ByteDance Seed: Seed-2.0-Mini", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-27", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "image", "text", - "video" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125 } }, - "bytedance-seed/seed-2.0-lite": { - "id": "bytedance-seed/seed-2.0-lite", - "name": "ByteDance Seed: Seed-2.0-Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-15", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ + "text", "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25 + } + } + } + }, + "llmtr": { + "id": "llmtr", + "env": [ + "LLMTR_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://llmtr.com/v1", + "name": "LLMTR", + "doc": "https://llmtr.com/docs", + "models": { + "muse-glimmer-30b-tr": { + "id": "muse-glimmer-30b-tr", + "name": "Muse Glimmer 30B (TR)", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", + "modalities": { + "input": [ "text", - "video" + "image" ], "output": [ "text" @@ -127893,24 +135236,25 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 131072, "output": 131072 }, "cost": { - "input": 0.25, - "output": 2 + "input": 2, + "output": 5, + "cache_read": 0.5 } }, - "anthracite-org/magnum-v4-72b": { - "id": "anthracite-org/magnum-v4-72b", - "name": "Magnum v4 72B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "magibu-11b-v8": { + "id": "magibu-11b-v8", + "name": "Magibu 11B v8", + "description": "Turkish-language chat model for instruction following and assistant flows", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2024-10-22", - "last_updated": "2026-03-15", + "release_date": "2026-06-05", + "last_updated": "2026-08-16", "modalities": { "input": [ "text" @@ -127919,75 +135263,68 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 2048 + "context": 8192, + "output": 8192 }, "cost": { - "input": 3, - "output": 5 + "input": 0.1, + "output": 0.5 } }, - "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free": { - "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free", - "name": "NVIDIA: Nemotron 3 Nano Omni (free)", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", + "gemma-4": { + "id": "gemma-4", + "name": "Gemma 4", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-05-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "audio", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 5, + "cache_read": 0.5 } }, - "nvidia/nemotron-3-nano-30b-a3b": { - "id": "nvidia/nemotron-3-nano-30b-a3b", - "name": "NVIDIA: Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "trendyol-asure-12b": { + "id": "trendyol-asure-12b", + "name": "Trendyol Asure 12B", + "description": "Turkish-language multimodal instruct model built on Gemma 3 12B for e-commerce text, chat, and image-text tasks", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2024-12", - "last_updated": "2026-02-04", + "release_date": "2026-02-19", + "last_updated": "2026-02-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -127995,63 +135332,68 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 52429 + "context": 40960, + "output": 40960 }, "cost": { - "input": 0.05, - "output": 0.2 + "input": 0.1, + "output": 0.5, + "cache_read": 0.025 } }, - "nvidia/llama-3.3-nemotron-super-49b-v1.5": { - "id": "nvidia/llama-3.3-nemotron-super-49b-v1.5", - "name": "NVIDIA: Llama 3.3 Nemotron Super 49B V1.5", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "medgemma-4b": { + "id": "medgemma-4b", + "name": "MedGemma 4B", + "description": "Multimodal medical-domain Gemma variant for text and image analysis", + "attachment": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-03-16", - "last_updated": "2025-03-16", + "release_date": "2026-04-26", + "last_updated": "2026-08-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 26215 + "context": 8192, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.4 + "input": 3, + "output": 5 } }, - "nvidia/nemotron-nano-9b-v2": { - "id": "nvidia/nemotron-nano-9b-v2", - "name": "NVIDIA: Nemotron Nano 9B V2", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", - "attachment": false, + "qwen3-6-35b": { + "id": "qwen3-6-35b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-18", - "last_updated": "2025-08-18", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -128059,19 +135401,19 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 26215 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0.04, - "output": 0.16 + "input": 5, + "output": 10 } }, - "nvidia/nemotron-3-super-120b-a12b:free": { - "id": "nvidia/nemotron-3-super-120b-a12b:free", - "name": "NVIDIA: Nemotron 3 Super (free)", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "poolside/laguna-xs-2.1": { + "id": "poolside/laguna-xs-2.1", + "name": "Laguna XS 2.1", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -128080,9 +135422,10 @@ } ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-03-12", - "last_updated": "2026-03-15", + "release_date": "2026-07-02", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" @@ -128094,29 +135437,24 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "output": 32768 }, "cost": { "input": 0, "output": 0 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "NVIDIA: Nemotron 3 Super", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "publicai/apertus-70b-instruct": { + "id": "publicai/apertus-70b-instruct", + "name": "Apertus 70B Instruct", + "description": "Fully open 70B multilingual LLM supporting 1800+ languages with 65K context. Trained on 15T tokens of compliant open data. Apache 2.0, EU AI Act compliant.", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-04-11", + "knowledge": "2025-09", + "release_date": "2025-09-02", + "last_updated": "2025-09-02", "modalities": { "input": [ "text" @@ -128127,37 +135465,28 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 65536, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.1 + "input": 0.82, + "output": 2.92 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "Xiaomi: MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "publicai/apertus-8b-instruct": { + "id": "publicai/apertus-8b-instruct", + "name": "Apertus 8B Instruct", + "description": "Fully open 8B multilingual LLM supporting 1800+ languages with 65K context. Trained on compliant open data. Apache 2.0, EU AI Act compliant.", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-09", + "release_date": "2025-09-02", + "last_updated": "2025-09-02", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -128165,83 +135494,87 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 65536, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08, - "tiers": [ - { - "input": 0.8, - "output": 4, - "cache_read": 0.16, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.8, - "output": 4, - "cache_read": 0.16 - } + "input": 0.1, + "output": 0.2 } }, - "xiaomi/mimo-v2-omni": { - "id": "xiaomi/mimo-v2-omni", - "name": "Xiaomi: MiMo-V2-Omni", - "description": "MiMo omni model for text, image, video, audio, and agents", - "family": "mimo", + "thinkingmachines/inkling-small": { + "id": "thinkingmachines/inkling-small", + "name": "Inkling Small", + "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", + "family": "ling", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", "image", - "audio", - "video", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "output": 262144 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08 + "input": 0.58, + "output": 1.44 } }, - "xiaomi/mimo-v2-flash": { - "id": "xiaomi/mimo-v2-flash", - "name": "Xiaomi: MiMo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, + "thinkingmachines/inkling": { + "id": "thinkingmachines/inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-12-01", - "release_date": "2025-12-16", - "last_updated": "2026-02-04", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -128250,27 +135583,26 @@ "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "output": 262144 }, "cost": { - "input": 0.09, - "output": 0.29, - "cache_read": 0.045 + "input": 1.87, + "output": 4.68 } }, - "xiaomi/mimo-v2-pro": { - "id": "xiaomi/mimo-v2-pro", - "name": "Xiaomi: MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", + "perplexity/sonar-deep-research": { + "id": "perplexity/sonar-deep-research", + "name": "Sonar Deep Research", + "description": "Sonar search model for autonomous research and citation-backed long-form reports", + "family": "sonar", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-02-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text" @@ -128281,105 +135613,88 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 32768 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 2, + "output": 8, + "reasoning": 3 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "Xiaomi: MiMo V2.5 Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "meta/muse-spark-1.2-contributor": { + "id": "meta/muse-spark-1.2-contributor", + "name": "Muse Spark 1.2 Contributor", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, "output": 131072 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 0.1, + "output": 0.2 } }, - "inception/mercury-2": { - "id": "inception/mercury-2", - "name": "Inception: Mercury 2", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "upstage/solar-pro4": { + "id": "upstage/solar-pro4", + "name": "Solar Pro 4", + "description": "Upstage's flagship model, specialized for agentic use", + "family": "solar-pro", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "knowledge": "2026-02", + "release_date": "2026-08-06", + "last_updated": "2026-08-06", "modalities": { "input": [ "text" @@ -128390,29 +135705,39 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 50000 + "context": 524288, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.025 + "input": 0.03, + "output": 0.12 } }, - "anthropic/claude-3.5-haiku": { - "id": "anthropic/claude-3.5-haiku", - "name": "Anthropic: Claude 3.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "attachment": true, - "reasoning": false, + "upstage/solar-pro3": { + "id": "upstage/solar-pro3", + "name": "Solar Pro 3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "solar-pro", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "knowledge": "2025-03", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -128420,48 +135745,37 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 131072, "output": 8192 }, "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "input": 0.15, + "output": 0.6 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Anthropic: Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "attachment": true, + "upstage/solar-pro2": { + "id": "upstage/solar-pro2", + "name": "Solar Pro 2", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "solar-pro", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", - "high", - "max" + "minimal", + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2026-03-15", + "knowledge": "2025-03", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ - "image", - "pdf", "text" ], "output": [ @@ -128470,49 +135784,35 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 65536, + "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.15, + "output": 0.6 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Anthropic: Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-05-22", - "last_updated": "2026-03-15", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -128520,146 +135820,104 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.1, + "output": 0.1 } }, - "anthropic/claude-opus-4.6-fast": { - "id": "anthropic/claude-opus-4.6-fast", - "name": "Anthropic: Claude Opus 4.6 (Fast)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "mimo/mimo-v2.5": { + "id": "mimo/mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-04-07", - "last_updated": "2026-04-11", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ + "text", "image", - "text" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 131072 }, "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 + "input": 0.14, + "output": 0.28 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Anthropic: Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "attachment": true, + "mimo/mimo-v2.5-pro": { + "id": "mimo/mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.435, + "output": 0.87 } }, - "anthropic/claude-opus-4.7-fast": { - "id": "anthropic/claude-opus-4.7-fast", - "name": "Anthropic: Claude Opus 4.7 (Fast)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "sakana/fugu-ultra": { + "id": "sakana/fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2026-05-12", - "last_updated": "2026-05-16", + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -128668,95 +135926,65 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 1000000 }, "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 + "input": 5, + "output": 30 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Anthropic: Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "mistral/voxtral-small-latest": { + "id": "mistral/voxtral-small-latest", + "name": "Voxtral Small (latest)", + "description": "Instruct model with native audio input for speech understanding and tool use", + "family": "voxtral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2026-04-16", - "last_updated": "2026-05-01", + "temperature": true, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", "modalities": { "input": [ "text", - "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 32000, + "output": 32000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.1, + "output": 0.3 } }, - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Anthropic: Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "attachment": true, + "qwen/qwen-flash": { + "id": "qwen/qwen-flash", + "name": "Qwen Flash", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "image", - "pdf", "text" ], "output": [ @@ -128765,50 +135993,38 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.05, + "output": 0.4 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Anthropic: Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "attachment": true, + "qwen/qwen3-vl-plus": { + "id": "qwen/qwen3-vl-plus", + "name": "Qwen3-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "release_date": "2025-11-24", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "image", - "pdf", - "text" + "text", + "image" ], "output": [ "text" @@ -128816,49 +136032,39 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 256000, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.2, + "output": 1.6 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Anthropic: Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "attachment": true, + "qwen/qwen3.5-plus": { + "id": "qwen/qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ + "text", "image", - "text" + "video" ], "output": [ "text" @@ -128867,27 +136073,36 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 65536 }, "cost": { - "input": 3, - "output": 15 + "input": 0.4, + "output": 2.4 } }, - "anthropic/claude-3-haiku": { - "id": "anthropic/claude-3-haiku", - "name": "Anthropic: Claude 3 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "attachment": true, - "reasoning": false, + "qwen/qwen-plus": { + "id": "qwen/qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "temperature": true, - "release_date": "2024-03-07", - "last_updated": "2024-03-07", + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -128895,49 +136110,36 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 4096 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.03, - "cache_write": 0.3 + "input": 0.4, + "output": 1.2 } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Anthropic: Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "attachment": true, + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "release_date": "2025-05-22", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "image", - "pdf", "text" ], "output": [ @@ -128946,20 +136148,19 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 256000, + "output": 65536 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 1.2, + "output": 6 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Anthropic: Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -128967,29 +136168,19 @@ "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -128998,26 +136189,25 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.25, + "output": 1.5 } }, - "tencent/hunyuan-a13b-instruct": { - "id": "tencent/hunyuan-a13b-instruct", - "name": "Tencent: Hunyuan A13B Instruct", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "qwen/qwen3-coder-flash": { + "id": "qwen/qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-11-25", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -129028,37 +136218,110 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 0.3, + "output": 1.5 } }, - "tencent/hy3-preview": { - "id": "tencent/hy3-preview", - "name": "Tencent: Hy3 Preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 1, + "output": 5 + } + }, + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-05-16", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 65536 + }, + "cost": { + "input": 0.6, + "output": 3.6 + } + }, + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" ], "output": [ "text" @@ -129066,26 +136329,86 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.066, - "output": 0.26, - "cache_read": 0.029 + "input": 0.5, + "output": 3 } }, - "deepcogito/cogito-v2.1-671b": { - "id": "deepcogito/cogito-v2.1-671b", - "name": "Deep Cogito: Cogito v2.1 671B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.4, + "output": 1.6 + } + } + } + }, + "alibaba": { + "id": "alibaba", + "env": [ + "DASHSCOPE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1", + "name": "Alibaba", + "doc": "https://www.alibabacloud.com/help/en/model-studio/models", + "models": { + "qwen-flash": { + "id": "qwen-flash", + "name": "Qwen Flash", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-11-14", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -129094,26 +136417,35 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 1000000, "output": 32768 }, "cost": { - "input": 1.25, - "output": 1.25 + "input": 0.05, + "output": 0.4 } }, - "cohere/command-a": { - "id": "cohere/command-a", - "name": "Cohere: Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -129122,26 +136454,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 1000000, + "output": 65536 }, "cost": { "input": 2.5, - "output": 10 + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "cohere/command-r-08-2024": { - "id": "cohere/command-r-08-2024", - "name": "Cohere: Command R (08-2024)", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "qwen2-5-32b-instruct": { + "id": "qwen2-5-32b-instruct", + "name": "Qwen2.5 32B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -129152,24 +136488,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.7, + "output": 2.8 } }, - "cohere/command-r7b-12-2024": { - "id": "cohere/command-r7b-12-2024", - "name": "Cohere: Command R7B (12-2024)", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "qwq-plus": { + "id": "qwq-plus", + "name": "QwQ Plus", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2024-12-02", - "last_updated": "2024-12-02", + "knowledge": "2024-04", + "release_date": "2025-03-05", + "last_updated": "2025-03-05", "modalities": { "input": [ "text" @@ -129178,29 +136517,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.0375, - "output": 0.15 + "input": 0.8, + "output": 2.4 } }, - "cohere/command-r-plus-08-2024": { - "id": "cohere/command-r-plus-08-2024", - "name": "Cohere: Command R+ (08-2024)", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "qwen2-5-vl-72b-instruct": { + "id": "qwen2-5-vl-72b-instruct", + "name": "Qwen2.5-VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -129208,65 +136550,80 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 2.5, - "output": 10 + "input": 2.8, + "output": 8.4 } }, - "gryphe/mythomax-l2-13b": { - "id": "gryphe/mythomax-l2-13b", - "name": "MythoMax 13B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "qwen3-vl-plus": { + "id": "qwen3-vl-plus", + "name": "Qwen3-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2024-04-25", - "last_updated": "2024-04-25", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 4096, - "output": 4096 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.06, - "output": 0.06 + "input": 0.2, + "output": 1.6, + "reasoning": 4.8 } }, - "stepfun/step-3.5-flash": { - "id": "stepfun/step-3.5-flash", - "name": "StepFun: Step 3.5 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, + "qwen3.5-27b": { + "id": "qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-01-29", - "last_updated": "2026-01-29", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -129274,26 +136631,42 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 0.3, + "output": 2.4 } }, - "prime-intellect/intellect-3": { - "id": "prime-intellect/intellect-3", - "name": "Prime Intellect: INTELLECT-3", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-11-26", - "last_updated": "2026-02-04", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -129304,24 +136677,27 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { "input": 0.2, - "output": 1.1 + "output": 0.4, + "cache_read": 0.04 } }, - "nex-agi/deepseek-v3.1-nex-n1": { - "id": "nex-agi/deepseek-v3.1-nex-n1", - "name": "Nex AGI: DeepSeek V3.1 Nex N1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "qwen-max": { + "id": "qwen-max", + "name": "Qwen Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-11-25", + "knowledge": "2024-04", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", "modalities": { "input": [ "text" @@ -129332,24 +136708,31 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 163840 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.27, - "output": 1 + "input": 1.6, + "output": 6.4 } }, - "undi95/remm-slerp-l2-13b": { - "id": "undi95/remm-slerp-l2-13b", - "name": "ReMM SLERP 13B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "qwen3-next-80b-a3b-thinking": { + "id": "qwen3-next-80b-a3b-thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2023-07-22", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ "text" @@ -129360,18 +136743,19 @@ }, "open_weights": true, "limit": { - "context": 6144, - "output": 4096 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.45, - "output": 0.65 + "input": 0.5, + "output": 6 } }, - "~openai/gpt-mini-latest": { - "id": "~openai/gpt-mini-latest", - "name": "OpenAI: GPT Mini Latest", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -129383,19 +136767,28 @@ "values": [ "low", "medium", - "high", "xhigh" ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 } ], "tool_call": true, - "temperature": false, - "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ @@ -129404,44 +136797,76 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "~openai/gpt-latest": { - "id": "~openai/gpt-latest", - "name": "OpenAI: GPT Latest", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": false, - "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "qwen-plus-character-ja": { + "id": "qwen-plus-character-ja", + "name": "Qwen Plus Character (Japanese)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01", + "last_updated": "2024-01", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -129449,34 +136874,39 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "output": 128000 + "context": 8192, + "output": 512 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.5, + "output": 1.4 } }, - "~moonshotai/kimi-latest": { - "id": "~moonshotai/kimi-latest", - "name": "MoonshotAI: Kimi Latest", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "attachment": true, + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -129484,52 +136914,63 @@ }, "open_weights": false, "limit": { - "context": 262142, - "output": 262142 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.74, - "output": 3.49, - "cache_read": 0.14 + "input": 0.4, + "output": 2.4, + "reasoning": 2.4 } }, - "relace/relace-search": { - "id": "relace/relace-search", - "name": "Relace: Relace Search", - "description": "Tool-capable chat model for instruction following and agentic application workflows", + "qwen3-livetranslate-flash-realtime": { + "id": "qwen3-livetranslate-flash-realtime", + "name": "Qwen3-LiveTranslate Flash Realtime", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2025-12-09", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 + "context": 53248, + "output": 4096 }, "cost": { - "input": 1, - "output": 3 + "input": 10, + "output": 10, + "input_audio": 10, + "output_audio": 38 } }, - "relace/relace-apply-3": { - "id": "relace/relace-apply-3", - "name": "Relace: Relace Apply 3", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "qwen2-5-72b-instruct": { + "id": "qwen2-5-72b-instruct", + "name": "Qwen2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "release_date": "2025-09-26", - "last_updated": "2026-03-15", + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -129538,26 +136979,36 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.85, - "output": 1.25 + "input": 1.4, + "output": 5.6 } }, - "ai21/jamba-large-1.7": { - "id": "ai21/jamba-large-1.7", - "name": "AI21: Jamba Large 1.7", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "temperature": true, - "release_date": "2025-08-09", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -129566,26 +137017,37 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 4096 + "context": 131072, + "output": 16384 }, "cost": { - "input": 2, - "output": 8 + "input": 0.7, + "output": 2.8, + "reasoning": 8.4 } }, - "arcee-ai/coder-large": { - "id": "arcee-ai/coder-large", - "name": "Arcee AI: Coder Large", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "qwen-plus": { + "id": "qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-05-06", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", "modalities": { "input": [ "text" @@ -129594,26 +137056,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, + "context": 1000000, "output": 32768 }, "cost": { - "input": 0.5, - "output": 0.8 + "input": 0.4, + "output": 1.2, + "reasoning": 4 } }, - "arcee-ai/virtuoso-large": { - "id": "arcee-ai/virtuoso-large", - "name": "Arcee AI: Virtuoso Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-05-06", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" @@ -129622,56 +137087,63 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 64000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.75, - "output": 1.2 + "input": 1.2, + "output": 6 } }, - "arcee-ai/spotlight": { - "id": "arcee-ai/spotlight", - "name": "Arcee AI: Spotlight", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "qwen-omni-turbo-realtime": { + "id": "qwen-omni-turbo-realtime", + "name": "Qwen-Omni Turbo Realtime", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-05-06", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-05-08", + "last_updated": "2025-05-08", "modalities": { "input": [ + "text", "image", - "text" + "audio" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65537 + "context": 32768, + "output": 2048 }, "cost": { - "input": 0.18, - "output": 0.18 + "input": 0.27, + "output": 1.07, + "input_audio": 4.44, + "output_audio": 8.89 } }, - "arcee-ai/trinity-large-thinking": { - "id": "arcee-ai/trinity-large-thinking", - "name": "Arcee AI: Trinity Large Thinking", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "qwen3-coder-480b-a35b-instruct": { + "id": "qwen3-coder-480b-a35b-instruct", + "name": "Qwen3-Coder 480B-A35B Instruct", + "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-11", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -129683,23 +137155,51 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "output": 65536 }, "cost": { - "input": 0.22, - "output": 0.85 + "input": 1.5, + "output": 7.5, + "tiers": [ + { + "input": 2.7, + "output": 13.5, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 4.5, + "output": 22.5, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "arcee-ai/maestro-reasoning": { - "id": "arcee-ai/maestro-reasoning", - "name": "Arcee AI: Maestro Reasoning", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "qwen-turbo": { + "id": "qwen-turbo", + "name": "Qwen Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-05-06", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2024-11-01", + "last_updated": "2025-04-28", "modalities": { "input": [ "text" @@ -129708,27 +137208,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32000 + "context": 1000000, + "output": 16384 }, "cost": { - "input": 0.9, - "output": 3.3 + "input": 0.05, + "output": 0.2, + "reasoning": 0.5 } }, - "arcee-ai/trinity-mini": { - "id": "arcee-ai/trinity-mini", - "name": "Arcee AI: Trinity Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-12", - "last_updated": "2026-01-28", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -129739,27 +137241,49 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.045, - "output": 0.15 + "input": 0.45, + "output": 2.25, + "tiers": [ + { + "input": 0.75, + "output": 3.75, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 1.2, + "output": 6, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "mancer/weaver": { - "id": "mancer/weaver", - "name": "Mancer: Weaver (alpha)", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "qvq-max": { + "id": "qvq-max", + "name": "QVQ Max", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qvq", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2023-08-02", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -129767,83 +137291,97 @@ }, "open_weights": false, "limit": { - "context": 8000, - "output": 2000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.75, - "output": 1 + "input": 1.2, + "output": 4.8 } }, - "perplexity/sonar-reasoning-pro": { - "id": "perplexity/sonar-reasoning-pro", - "name": "Perplexity: Sonar Reasoning Pro", - "description": "Web-grounded reasoning model for multi-step research and cited answers", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "qwen2-5-7b-instruct": { + "id": "qwen2-5-7b-instruct", + "name": "Qwen2.5 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 25600 + "context": 131072, + "output": 8192 }, "cost": { - "input": 2, - "output": 8 + "input": 0.175, + "output": 0.7 } }, - "perplexity/sonar": { - "id": "perplexity/sonar", - "name": "Perplexity: Sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "attachment": true, - "reasoning": false, - "tool_call": false, + "qwen3-omni-flash": { + "id": "qwen3-omni-flash", + "name": "Qwen3-Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "knowledge": "2024-04", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 127072, - "output": 25415 + "context": 65536, + "output": 16384 }, "cost": { - "input": 1, - "output": 1 + "input": 0.43, + "output": 1.66, + "input_audio": 3.81, + "output_audio": 15.11 } }, - "perplexity/sonar-pro": { - "id": "perplexity/sonar-pro", - "name": "Perplexity: Sonar Pro", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "attachment": true, + "qwen-vl-max": { + "id": "qwen-vl-max", + "name": "Qwen-VL Max", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "knowledge": "2024-04", + "release_date": "2024-04-08", + "last_updated": "2025-08-13", "modalities": { "input": [ "text", @@ -129855,55 +137393,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8000 - }, - "cost": { - "input": 3, - "output": 15 - } - }, - "perplexity/sonar-pro-search": { - "id": "perplexity/sonar-pro-search", - "name": "Perplexity: Sonar Pro Search", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2025-10-31", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "image", - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 3, - "output": 15 + "input": 0.8, + "output": 3.2 } }, - "perplexity/sonar-deep-research": { - "id": "perplexity/sonar-deep-research", - "name": "Perplexity: Sonar Deep Research", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "qwen3-235b-a22b": { + "id": "qwen3-235b-a22b", + "name": "Qwen3 235B-A22B", + "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-01-27", - "last_updated": "2025-01-27", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -129912,144 +137429,148 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 25600 + "context": 131072, + "output": 16384 }, "cost": { - "input": 2, - "output": 8 + "input": 0.7, + "output": 2.8, + "reasoning": 8.4 } }, - "switchpoint/router": { - "id": "switchpoint/router", - "name": "Switchpoint Router", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "qwen3-vl-30b-a3b": { + "id": "qwen3-vl-30b-a3b", + "name": "Qwen3-VL 30B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-07-12", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, "output": 32768 }, "cost": { - "input": 0.85, - "output": 3.4 - } - }, - "openrouter/bodybuilder": { - "id": "openrouter/bodybuilder", - "name": "Body Builder (beta)", - "description": "Preview model for early access evaluation, prototyping, and compatibility testing", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-03-15", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 32768 - }, - "status": "beta", - "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 0.8, + "reasoning": 2.4 } }, - "openrouter/free": { - "id": "openrouter/free", - "name": "Free Models Router", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "qwen3.5-122b-a10b": { + "id": "qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-01", - "last_updated": "2026-03-15", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ + "text", "image", - "text" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32768 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 3.2 } }, - "openrouter/owl-alpha": { - "id": "openrouter/owl-alpha", - "name": "Owl Alpha", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-30", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "output": 262144 + "context": 262144, + "output": 65536 }, - "status": "alpha", "cost": { - "input": 0, - "output": 0 + "input": 0.248, + "output": 1.485 } }, - "openrouter/pareto-code": { - "id": "openrouter/pareto-code", - "name": "Pareto Code Router", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-05-01", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ "text" @@ -130058,54 +137579,52 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.5, + "output": 2 } }, - "openrouter/auto": { - "id": "openrouter/auto", - "name": "Auto Router", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen-vl-plus": { + "id": "qwen-vl-plus", + "name": "Qwen-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-08-15", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { - "context": 2000000, - "output": 32768 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.21, + "output": 0.63 } }, - "qwen/qwen3.5-plus-20260420": { - "id": "qwen/qwen3.5-plus-20260420", - "name": "Qwen: Qwen3.5 Plus 2026-04-20", + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -130113,15 +137632,14 @@ "type": "toggle" }, { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", @@ -130138,80 +137656,88 @@ "output": 65536 }, "cost": { - "input": 0.4, - "output": 2.4 + "input": 0.1875, + "output": 1.125, + "cache_write": 0.234375 } }, - "qwen/qwen3-vl-235b-a22b-thinking": { - "id": "qwen/qwen3-vl-235b-a22b-thinking", - "name": "Qwen: Qwen3 VL 235B A22B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen3-coder-flash": { + "id": "qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-09-24", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.26, - "output": 2.6 + "input": 0.3, + "output": 1.5 } }, - "qwen/qwen3-vl-30b-a3b-thinking": { - "id": "qwen/qwen3-vl-30b-a3b-thinking", - "name": "Qwen: Qwen3 VL 30B A3B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen2-5-omni-7b": { + "id": "qwen2-5-omni-7b", + "name": "Qwen2.5-Omni 7B", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-10-11", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2024-12", + "last_updated": "2024-12", "modalities": { "input": [ + "text", "image", - "text" + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 32768, + "output": 2048 }, "cost": { - "input": 0.13, - "output": 1.56 + "input": 0.1, + "output": 0.4, + "input_audio": 6.76 } }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen: Qwen3 Coder Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "qwen-mt-plus": { + "id": "qwen-mt-plus", + "name": "Qwen-MT Plus", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2025-07-01", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-01", + "last_updated": "2025-01", "modalities": { "input": [ "text" @@ -130220,56 +137746,65 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 16384, + "output": 8192 }, "cost": { - "input": 0.65, - "output": 3.25, - "cache_read": 0.2 + "input": 2.46, + "output": 7.37 } }, - "qwen/qwen-plus": { - "id": "qwen/qwen-plus", - "name": "Qwen: Qwen-Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen3-vl-235b-a22b": { + "id": "qwen3-vl-235b-a22b", + "name": "Qwen3-VL 235B-A22B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], "tool_call": true, "temperature": true, - "release_date": "2024-01-25", - "last_updated": "2025-09-11", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 131072, "output": 32768 }, "cost": { - "input": 0.4, - "output": 1.2, - "cache_read": 0.08 + "input": 0.7, + "output": 2.8, + "reasoning": 8.4 } }, - "qwen/qwen3-coder-30b-a3b-instruct": { - "id": "qwen/qwen3-coder-30b-a3b-instruct", - "name": "Qwen: Qwen3 Coder 30B A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-07-31", - "last_updated": "2025-07-31", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -130280,29 +137815,26 @@ }, "open_weights": true, "limit": { - "context": 160000, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.07, - "output": 0.27 + "input": 1, + "output": 5 } }, - "qwen/qwen3-32b": { - "id": "qwen/qwen3-32b", - "name": "Qwen: Qwen3 32B", + "qwen2-5-14b-instruct": { + "id": "qwen2-5-14b-instruct", + "name": "Qwen2.5 14B Instruct", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2026-02-04", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -130313,57 +137845,83 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 40960 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.08, - "output": 0.24, - "cache_read": 0.04 + "input": 0.35, + "output": 1.4 } }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen: Qwen3 Next 80B A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-09-11", - "last_updated": "2026-03-15", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 52429 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.09, - "output": 1.1 + "input": 0.15, + "output": 0.47, + "cache_read": 0.016, + "cache_write": 0.2 } }, - "qwen/qwen3-vl-8b-instruct": { - "id": "qwen/qwen3-vl-8b-instruct", - "name": "Qwen: Qwen3 VL 8B Instruct", + "qwen2-5-vl-7b-instruct": { + "id": "qwen2-5-vl-7b-instruct", + "name": "Qwen2.5-VL 7B Instruct", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-11-25", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -130372,53 +137930,59 @@ "open_weights": true, "limit": { "context": 131072, - "output": 32768 + "output": 8192 }, "cost": { - "input": 0.08, - "output": 0.5 + "input": 0.35, + "output": 1.05 } }, - "qwen/qwen3.6-35b-a3b": { - "id": "qwen/qwen3.6-35b-a3b", - "name": "Qwen: Qwen3.6 35B A3B", + "qwen3.5-35b-a3b": { + "id": "qwen3.5-35b-a3b", + "name": "Qwen3.5 35B-A3B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens" } ], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, "output": 65536 }, "cost": { - "input": 0.1612, - "output": 0.96525, - "cache_read": 0.1612 + "input": 0.25, + "output": 2 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen: Qwen3.7 Max", + "qwen3.6-max-preview": { + "id": "qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -130426,15 +137990,14 @@ "type": "toggle" }, { - "type": "budget_tokens", - "min": 1, - "max": 262144 + "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "release_date": "2025-08-26", - "last_updated": "2026-05-27", + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text" @@ -130445,26 +138008,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 262144, "output": 65536 }, "cost": { - "input": 1.625, - "output": 4.875, - "cache_read": 0.1625, - "cache_write": 2.03125 + "input": 1.3, + "output": 7.8, + "cache_read": 0.13, + "cache_write": 1.625 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen: Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "qwen-mt-turbo": { + "id": "qwen-mt-turbo", + "name": "Qwen-MT Turbo", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2025-09-05", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-01", + "last_updated": "2025-01", "modalities": { "input": [ "text" @@ -130475,33 +138040,40 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 16384, + "output": 8192 }, "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24 + "input": 0.16, + "output": 0.49 } }, - "qwen/qwen3-8b": { - "id": "qwen/qwen3-8b", - "name": "Qwen: Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-04", - "last_updated": "2026-03-15", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -130509,122 +138081,34 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 8192 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.05 + "input": 0.6, + "output": 3.6 } }, - "qwen/qwen-plus-2025-07-28": { - "id": "qwen/qwen-plus-2025-07-28", - "name": "Qwen: Qwen Plus 0728", + "qwen3-8b": { + "id": "qwen3-8b", + "name": "Qwen3 8B", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-09-09", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 0.26, - "output": 0.78 - } - }, - "qwen/qwen3.5-flash-02-23": { - "id": "qwen/qwen3.5-flash-02-23", - "name": "Qwen: Qwen3.5-Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "image", - "text", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.4 - } - }, - "qwen/qwen3-30b-a3b-instruct-2507": { - "id": "qwen/qwen3-30b-a3b-instruct-2507", - "name": "Qwen: Qwen3 30B A3B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-07-29", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.09, - "output": 0.3, - "cache_read": 0.04 - } - }, - "qwen/qwen-2.5-coder-32b-instruct": { - "id": "qwen/qwen-2.5-coder-32b-instruct", - "name": "Qwen2.5 Coder 32B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-11-11", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -130635,26 +138119,44 @@ }, "open_weights": true, "limit": { - "context": 32768, + "context": 131072, "output": 8192 }, "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.015 + "input": 0.18, + "output": 0.7, + "reasoning": 2.1 } }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen: Qwen3 Next 80B A3B Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-09-11", - "last_updated": "2026-03-15", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -130665,105 +138167,123 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.0975, - "output": 0.78 + "input": 1.4, + "output": 4.4, + "cache_read": 0.28, + "cache_write": 0 } }, - "qwen/qwen3-235b-a22b-thinking-2507": { - "id": "qwen/qwen3-235b-a22b-thinking-2507", - "name": "Qwen: Qwen3 235B A22B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "qwen-omni-turbo": { + "id": "qwen-omni-turbo", + "name": "Qwen-Omni Turbo", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-07-25", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2025-01-19", + "last_updated": "2025-03-26", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 32768, + "output": 2048 }, "cost": { - "input": 0.11, - "output": 0.6 + "input": 0.07, + "output": 0.27, + "input_audio": 4.44, + "output_audio": 8.89 } }, - "qwen/qwen3-vl-32b-instruct": { - "id": "qwen/qwen3-vl-32b-instruct", - "name": "Qwen: Qwen3 VL 32B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "qwen3-omni-flash-realtime": { + "id": "qwen3-omni-flash-realtime", + "name": "Qwen3-Omni Flash Realtime", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-10-21", - "last_updated": "2025-11-25", + "knowledge": "2024-04", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 65536, + "output": 16384 }, "cost": { - "input": 0.104, - "output": 0.416 + "input": 0.52, + "output": 1.99, + "input_audio": 4.57, + "output_audio": 18.13 } }, - "qwen/qwen3-coder": { - "id": "qwen/qwen3-coder", - "name": "Qwen: Qwen3 Coder 480B A35B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "qwen3-asr-flash": { + "id": "qwen3-asr-flash", + "name": "Qwen3-ASR Flash", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "tool_call": false, + "temperature": false, + "knowledge": "2024-04", + "release_date": "2025-09-08", + "last_updated": "2025-09-08", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 52429 + "context": 53248, + "output": 4096 }, "cost": { - "input": 0.22, - "output": 1, - "cache_read": 0.022 + "input": 0.035, + "output": 0.035 } }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen: Qwen3.6 Flash", + "qwen3.6-27b": { + "id": "qwen3.6-27b", + "name": "Qwen3.6 27B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -130771,60 +138291,59 @@ "type": "toggle" }, { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 262144, "output": 65536 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_write": 0.3125 + "input": 0.6, + "output": 3.6 } }, - "qwen/qwen3.5-plus-02-15": { - "id": "qwen/qwen3.5-plus-02-15", - "name": "Qwen: Qwen3.5 Plus 2026-02-15", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "image", "text", + "image", "video" ], "output": [ @@ -130837,20 +138356,50 @@ "output": 65536 }, "cost": { - "input": 0.26, - "output": 1.56 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "qwen/qwen-2.5-7b-instruct": { - "id": "qwen/qwen-2.5-7b-instruct", - "name": "Qwen: Qwen2.5 7B Instruct", + "qwen3-14b": { + "id": "qwen3-14b", + "name": "Qwen3 14B", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "temperature": true, - "release_date": "2024-09", - "last_updated": "2025-04-16", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -130861,29 +138410,40 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 6554 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.04, - "output": 0.1 + "input": 0.35, + "output": 1.4, + "reasoning": 4.2 } }, - "qwen/qwen3-vl-8b-thinking": { - "id": "qwen/qwen3-vl-8b-thinking", - "name": "Qwen: Qwen3 VL 8B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-11-25", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-04", "modalities": { "input": [ + "text", "image", - "text" + "video" ], "output": [ "text" @@ -130891,34 +138451,50 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.117, - "output": 1.365 - } - }, - "qwen/qwen3-max-thinking": { - "id": "qwen/qwen3-max-thinking", - "name": "Qwen: Qwen3 Max Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1, - "max": 81920 - } - ], - "tool_call": true, + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } + } + }, + "qwen-vl-ocr": { + "id": "qwen-vl-ocr", + "name": "Qwen-VL OCR", + "description": "OCR model for extracting structured text from documents and screenshots", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-01-23", - "last_updated": "2026-03-15", + "knowledge": "2024-04", + "release_date": "2024-10-28", + "last_updated": "2025-04-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -130926,56 +138502,92 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 34096, + "output": 4096 }, "cost": { - "input": 0.78, - "output": 3.9 + "input": 0.72, + "output": 0.72 } - }, - "qwen/qwen3-30b-a3b-thinking-2507": { - "id": "qwen/qwen3-30b-a3b-thinking-2507", - "name": "Qwen: Qwen3 30B A3B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "attachment": false, + } + } + }, + "auriko": { + "id": "auriko", + "env": [ + "AURIKO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.auriko.ai/v1", + "name": "Auriko", + "doc": "https://docs.auriko.ai", + "models": { + "qwen-3.6-plus": { + "id": "qwen-3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 6554 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.051, - "output": 0.34 + "input": 0.5, + "output": 3, + "cache_read": 0.1, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "qwen/qwen2.5-vl-72b-instruct": { - "id": "qwen/qwen2.5-vl-72b-instruct", - "name": "Qwen: Qwen2.5 VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, - "reasoning": false, - "tool_call": false, + "minimax-m2-7-highspeed": { + "id": "minimax-m2-7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-02-01", - "last_updated": "2026-03-15", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -130984,34 +138596,50 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.8, - "output": 0.8, - "cache_read": 0.075 + "input": 0.6, + "output": 2.4, + "cache_write": 0.375 } }, - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen: Qwen3.5-27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-03-15", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "image", "text", + "image", "video" ], "output": [ @@ -131021,128 +138649,174 @@ "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "output": 262144 }, "cost": { - "input": 0.195, - "output": 1.56 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "qwen/qwen3-235b-a22b": { - "id": "qwen/qwen3-235b-a22b", - "name": "Qwen: Qwen3 235B A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "grok-4.3": { + "id": "grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens", - "min": 1, - "max": 38912 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2026-03-15", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.455, - "output": 1.82, - "cache_read": 0.15 - } - }, - "qwen/qwen-2.5-72b-instruct": { - "id": "qwen/qwen-2.5-72b-instruct", - "name": "Qwen2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-09", - "last_updated": "2026-01-10", - "modalities": { - "input": [ - "text" + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 0.12, - "output": 0.39 + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "qwen/qwen-plus-2025-07-28:thinking": { - "id": "qwen/qwen-plus-2025-07-28:thinking", - "name": "Qwen: Qwen Plus 0728 (thinking)", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "attachment": false, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-09-09", - "last_updated": "2026-03-15", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.26, - "output": 0.78 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "Qwen: Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "attachment": false, - "reasoning": false, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-02", - "last_updated": "2026-03-15", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -131151,34 +138825,45 @@ "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "output": 262144 }, "cost": { - "input": 0.12, - "output": 0.75, - "cache_read": 0.035 + "input": 0.5, + "output": 2.8 } }, - "qwen/qwen3.6-27b": { - "id": "qwen/qwen3.6-27b", - "name": "Qwen: Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -131186,69 +138871,97 @@ }, "open_weights": false, "limit": { - "context": 256000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.325, - "output": 3.25 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen: Qwen3.5-35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-03-15", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "image", "text", - "video" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.1625, - "output": 1.3 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "qwen/qwen3.5-9b": { - "id": "qwen/qwen3.5-9b", - "name": "Qwen: Qwen3.5-9B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "minimax-m2-7": { + "id": "minimax-m2-7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-15", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" @@ -131256,34 +138969,46 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.15 + "input": 0.3, + "output": 1.2, + "cache_write": 0.375 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen: Qwen3.5 397B A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-03-15", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "image", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -131291,53 +139016,32 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.39, - "output": 2.34 - } - }, - "qwen/qwen3-vl-30b-a3b-instruct": { - "id": "qwen/qwen3-vl-30b-a3b-instruct", - "name": "Qwen: Qwen3 VL 30B A3B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-10-05", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.13, - "output": 0.52 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen/qwen3-235b-a22b-2507": { - "id": "qwen/qwen3-235b-a22b-2507", - "name": "Qwen: Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-04", - "last_updated": "2026-01", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -131348,24 +139052,46 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 52429 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.071, - "output": 0.1 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "qwen/qwen3-coder-flash": { - "id": "qwen/qwen3-coder-flash", - "name": "Qwen: Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-07-23", - "last_updated": "2026-03-15", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -131374,96 +139100,165 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 65536 + "output": 384000 }, "cost": { - "input": 0.195, - "output": 0.975, - "cache_read": 0.06 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "qwen/qwen3-14b": { - "id": "qwen/qwen3-14b", - "name": "Qwen: Qwen3 14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, "temperature": true, - "release_date": "2025-04", - "last_updated": "2026-03-15", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 40960 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.025 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "qwen/qwen3-vl-235b-a22b-instruct": { - "id": "qwen/qwen3-vl-235b-a22b-instruct", - "name": "Qwen: Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-09-23", - "last_updated": "2026-01-10", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 52429 + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 0.2, - "output": 0.88, - "cache_read": 0.11 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen/qwen3-30b-a3b": { - "id": "qwen/qwen3-30b-a3b", - "name": "Qwen: Qwen3 30B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-04", - "last_updated": "2026-03-15", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -131474,38 +139269,53 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 40960 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.08, + "input": 0.14, "output": 0.28, - "cache_read": 0.03 + "cache_read": 0.0028 } - }, - "qwen/qwen3.6-max-preview": { - "id": "qwen/qwen3.6-max-preview", - "name": "Qwen: Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "attachment": false, + } + } + }, + "zenmux": { + "id": "zenmux", + "env": [ + "ZENMUX_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://zenmux.ai/api/v1", + "name": "ZenMux", + "doc": "https://docs.zenmux.ai", + "models": { + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 131072 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-05-01", + "knowledge": "2025-01-01", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ "text" @@ -131513,70 +139323,88 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 1.04, - "output": 6.24, - "cache_write": 1.3 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen: Qwen3.5-122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-03-15", + "knowledge": "2025-01-01", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ + "pdf", "image", - "text", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.26, - "output": 2.08 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen: Qwen3.6 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "release_date": "2025-08-26", - "last_updated": "2026-04-11", + "knowledge": "2025-05-31", + "release_date": "2026-02-06", + "last_updated": "2026-02-06", "modalities": { "input": [ "image", @@ -131589,29 +139417,34 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 65536 + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.325, - "output": 1.95, - "cache_read": 0.0325, - "cache_write": 0.40625 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "amazon/nova-lite-v1": { - "id": "amazon/nova-lite-v1", - "name": "Amazon: Nova Lite 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "anthropic/claude-3.5-haiku": { + "id": "anthropic/claude-3.5-haiku", + "name": "Claude 3.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-12-06", - "last_updated": "2026-03-15", + "knowledge": "2025-01-01", + "release_date": "2024-11-04", + "last_updated": "2024-11-04", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -131619,28 +139452,47 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 5120 + "context": 200000, + "output": 64000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.06, - "output": 0.24 + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 } }, - "amazon/nova-premier-v1": { - "id": "amazon/nova-premier-v1", - "name": "Amazon: Nova Premier 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-11-01", - "last_updated": "2026-03-15", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -131649,27 +139501,46 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 32000 + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 2.5, - "output": 12.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "amazon/nova-pro-v1": { - "id": "amazon/nova-pro-v1", - "name": "Amazon: Nova Pro 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -131677,59 +139548,47 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 5120 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.8, - "output": 3.2 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "amazon/nova-micro-v1": { - "id": "amazon/nova-micro-v1", - "name": "Amazon: Nova Micro 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-12-06", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 5120 - }, - "cost": { - "input": 0.035, - "output": 0.14 - } - }, - "amazon/nova-2-lite-v1": { - "id": "amazon/nova-2-lite-v1", - "name": "Amazon: Nova 2 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "anthropic/claude-sonnet-5-free": { + "id": "anthropic/claude-sonnet-5-free", + "name": "Claude Sonnet 5 (Free)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2026-03-15", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "image", - "pdf", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -131738,25 +139597,33 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 65535 + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "aion-labs/aion-rp-llama-3.1-8b": { - "id": "aion-labs/aion-rp-llama-3.1-8b", - "name": "AionLabs: Aion-RP 1.0 (8B)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-02-05", - "last_updated": "2026-03-15", + "knowledge": "2025-01-01", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ + "image", "text" ], "output": [ @@ -131765,28 +139632,46 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.8, - "output": 1.6 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "aion-labs/aion-1.0-mini": { - "id": "aion-labs/aion-1.0-mini", - "name": "AionLabs: Aion-1.0-Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-02-05", - "last_updated": "2026-03-15", + "knowledge": "2025-01-01", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -131794,28 +139679,47 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.7, - "output": 1.4 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "aion-labs/aion-2.0": { - "id": "aion-labs/aion-2.0", - "name": "AionLabs: Aion-2.0", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2026-02-24", - "last_updated": "2026-03-15", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -131823,28 +139727,46 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.8, - "output": 1.6 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "aion-labs/aion-1.0": { - "id": "aion-labs/aion-1.0", - "name": "AionLabs: Aion-1.0", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "anthropic/claude-3.7-sonnet": { + "id": "anthropic/claude-3.7-sonnet", + "name": "Claude 3.7 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-02-05", - "last_updated": "2026-03-15", + "knowledge": "2025-01-01", + "release_date": "2025-02-24", + "last_updated": "2025-02-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -131852,55 +139774,47 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 4, - "output": 8 - } - }, - "inflection/inflection-3-pi": { - "id": "inflection/inflection-3-pi", - "name": "Inflection: Inflection 3 Pi", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-10-11", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 200000, + "output": 64000 }, - "open_weights": false, - "limit": { - "context": 8000, - "output": 1024 + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 2.5, - "output": 10 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "inflection/inflection-3-productivity": { - "id": "inflection/inflection-3-productivity", - "name": "Inflection: Inflection 3 Productivity", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "temperature": true, - "release_date": "2024-10-11", - "last_updated": "2026-03-15", + "knowledge": "2025-01-01", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ "text" @@ -131908,165 +139822,172 @@ }, "open_weights": false, "limit": { - "context": 8000, - "output": 1024 + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 2.5, - "output": 10 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "sao10k/l3.1-euryale-70b": { - "id": "sao10k/l3.1-euryale-70b", - "name": "Sao10K: Llama 3.1 Euryale 70B v2.2", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2024-08-28", - "last_updated": "2026-03-15", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.85, - "output": 0.85 - } - }, - "sao10k/l3.3-euryale-70b": { - "id": "sao10k/l3.3-euryale-70b", - "name": "Sao10K: Llama 3.3 Euryale 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-12-18", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 128000 }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.65, - "output": 0.75 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 4 } }, - "sao10k/l3-lunaris-8b": { - "id": "sao10k/l3-lunaris-8b", - "name": "Sao10K: Llama 3 8B Lunaris", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": false, + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "temperature": true, - "release_date": "2024-08-13", - "last_updated": "2026-03-15", + "knowledge": "2025-08-31", + "release_date": "2026-02-18", + "last_updated": "2026-02-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.04, - "output": 0.05 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "sao10k/l3-euryale-70b": { - "id": "sao10k/l3-euryale-70b", - "name": "Sao10k: Llama 3 Euryale 70B v2.1", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "release_date": "2024-06-18", - "last_updated": "2026-03-15", + "knowledge": "2025-01-01", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 1.48, - "output": 1.48 - } - }, - "sao10k/l3.1-70b-hanami-x1": { - "id": "sao10k/l3.1-70b-hanami-x1", - "name": "Sao10K: Llama 3.1 70B Hanami x1", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-01-08", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 200000, + "output": 32000 }, - "open_weights": true, - "limit": { - "context": 16000, - "output": 16000 + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 3, - "output": 3 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "upstage/solar-pro-3": { - "id": "upstage/solar-pro-3", - "name": "Upstage: Solar Pro 3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "deepseek/deepseek-chat": { + "id": "deepseek/deepseek-chat", + "name": "DeepSeek-V3.2 (Non-thinking Mode)", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-03-15", + "knowledge": "2025-01-01", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -132078,24 +139999,31 @@ "open_weights": false, "limit": { "context": 128000, - "output": 32768 + "output": 64000 }, + "status": "deprecated", "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.28, + "output": 0.42, + "cache_read": 0.03 } }, - "allenai/olmo-3-32b-think": { - "id": "allenai/olmo-3-32b-think", - "name": "AllenAI: Olmo 3 32B Think", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "deepseek/deepseek-v3.2-exp": { + "id": "deepseek/deepseek-v3.2-exp", + "name": "DeepSeek-V3.2-Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-11-22", - "last_updated": "2026-03-15", + "knowledge": "2025-01-01", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text" @@ -132104,26 +140032,45 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 65536 + "context": 163000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.5 + "input": 0.22, + "output": 0.33 } }, - "essentialai/rnj-1-instruct": { - "id": "essentialai/rnj-1-instruct", - "name": "EssentialAI: Rnj 1 Instruct", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-12-05", - "last_updated": "2026-03-15", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -132134,25 +140081,31 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 6554 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "deepseek/deepseek-r1-0528": { - "id": "deepseek/deepseek-r1-0528", - "name": "DeepSeek: R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2026-03-15", + "knowledge": "2025-01-01", + "release_date": "2025-12-05", + "last_updated": "2025-12-05", "modalities": { "input": [ "text" @@ -132161,21 +140114,21 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 65536 + "context": 128000, + "output": 64000 }, "cost": { - "input": 0.45, - "output": 2.15, - "cache_read": 0.2 + "input": 0.28, + "output": 0.43 } }, "deepseek/deepseek-v4-flash": { "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek: DeepSeek V4 Flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -132185,15 +140138,21 @@ { "type": "effort", "values": [ - "high", - "xhigh" + "low", + "medium", + "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, + "knowledge": "2025-05", "release_date": "2026-04-24", - "last_updated": "2026-05-01", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -132202,9 +140161,9 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 1000000, "output": 384000 }, "cost": { @@ -132213,21 +140172,27 @@ "cache_read": 0.0028 } }, - "deepseek/deepseek-v3.1-terminus": { - "id": "deepseek/deepseek-v3.1-terminus", - "name": "DeepSeek: DeepSeek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "tencent/hy3-preview": { + "id": "tencent/hy3-preview", + "name": "Hy3 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "high" + ] } ], "tool_call": true, "temperature": true, - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text" @@ -132238,33 +140203,38 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 32768 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.21, - "output": 0.79, - "cache_read": 0.13 + "input": 0.172, + "output": 0.572, + "cache_read": 0.058, + "cache_write": 0 } }, - "deepseek/deepseek-r1-distill-llama-70b": { - "id": "deepseek/deepseek-r1-distill-llama-70b", - "name": "DeepSeek: R1 Distill Llama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "attachment": false, + "z-ai/glm-4.6v": { + "id": "z-ai/glm-4.6v", + "name": "GLM 4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" } ], - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-01-23", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -132272,37 +140242,44 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.7, - "output": 0.8, - "cache_read": 0.015 + "input": 0.1456, + "output": 0.4367, + "cache_read": 0.0291, + "tiers": [ + { + "input": 0.2911, + "output": 0.8734, + "cache_read": 0.0582, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek: DeepSeek V4 Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "z-ai/glm-4.5": { + "id": "z-ai/glm-4.5", + "name": "GLM 4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-24", - "last_updated": "2026-05-01", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -132311,60 +140288,111 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 384000 + "context": 128000, + "output": 96000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0.2911, + "output": 1.1645, + "cache_read": 0.0582, + "tiers": [ + { + "input": 0.5823, + "output": 2.3291, + "cache_read": 0.1165, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "deepseek/deepseek-r1": { - "id": "deepseek/deepseek-r1", - "name": "DeepSeek: R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "attachment": false, + "z-ai/glm-5v-turbo": { + "id": "z-ai/glm-5v-turbo", + "name": "GLM 5V Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 64000, - "output": 16000 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.7, - "output": 2.5 + "input": 0.726, + "output": 3.1946, + "cache_read": 0.1743, + "tiers": [ + { + "input": 1.0165, + "output": 3.7754, + "cache_read": 0.2614, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "deepseek/deepseek-v3.2-speciale": { - "id": "deepseek/deepseek-v3.2-speciale", - "name": "DeepSeek: DeepSeek V3.2 Speciale", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "z-ai/glm-5.3-flash": { + "id": "z-ai/glm-5.3-flash", + "name": "GLM 5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2026-03-15", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" @@ -132372,19 +140400,20 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 1.2, - "cache_read": 0.135 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "deepseek/deepseek-v3.2-exp": { - "id": "deepseek/deepseek-v3.2-exp", - "name": "DeepSeek: DeepSeek V3.2 Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "GLM 4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -132393,9 +140422,11 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-09-29", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -132406,24 +140437,46 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 65536 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 0.41 + "input": 0.2911, + "output": 1.1645, + "cache_read": 0.0582, + "tiers": [ + { + "input": 0.5823, + "output": 2.3291, + "cache_read": 0.1165, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "deepseek/deepseek-chat-v3-0324": { - "id": "deepseek/deepseek-chat-v3-0324", - "name": "DeepSeek: DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM 5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-03-24", - "last_updated": "2026-03-15", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -132434,29 +140487,47 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 65536 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.77, - "cache_read": 0.095 + "input": 0.58, + "output": 2.6, + "cache_read": 0.14, + "tiers": [ + { + "input": 0.87, + "output": 3.18, + "cache_read": 0.22, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "deepseek/deepseek-r1-distill-qwen-32b": { - "id": "deepseek/deepseek-r1-distill-qwen-32b", - "name": "DeepSeek: R1 Distill Qwen 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "z-ai/glm-4.6v-flash": { + "id": "z-ai/glm-4.6v-flash", + "name": "GLM 4.6V FlashX", + "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-11-25", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -132464,27 +140535,57 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.29, - "output": 0.29 + "input": 0.0218, + "output": 0.2184, + "cache_read": 0.0044, + "tiers": [ + { + "input": 0.0437, + "output": 0.4367, + "cache_read": 0.0044, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "deepseek/deepseek-chat": { - "id": "deepseek/deepseek-chat", - "name": "DeepSeek: DeepSeek V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, - "reasoning": false, + "z-ai/glm-5.3-flashx": { + "id": "z-ai/glm-5.3-flashx", + "name": "GLM 5.3 FlashX", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2026-03-15", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" @@ -132492,19 +140593,20 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.32, - "output": 0.89, - "cache_read": 0.15 + "input": 0.375, + "output": 1.25, + "cache_read": 0.075 } }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek: DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "GLM 4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -132513,9 +140615,14 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2026-03-15", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -132526,19 +140633,31 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 65536 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.26, - "output": 0.38, - "cache_read": 0.125 + "input": 0.2911, + "output": 1.1645, + "cache_read": 0.0582, + "tiers": [ + { + "input": 0.5823, + "output": 2.3291, + "cache_read": 0.1165, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "deepseek/deepseek-chat-v3.1": { - "id": "deepseek/deepseek-chat-v3.1", - "name": "DeepSeek: DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "z-ai/glm-4.7-flash-free": { + "id": "z-ai/glm-4.7-flash-free", + "name": "GLM 4.7 Flash (Free)", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -132547,9 +140666,13 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -132560,25 +140683,39 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 7168 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.75 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax: MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-03-15", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -132589,26 +140726,35 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 196608 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1.2, - "cache_read": 0.029 + "input": 0.98, + "output": 3.08, + "cache_read": 0.182 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax: MiniMax M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -132619,29 +140765,49 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 39322 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 0.95, - "cache_read": 0.03 + "input": 0.8781, + "output": 3.5126, + "cache_read": 0.1903, + "tiers": [ + { + "input": 1.1709, + "output": 4.098, + "cache_read": 0.2927, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "minimax/minimax-01": { - "id": "minimax/minimax-01", - "name": "MiniMax: MiniMax-01", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "attachment": true, - "reasoning": false, - "tool_call": false, + "z-ai/glm-4.7-flashx": { + "id": "z-ai/glm-4.7-flashx", + "name": "GLM 4.7 FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-01-15", - "last_updated": "2025-01-15", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -132649,20 +140815,21 @@ }, "open_weights": true, "limit": { - "context": 1000192, - "output": 1000192 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.1 + "input": 0.0728, + "output": 0.4367, + "cache_read": 0.0146 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax: MiniMax M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "GLM 5 Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -132670,44 +140837,65 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-30", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 512000 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.73, + "output": 3.19, + "cache_read": 0.174, + "tiers": [ + { + "input": 1.02, + "output": 3.77, + "cache_read": 0.261, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax: MiniMax M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": false, + "z-ai/glm-4.6v-flash-free": { + "id": "z-ai/glm-4.6v-flash-free", + "name": "GLM 4.6V Flash (Free)", + "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "release_date": "2025-10-23", - "last_updated": "2026-03-15", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" @@ -132715,55 +140903,67 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 196608 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.255, - "output": 1, - "cache_read": 0.03 + "input": 0, + "output": 0, + "cache_read": 0, + "tiers": [ + { + "input": 0, + "output": 0, + "cache_read": 0, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "minimax/minimax-m2-her": { - "id": "minimax/minimax-m2-her", - "name": "MiniMax: MiniMax M2-her", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "z-ai/glm-image": { + "id": "z-ai/glm-image", + "name": "GLM-Image", + "description": "GLM-Image is an image generation model adopts a hybrid autoregressive + diffusion decoder architecture. In general image generation quality, GLM‑Image aligns with mainstream latent diffusion approaches, but it shows significant advantages in text-rendering and knowledge‑intensive generation scenarios. It performs especially well in tasks requiring precise semantic understanding and complex information expression, while maintaining strong capabilities in high‑fidelity and fine‑grained detail generation. In addition to text‑to‑image generation, GLM‑Image also supports a rich set of image‑to‑image tasks including image editing, style transfer, identity‑preserving generation, and multi‑subject consistency.", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-01-23", - "last_updated": "2026-03-15", + "temperature": false, + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": true, "limit": { - "context": 65536, - "output": 2048 - }, - "cost": { - "input": 0.3, - "output": 1.2 + "context": 10240, + "output": 0 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax: MiniMax M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax-m2.7", + "z-ai/glm-4.5-air": { + "id": "z-ai/glm-4.5-air", + "name": "GLM 4.5 Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -132774,26 +140974,51 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 96000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.1165, + "output": 0.2911, + "cache_read": 0.0233, + "tiers": [ + { + "input": 0.1747, + "output": 1.1645, + "cache_read": 0.0349, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "minimax/minimax-m1": { - "id": "minimax/minimax-m1", - "name": "MiniMax: MiniMax M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "z-ai/glm-5.3": { + "id": "z-ai/glm-5.3", + "name": "GLM 5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -132805,42 +141030,26 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 40000 + "output": 128000 }, "cost": { - "input": 0.4, - "output": 2.2 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "stealth/claude-opus-4.7": { - "id": "stealth/claude-opus-4.7", - "name": "Stealth: Claude Opus 4.7 (20% off)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "kuaishou/kat-coder-pro-v2": { + "id": "kuaishou/kat-coder-pro-v2", + "name": "KAT-Coder-Pro-V2", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2026-04-16", - "last_updated": "2026-05-27", + "temperature": true, + "release_date": "2026-03-30", + "last_updated": "2026-03-30", "modalities": { "input": [ - "image", - "pdf", "text" ], "output": [ @@ -132849,45 +141058,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "output": 80000 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "stealth/claude-sonnet-4.6": { - "id": "stealth/claude-sonnet-4.6", - "name": "Stealth: Claude Sonnet 4.6 (20% off)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "baidu/ernie-5.0-thinking-preview": { + "id": "baidu/ernie-5.0-thinking-preview", + "name": "ERNIE 5.0", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-02-17", - "last_updated": "2026-05-27", + "knowledge": "2025-01-01", + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "video" ], "output": [ "text" @@ -132895,46 +141091,42 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 128000, "output": 64000 }, "cost": { - "input": 2.4, - "output": 12, - "cache_read": 0.24, - "cache_write": 3 + "input": 0.84, + "output": 3.37 } }, - "stealth/claude-opus-4.6": { - "id": "stealth/claude-opus-4.6", - "name": "Stealth: Claude Opus 4.6 (20% off)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-05", - "last_updated": "2026-05-27", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ + "text", "image", - "pdf", - "text" + "pdf" ], "output": [ "text" @@ -132943,223 +141135,254 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 1000000 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "kwaipilot/kat-coder-pro-v2": { - "id": "kwaipilot/kat-coder-pro-v2", - "name": "Kwaipilot: KAT-Coder-Pro V2", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, - "reasoning": false, + "x-ai/grok-4.5": { + "id": "x-ai/grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-04-11", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 80000 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } }, - "nousresearch/hermes-2-pro-llama-3-8b": { - "id": "nousresearch/hermes-2-pro-llama-3-8b", - "name": "NousResearch: Hermes 2 Pro - Llama-3 8B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "x-ai/grok-voice-tts-1.0": { + "id": "x-ai/grok-voice-tts-1.0", + "name": "Grok Voice TTS 1.0", + "description": "Convert text into spoken audio with a single API call. The API supports a rich set of expressive voices, inline speech tags for fine-grained delivery control, and output formats from high-fidelity MP3 to telephony-optimized μ-law.", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-05-27", - "last_updated": "2024-06-27", + "temperature": false, + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.14, - "output": 0.14 + "context": 15000, + "output": 15000 } }, - "nousresearch/hermes-4-405b": { - "id": "nousresearch/hermes-4-405b", - "name": "Nous: Hermes 4 405B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "x-ai/grok-imagine-image-2.0": { + "id": "x-ai/grok-imagine-image-2.0", + "name": "Grok Imagine Image 2.0", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": true, + "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-08-25", - "last_updated": "2025-08-25", + "temperature": false, + "release_date": "2026-08-07", + "last_updated": "2026-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 26215 - }, - "cost": { - "input": 1, - "output": 3 + "context": 66000, + "output": 0 } }, - "nousresearch/hermes-3-llama-3.1-70b": { - "id": "nousresearch/hermes-3-llama-3.1-70b", - "name": "Nous: Hermes 3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, + "x-ai/grok-voice-stt-1.0": { + "id": "x-ai/grok-voice-stt-1.0", + "name": "Grok Voice STT 1.0", + "description": "Grok Voice STT 1.0 is xAI's speech-to-text model. It supports transcription with word-level timestamps, optional speaker diarization, and multichannel audio.", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-08-18", - "last_updated": "2026-03-15", + "temperature": false, + "release_date": "2026-08-04", + "last_updated": "2026-08-04", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 0.3 + "context": 15000, + "output": 15000 } }, - "nousresearch/hermes-3-llama-3.1-405b": { - "id": "nousresearch/hermes-3-llama-3.1-405b", - "name": "Nous: Hermes 3 405B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, + "x-ai/grok-build-0.1": { + "id": "x-ai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-08-16", - "last_updated": "2024-08-16", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 256000, + "output": 256000 }, "cost": { "input": 1, - "output": 1 + "output": 2, + "cache_read": 0.2 } }, - "nousresearch/hermes-4-70b": { - "id": "nousresearch/hermes-4-70b", - "name": "Nous: Hermes 4 70B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": false, + "x-ai/grok-4.2-fast": { + "id": "x-ai/grok-4.2-fast", + "name": "Grok 4.2 Fast", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-25", - "last_updated": "2026-03-15", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.13, - "output": 0.4, - "cache_read": 0.055 + "input": 2, + "output": 6, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 0.2, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } - } - } - }, - "lucidquery": { - "id": "lucidquery", - "env": [ - "LUCIDQUERY_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.lucidquery.com/v1", - "name": "LucidQuery", - "doc": "https://lucidquery.com/docs", - "models": { - "lucidnova-rf1-100b": { - "id": "lucidnova-rf1-100b", - "name": "LucidNova RF1 100B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "nova", + }, + "x-ai/grok-4.2-fast-non-reasoning": { + "id": "x-ai/grok-4.2-fast-non-reasoning", + "name": "Grok 4.2 Fast Non Reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2025-09-16", - "release_date": "2024-12-28", - "last_updated": "2025-09-10", + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -133167,30 +141390,54 @@ }, "open_weights": false, "limit": { - "context": 120000, - "output": 8000 + "context": 2000000, + "output": 30000 }, "cost": { "input": 2, - "output": 5 + "output": 6, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 0.2, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "lucidquery-nexus-coder": { - "id": "lucidquery-nexus-coder", - "name": "LucidQuery Nexus Coder", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "lucid", + "x-ai/grok-4.6": { + "id": "x-ai/grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2025-08-01", - "release_date": "2025-09-01", - "last_updated": "2025-09-01", + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -133198,31 +141445,57 @@ }, "open_weights": false, "limit": { - "context": 250000, - "output": 60000 + "context": 500000, + "output": 500000 }, "cost": { "input": 2, - "output": 5 + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } }, - "lucidquery-agi-01-swift": { - "id": "lucidquery-agi-01-swift", - "name": "AGI-01 Swift", + "volcengine/doubao-seed-2.0-mini": { + "id": "volcengine/doubao-seed-2.0-mini", + "name": "Doubao-Seed-2.0-mini", "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "agi", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2026-06-05", - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "knowledge": "2026-02-14", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -133230,31 +141503,42 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 120000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 2.5, - "output": 15 + "input": 0.03, + "output": 0.28, + "cache_read": 0.01, + "cache_write": 0.0024 } }, - "lucidquery-agi-01-frontier": { - "id": "lucidquery-agi-01-frontier", - "name": "AGI-01 Frontier", + "volcengine/doubao-seed-2.0-lite": { + "id": "volcengine/doubao-seed-2.0-lite", + "name": "Doubao-Seed-2.0-lite", "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "agi", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2026-06-05", - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "knowledge": "2026-02-14", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -133262,172 +141546,200 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 120000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 4.5, - "output": 22 + "input": 0.09, + "output": 0.51, + "cache_read": 0.02, + "cache_write": 0.0024 } - } - } - }, - "meganova": { - "id": "meganova", - "env": [ - "MEGANOVA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.meganova.ai/v1", - "name": "Meganova", - "doc": "https://docs.meganova.ai", - "models": { - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + }, + "volcengine/doubao-seed-code": { + "id": "volcengine/doubao-seed-code", + "name": "Doubao-Seed-Code", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-01-01", + "release_date": "2025-11-11", + "last_updated": "2025-11-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 256000, + "output": 64000 }, + "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.17, + "output": 1.12, + "cache_read": 0.03 } }, - "moonshotai/Kimi-K2-Thinking": { - "id": "moonshotai/Kimi-K2-Thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, + "volcengine/doubao-seed-2.0-pro": { + "id": "volcengine/doubao-seed-2.0-pro", + "name": "Doubao-Seed-2.0-pro", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "knowledge": "2026-02-14", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 2.6 + "input": 0.45, + "output": 2.24, + "cache_read": 0.09, + "cache_write": 0.0024 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": false, + "volcengine/doubao-seed-1.8": { + "id": "volcengine/doubao-seed-1.8", + "name": "Doubao-Seed-1.8", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "knowledge": "2025-01-01", + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.45, - "output": 2.8 + "input": 0.11, + "output": 0.28, + "cache_read": 0.02, + "cache_write": 0.0024 } }, - "Qwen/Qwen2.5-VL-32B-Instruct": { - "id": "Qwen/Qwen2.5-VL-32B-Instruct", - "name": "Qwen2.5 VL 32B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "volcengine/doubao-seed-2.0-code": { + "id": "volcengine/doubao-seed-2.0-code", + "name": "Doubao Seed 2.0 Code", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "knowledge": "2025-01-01", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 0.9, + "output": 4.48 } }, - "Qwen/Qwen3.5-Plus": { - "id": "Qwen/Qwen3.5-Plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02", - "last_updated": "2026-02", + "knowledge": "2026-02-19", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", + "pdf", + "audio", "video" ], "output": [ @@ -133436,288 +141748,346 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 1048000, + "output": 64000 }, "cost": { - "input": 0.4, - "output": 2.4, - "reasoning": 2.4 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 4.5 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.09, - "output": 0.6 + "input": 1.5, + "output": 9, + "cache_read": 0.15 } }, - "XiaomiMiMo/MiMo-V2-Flash": { - "id": "XiaomiMiMo/MiMo-V2-Flash", - "name": "MiMo V2 Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-12-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2025-01-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "pdf", + "image", + "text", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32000 + "context": 1048000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 1.25, + "output": 10, + "cache_read": 0.31, + "cache_write": 4.5 } }, - "mistralai/Mistral-Small-3.2-24B-Instruct-2506": { - "id": "mistralai/Mistral-Small-3.2-24B-Instruct-2506", - "name": "Mistral Small 3.2 24B Instruct", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", + "knowledge": "2025-01-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ + "pdf", + "image", "text", - "image" + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 1048000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 2.5, + "cache_read": 0.07, + "cache_write": 1 } }, - "mistralai/Mistral-Nemo-Instruct-2407": { - "id": "mistralai/Mistral-Nemo-Instruct-2407", - "name": "Mistral Nemo Instruct 2407", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": false, - "reasoning": false, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2025-01-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 1048000, + "output": 64000 }, "cost": { - "input": 0.02, - "output": 0.04 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 1 } }, - "zai-org/GLM-4.6": { - "id": "zai-org/GLM-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2025-01-01", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ - "text" + "pdf", + "image", + "text", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1048000, + "output": 64000 }, "cost": { - "input": 0.45, - "output": 1.9 + "input": 0.1, + "output": 0.4, + "cache_read": 0.03, + "cache_write": 1 } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "release_date": "2025-03-20", + "last_updated": "2025-03-20", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1050000, + "output": 65530 }, "cost": { - "input": 0.8, - "output": 2.56 + "input": 0.25, + "output": 1.5 } }, - "zai-org/GLM-4.7": { - "id": "zai-org/GLM-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025 } }, - "deepseek-ai/DeepSeek-V3-0324": { - "id": "deepseek-ai/DeepSeek-V3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "sapiens-ai/agnes-1.5-lite": { + "id": "sapiens-ai/agnes-1.5-lite", + "name": "Agnes 1.5 Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "release_date": "2026-03-26", + "last_updated": "2026-03-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.25, - "output": 0.88 + "input": 0.12, + "output": 0.6 } }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "sapiens-ai/agnes-1.5-pro": { + "id": "sapiens-ai/agnes-1.5-pro", + "name": "Agnes 1.5 Pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "release_date": "2026-03-21", + "last_updated": "2026-03-21", "modalities": { "input": [ "text" @@ -133726,28 +142096,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 64000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.5, - "output": 2.15 + "input": 0.16, + "output": 0.8 } }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "inclusionai/ring-2.6-1t": { + "id": "inclusionai/ring-2.6-1t", + "name": "inclusionAI: Ring-2.6-1T", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-25", - "last_updated": "2025-08-25", + "knowledge": "2025-12-31", + "release_date": "2026-05-07", + "last_updated": "2026-05-14", "modalities": { "input": [ "text" @@ -133758,26 +142128,27 @@ }, "open_weights": true, "limit": { - "context": 164000, - "output": 164000 + "context": 262000, + "output": 65000 }, "cost": { - "input": 0.27, - "output": 1 + "input": 0.3, + "output": 2.5, + "cache_read": 0.06 } }, - "deepseek-ai/DeepSeek-V3.2-Exp": { - "id": "deepseek-ai/DeepSeek-V3.2-Exp", - "name": "DeepSeek V3.2 Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "inclusionai/ring-1t": { + "id": "inclusionai/ring-1t", + "name": "Ring-1T", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-10-10", - "last_updated": "2025-10-10", + "knowledge": "2025-01-01", + "release_date": "2025-10-12", + "last_updated": "2025-10-12", "modalities": { "input": [ "text" @@ -133786,28 +142157,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 128000, + "output": 64000 }, + "status": "deprecated", "cost": { - "input": 0.27, - "output": 0.4 + "input": 0.56, + "output": 2.24, + "cache_read": 0.11 } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "inclusionai/ling-1t": { + "id": "inclusionai/ling-1t", + "name": "Ling-1T", + "description": "Tool-capable chat model for instruction following and agentic application workflows", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", + "knowledge": "2025-01-01", + "release_date": "2025-10-09", + "last_updated": "2025-10-09", "modalities": { "input": [ "text" @@ -133816,34 +142188,49 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 128000, + "output": 64000 }, + "status": "deprecated", "cost": { - "input": 0.26, - "output": 0.38 + "input": 0.56, + "output": 2.24, + "cache_read": 0.11 } }, - "MiniMaxAI/MiniMax-M2.1": { - "id": "MiniMaxAI/MiniMax-M2.1", - "name": "MiniMax M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "max" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -133851,112 +142238,67 @@ }, "open_weights": true, "limit": { - "context": 196608, + "context": 1048576, "output": 131072 }, "cost": { - "input": 0.28, - "output": 1.2 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - } - } - }, - "perplexity": { - "id": "perplexity", - "env": [ - "PERPLEXITY_API_KEY" - ], - "npm": "@ai-sdk/perplexity", - "name": "Perplexity", - "doc": "https://docs.perplexity.ai", - "models": { - "sonar-reasoning-pro": { - "id": "sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", - "family": "sonar-reasoning", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2025-01-01", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262140, + "output": 262140 }, "cost": { - "input": 2, - "output": 8 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "sonar": { - "id": "sonar", - "name": "Sonar", - "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", - "family": "sonar", + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "knowledge": "2025-01-01", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ "text" @@ -133967,30 +142309,40 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262000, + "output": 64000 }, + "status": "deprecated", "cost": { - "input": 1, - "output": 1 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "sonar-pro": { - "id": "sonar-pro", - "name": "Sonar Pro", - "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", - "family": "sonar-pro", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2025-01-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -133998,93 +142350,76 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 262000, + "output": 64000 }, "cost": { - "input": 3, - "output": 15 + "input": 0.58, + "output": 3.02, + "cache_read": 0.1 } }, - "sonar-deep-research": { - "id": "sonar-deep-research", - "name": "Perplexity Sonar Deep Research", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "attachment": false, + "moonshotai/kimi-k3-free": { + "id": "moonshotai/kimi-k3-free", + "name": "Kimi K3 (Free)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", - "low", - "medium", - "high" + "max" ] } ], - "tool_call": false, + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-02-01", - "last_updated": "2025-09-01", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2, - "output": 8, - "reasoning": 3 + "input": 0, + "output": 0, + "cache_read": 0 } - } - } - }, - "amazon-bedrock": { - "id": "amazon-bedrock", - "env": [ - "AWS_ACCESS_KEY_ID", - "AWS_SECRET_ACCESS_KEY", - "AWS_REGION", - "AWS_BEARER_TOKEN_BEDROCK" - ], - "npm": "@ai-sdk/amazon-bedrock", - "name": "Amazon Bedrock", - "doc": "https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html", - "models": { - "global.anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "global.anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5 (Global)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + }, + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-01-01", + "release_date": "2025-09-04", + "last_updated": "2025-09-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -134092,81 +142427,66 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 262000, "output": 64000 }, + "status": "deprecated", "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "global.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "global.anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5 (Global)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "moonshotai/kimi-k2.7-code-free": { + "id": "moonshotai/kimi-k2.7-code-free", + "name": "Kimi K2.7 Code (Free)", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "jp.anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "jp.anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5 (JP)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "moonshotai/kimi-k2-thinking-turbo": { + "id": "moonshotai/kimi-k2-thinking-turbo", + "name": "Kimi K2 Thinking Turbo", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-01-01", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -134174,32 +142494,35 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 262000, "output": 64000 }, + "status": "deprecated", "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 1.15, + "output": 8, + "cache_read": 0.15 } }, - "us.meta.llama4-scout-17b-instruct-v1:0": { - "id": "us.meta.llama4-scout-17b-instruct-v1:0", - "name": "Llama 4 Scout 17B Instruct (US)", - "description": "Open Llama with long-context vision for efficient multimodal agents", - "family": "llama", + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -134207,27 +142530,32 @@ }, "open_weights": true, "limit": { - "context": 3500000, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.17, - "output": 0.66 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "minimax.minimax-m2": { - "id": "minimax.minimax-m2", - "name": "MiniMax M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "xiaomi/mimo-v2-flash": { + "id": "xiaomi/mimo-v2-flash", + "name": "MiMo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "knowledge": "2024-12-01", + "release_date": "2025-12-16", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -134238,92 +142566,96 @@ }, "open_weights": true, "limit": { - "context": 204608, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01 } }, - "anthropic.claude-opus-4-7": { - "id": "anthropic.claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.4, + "output": 2, + "cache_read": 0.08, + "tiers": [ + { + "input": 0.8, + "output": 4, + "cache_read": 0.16, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.8, + "output": 4, + "cache_read": 0.16 + } } }, - "eu.anthropic.claude-sonnet-4-6": { - "id": "eu.anthropic.claude-sonnet-4-6", - "name": "Claude Sonnet 4.6 (EU)", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "xiaomi/mimo-v2-omni": { + "id": "xiaomi/mimo-v2-omni", + "name": "MiMo V2 Omni", + "description": "MiMo omni model for text, image, video, audio, and agents", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -134332,63 +142664,90 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 265000, + "output": 265000 }, "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 + "input": 0.4, + "output": 2, + "cache_read": 0.08 } }, - "mistral.voxtral-small-24b-2507": { - "id": "mistral.voxtral-small-24b-2507", - "name": "Voxtral Small 24B 2507", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral", - "attachment": true, - "reasoning": false, + "xiaomi/mimo-v2-pro": { + "id": "xiaomi/mimo-v2-pro", + "name": "MiMo V2 Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 8192 + "context": 1000000, + "output": 256000 }, "cost": { - "input": 0.15, - "output": 0.35 - } - }, - "mistral.ministral-3-3b-instruct": { - "id": "mistral.ministral-3-3b-instruct", - "name": "Ministral 3 3B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "input": 1, + "output": 3, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } + } + }, + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -134396,36 +142755,43 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 8192 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 1, + "output": 3, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "openai.gpt-oss-20b": { - "id": "openai.gpt-oss-20b", - "name": "gpt-oss-20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "minimax/minimax-m2.5-lightning": { + "id": "minimax/minimax-m2.5-lightning", + "name": "MiniMax M2.5 highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01-01", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -134436,52 +142802,35 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 131072 }, "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/v1", - "shape": "responses" + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.07, - "output": 0.3 + "input": 0.6, + "output": 4.8, + "cache_read": 0.06, + "cache_write": 0.75 } }, - "anthropic.claude-opus-4-6-v1": { - "id": "anthropic.claude-opus-4-6-v1", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2025-01-01", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -134489,28 +142838,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 204800, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "openai.gpt-oss-safeguard-20b": { - "id": "openai.gpt-oss-safeguard-20b", - "name": "GPT OSS Safeguard 20B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, - "reasoning": false, + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "knowledge": "2025-01-01", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ "text" @@ -134521,92 +142874,73 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 131070 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.07, - "output": 0.2 + "input": 0.3055, + "output": 1.2219 } }, - "anthropic.claude-opus-4-5-20251101-v1:0": { - "id": "anthropic.claude-opus-4-5-20251101-v1:0", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-08-01", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 512000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.6, + "output": 2.4 } }, - "global.anthropic.claude-fable-5": { - "id": "global.anthropic.claude-fable-5", - "name": "Claude Fable 5 (Global)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax M2.7 highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -134614,38 +142948,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 204800, + "output": 131070 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.611, + "output": 2.4439 } }, - "openai.gpt-oss-120b-1:0": { - "id": "openai.gpt-oss-120b-1:0", - "name": "gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01-01", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -134656,38 +142982,35 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 204000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.38 } }, - "anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-01-01", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -134695,33 +143018,36 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 204000, "output": 64000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" + }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.38 } }, - "amazon.nova-pro-v1:0": { - "id": "amazon.nova-pro-v1:0", - "name": "Nova Pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", + "stepfun/step-3": { + "id": "stepfun/step-3", + "name": "Step-3", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "knowledge": "2025-01-01", + "release_date": "2025-07-31", + "last_updated": "2025-07-31", "modalities": { "input": [ - "text", "image", - "video" + "text" ], "output": [ "text" @@ -134729,28 +143055,26 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 8192 + "context": 65536, + "output": 64000 }, + "status": "deprecated", "cost": { - "input": 0.8, - "output": 3.2, - "cache_read": 0.2 + "input": 0.21, + "output": 0.57 } }, - "qwen.qwen3-coder-next": { - "id": "qwen.qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "stepfun/step-3.5-flash": { + "id": "stepfun/step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "knowledge": "2025-01-01", + "release_date": "2026-02-02", + "last_updated": "2026-02-02", "modalities": { "input": [ "text" @@ -134759,21 +143083,62 @@ "text" ] }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "stepfun/step-3.7-flash": { + "id": "stepfun/step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.22, - "output": 1.8 + "input": 0.2, + "output": 1.15 } }, - "us.anthropic.claude-opus-4-7": { - "id": "us.anthropic.claude-opus-4-7", - "name": "Claude Opus 4.7 (US)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "stepfun/step-3.7-flash-free": { + "id": "stepfun/step-3.7-flash-free", + "name": "Step 3.7 Flash (Free)", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -134782,54 +143147,51 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0 } }, - "nvidia.nemotron-nano-9b-v2": { - "id": "nvidia.nemotron-nano-9b-v2", - "name": "NVIDIA Nemotron Nano 9B v2", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", - "attachment": false, + "qwen/qwen3.5-flash": { + "id": "qwen/qwen3.5-flash", + "name": "Qwen3.5 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "knowledge": "2025-01-01", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -134837,28 +143199,30 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1020000, + "output": 1020000 }, "cost": { - "input": 0.06, - "output": 0.23 + "input": 0.1, + "output": 0.4 } }, - "qwen.qwen3-32b-v1:0": { - "id": "qwen.qwen3-32b-v1:0", - "name": "Qwen3 32B (dense)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-18", - "last_updated": "2025-09-18", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -134867,21 +143231,22 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "jp.anthropic.claude-sonnet-4-6": { - "id": "jp.anthropic.claude-sonnet-4-6", - "name": "Claude Sonnet 4.6 (JP)", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "qwen/qwen3.5-plus": { + "id": "qwen/qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -134890,26 +143255,19 @@ "values": [ "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-01-01", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -134921,25 +143279,22 @@ "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.8, + "output": 4.8 } }, - "deepseek.r1-v1:0": { - "id": "deepseek.r1-v1:0", - "name": "DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3-Max-Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "knowledge": "2025-01-01", + "release_date": "2026-01-23", + "last_updated": "2026-01-23", "modalities": { "input": [ "text" @@ -134950,110 +143305,120 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 256000, + "output": 64000 }, "cost": { - "input": 1.35, - "output": 5.4 + "input": 1.2, + "output": 6 } }, - "mistral.mistral-large-3-675b-instruct": { - "id": "mistral.mistral-large-3-675b-instruct", - "name": "Mistral Large 3", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral", + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen3-Coder-Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "knowledge": "2025-01-01", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "google.gemma-3-27b-it": { - "id": "google.gemma-3-27b-it", - "name": "Google Gemma 3 27B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": false, + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6-Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-27", - "last_updated": "2025-07-27", + "release_date": "2026-03-30", + "last_updated": "2026-03-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 8192 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.12, - "output": 0.2 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "anthropic.claude-sonnet-4-6": { - "id": "anthropic.claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -135065,23 +143430,37 @@ "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "cache_write": 0.5, + "tiers": [ + { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24, + "cache_write": 1.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24, + "cache_write": 1.5 + } } }, - "amazon.nova-2-lite-v1:0": { - "id": "amazon.nova-2-lite-v1:0", - "name": "Nova 2 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova", - "attachment": false, + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -135093,13 +143472,13 @@ ], "tool_call": true, "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -135107,29 +143486,43 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1050000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 0.33, - "output": 2.75 + "input": 3.75, + "output": 18.75 } }, - "openai.gpt-oss-safeguard-120b": { - "id": "openai.gpt-oss-safeguard-120b", - "name": "GPT OSS Safeguard 120B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, - "reasoning": false, + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -135137,26 +143530,29 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1050000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 45, + "output": 225 } }, - "mistral.ministral-3-8b-instruct": { - "id": "mistral.ministral-3-8b-instruct", - "name": "Ministral 3 8B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, + "openai/gpt-5.3-chat": { + "id": "openai/gpt-5.3-chat", + "name": "GPT-5.3 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ "text" @@ -135168,41 +143564,40 @@ "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 16380 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 1.75, + "output": 14 } }, - "eu.anthropic.claude-opus-4-6-v1": { - "id": "eu.anthropic.claude-opus-4-6-v1", - "name": "Claude Opus 4.6 (EU)", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", "high", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -135215,49 +143610,43 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "au.anthropic.claude-opus-4-6-v1": { - "id": "au.anthropic.claude-opus-4-6-v1", - "name": "AU Anthropic Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -135265,44 +143654,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" + }, "cost": { - "input": 16.5, - "output": 82.5, - "cache_read": 1.65, - "cache_write": 20.625 + "input": 0.2, + "output": 1.25 } }, - "jp.anthropic.claude-sonnet-5": { - "id": "jp.anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5 (JP)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "knowledge": "2025-01-01", + "release_date": "2026-01-15", + "last_updated": "2026-01-15", "modalities": { "input": [ "text", @@ -135315,22 +143699,24 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 400000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 1.75, + "output": 14, + "cache_read": 0.17 } }, - "openai.gpt-oss-120b": { - "id": "openai.gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -135343,13 +143729,14 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01-01", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -135357,52 +143744,44 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 400000, + "output": 64000 }, "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/v1", - "shape": "responses" + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 1.25, + "output": 10, + "cache_read": 0.12 } }, - "global.anthropic.claude-sonnet-5": { - "id": "global.anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5 (Global)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "GPT-5 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -135410,40 +143789,40 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 400000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 2, + "input": 1.25, "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "cache_read": 0.12 } }, - "global.anthropic.claude-opus-4-8": { - "id": "global.anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8 (Global)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2-Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -135456,72 +143835,83 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" + }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 21, + "output": 168 } }, - "us.meta.llama4-maverick-17b-instruct-v1:0": { - "id": "us.meta.llama4-maverick-17b-instruct-v1:0", - "name": "Llama 4 Maverick 17B Instruct (US)", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 16384 + "context": 400000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 0.24, - "output": 0.97 + "input": 1.75, + "output": 14 } }, - "openai.gpt-5.4": { - "id": "openai.gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "openai/gpt-5.5-instant": { + "id": "openai/gpt-5.5-instant", + "name": "GPT-5.5 Instant", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-06-01", + "temperature": true, + "knowledge": "2025-12-01", + "release_date": "2026-05-05", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -135534,52 +143924,41 @@ }, "open_weights": false, "limit": { - "context": 272000, + "context": 400000, + "input": 400000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, "cost": { - "input": 2.75, - "output": 16.5, - "cache_read": 0.275 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "anthropic.claude-sonnet-5": { - "id": "anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ @@ -135588,28 +143967,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 400000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 2, + "input": 1.25, "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "cache_read": 0.12 } }, - "mistral.devstral-2-123b": { - "id": "mistral.devstral-2-123b", - "name": "Devstral 2 123B", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ "text" @@ -135618,62 +143999,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.4, - "output": 2 - } - }, - "zai.glm-4.7": { - "id": "zai.glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 400000, + "output": 128000 }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 0.75, + "output": 4.5 } }, - "anthropic.claude-fable-5": { - "id": "anthropic.claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -135683,10 +144034,11 @@ } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -135699,31 +144051,61 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 1, + "output": 6, + "cache_read": 0.1, + "cache_write": 1.25, + "tiers": [ + { + "input": 2, + "output": 9, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 9, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "writer.palmyra-x4-v1:0": { - "id": "writer.palmyra-x4-v1:0", - "name": "Palmyra X4", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "palmyra", - "attachment": false, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "temperature": false, + "knowledge": "2025-01-01", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ "text" @@ -135731,92 +144113,164 @@ }, "open_weights": false, "limit": { - "context": 122880, - "output": 8192 + "context": 400000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 2.5, - "output": 10 + "input": 1.75, + "output": 14, + "cache_read": 0.17 } }, - "mistral.magistral-small-2509": { - "id": "mistral.magistral-small-2509", - "name": "Magistral Small 1.2", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral", - "attachment": false, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 40000 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "qwen.qwen3-coder-480b-a35b-v1:0": { - "id": "qwen.qwen3-coder-480b-a35b-v1:0", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "openai/gpt-5.1-chat": { + "id": "openai/gpt-5.1-chat", + "name": "GPT-5.1 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-18", - "last_updated": "2025-09-18", + "knowledge": "2025-01-01", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 128000, + "output": 64000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 0.22, - "output": 1.8 + "input": 1.25, + "output": 10, + "cache_read": 0.12 } }, - "amazon.nova-micro-v1:0": { - "id": "amazon.nova-micro-v1:0", - "name": "Nova Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -135824,30 +144278,60 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.035, - "output": 0.14, - "cache_read": 0.00875 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25 + } } }, - "mistral.pixtral-large-2502-v1:0": { - "id": "mistral.pixtral-large-2502-v1:0", - "name": "Pixtral Large (25.02)", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "mistral", - "attachment": false, - "reasoning": false, + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1-Codex-Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "release_date": "2025-04-08", - "last_updated": "2025-04-08", + "knowledge": "2025-01-01", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" @@ -135855,19 +144339,23 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 400000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 2, - "output": 6 + "input": 0.25, + "output": 2, + "cache_read": 0.03 } }, - "us.anthropic.claude-opus-4-6-v1": { - "id": "us.anthropic.claude-opus-4-6-v1", - "name": "Claude Opus 4.6 (US)", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -135876,21 +144364,15 @@ "values": [ "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2025-01-01", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -135903,27 +144385,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 400000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.25, + "output": 10, + "cache_read": 0.12 } }, - "jp.anthropic.claude-opus-4-7": { - "id": "jp.anthropic.claude-opus-4-7", - "name": "Claude Opus 4.7 (JP)", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -135933,10 +144419,11 @@ } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -135949,35 +144436,61 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { "input": 5, - "output": 25, + "output": 30, "cache_read": 0.5, - "cache_write": 6.25 + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5 + } } - }, - "au.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "au.anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5 (AU)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + } + } + }, + "unorouter": { + "id": "unorouter", + "env": [ + "UNOROUTER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.unorouter.com/v1", + "name": "UnoRouter", + "doc": "https://unorouter.com/models", + "models": { + "gpt-5.5:free": { + "id": "gpt-5.5:free", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -135990,71 +144503,66 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0 } }, - "deepseek.v3-v1:0": { - "id": "deepseek.v3-v1:0", - "name": "DeepSeek-V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-09-18", - "last_updated": "2025-09-18", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 81920 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.58, - "output": 1.68 + "input": 1.8, + "output": 10.8 } }, - "anthropic.claude-opus-4-1-20250805-v1:0": { - "id": "anthropic.claude-opus-4-1-20250805-v1:0", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "glm-4.5-flash:free": { + "id": "glm-4.5-flash:free", + "name": "GLM-4.5-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -136062,144 +144570,129 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 131072, + "output": 98304 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0, + "output": 0 } }, - "jp.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "jp.anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5 (JP)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "gemma-4-31b-it:free": { + "id": "gemma-4-31b-it:free", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0 } }, - "google.gemma-3-4b-it": { - "id": "google.gemma-3-4b-it", - "name": "Gemma 3 4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, - "reasoning": false, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.04, - "output": 0.08 + "input": 1.2675, + "output": 5.3368 } }, - "eu.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "eu.anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5 (EU)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "nemotron-3-ultra-550b-a55b:free": { + "id": "nemotron-3-ultra-550b-a55b:free", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 + "input": 0, + "output": 0 } }, - "qwen.qwen3-vl-235b-a22b": { - "id": "qwen.qwen3-vl-235b-a22b", - "name": "Qwen/Qwen3-VL-235B-A22B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-5.4:free": { + "id": "gpt-5.4:free", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-04", - "last_updated": "2025-11-25", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -136207,26 +144700,27 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 0, + "output": 0 } }, - "writer.palmyra-x5-v1:0": { - "id": "writer.palmyra-x5-v1:0", - "name": "Palmyra X5", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "palmyra", + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -136235,48 +144729,36 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1040000, - "output": 8192 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 6 + "input": 0.819, + "output": 3.276 } }, - "us.anthropic.claude-sonnet-4-6": { - "id": "us.anthropic.claude-sonnet-4-6", - "name": "Claude Sonnet 4.6 (US)", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -136285,40 +144767,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.1857, + "output": 1.1142 } }, - "au.anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "au.anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5 (AU)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -136326,104 +144800,96 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 1.05, + "output": 8.4 } }, - "meta.llama3-3-70b-instruct-v1:0": { - "id": "meta.llama3-3-70b-instruct-v1:0", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 0.425, + "output": 2.125 } }, - "zai.glm-5": { - "id": "zai.glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 101376 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1, - "output": 3.2 + "input": 0.1875, + "output": 1.125 } }, - "us.anthropic.claude-opus-4-8": { - "id": "us.anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8 (US)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", @@ -136440,152 +144906,115 @@ "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.44, + "output": 7.2 } }, - "global.anthropic.claude-opus-4-7": { - "id": "global.anthropic.claude-opus-4-7", - "name": "Claude Opus 4.7 (Global)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.6001, + "output": 5.0288 } }, - "us.anthropic.claude-fable-5": { - "id": "us.anthropic.claude-fable-5", - "name": "Claude Fable 5 (US)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "step-3.7-flash:free": { + "id": "step-3.7-flash:free", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0, + "output": 0 } }, - "amazon.nova-lite-v1:0": { - "id": "amazon.nova-lite-v1:0", - "name": "Nova Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", - "attachment": true, - "reasoning": false, + "deepseek-v4-flash:free": { + "id": "deepseek-v4-flash:free", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 300000, - "output": 8192 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.015 + "input": 0, + "output": 0 } }, - "us.anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "us.anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5 (US)", + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", "family": "claude-haiku", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-02-28", "release_date": "2025-10-15", @@ -136606,105 +145035,87 @@ "output": 64000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 1.2, + "output": 6 } }, - "xai.grok-4.3": { - "id": "xai.grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + "deepseek-v4-pro:free": { + "id": "deepseek-v4-pro:free", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-06-28", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" + "output": 384000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0, + "output": 0 } }, - "mistral.voxtral-mini-3b-2507": { - "id": "mistral.voxtral-mini-3b-2507", - "name": "Voxtral Mini 3B 2507", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral", + "glm-5.2:free": { + "id": "glm-5.2:free", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "audio", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.04, - "output": 0.04 + "input": 0, + "output": 0 } }, - "moonshot.kimi-k2-thinking": { - "id": "moonshot.kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -136715,26 +145126,60 @@ }, "open_weights": true, "limit": { - "context": 262143, - "output": 16000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.6, - "output": 2.5 + "input": 0.8999, + "output": 1.7999 } }, - "meta.llama3-1-70b-instruct-v1:0": { - "id": "meta.llama3-1-70b-instruct-v1:0", - "name": "Llama 3.1 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "qwen3.5-397b-a17b:free": { + "id": "qwen3.5-397b-a17b:free", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "minimax-m2.7:free": { + "id": "minimax-m2.7:free", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -136745,27 +145190,28 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 0, + "output": 0 } }, - "us.deepseek.r1-v1:0": { - "id": "us.deepseek.r1-v1:0", - "name": "DeepSeek-R1 (US)", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -136776,87 +145222,92 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1.35, - "output": 5.4 + "input": 0.0625, + "output": 0.125 } - }, - "global.anthropic.claude-sonnet-4-6": { - "id": "global.anthropic.claude-sonnet-4-6", - "name": "Claude Sonnet 4.6 (Global)", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + } + } + }, + "salad-cloud": { + "id": "salad-cloud", + "env": [ + "SALAD_CLOUD_API_KEY" + ], + "npm": "@saladtechnologies-oss/ai-sdk-provider", + "name": "SaladCloud AI Gateway", + "doc": "https://docs.salad.com/ai-gateway/explanation/overview", + "models": { + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Qwen MoE for agentic tasks, complex reasoning, code generation, and instruction following", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "input": 262144, + "output": 262144 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.09, + "output": 0.6 } - }, - "anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + } + } + }, + "vispark": { + "id": "vispark", + "env": [ + "VISPARK_LAB_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.lab.vispark.in/v1", + "name": "Vispark", + "doc": "https://lab.vispark.in/#vision", + "models": { + "vispark/vision-large": { + "id": "vispark/vision-large", + "name": "Vision Large", + "description": "Most capable Vision model for complex reasoning, detailed media analysis, and structured output over a 1M-token context window.", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2024-05-15", + "last_updated": "2026-09", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -136865,77 +145316,66 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 7.37, + "output": 22.11 } }, - "moonshotai.kimi-k2.5": { - "id": "moonshotai.kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", - "attachment": false, + "vispark/vision-small": { + "id": "vispark/vision-small", + "name": "Vision Small", + "description": "Fast, low-cost multimodal model for understanding text, images, audio, video, and PDFs, with tool calling and a 1M-token context window.", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "release_date": "2024-05-15", + "last_updated": "2026-09", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262143, - "output": 16000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 3 + "input": 1.05, + "output": 3.16 } }, - "au.anthropic.claude-opus-4-8": { - "id": "au.anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8 (AU)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "vispark/vision-medium": { + "id": "vispark/vision-medium", + "name": "Vision Medium", + "description": "Balanced multimodal model pairing a 1M-token context window with deeper reasoning for analysis, content creation, and tool use across text, image, audio, video, and PDF inputs.", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "temperature": true, + "release_date": "2024-05-15", + "last_updated": "2026-09", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -136945,31 +145385,40 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 4.21, + "output": 12.63 } - }, - "nvidia.nemotron-nano-12b-v2": { - "id": "nvidia.nemotron-nano-12b-v2", - "name": "NVIDIA Nemotron Nano 12B v2 VL BF16", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", + } + } + }, + "siliconflow-cn": { + "id": "siliconflow-cn", + "env": [ + "SILICONFLOW_CN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.siliconflow.cn/v1", + "name": "SiliconFlow (China)", + "doc": "https://cloud.siliconflow.com/models", + "models": { + "ByteDance-Seed/Seed-OSS-36B-Instruct": { + "id": "ByteDance-Seed/Seed-OSS-36B-Instruct", + "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "seed", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "release_date": "2025-09-04", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -136977,28 +145426,31 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 0.21, + "output": 0.57 } }, - "zai.glm-4.7-flash": { - "id": "zai.glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "tencent/Hunyuan-A13B-Instruct": { + "id": "tencent/Hunyuan-A13B-Instruct", + "name": "tencent/Hunyuan-A13B-Instruct", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-06-30", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -137007,32 +145459,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.07, - "output": 0.4 + "input": 0.14, + "output": 0.57 } }, - "meta.llama4-scout-17b-instruct-v1:0": { - "id": "meta.llama4-scout-17b-instruct-v1:0", - "name": "Llama 4 Scout 17B Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": true, - "reasoning": false, + "Qwen/Qwen3.5-4B": { + "id": "Qwen/Qwen3.5-4B", + "name": "Qwen/Qwen3.5-4B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-04", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -137040,80 +145498,68 @@ }, "open_weights": true, "limit": { - "context": 3500000, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.17, - "output": 0.66 + "input": 0, + "output": 0 } }, - "au.anthropic.claude-sonnet-5": { - "id": "au.anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5 (AU)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "Qwen/Qwen3.5-27B": { + "id": "Qwen/Qwen3.5-27B", + "name": "Qwen/Qwen3.5-27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-25", + "last_updated": "2026-02-25", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.26, + "output": 2.09 } }, - "qwen.qwen3-235b-a22b-2507-v1:0": { - "id": "qwen.qwen3-235b-a22b-2507-v1:0", - "name": "Qwen3 235B A22B 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen/Qwen3.6-35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-18", - "last_updated": "2025-09-18", + "knowledge": "2025-04", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -137122,37 +145568,35 @@ "open_weights": true, "limit": { "context": 262144, - "output": 131072 + "output": 65536 }, "cost": { - "input": 0.22, - "output": 0.88 + "input": 0.23, + "output": 1.86 } }, - "eu.anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "eu.anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5 (EU)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "budget_tokens", - "min": 1024 + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2025-07-28", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -137160,38 +145604,36 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 1.1, - "output": 5.5, - "cache_read": 0.11, - "cache_write": 1.375 + "input": 0.13, + "output": 0.6 } }, - "openai.gpt-oss-20b-1:0": { - "id": "openai.gpt-oss-20b-1:0", - "name": "gpt-oss-20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "Qwen/Qwen3-14B": { + "id": "Qwen/Qwen3-14B", + "name": "Qwen/Qwen3-14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -137202,43 +145644,39 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 131000, + "output": 131000 }, "cost": { "input": 0.07, - "output": 0.3 + "output": 0.28 } }, - "jp.anthropic.claude-opus-4-8": { - "id": "jp.anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8 (JP)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "Qwen/Qwen3-8B": { + "id": "Qwen/Qwen3-8B", + "name": "Qwen/Qwen3-8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "temperature": true, + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -137246,45 +145684,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.06, + "output": 0.06 } }, - "anthropic.claude-opus-4-8": { - "id": "anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen/Qwen3-32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "temperature": true, + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -137292,161 +145724,177 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.14, + "output": 0.57 } }, - "qwen.qwen3-coder-30b-a3b-v1:0": { - "id": "qwen.qwen3-coder-30b-a3b-v1:0", - "name": "Qwen3 Coder 30B A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "Qwen/Qwen3.5-35B-A3B": { + "id": "Qwen/Qwen3.5-35B-A3B", + "name": "Qwen/Qwen3.5-35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-18", - "last_updated": "2025-09-18", + "knowledge": "2025-04", + "release_date": "2026-02-25", + "last_updated": "2026-02-25", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "output": 131072 + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.23, + "output": 1.86 } }, - "qwen.qwen3-next-80b-a3b": { - "id": "qwen.qwen3-next-80b-a3b", - "name": "Qwen/Qwen3-Next-80B-A3B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "Qwen/Qwen3.5-122B-A10B": { + "id": "Qwen/Qwen3.5-122B-A10B", + "name": "Qwen/Qwen3.5-122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-09-18", - "last_updated": "2025-11-25", + "knowledge": "2025-04", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 1.4 + "input": 0.29, + "output": 2.32 } }, - "openai.gpt-5.5": { - "id": "openai.gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "Qwen/Qwen3.5-9B": { + "id": "Qwen/Qwen3.5-9B", + "name": "Qwen/Qwen3.5-9B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-06-01", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" + "context": 262144, + "output": 65536 }, "cost": { - "input": 5.5, - "output": 33, - "cache_read": 0.55 + "input": 0.22, + "output": 1.74 } }, - "au.anthropic.claude-sonnet-4-6": { - "id": "au.anthropic.claude-sonnet-4-6", - "name": "AU Anthropic Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen/Qwen3.5-397B-A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", "image", - "pdf" + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.29, + "output": 1.74 + } + }, + "Qwen/Qwen3-Coder-480B-A35B-Instruct": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "name": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-31", + "last_updated": "2025-11-25", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -137454,49 +145902,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 + "input": 0.25, + "output": 1 } }, - "us.anthropic.claude-sonnet-5": { - "id": "us.anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5 (US)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "Qwen/Qwen3-VL-32B-Instruct": { + "id": "Qwen/Qwen3-VL-32B-Instruct", + "name": "Qwen/Qwen3-VL-32B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "release_date": "2025-10-21", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -137504,48 +145933,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.2, + "output": 0.6 } }, - "us.anthropic.claude-opus-4-5-20251101-v1:0": { - "id": "us.anthropic.claude-opus-4-5-20251101-v1:0", - "name": "Claude Opus 4.5 (US)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "Qwen/Qwen2.5-7B-Instruct": { + "id": "Qwen/Qwen2.5-7B-Instruct", + "name": "Qwen/Qwen2.5-7B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-08-01", + "release_date": "2024-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -137553,29 +145963,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 33000, + "output": 4000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.05, + "output": 0.05 } }, - "minimax.minimax-m2.5": { - "id": "minimax.minimax-m2.5", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "Qwen/Qwen3-Coder-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "name": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2025-08-01", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -137584,45 +145991,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 98304 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.07, + "output": 0.28 } }, - "eu.anthropic.claude-opus-4-8": { - "id": "eu.anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8 (EU)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "Qwen/Qwen2.5-72B-Instruct": { + "id": "Qwen/Qwen2.5-72B-Instruct", + "name": "Qwen/Qwen2.5-72B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "temperature": true, + "release_date": "2024-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -137630,45 +146023,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 33000, + "output": 4000 }, "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 + "input": 0.59, + "output": 0.59 } }, - "eu.anthropic.claude-opus-4-7": { - "id": "eu.anthropic.claude-opus-4-7", - "name": "Claude Opus 4.7 (EU)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "Qwen/Qwen3-VL-30B-A3B-Thinking": { + "id": "Qwen/Qwen3-VL-30B-A3B-Thinking", + "name": "Qwen/Qwen3-VL-30B-A3B-Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "release_date": "2025-10-11", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -137676,28 +146055,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 + "input": 0.29, + "output": 1 } }, - "meta.llama3-1-8b-instruct-v1:0": { - "id": "meta.llama3-1-8b-instruct-v1:0", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "Qwen/Qwen3-30B-A3B-Instruct-2507": { + "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2025-07-30", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -137706,39 +146083,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.22, - "output": 0.22 + "input": 0.09, + "output": 0.3 } }, - "us.anthropic.claude-opus-4-1-20250805-v1:0": { - "id": "us.anthropic.claude-opus-4-1-20250805-v1:0", - "name": "Claude Opus 4.1 (US)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "Qwen/Qwen3-VL-32B-Thinking": { + "id": "Qwen/Qwen3-VL-32B-Thinking", + "name": "Qwen/Qwen3-VL-32B-Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-10-21", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -137746,28 +146117,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.2, + "output": 1.5 } }, - "meta.llama4-maverick-17b-instruct-v1:0": { - "id": "meta.llama4-maverick-17b-instruct-v1:0", - "name": "Llama 4 Maverick 17B Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "Qwen/Qwen3-VL-8B-Instruct": { + "id": "Qwen/Qwen3-VL-8B-Instruct", + "name": "Qwen/Qwen3-VL-8B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2025-10-15", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -137777,48 +146146,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 16384 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.24, - "output": 0.97 + "input": 0.18, + "output": 0.68 } }, - "global.anthropic.claude-opus-4-5-20251101-v1:0": { - "id": "global.anthropic.claude-opus-4-5-20251101-v1:0", - "name": "Claude Opus 4.5 (Global)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "Qwen/Qwen3-VL-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", + "name": "Qwen/Qwen3-VL-30B-A3B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-08-01", + "release_date": "2025-10-05", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -137826,29 +146179,33 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.29, + "output": 1 } }, - "nvidia.nemotron-super-3-120b": { - "id": "nvidia.nemotron-super-3-120b", - "name": "NVIDIA Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "deepseek-ai/DeepSeek-R1": { + "id": "deepseek-ai/DeepSeek-R1", + "name": "deepseek-ai/DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "release_date": "2025-05-28", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -137857,82 +146214,72 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.15, - "output": 0.65 + "input": 0.5, + "output": 2.18 } }, - "eu.anthropic.claude-sonnet-5": { - "id": "eu.anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5 (EU)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 384000 }, "cost": { - "input": 2.2, - "output": 11, - "cache_read": 0.22, - "cache_write": 2.75 + "input": 0.14, + "output": 0.28, + "cache_read": 0.003 } }, - "nvidia.nemotron-nano-3-30b": { - "id": "nvidia.nemotron-nano-3-30b", - "name": "NVIDIA Nemotron Nano 3 30B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "deepseek-ai/DeepSeek-OCR": { + "id": "deepseek-ai/DeepSeek-OCR", + "name": "deepseek-ai/DeepSeek-OCR", + "description": "OCR model for extracting structured text from documents and screenshots", + "attachment": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2025-10-20", + "last_updated": "2025-10-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -137940,96 +146287,66 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.06, - "output": 0.24 + "input": 0, + "output": 0 } }, - "eu.anthropic.claude-opus-4-5-20251101-v1:0": { - "id": "eu.anthropic.claude-opus-4-5-20251101-v1:0", - "name": "Claude Opus 4.5 (EU)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "deepseek-ai/DeepSeek-V4-Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-08-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1049000, + "output": 393000 }, "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 + "input": 1.74, + "output": 3.48, + "cache_read": 0.145 } }, - "global.anthropic.claude-opus-4-6-v1": { - "id": "global.anthropic.claude-opus-4-6-v1", - "name": "Claude Opus 4.6 (Global)", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "deepseek-ai/DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2025-12-03", + "last_updated": "2025-12-03", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -138037,40 +146354,34 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 164000, + "output": 164000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.27, + "output": 0.42 } }, - "us.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "us.anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5 (US)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "deepseek-ai/DeepSeek-V3.1-Terminus": { + "id": "deepseek-ai/DeepSeek-V3.1-Terminus", + "name": "deepseek-ai/DeepSeek-V3.1-Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -138078,33 +146389,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 164000, + "output": 164000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.27, + "output": 1 } }, - "google.gemma-3-12b-it": { - "id": "google.gemma-3-12b-it", - "name": "Google Gemma 3 12B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "deepseek-ai/DeepSeek-V3": { + "id": "deepseek-ai/DeepSeek-V3", + "name": "deepseek-ai/DeepSeek-V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "release_date": "2024-12-26", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -138112,27 +146419,27 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.049999999999999996, - "output": 0.09999999999999999 + "input": 0.25, + "output": 1 } }, - "minimax.minimax-m2.1": { - "id": "minimax.minimax-m2.1", - "name": "MiniMax M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "stepfun-ai/Step-3.5-Flash": { + "id": "stepfun-ai/Step-3.5-Flash", + "name": "stepfun-ai/Step-3.5-Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "family": "step", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -138141,45 +146448,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.1, + "output": 0.3 } }, - "eu.anthropic.claude-fable-5": { - "id": "eu.anthropic.claude-fable-5", - "name": "Claude Fable 5 (EU)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "inclusionAI/Ling-flash-2.0": { + "id": "inclusionAI/Ling-flash-2.0", + "name": "inclusionAI/Ling-flash-2.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "structured_output": true, + "temperature": true, + "release_date": "2025-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -138187,30 +146480,38 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 11, - "output": 55, - "cache_read": 1.1, - "cache_write": 13.75 + "input": 0.14, + "output": 0.57 } }, - "deepseek.v3.2": { - "id": "deepseek.v3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -138221,26 +146522,28 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 81920 + "context": 1049000, + "output": 262000 }, "cost": { - "input": 0.62, - "output": 1.85 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "mistral.ministral-3-14b-instruct": { - "id": "mistral.ministral-3-14b-instruct", - "name": "Ministral 14B 3.0", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "zai-org/GLM-4.5-Air": { + "id": "zai-org/GLM-4.5-Air", + "name": "zai-org/GLM-4.5-Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "release_date": "2025-07-28", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -138251,69 +146554,56 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 0.14, + "output": 0.86 } - } - } - }, - "umans-ai": { - "id": "umans-ai", - "env": [ - "UMANS_AI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.code.umans.ai/v1", - "name": "Umans AI", - "doc": "https://app.umans.ai/offers/code/docs/orgs", - "models": { - "umans-kimi-k2.7": { - "id": "umans-kimi-k2.7", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + }, + "Pro/deepseek-ai/DeepSeek-R1": { + "id": "Pro/deepseek-ai/DeepSeek-R1", + "name": "Pro/deepseek-ai/DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2025-05-28", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.5, + "output": 2.18 } }, - "umans-glm-5.1": { - "id": "umans-glm-5.1", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "Pro/deepseek-ai/DeepSeek-V3.2": { + "id": "Pro/deepseek-ai/DeepSeek-V3.2", + "name": "Pro/deepseek-ai/DeepSeek-V3.2", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -138322,145 +146612,142 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-12-03", + "last_updated": "2025-12-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 164000, + "output": 164000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.29 + "input": 0.27, + "output": 0.42 } }, - "umans-coder": { - "id": "umans-coder", - "name": "Umans Coder", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Pro/deepseek-ai/DeepSeek-V3": { + "id": "Pro/deepseek-ai/DeepSeek-V3", + "name": "Pro/deepseek-ai/DeepSeek-V3", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2024-12-26", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.25, + "output": 1 } }, - "umans-flash": { - "id": "umans-flash", - "name": "Umans Flash", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, + "Pro/deepseek-ai/DeepSeek-V3.1-Terminus": { + "id": "Pro/deepseek-ai/DeepSeek-V3.1-Terminus", + "name": "Pro/deepseek-ai/DeepSeek-V3.1-Terminus", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] } ], "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-09-29", + "last_updated": "2025-11-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 164000, + "output": 164000 + }, + "cost": { + "input": 0.27, + "output": 1 + } + }, + "Pro/MiniMaxAI/MiniMax-M2.5": { + "id": "Pro/MiniMaxAI/MiniMax-M2.5", + "name": "Pro/MiniMaxAI/MiniMax-M2.5", + "description": "Frontier MiniMax model for engineering, office tasks, and agentic reasoning", + "family": "minimax", + "attachment": false, + "reasoning": false, + "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, - "temperature": false, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": true, + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 192000, + "output": 131000 }, "cost": { - "input": 0.15, - "output": 1, - "cache_read": 0.05 + "input": 0.3, + "output": 1.22 } }, - "umans-glm-5.2": { - "id": "umans-glm-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "Pro/moonshotai/Kimi-K2.6": { + "id": "Pro/moonshotai/Kimi-K2.6", + "name": "Pro/moonshotai/Kimi-K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -138472,40 +146759,35 @@ }, "open_weights": true, "limit": { - "context": 405504, - "output": 131072 - }, + "context": 262000, + "output": 262000 + }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } - } - } - }, - "togetherai": { - "id": "togetherai", - "env": [ - "TOGETHER_API_KEY" - ], - "npm": "@ai-sdk/togetherai", - "name": "Together AI", - "doc": "https://docs.together.ai/docs/serverless-models", - "models": { - "LiquidAI/LFM2-24B-A2B": { - "id": "LiquidAI/LFM2-24B-A2B", - "name": "LFM2-24B-A2B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "liquid", + }, + "Pro/moonshotai/Kimi-K2.5": { + "id": "Pro/moonshotai/Kimi-K2.5", + "name": "Pro/moonshotai/Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-25", - "last_updated": "2026-02-25", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -138513,25 +146795,31 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.03, - "output": 0.12 + "input": 0.45, + "output": 2.25, + "cache_read": 0.07 } }, - "meta-llama/Meta-Llama-3-8B-Instruct-Lite": { - "id": "meta-llama/Meta-Llama-3-8B-Instruct-Lite", - "name": "Meta Llama 3 8B Instruct Lite", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", + "Pro/zai-org/GLM-5.1": { + "id": "Pro/zai-org/GLM-5.1", + "name": "Pro/zai-org/GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2024-04-18", - "last_updated": "2024-04-18", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ "text" @@ -138542,26 +146830,36 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 205000, + "output": 205000 }, "cost": { - "input": 0.14, - "output": 0.14 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "meta-llama/Llama-3.3-70B-Instruct-Turbo": { - "id": "meta-llama/Llama-3.3-70B-Instruct-Turbo", - "name": "Llama 3.3 70B", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", + "Pro/zai-org/GLM-5": { + "id": "Pro/zai-org/GLM-5", + "name": "Pro/zai-org/GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2026-07-02", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -138572,19 +146870,60 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 205000, + "output": 205000 }, "cost": { - "input": 1.04, - "output": 1.04 + "input": 1, + "output": 3.2 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "PaddlePaddle/PaddleOCR-VL-1.5": { + "id": "PaddlePaddle/PaddleOCR-VL-1.5", + "name": "PaddlePaddle/PaddleOCR-VL-1.5", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-01-29", + "last_updated": "2026-01-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 16384, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + } + } + }, + "regolo-ai": { + "id": "regolo-ai", + "env": [ + "REGOLO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.regolo.ai/v1", + "name": "Regolo AI", + "doc": "https://docs.regolo.ai/", + "models": { + "qwen3.8-27b": { + "id": "qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -138595,9 +146934,8 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", @@ -138610,33 +146948,57 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131000 + "context": 120000, + "output": 120000 }, "cost": { - "input": 1.2, - "output": 4.5, - "cache_read": 0.2 + "input": 0.58, + "output": 2.42 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-k2", - "attachment": false, + "brick-v1-beta": { + "id": "brick-v1-beta", + "name": "Brick v1 Beta", + "description": "Semantic router by Regolo.ai that directs each request to the most suitable model, optimizing costs and performance", + "family": "model-router", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-02-06", + "last_updated": "2026-02-06", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 100000, + "output": 15000 + }, + "status": "beta", + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen3.5-122b": { + "id": "qwen3.5-122b", + "name": "Qwen3.5-122B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ "text", @@ -138649,27 +147011,25 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "output": 16384 }, - "status": "deprecated", "cost": { - "input": 0.5, - "output": 2.8 + "input": 0.9, + "output": 3.6 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Kimi coding model for software agents, refactors, and repository reasoning", - "family": "kimi-k2", + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "Qwen3-Coder-Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-14", - "last_updated": "2026-06-14", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ "text" @@ -138681,32 +147041,66 @@ "open_weights": true, "limit": { "context": 262144, - "output": 131072 + "output": 16384 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.3, + "output": 1.2 } }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "apertus-70b": { + "id": "apertus-70b", + "name": "Apertus 70B", + "description": "Fully open 70B multilingual LLM supporting 1800+ languages with 65K context. Trained on 15T tokens of compliant open data. Apache 2.0, EU AI Act compliant.", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-09-02", + "last_updated": "2025-09-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 30000, + "output": 30000 + }, + "cost": { + "input": 0.46, + "output": 2.42 + } + }, + "glm5.2": { + "id": "glm5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -138714,26 +147108,25 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 96000, + "output": 96000 }, "cost": { - "input": 0.39, - "output": 0.97 + "input": 2.31, + "output": 6 } }, - "google/gemma-3n-E4B-it": { - "id": "google/gemma-3n-E4B-it", - "name": "Gemma 3N E4B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "qwen3-reranker-4b": { + "id": "qwen3-reranker-4b", + "name": "Qwen3-Reranker-4B", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "temperature": false, + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ "text" @@ -138745,58 +147138,52 @@ "open_weights": true, "limit": { "context": 32768, - "output": 32768 + "output": 8192 }, "cost": { - "input": 0.06, + "input": 0.12, "output": 0.12 } }, - "Qwen/Qwen3.7-Max": { - "id": "Qwen/Qwen3.7-Max", - "name": "Qwen3.7 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "deepseek-ocr-2": { + "id": "deepseek-ocr-2", + "name": "DeepSeek OCR 2", + "description": "High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes", + "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-07-02", + "tool_call": false, + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 500000 + "context": 4000, + "output": 4000 }, "cost": { - "input": 1.25, - "output": 3.75 + "input": 0, + "output": 0 } }, - "Qwen/Qwen3.6-Plus": { - "id": "Qwen/Qwen3.6-Plus", - "name": "Qwen3.6 Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ "text" @@ -138805,22 +147192,22 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 500000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.6, + "output": 2.7 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen3.5 397B A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "gemma4-31b": { + "id": "gemma4-31b", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -138828,9 +147215,10 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-16", - "last_updated": "2026-06-15", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -138842,57 +147230,64 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 130000 + "context": 100000, + "output": 100000 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.46, + "output": 2.42 } }, - "Qwen/Qwen3-Coder-Next-FP8": { - "id": "Qwen/Qwen3-Coder-Next-FP8", - "name": "Qwen3 Coder Next FP8", - "description": "Legacy model retained for compatibility with older integrations", - "family": "qwen", - "attachment": false, + "brick-complexity-pro": { + "id": "brick-complexity-pro", + "name": "Brick Complexity Pro", + "description": "Complexity classifier that powers the Brick semantic router by extracting query difficulty", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2026-02-03", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", + "release_date": "2026-02-06", + "last_updated": "2026-02-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 100000, + "output": 15000 }, - "status": "deprecated", "cost": { - "input": 0.5, - "output": 1.2 + "input": 0.12, + "output": 0.46 } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Legacy model retained for compatibility with older integrations", - "family": "qwen", + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT-OSS-20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ "text" @@ -138903,30 +147298,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 16384 }, - "status": "deprecated", "cost": { - "input": 2, - "output": 2 + "input": 0.4, + "output": 1.8 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507-tput": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507-tput", - "name": "Qwen3 235B A22B Instruct 2507 FP8", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5-9B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -138935,25 +147330,24 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "output": 8192 }, "cost": { - "input": 0.2, + "input": 0.15, "output": 0.6 } }, - "Qwen/Qwen2.5-7B-Instruct-Turbo": { - "id": "Qwen/Qwen2.5-7B-Instruct-Turbo", - "name": "Qwen 2.5 7B Instruct Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "qwen3-embedding-8b": { + "id": "qwen3-embedding-8b", + "name": "Qwen3-Embedding-8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "tool_call": false, + "temperature": false, + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ "text" @@ -138965,30 +147359,25 @@ "open_weights": true, "limit": { "context": 32768, - "output": 32768 + "output": 8192 }, "cost": { - "input": 0.3, - "output": 0.3 + "input": 0.1, + "output": 0.1 } }, - "Qwen/Qwen3.5-9B": { - "id": "Qwen/Qwen3.5-9B", - "name": "Qwen3.5 9B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "mistral-small-4-119b": { + "id": "mistral-small-4-119b", + "name": "Mistral Small 4 119B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2026-03-15", + "last_updated": "2026-03-15", "modalities": { "input": [ "text", @@ -138998,59 +147387,48 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0.17, - "output": 0.25 + "input": 0.75, + "output": 3 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "qwen-image": { + "id": "qwen-image", + "name": "Qwen-Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 8192, + "output": 4096 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.5, + "output": 2 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT-OSS-120B", "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "family": "gpt-oss", "attachment": false, @@ -139066,7 +147444,6 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, "release_date": "2025-08-05", "last_updated": "2025-08-05", @@ -139078,52 +147455,62 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.2 + "input": 1, + "output": 4.2 } }, - "pearl-ai/gemma-4-31b-it": { - "id": "pearl-ai/gemma-4-31b-it", - "name": "Pearl AI Gemma 4 31B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "faster-whisper-large-v3": { + "id": "faster-whisper-large-v3", + "name": "Faster Whisper Large v3", + "description": "Open Whisper checkpoint for robust multilingual transcription and captioning", + "family": "whisper", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2024-10-01", + "last_updated": "2024-10-01", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "output": 32000 + "context": 448, + "output": 4096 }, "cost": { - "input": 0.28, - "output": 0.86 + "input": 0, + "output": 0 } - }, - "nvidia/nemotron-3-ultra-550b-a55b": { - "id": "nvidia/nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, + } + } + }, + "xiaomi-token-plan-ams": { + "id": "xiaomi-token-plan-ams", + "env": [ + "XIAOMI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://token-plan-ams.xiaomimimo.com/v1", + "name": "Xiaomi Token Plan (Europe)", + "doc": "https://platform.xiaomimimo.com/#/docs", + "models": { + "mimo-v2.6-pro": { + "id": "mimo-v2.6-pro", + "name": "MiMo-V2.6-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -139131,13 +147518,18 @@ } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -139145,54 +147537,89 @@ }, "open_weights": true, "limit": { - "context": 512300, - "output": 512300 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 3.6, - "cache_read": 0.2 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "deepcogito/cogito-v2-1-671b": { - "id": "deepcogito/cogito-v2-1-671b", - "name": "Cogito v2.1 671B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "cogito", - "attachment": false, + "mimo-v2.6-flash": { + "id": "mimo-v2.6-flash", + "name": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" } ], - "tool_call": false, + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 1.25 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "mimo-v2.5-tts-voiceclone": { + "id": "mimo-v2.5-tts-voiceclone", + "name": "MiMo-V2.5-TTS-VoiceClone", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -139204,10 +147631,10 @@ "interleaved": { "field": "reasoning_content" }, - "structured_output": true, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -139218,41 +147645,63 @@ }, "open_weights": true, "limit": { - "context": 202752, + "context": 1048576, "output": 131072 }, "cost": { - "input": 1, - "output": 3.2 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "mimo-v2-tts": { + "id": "mimo-v2-tts", + "name": "MiMo-V2-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -139261,91 +147710,80 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 164000 + "context": 1048576, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "mimo-v2.5-tts": { + "id": "mimo-v2.5-tts", + "name": "MiMo-V2.5-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-11", - "release_date": "2026-04-07", - "last_updated": "2026-07-02", + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 8192, + "output": 8192 }, "cost": { - "input": 1.4, - "output": 4.4 + "input": 0, + "output": 0 } }, - "deepseek-ai/DeepSeek-R1": { - "id": "deepseek-ai/DeepSeek-R1", - "name": "DeepSeek-R1", - "description": "Legacy model retained for compatibility with older integrations", - "family": "deepseek-thinking", + "mimo-v2.5-tts-voicedesign": { + "id": "mimo-v2.5-tts-voicedesign", + "name": "MiMo-V2.5-TTS-VoiceDesign", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-03-24", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": true, "limit": { - "context": 163839, - "output": 163839 + "context": 8192, + "output": 8192 }, - "status": "deprecated", "cost": { - "input": 3, - "output": 7 + "input": 0, + "output": 0 } }, - "deepseek-ai/DeepSeek-V3-1": { - "id": "deepseek-ai/DeepSeek-V3-1", - "name": "DeepSeek V3.1", - "description": "Legacy model retained for compatibility with older integrations", - "family": "deepseek", - "attachment": false, + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -139353,13 +147791,19 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -139367,26 +147811,35 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 1048576, "output": 131072 }, - "status": "deprecated", "cost": { - "input": 0.6, - "output": 1.7 + "input": 0, + "output": 0, + "cache_read": 0 } - }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", + } + } + }, + "inceptron": { + "id": "inceptron", + "env": [ + "INCEPTRON_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.inceptron.io/v1", + "name": "Inceptron", + "doc": "https://docs.inceptron.io", + "models": { + "deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -139401,8 +147854,9 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -139413,30 +147867,37 @@ }, "open_weights": true, "limit": { - "context": 512000, - "output": 384000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.2 + "input": 0.13, + "output": 0.28, + "cache_read": 0.03, + "cache_write": 0 } }, - "deepseek-ai/DeepSeek-V3": { - "id": "deepseek-ai/DeepSeek-V3", - "name": "DeepSeek-V3", - "description": "Legacy model retained for compatibility with older integrations", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-12-26", - "last_updated": "2025-05-29", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -139444,58 +147905,31 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 262144 }, - "status": "deprecated", "cost": { - "input": 1.25, - "output": 1.25 + "input": 0.53, + "output": 3.39, + "cache_read": 0.17, + "cache_write": 0 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "MiniMaxAI/MiniMax-M3": { - "id": "MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, + "knowledge": "2025-01", "release_date": "2026-06-12", "last_updated": "2026-06-12", "modalities": { @@ -139509,28 +147943,40 @@ }, "open_weights": true, "limit": { - "context": 524288, - "output": 250000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.66, + "output": 3.4, + "cache_read": 0.18, + "cache_write": 0 } }, - "MiniMaxAI/MiniMax-M2.7": { - "id": "MiniMaxAI/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -139541,27 +147987,55 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.71, + "output": 2.35, + "cache_read": 0.12, + "cache_write": 0 } - }, - "essentialai/Rnj-1-Instruct": { - "id": "essentialai/Rnj-1-Instruct", - "name": "Rnj-1 Instruct", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "rnj", + } + } + }, + "upstage": { + "id": "upstage", + "env": [ + "UPSTAGE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.upstage.ai/v1/solar", + "name": "Upstage", + "doc": "https://developers.upstage.ai/docs/apis/chat", + "models": { + "solar-pro4": { + "id": "solar-pro4", + "name": "Solar Pro 4", + "description": "Upstage's flagship model, specialized for agentic use", + "family": "solar-pro", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-05", - "last_updated": "2025-12-05", + "knowledge": "2026-02", + "release_date": "2026-08-06", + "last_updated": "2026-08-06", "modalities": { "input": [ "text" @@ -139570,40 +148044,39 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 524288, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } - } - } - }, - "frogbot": { - "id": "frogbot", - "env": [ - "FROGBOT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://app.frogbot.ai/api/v1", - "name": "FrogBot", - "doc": "https://docs.frogbot.ai", - "models": { - "minimax-m2-5": { - "id": "minimax-m2-5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": true, - "reasoning": false, + }, + "solar-pro3": { + "id": "solar-pro3", + "name": "solar-pro3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "solar-pro", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2025-01-15", - "last_updated": "2025-02-22", + "knowledge": "2025-03", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -139614,31 +148087,29 @@ }, "open_weights": false, "limit": { - "context": 192000, + "context": 131072, "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.25, + "output": 0.25 } }, - "kimi-k2-6": { - "id": "kimi-k2-6", - "name": "Kimi-K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "solar-mini": { + "id": "solar-mini", + "name": "solar-mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "solar-mini", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "knowledge": "2024-09", + "release_date": "2024-06-12", + "last_updated": "2025-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -139646,27 +148117,35 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 + "context": 32768, + "output": 4096 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.15, + "output": 0.15 } }, - "zai-glm-5-1": { - "id": "zai-glm-5-1", - "name": "Z.AI GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": true, - "reasoning": false, + "solar-pro2": { + "id": "solar-pro2", + "name": "solar-pro2", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "solar-pro", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-01-20", - "last_updated": "2025-02-22", + "knowledge": "2025-03", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text" @@ -139675,30 +148154,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 198000, + "context": 65536, "output": 8192 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.25, + "output": 0.25 } - }, - "grok-code-fast-1": { - "id": "grok-code-fast-1", - "name": "Grok 4.1 Fast (Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + } + } + }, + "vultr": { + "id": "vultr", + "env": [ + "VULTR_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.vultrinference.com/v1", + "name": "Vultr", + "doc": "https://api.vultrinference.com/", + "models": { + "XiaomiMiMo/MiMo-V2.5-Pro": { + "id": "XiaomiMiMo/MiMo-V2.5-Pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -139707,223 +148197,197 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 1.5, - "cache_read": 0.02 + "input": 0.55, + "output": 1.65 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-05", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", "image", - "audio", "video", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 262144, "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.31 + "input": 0.3, + "output": 2 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-07-17", - "last_updated": "2025-07-17", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "audio", "video", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 262144, "output": 65536 }, "cost": { "input": 0.3, - "output": 2.5, - "cache_read": 0.075 + "output": 2 } }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.3, + "output": 1 } }, - "qwen-3-6-plus": { - "id": "qwen-3-6-plus", - "name": "Qwen 3.6 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "MiniMaxAI/MiniMax-M2.7": { + "id": "MiniMaxAI/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-03", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1 + "input": 0.3, + "output": 1.2 } }, - "grok-4-3": { - "id": "grok-4-3", - "name": "Grok 4.3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-11", - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.3, + "output": 1.2 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek v4 Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", - "attachment": true, - "reasoning": false, + "zai-org/GLM-5.2-FP8": { + "id": "zai-org/GLM-5.2-FP8", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -139932,105 +148396,91 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 393216, + "output": 131072 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.14 + "input": 0.85, + "output": 3.1 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "nvidia/DeepSeek-V3.2-NVFP4": { + "id": "nvidia/DeepSeek-V3.2-NVFP4", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.55, + "output": 1.65 } }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "Grok 4.1 Fast (Non-Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": false, + "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16": { + "id": "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16", + "name": "NVIDIA Nemotron 3 Nano Omni", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "knowledge": "2025-05", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.13, + "output": 0.38 } }, - "minimax-m2-7": { - "id": "minimax-m2-7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": true, - "reasoning": false, + "nvidia/Nemotron-Cascade-2-30B-A3B": { + "id": "nvidia/Nemotron-Cascade-2-30B-A3B", + "name": "NVIDIA Nemotron Cascade 2", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -140039,199 +148489,229 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 192000, - "output": 8192 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.15, + "output": 0.6 } - }, - "gpt-5-3-codex": { - "id": "gpt-5-3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", - "attachment": true, + } + } + }, + "huggingface": { + "id": "huggingface", + "env": [ + "HF_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://router.huggingface.co/v1", + "name": "Hugging Face", + "doc": "https://huggingface.co/docs/inference-providers", + "models": { + "tencent/Hy3": { + "id": "tencent/Hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "structured_output": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, + "context": 262144, + "input": 192000, "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.14, + "output": 0.58 } }, - "gpt-5-4-nano": { - "id": "gpt-5-4-nano", - "name": "GPT-5.4 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, + "tencent/Hy4-preview": { + "id": "tencent/Hy4-preview", + "name": "Hy4 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-08-28", + "last_updated": "2026-08-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.834, + "output": 2.501 } }, - "gpt-5-4-mini": { - "id": "gpt-5-4-mini", - "name": "GPT-5.4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "meta-llama/Llama-3.1-8B-Instruct": { + "id": "meta-llama/Llama-3.1-8B-Instruct", + "name": "Llama-3.1-8B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 4096 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.06, + "output": 0.06 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 4096 + }, + "cost": { + "input": 0.59, + "output": 0.79 + } + }, + "XiaomiMiMo/MiMo-V2-Flash": { + "id": "XiaomiMiMo/MiMo-V2-Flash", + "name": "MiMo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "knowledge": "2024-12", + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 128000 + "context": 262144, + "output": 4096 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 0.1, + "output": 0.3 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "XiaomiMiMo/MiMo-V2.5-Pro": { + "id": "XiaomiMiMo/MiMo-V2.5-Pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -140242,101 +148722,96 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 1, + "output": 3 } }, - "gpt-5-5": { - "id": "gpt-5-5", - "name": "GPT-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, + "XiaomiMiMo/MiMo-V2.5": { + "id": "XiaomiMiMo/MiMo-V2.5", + "name": "MiMo-V2.5", + "description": "MiMo model for long-context reasoning, perception, and agentic tasks", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 272000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 0.4, + "output": 2 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "thinkingmachines/Inkling-Small": { + "id": "thinkingmachines/Inkling-Small", + "name": "Inkling Small", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 524288, + "output": 1048576 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.5, + "output": 1.2 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "thinkingmachines/Inkling": { + "id": "thinkingmachines/Inkling", + "name": "Inkling", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -140347,98 +148822,77 @@ "medium", "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1, + "output": 4.05 } }, - "gemini-3-1-pro-preview": { - "id": "gemini-3-1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma 3 27B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-02-18", - "last_updated": "2026-02-18", + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 0.08, + "output": 0.16 } }, - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "Grok 4.1 Fast (Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "google/gemma-3-12b-it": { + "id": "google/gemma-3-12b-it", + "name": "Gemma 3 12B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text", @@ -140448,136 +148902,97 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.05, + "output": 0.15 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "google/gemma-3-4b-it": { + "id": "google/gemma-3-4b-it", + "name": "Gemma 3 4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.05, + "output": 0.1 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 0.14, + "output": 0.4 } }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "google/gemma-4-26B-A4B-it": { + "id": "google/gemma-4-26B-A4B-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -140585,109 +149000,99 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 262144, "output": 32768 }, "cost": { - "input": 0.07, - "output": 0.2 + "input": 0.13, + "output": 0.4 } - } - } - }, - "openrouter": { - "id": "openrouter", - "env": [ - "OPENROUTER_API_KEY" - ], - "npm": "@openrouter/ai-sdk-provider", - "api": "https://openrouter.ai/api/v1", - "name": "OpenRouter", - "doc": "https://openrouter.ai/models", - "models": { - "inclusionai/ling-2.6-1t": { - "id": "inclusionai/ling-2.6-1t", - "name": "Ling-2.6-1T", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "ling", - "attachment": false, + }, + "Qwen/Qwen3-VL-235B-A22B-Instruct": { + "id": "Qwen/Qwen3-VL-235B-A22B-Instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, + "context": 131072, "output": 32768 }, "cost": { - "input": 0.075, - "output": 0.625, - "cache_read": 0.015 + "input": 0.3, + "output": 1.5 } }, - "inclusionai/ring-2.6-1t": { - "id": "inclusionai/ring-2.6-1t", - "name": "Ring-2.6-1T", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "ring", - "attachment": false, + "Qwen/Qwen3.8-27B": { + "id": "Qwen/Qwen3.8-27B", + "name": "Qwen3.8 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", + "low", + "medium", "xhigh" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-05-08", - "last_updated": "2026-05-08", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "output": 32768 }, "cost": { - "input": 0.075, - "output": 0.625, - "cache_read": 0.015 + "input": 0.4, + "output": 3 } }, - "inclusionai/ling-2.6-flash": { - "id": "inclusionai/ling-2.6-flash", - "name": "Ling-2.6-flash", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", + "Qwen/Qwen3-Next-80B-A3B-Instruct": { + "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", + "name": "Qwen3-Next-80B-A3B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-04", + "release_date": "2025-09-11", + "last_updated": "2025-09-11", "modalities": { "input": [ "text" @@ -140696,29 +149101,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 66536 }, "cost": { - "input": 0.01, - "output": 0.03, - "cache_read": 0.002 + "input": 0.25, + "output": 1 } }, - "ibm-granite/granite-4.0-h-micro": { - "id": "ibm-granite/granite-4.0-h-micro", - "name": "Granite 4.0 Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "granite", + "Qwen/Qwen2.5-Coder-32B-Instruct": { + "id": "Qwen/Qwen2.5-Coder-32B-Instruct", + "name": "Qwen2.5-Coder-32B-Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-10-20", - "last_updated": "2025-10-20", + "release_date": "2024-11-12", + "last_updated": "2024-11-12", "modalities": { "input": [ "text" @@ -140729,26 +149133,26 @@ }, "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.017, - "output": 0.112 + "input": 0.06, + "output": 0.2 } }, - "ibm-granite/granite-4.1-8b": { - "id": "ibm-granite/granite-4.1-8b", - "name": "Granite 4.1 8B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "granite", + "Qwen/Qwen3-Coder-Next": { + "id": "Qwen/Qwen3-Coder-Next", + "name": "Qwen3-Coder-Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "knowledge": "2025-04", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ "text" @@ -140759,31 +149163,41 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.05, - "output": 0.1, - "cache_read": 0.05 + "input": 0.2, + "output": 1.5 } }, - "meta-llama/llama-3.1-8b-instruct": { - "id": "meta-llama/llama-3.1-8b-instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "Qwen/Qwen3-VL-235B-A22B-Thinking": { + "id": "Qwen/Qwen3-VL-235B-A22B-Thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -140792,29 +149206,30 @@ "open_weights": true, "limit": { "context": 131072, - "output": 16384 + "output": 32768 }, "cost": { - "input": 0.02, - "output": 0.03 + "input": 0.98, + "output": 3.95 } }, - "meta-llama/llama-3.1-70b-instruct": { - "id": "meta-llama/llama-3.1-70b-instruct", - "name": "Llama 3.1 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "Qwen/Qwen3.5-27B": { + "id": "Qwen/Qwen3.5-27B", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -140822,27 +149237,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 0.4 + "input": 0.3, + "output": 2.4 } }, - "meta-llama/llama-3.2-1b-instruct": { - "id": "meta-llama/llama-3.2-1b-instruct", - "name": "Llama 3.2 1B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "Qwen/Qwen3-Next-80B-A3B-Thinking": { + "id": "Qwen/Qwen3-Next-80B-A3B-Thinking", + "name": "Qwen3-Next-80B-A3B-Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2025-04", + "release_date": "2025-09-11", + "last_updated": "2025-09-11", "modalities": { "input": [ "text" @@ -140853,27 +149267,27 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 60000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.027, - "output": 0.201 + "input": 0.3, + "output": 2 } }, - "meta-llama/llama-4-maverick": { - "id": "meta-llama/llama-4-maverick", - "name": "Llama 4 Maverick", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", @@ -140885,31 +149299,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { "input": 0.15, - "output": 0.6 + "output": 0.95 } }, - "meta-llama/llama-3.2-11b-vision-instruct": { - "id": "meta-llama/llama-3.2-11b-vision-instruct", - "name": "Llama 3.2 11B Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen3-235B-A22B-Thinking-2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2025-04", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -140917,27 +149330,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.345, - "output": 0.345 + "input": 0.3, + "output": 3 } }, - "meta-llama/llama-3.3-70b-instruct:free": { - "id": "meta-llama/llama-3.3-70b-instruct:free", - "name": "Llama 3.3 70B Instruct (free)", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B-A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2025-07-21", + "last_updated": "2025-07-21", "modalities": { "input": [ "text" @@ -140948,27 +149360,28 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.855, + "output": 2.565 } }, - "meta-llama/llama-3.3-70b-instruct": { - "id": "meta-llama/llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "Qwen/Qwen3-235B-A22B": { + "id": "Qwen/Qwen3-235B-A22B", + "name": "Qwen3 235B-A22B", + "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -140979,27 +149392,28 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 40960, "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.32 + "input": 0.2, + "output": 0.8 } }, - "meta-llama/llama-3.2-3b-instruct:free": { - "id": "meta-llama/llama-3.2-3b-instruct:free", - "name": "Llama 3.2 3B Instruct (free)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -141011,29 +149425,28 @@ "open_weights": true, "limit": { "context": 131072, - "output": 131072 + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.29, + "output": 0.59 } }, - "meta-llama/llama-guard-4-12b": { - "id": "meta-llama/llama-guard-4-12b", - "name": "Llama Guard 4 12B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": true, + "Qwen/Qwen3-Coder-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-04-30", - "last_updated": "2025-04-30", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -141042,27 +149455,26 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.18, - "output": 0.18 + "input": 0.07, + "output": 0.26 } }, - "meta-llama/llama-3-8b-instruct": { - "id": "meta-llama/llama-3-8b-instruct", - "name": "Llama 3 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "Qwen/Qwen3-Embedding-8B": { + "id": "Qwen/Qwen3-Embedding-8B", + "name": "Qwen 3 Embedding 8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-04-18", - "last_updated": "2024-04-18", + "temperature": false, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -141073,27 +149485,27 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 32000, + "output": 4096 }, "cost": { - "input": 0.14, - "output": 0.14 + "input": 0.01, + "output": 0 } }, - "meta-llama/llama-4-scout": { - "id": "meta-llama/llama-4-scout", - "name": "Llama 4 Scout", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + "Qwen/Qwen3.5-35B-A3B": { + "id": "Qwen/Qwen3.5-35B-A3B", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", @@ -141105,30 +149517,31 @@ }, "open_weights": true, "limit": { - "context": 10000000, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.25, + "output": 2 } }, - "meta-llama/llama-3.2-3b-instruct": { - "id": "meta-llama/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + "Qwen/Qwen3.5-122B-A10B": { + "id": "Qwen/Qwen3.5-122B-A10B", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -141136,219 +149549,143 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.05, - "output": 0.33 + "input": 0.4, + "output": 3.2 } }, - "~anthropic/claude-haiku-latest": { - "id": "~anthropic/claude-haiku-latest", - "name": "Anthropic Claude Haiku Latest", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "Qwen/Qwen3.5-9B": { + "id": "Qwen/Qwen3.5-9B", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.17, + "output": 0.25 } }, - "~anthropic/claude-fable-latest": { - "id": "~anthropic/claude-fable-latest", - "name": "Claude Fable Latest", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen3.5-397B-A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.6, + "output": 3.6 } }, - "~anthropic/claude-sonnet-latest": { - "id": "~anthropic/claude-sonnet-latest", - "name": "Anthropic Claude Sonnet Latest", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 - } - ], - "tool_call": true, - "structured_output": true, + "Qwen/Qwen3-Embedding-4B": { + "id": "Qwen/Qwen3-Embedding-4B", + "name": "Qwen 3 Embedding 4B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 32000, + "output": 2048 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.01, + "output": 0 } }, - "~anthropic/claude-opus-latest": { - "id": "~anthropic/claude-opus-latest", - "name": "Claude Opus Latest", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "Qwen/Qwen3.8-2.4T-A95B": { + "id": "Qwen/Qwen3.8-2.4T-A95B", + "name": "Qwen3.8 2.4T A95B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "moonshotai/kimi-k2": { - "id": "moonshotai/kimi-k2", - "name": "Kimi K2 0711", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-12-31", - "release_date": "2025-07-11", - "last_updated": "2025-07-11", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text" @@ -141359,32 +149696,30 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 100352 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.57, - "output": 2.3 + "input": 2.5, + "output": 6.25 } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "Qwen/Qwen3-30B-A3B": { + "id": "Qwen/Qwen3-30B-A3B", + "name": "Qwen3 30B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -141392,32 +149727,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 40960, + "output": 16384 }, "cost": { - "input": 0.72, - "output": 3.49, - "cache_read": 0.159 + "input": 0.12, + "output": 0.5 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", + "Qwen/Qwen3-Coder-480B-A35B-Instruct": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "name": "Qwen3-Coder-480B-A35B-Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -141429,31 +149758,26 @@ "open_weights": true, "limit": { "context": 262144, - "output": 100352 + "output": 66536 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 2, + "output": 2 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", @@ -141466,35 +149790,29 @@ "open_weights": true, "limit": { "context": 262144, - "output": 256000 + "output": 65536 }, "cost": { - "input": 0.375, - "output": 2.025, - "cache_read": 0.203 + "input": 0.47, + "output": 3.19 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek-V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -141502,28 +149820,39 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.66, - "output": 3.41, - "cache_read": 0.15 + "input": 0.27, + "output": 1 } }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-12-31", - "release_date": "2025-09-04", - "last_updated": "2025-09-04", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -141534,31 +149863,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 100352 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0.6, - "output": 2.5 + "input": 0.14, + "output": 0.28 } }, - "baidu/ernie-4.5-vl-424b-a47b": { - "id": "baidu/ernie-4.5-vl-424b-a47b", - "name": "ERNIE 4.5 VL 424B A47B ", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ernie", - "attachment": true, + "deepseek-ai/DeepSeek-R1": { + "id": "deepseek-ai/DeepSeek-R1", + "name": "DeepSeek-R1", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -141567,149 +149895,154 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16000 + "context": 64000, + "output": 32768 }, "cost": { - "input": 0.42, - "output": 1.25 + "input": 0.7, + "output": 2.5 } }, - "perceptron/perceptron-mk1": { - "id": "perceptron/perceptron-mk1", - "name": "Perceptron Mk1", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Pro-0813": { + "id": "deepseek-ai/DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 8192 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 1.32, + "output": 3.96 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "deepseek-ai/DeepSeek-V4.1-Flash": { + "id": "deepseek-ai/DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", - "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 384000 }, "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.083333 + "input": 0.3, + "output": 1.2 } }, - "google/gemini-3.1-flash-image": { - "id": "google/gemini-3.1-flash-image", - "name": "Nano Banana 2 (Gemini 3.1 Flash Image)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "minimal" - ] - } - ], - "tool_call": false, + "reasoning_options": [], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-18", - "last_updated": "2026-06-18", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "image", "text" ], "output": [ - "image", "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.14, + "output": 0.28 } }, - "google/gemma-3n-e4b-it": { - "id": "google/gemma-3n-e4b-it", - "name": "Gemma 3n 4B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -141720,36 +150053,31 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.06, - "output": 0.12 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "google/gemma-4-26b-a4b-it:free": { - "id": "google/gemma-4-26b-a4b-it:free", - "name": "Gemma 4 26B A4B (free)", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" @@ -141757,231 +150085,150 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 163840, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.28, + "output": 0.4 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", + "name": "DeepSeek-R1-0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2025-05", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 163840, + "output": 163840 }, "cost": { - "input": 1.25, - "output": 10, - "reasoning": 10, - "cache_read": 0.125, - "cache_write": 0.375 + "input": 3, + "output": 5 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], + "deepseek-ai/DeepSeek-V3": { + "id": "deepseek-ai/DeepSeek-V3", + "name": "DeepSeek-V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2024-12-26", + "last_updated": "2024-12-26", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65535 + "context": 64000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333 + "input": 0.4, + "output": 1.3 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "deepseek-ai/DeepSeek-V3-0324": { + "id": "deepseek-ai/DeepSeek-V3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 163840, + "output": 163840 }, "cost": { - "input": 1.5, - "output": 9, - "reasoning": 9, - "cache_read": 0.15, - "cache_write": 0.083333 + "input": 0.27, + "output": 1.12 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp": { + "id": "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ - "image", "text", - "video" + "image" ], "output": [ "text" ] }, "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.12, - "output": 0.35, - "cache_read": 0.09 - } - }, - "google/lyria-3-clip-preview": { - "id": "google/lyria-3-clip-preview", - "name": "Lyria 3 Clip Preview", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "lyria", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": false, "limit": { "context": 1048576, - "output": 65536 + "output": 384000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.44, + "output": 1.32 } }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "stepfun-ai/Step-3.7-Flash": { + "id": "stepfun-ai/Step-3.7-Flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -141995,393 +150242,288 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "output": 65536 + "context": 262144, + "output": 256000 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.2, + "output": 1.15 } }, - "google/gemini-3.1-flash-lite-image": { - "id": "google/gemini-3.1-flash-lite-image", - "name": "Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini", - "attachment": true, + "stepfun-ai/Step-3.5-Flash": { + "id": "stepfun-ai/Step-3.5-Flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", "modalities": { "input": [ - "image", "text" ], "output": [ - "image", "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 66000 + "context": 262144, + "output": 256000 }, "cost": { - "input": 0.25, - "output": 1.5 + "input": 0.1, + "output": 0.3 } }, - "google/gemini-3-pro-image-preview": { - "id": "google/gemini-3-pro-image-preview", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", + "MiniMaxAI/MiniMax-M3": { + "id": "MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 32768 + "context": 524288, + "output": 512000 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "input": 0.3, + "output": 1.2 } }, - "google/gemini-2.5-flash-image": { - "id": "google/gemini-2.5-flash-image", - "name": "Nano Banana", - "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", - "family": "gemini-flash", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "MiniMaxAI/MiniMax-M2.1": { + "id": "MiniMaxAI/MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "knowledge": "2025-10", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333 + "output": 1.2 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65535 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.4, - "reasoning": 0.4, - "cache_read": 0.01, - "cache_write": 0.083333 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "google/gemini-3.1-flash-image-preview": { - "id": "google/gemini-3.1-flash-image-preview", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, + "MiniMaxAI/MiniMax-M2.7": { + "id": "MiniMaxAI/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "image", "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "google/gemini-2.5-pro-preview-05-06": { - "id": "google/gemini-2.5-pro-preview-05-06", - "name": "Gemini 2.5 Pro Preview 05-06", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "MiniMaxAI/MiniMax-M2": { + "id": "MiniMaxAI/MiniMax-M2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01-31", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image", - "pdf", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65535 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "reasoning": 10, - "cache_read": 0.125, - "cache_write": 0.375 + "input": 0.3, + "output": 1.2 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi-K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { - "field": "reasoning_details" + "field": "reasoning_content" }, - "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "moonshotai/Kimi-K3": { + "id": "moonshotai/Kimi-K3", + "name": "Kimi K3", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "image", "text", - "video" + "image" ], "output": [ "text" @@ -142389,127 +150531,98 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.06, - "output": 0.33 + "input": 3, + "output": 15 } }, - "google/gemini-2.5-pro-preview": { - "id": "google/gemini-2.5-pro-preview", - "name": "Gemini 2.5 Pro Preview 06-05", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "moonshotai/Kimi-K2-Instruct": { + "id": "moonshotai/Kimi-K2-Instruct", + "name": "Kimi-K2-Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01-31", - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "knowledge": "2024-10", + "release_date": "2025-07-14", + "last_updated": "2025-07-14", "modalities": { "input": [ - "pdf", - "image", - "text", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "reasoning": 10, - "cache_read": 0.125, - "cache_write": 0.375 + "input": 1, + "output": 3 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65535 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.5, - "output": 3, - "reasoning": 3, - "cache_read": 0.05, - "cache_write": 0.083333 + "input": 0.95, + "output": 4 } }, - "google/gemma-3-12b-it": { - "id": "google/gemma-3-12b-it", - "name": "Gemma 3 12B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "knowledge": "2025-01", + "release_date": "2026-01-01", + "last_updated": "2026-01-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -142517,31 +150630,34 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.05, - "output": 0.15 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "google/gemma-3-4b-it": { - "id": "google/gemma-3-4b-it", - "name": "Gemma 3 4B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "moonshotai/Kimi-K2-Thinking": { + "id": "moonshotai/Kimi-K2-Thinking", + "name": "Kimi-K2-Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -142549,31 +150665,30 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.05, - "output": 0.1 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma 3 27B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "moonshotai/Kimi-K2-Instruct-0905": { + "id": "moonshotai/Kimi-K2-Instruct-0905", + "name": "Kimi-K2-Instruct-0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "knowledge": "2024-10", + "release_date": "2025-09-04", + "last_updated": "2025-09-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -142581,68 +150696,67 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 262144, "output": 16384 }, "cost": { - "input": 0.08, - "output": 0.16 + "input": 1, + "output": 3 } }, - "google/lyria-3-pro-preview": { - "id": "google/lyria-3-pro-preview", - "name": "Lyria 3 Pro Preview", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "lyria", + "zai-org/GLM-4.6V-Flash": { + "id": "zai-org/GLM-4.6V-Flash", + "name": "GLM-4.6V-Flash", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 0.9 } }, - "google/gemma-4-31b-it:free": { - "id": "google/gemma-4-31b-it:free", - "name": "Gemma 4 31B (free)", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-04-03", + "last_updated": "2026-04-03", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" @@ -142650,27 +150764,28 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 8192 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "google/gemma-2-27b-it": { - "id": "google/gemma-2-27b-it", - "name": "Gemma 2 27B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2024-07-13", - "last_updated": "2024-07-13", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -142681,113 +150796,200 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 2048 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.65, - "output": 0.65 + "input": 1.4, + "output": 4.4 } }, - "google/gemini-3-pro-image": { - "id": "google/gemini-3-pro-image", - "name": "Nano Banana Pro (Gemini 3 Pro Image)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini", - "attachment": true, + "zai-org/GLM-4.5-Air": { + "id": "zai-org/GLM-4.5-Air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-18", - "last_updated": "2026-06-18", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "image", "text" ], "output": [ - "image", "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 32768 + "context": 131072, + "output": 98304 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "input": 0.13, + "output": 0.85 } }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "zai-org/GLM-4.7": { + "id": "zai-org/GLM-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 + } + }, + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.2 + } + }, + "zai-org/GLM-5.3-Flash": { + "id": "zai-org/GLM-5.3-Flash", + "name": "GLM-5.3-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.083333 + "input": 0.15, + "output": 0.5 } }, - "liquid/lfm-2.5-1.2b-thinking:free": { - "id": "liquid/lfm-2.5-1.2b-thinking:free", - "name": "LFM2.5-1.2B-Thinking (free)", - "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", - "family": "liquid", + "zai-org/GLM-4.5V": { + "id": "zai-org/GLM-4.5V", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 65536, + "output": 16384 + }, + "cost": { + "input": 0.6, + "output": 1.8 + } + }, + "zai-org/GLM-4.6": { + "id": "zai-org/GLM-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-01-20", - "last_updated": "2026-01-20", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -142798,27 +151000,70 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.55, + "output": 2.2 + } + }, + "zai-org/GLM-4.7-Flash": { + "id": "zai-org/GLM-4.7-Flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 }, "cost": { "input": 0, "output": 0 } }, - "liquid/lfm-2.5-1.2b-instruct:free": { - "id": "liquid/lfm-2.5-1.2b-instruct:free", - "name": "LFM2.5-1.2B-Instruct (free)", - "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", - "family": "liquid", + "zai-org/GLM-5.3": { + "id": "zai-org/GLM-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-01-20", - "last_updated": "2026-01-20", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -142829,61 +151074,56 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4 } }, - "x-ai/grok-4.20": { - "id": "x-ai/grok-4.20", - "name": "Grok 4.20", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "zai-org/GLM-4.5": { + "id": "zai-org/GLM-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 131072, + "output": 98304 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.6, + "output": 2.2 } }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -142893,51 +151133,32 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.1, + "output": 0.5 } }, - "x-ai/grok-4.5": { - "id": "x-ai/grok-4.5", - "name": "Grok 4.5", - "description": "xAI's latest Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -142952,50 +151173,43 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "input": 0.25, + "output": 0.69 } - }, - "x-ai/grok-4.20-multi-agent": { - "id": "x-ai/grok-4.20-multi-agent", - "name": "Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + } + } + }, + "volcengine": { + "id": "volcengine", + "env": [ + "ARK_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://ark.cn-beijing.volces.com/api/v3", + "name": "Volcengine Ark", + "doc": "https://www.volcengine.com/docs/82379/1330310", + "models": { + "glm-5-3-flash-260828": { + "id": "glm-5-3-flash-260828", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -143003,55 +151217,73 @@ "type": "effort", "values": [ "low", - "high" + "high", + "max" ] } ], - "tool_call": false, + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.11875, + "output": 0.41563, + "cache_read": 0.03414 } }, - "x-ai/grok-build-0.1": { - "id": "x-ai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, + "doubao-seed-1-6-251015": { + "id": "doubao-seed-1-6-251015", + "name": "Seed 1.6", + "description": "ByteDance Seed model for long-context reasoning, instruction following, and tool-assisted tasks", + "family": "seed", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -143060,42 +151292,67 @@ "open_weights": false, "limit": { "context": 256000, - "output": 256000 + "output": 64000 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 + "input": 0.11875, + "output": 1.18747, + "cache_read": 0.02375, + "tiers": [ + { + "input": 0.17812, + "output": 2.37494, + "cache_read": 0.02375, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.35624, + "output": 3.56241, + "cache_read": 0.02375, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "~google/gemini-pro-latest": { - "id": "~google/gemini-pro-latest", - "name": "Google Gemini Pro Latest", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "doubao-seed-2-0-pro-260215": { + "id": "doubao-seed-2-0-pro-260215", + "name": "Seed 2.0 Pro", + "description": "Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "audio", - "pdf", - "image", "text", + "image", "video" ], "output": [ @@ -143104,48 +151361,69 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "input": 0.47499, + "output": 2.37494, + "cache_read": 0.095, + "tiers": [ + { + "input": 0.71248, + "output": 3.56241, + "cache_read": 0.1425, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 1.42496, + "output": 7.12482, + "cache_read": 0.28499, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "~google/gemini-flash-latest": { - "id": "~google/gemini-flash-latest", - "name": "Google Gemini Flash Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "doubao-seed-2-0-lite-260428": { + "id": "doubao-seed-2-0-lite-260428", + "name": "Seed 2.0 Lite", + "description": "Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", "image", - "video", - "pdf", - "audio" + "video" ], "output": [ "text" @@ -143153,188 +151431,363 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 9, - "reasoning": 9, - "cache_read": 0.15, - "cache_write": 0.083333 + "input": 0.08906, + "output": 0.53436, + "cache_read": 0.01781, + "tiers": [ + { + "input": 0.13359, + "output": 0.80154, + "cache_read": 0.02672, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.26718, + "output": 1.60308, + "cache_read": 0.05344, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "microsoft/phi-4": { - "id": "microsoft/phi-4", - "name": "Phi 4", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", - "attachment": false, - "reasoning": false, - "tool_call": false, + "doubao-seed-2-1-pro-260628": { + "id": "doubao-seed-2-1-pro-260628", + "name": "Seed 2.1 Pro", + "description": "Flagship ByteDance Seed 2.1 model for complex multimodal reasoning, coding, and agents", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2025-01-10", - "last_updated": "2025-01-10", + "release_date": "2026-06-23", + "last_updated": "2026-06-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.07, - "output": 0.14 + "input": 0.8906, + "output": 4.45301, + "cache_read": 0.17812 } }, - "microsoft/wizardlm-2-8x22b": { - "id": "microsoft/wizardlm-2-8x22b", - "name": "WizardLM-2 8x22B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "doubao-seed-1-6-vision-250815": { + "id": "doubao-seed-1-6-vision-250815", + "name": "Seed 1.6 Vision", + "description": "ByteDance Seed multimodal model for image understanding, visual reasoning, and tool-assisted tasks", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2024-04-16", - "last_updated": "2024-04-16", + "release_date": "2025-08-15", + "last_updated": "2025-08-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 8000 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.62, - "output": 0.62 + "input": 0.11875, + "output": 1.18747, + "cache_read": 0.02375, + "tiers": [ + { + "input": 0.17812, + "output": 2.37494, + "cache_read": 0.02375, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.35624, + "output": 3.56241, + "cache_read": 0.02375, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "poolside/laguna-xs-2.1:free": { - "id": "poolside/laguna-xs-2.1:free", - "name": "Laguna XS 2.1 (free)", - "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", - "family": "laguna", - "attachment": false, + "doubao-seed-2-0-mini-260428": { + "id": "doubao-seed-2-0-mini-260428", + "name": "Seed 2.0 Mini", + "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-07-02", - "last_updated": "2026-07-02", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 256000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.02969, + "output": 0.29687, + "cache_read": 0.00594, + "tiers": [ + { + "input": 0.05937, + "output": 0.59374, + "cache_read": 0.01187, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.11875, + "output": 1.18747, + "cache_read": 0.02375, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "poolside/laguna-m.1": { - "id": "poolside/laguna-m.1", - "name": "Laguna M.1", - "description": "Poolside's flagship agentic coding model for long-horizon work", - "family": "laguna", - "attachment": false, + "doubao-seed-2-0-code-preview-260215": { + "id": "doubao-seed-2-0-code-preview-260215", + "name": "Seed 2.0 Code", + "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-06-13", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 32768 + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.4, - "cache_read": 0.1 + "input": 0.47499, + "output": 2.37494, + "cache_read": 0.095, + "tiers": [ + { + "input": 0.71248, + "output": 3.56241, + "cache_read": 0.1425, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 1.42496, + "output": 7.12482, + "cache_read": 0.28499, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "poolside/laguna-xs-2.1": { - "id": "poolside/laguna-xs-2.1", - "name": "Laguna XS 2.1", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", - "attachment": false, + "doubao-seed-2-1-turbo-260628": { + "id": "doubao-seed-2-1-turbo-260628", + "name": "Seed 2.1 Turbo", + "description": "Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-07-02", - "last_updated": "2026-07-02", + "release_date": "2026-06-23", + "last_updated": "2026-06-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.06, - "output": 0.12, - "cache_read": 0.03 + "input": 0.4453, + "output": 2.22651, + "cache_read": 0.08906 } }, - "poolside/laguna-m.1:free": { - "id": "poolside/laguna-m.1:free", - "name": "Laguna M.1 (free)", - "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", - "family": "laguna", + "glm-5-2-260617": { + "id": "glm-5-2-260617", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-04-28", + "release_date": "2026-06-13", "last_updated": "2026-06-13", "modalities": { "input": [ @@ -143346,29 +151799,48 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.18747, + "output": 4.15615, + "cache_read": 0.29687 } }, - "writer/palmyra-x5": { - "id": "writer/palmyra-x5", - "name": "Palmyra X5", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "palmyra", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "doubao-seed-1-8-251228": { + "id": "doubao-seed-1-8-251228", + "name": "Seed 1.8", + "description": "ByteDance Seed model for multimodal reasoning, long-context analysis, and agent workflows", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-01-21", - "last_updated": "2026-01-21", + "release_date": "2025-12-28", + "last_updated": "2025-12-28", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -143376,55 +151848,90 @@ }, "open_weights": false, "limit": { - "context": 1040000, - "output": 8192 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 6 + "input": 0.11875, + "output": 1.18747, + "cache_read": 0.02375, + "tiers": [ + { + "input": 0.17812, + "output": 2.37494, + "cache_read": 0.02375, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.35624, + "output": 3.56241, + "cache_read": 0.02375, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, + "doubao-seed-evolving": { + "id": "doubao-seed-evolving", + "name": "Seed Evolving", + "description": "Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "interleaved": { - "field": "reasoning_details" + "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-06-23", + "last_updated": "2026-06-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.4, - "output": 1.75, - "cache_read": 0.08 + "input": 0.8906, + "output": 4.45301, + "cache_read": 0.17812 } }, - "z-ai/glm-4.5v": { - "id": "z-ai/glm-4.5v", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "doubao-seed-1-6-flash-250828": { + "id": "doubao-seed-1-6-flash-250828", + "name": "Seed 1.6 Flash", + "description": "Low-latency ByteDance Seed model for high-throughput chat, extraction, and lightweight tool use", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -143433,11 +151940,13 @@ } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ "text", @@ -143447,70 +151956,126 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11 + "input": 0.02227, + "output": 0.22265, + "cache_read": 0.00445, + "tiers": [ + { + "input": 0.04453, + "output": 0.4453, + "cache_read": 0.00445, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.08906, + "output": 0.8906, + "cache_read": 0.00445, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "z-ai/glm-4.5": { - "id": "z-ai/glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, + "doubao-seed-character-260628": { + "id": "doubao-seed-character-260628", + "name": "Seed Character", + "description": "ByteDance Seed model optimized for character-driven dialogue and consistent conversational behavior", + "family": "seed", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "high" + ] } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-06-23", + "last_updated": "2026-06-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 - } - }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "input": 0.11875, + "output": 0.29687, + "cache_read": 0.02375, + "tiers": [ + { + "input": 0.17812, + "output": 0.8906, + "cache_read": 0.02375, + "tier": { + "type": "context", + "size": 32000 + } + } + ] + } + }, + "deepseek-v4-pro-ga-260813": { + "id": "deepseek-v4-pro-ga-260813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -143521,29 +152086,45 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.966, - "output": 3.036, - "cache_read": 0.1794 + "input": 1.3359, + "output": 4.00771, + "cache_read": 0.04453 } }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", + "deepseek-v4-flash-ga-260731": { + "id": "deepseek-v4-flash-ga-260731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -143554,39 +152135,49 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.43, - "output": 1.74, - "cache_read": 0.08 + "input": 0.4453, + "output": 1.3359, + "cache_read": 0.01484 } - }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + } + } + }, + "impossibl": { + "id": "impossibl", + "env": [ + "IMPOSSIBL_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.impossibl.com/v1", + "name": "Impossibl", + "doc": "https://impossibl.com/docs/models", + "models": { + "cerebras/gpt-oss-120b": { + "id": "cerebras/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", - "xhigh" + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -143597,34 +152188,39 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 101376 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.54, - "output": 1.76, - "cache_read": 0.1 + "input": 0.35, + "output": 0.75 } }, - "z-ai/glm-4.6v": { - "id": "z-ai/glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, + "groq/gpt-oss-20b": { + "id": "groq/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -143636,50 +152232,57 @@ "output": 32768 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.055 + "input": 0.075, + "output": 0.3, + "cache_read": 0.0375 } }, - "z-ai/glm-5v-turbo": { - "id": "z-ai/glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, + "groq/gpt-oss-120b": { + "id": "groq/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "z-ai/glm-4.5-air": { - "id": "z-ai/glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "zai/glm-4.5": { + "id": "zai/glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -143688,7 +152291,6 @@ } ], "tool_call": true, - "structured_output": false, "temperature": true, "knowledge": "2025-04", "release_date": "2025-07-28", @@ -143707,28 +152309,28 @@ "output": 98304 }, "cost": { - "input": 0.13, - "output": 0.85, - "cache_read": 0.025 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "z-ai/glm-4.7-flash": { - "id": "z-ai/glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", + "zai/glm-4.6": { + "id": "zai/glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -143739,28 +152341,28 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.06, - "output": 0.4, - "cache_read": 0.01 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", + "zai/glm-5": { + "id": "zai/glm-5", "name": "GLM-5", "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, "release_date": "2026-02-12", "last_updated": "2026-02-12", @@ -143774,31 +152376,32 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 1.92, - "cache_read": 0.12 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", + "zai/glm-4.7": { + "id": "zai/glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -143807,73 +152410,66 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, + "context": 204800, "output": 131072 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "openai/gpt-4o-mini-2024-07-18": { - "id": "openai/gpt-4o-mini-2024-07-18", - "name": "GPT-4o-mini (2024-07-18)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "o-mini", - "attachment": true, - "reasoning": false, + "zai/glm-5.2": { + "id": "zai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-10-31", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai/gpt-oss-safeguard-20b": { - "id": "openai/gpt-oss-safeguard-20b", - "name": "gpt-oss-safeguard-20b", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", + "zai/glm-5.1": { + "id": "zai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -143884,28 +152480,32 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.075, - "output": 0.3, - "cache_read": 0.0375 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai/gpt-3.5-turbo-instruct": { - "id": "openai/gpt-3.5-turbo-instruct", - "name": "GPT-3.5 Turbo Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "zai/glm-5-turbo": { + "id": "zai/glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2021-09-30", - "release_date": "2023-09-28", - "last_updated": "2023-09-28", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text" @@ -143916,74 +152516,72 @@ }, "open_weights": false, "limit": { - "context": 4095, - "output": 4096 + "context": 200000, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 2 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "openai/gpt-5.2-chat": { - "id": "openai/gpt-5.2-chat", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": false, + "zai/glm-4.5-air": { + "id": "zai/glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-10", - "last_updated": "2025-12-10", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "pdf", - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 98304 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.2, + "output": 1.1, + "cache_read": 0.03 } }, - "openai/gpt-5.6-luna-pro": { - "id": "openai/gpt-5.6-luna-pro", - "name": "GPT-5.6 Luna Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-nano", + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", @@ -143996,25 +152594,31 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { "input": 1, - "output": 6, + "output": 5, "cache_read": 0.1, "cache_write": 1.25 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "anthropic/claude-opus-4-5": { + "id": "anthropic/claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + }, { "type": "effort", "values": [ @@ -144025,11 +152629,10 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", @@ -144043,40 +152646,36 @@ "open_weights": false, "limit": { "context": 200000, - "output": 100000 + "output": 64000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/gpt-5.6-sol-pro": { - "id": "openai/gpt-5.6-sol-pro", - "name": "GPT-5.6 Sol Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "anthropic/claude-sonnet-4-5": { + "id": "anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", @@ -144089,42 +152688,47 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openai/o4-mini-high": { - "id": "openai/o4-mini-high", - "name": "o4 Mini High", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "high" + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-06-30", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "image", "text", + "image", "pdf" ], "output": [ @@ -144133,74 +152737,96 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "openai/gpt-audio": { - "id": "openai/gpt-audio", - "name": "GPT Audio", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", + "anthropic/claude-opus-4-8": { + "id": "anthropic/claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "audio" + "image", + "pdf" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2 Pro", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "image", "text", + "image", "pdf" ], "output": [ @@ -144209,31 +152835,51 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 21, - "output": 168 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "openai/gpt-4o-mini-search-preview": { - "id": "openai/gpt-4o-mini-search-preview", - "name": "GPT-4o-mini Search Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "o-mini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": false, - "knowledge": "2023-10-31", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -144241,40 +152887,46 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/gpt-5.6-terra-pro": { - "id": "openai/gpt-5.6-terra-pro", - "name": "GPT-5.6 Terra Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-mini", + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + }, { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -144287,41 +152939,43 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 2.5, + "input": 3, "output": 15, - "cache_read": 0.25, - "cache_write": 3.125 + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -144334,63 +152988,63 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/gpt-5-chat": { - "id": "openai/gpt-5-chat", - "name": "GPT-5 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": false, - "tool_call": false, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "pdf", - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.74, + "output": 3.48, + "cache_read": 0.145 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -144399,75 +153053,91 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16385, - "output": 4096 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 0.19, + "output": 0.51, + "cache_read": 0.028 } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", + "thinkingmachines/inkling": { + "id": "thinkingmachines/inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high" + "none", + "minimal", + "low", + "medium", + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 15, - "output": 120 + "input": 1.87, + "output": 4.68, + "cache_read": 0.374 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "google/gemini-3.6-flash": { + "id": "google/gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -144476,30 +153146,46 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 10 + "input": 1.5, + "output": 7.5, + "cache_read": 0.15 } }, - "openai/gpt-4": { - "id": "openai/gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, + "google/gemini-3.5-flash-lite": { + "id": "google/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -144507,19 +153193,20 @@ }, "open_weights": false, "limit": { - "context": 8191, - "output": 4096 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 30, - "output": 60 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -144534,14 +153221,16 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "image", "text", + "image", + "video", + "audio", "pdf" ], "output": [ @@ -144550,31 +153239,62 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 - } - }, - "openai/gpt-3.5-turbo-16k": { - "id": "openai/gpt-3.5-turbo-16k", - "name": "GPT-3.5 Turbo 16k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2021-09-30", - "release_date": "2023-08-28", - "last_updated": "2023-08-28", + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -144582,42 +153302,42 @@ }, "open_weights": false, "limit": { - "context": 16385, - "output": 4096 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 4 + "input": 1.5, + "output": 9, + "cache_read": 0.15 } }, - "openai/o3-pro": { - "id": "openai/o3-pro", - "name": "o3-pro", - "description": "High-effort o3 tier for difficult technical reasoning and careful answers", - "family": "o-pro", + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "pdf", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -144625,32 +153345,61 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 20, - "output": 80 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "openai/gpt-5.1-chat": { - "id": "openai/gpt-5.1-chat", - "name": "GPT-5.1 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -144658,32 +153407,44 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "openai/gpt-4o-2024-05-13": { - "id": "openai/gpt-4o-2024-05-13", - "name": "GPT-4o (2024-05-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 512, + "max": 24576 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -144692,44 +153453,80 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5, - "output": 15 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "pdf", + "text", "image", + "video", + "audio", + "pdf" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025 + } + }, + "xai/grok-4.3": { + "id": "xai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" ], "output": [ "text" @@ -144737,28 +153534,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "openai/gpt-5.3-chat": { - "id": "openai/gpt-5.3-chat", - "name": "GPT-5.3 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "xai/grok-4.20-0309-reasoning": { + "id": "xai/grok-4.20-0309-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", @@ -144771,31 +153568,65 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "openai/gpt-3.5-turbo-0613": { - "id": "openai/gpt-3.5-turbo-0613", - "name": "GPT-3.5 Turbo (older v0613)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, + "xai/grok-4.5": { + "id": "xai/grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2021-09-30", - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.3 + } + }, + "xai/grok-4.20-0309-non-reasoning": { + "id": "xai/grok-4.20-0309-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" ], "output": [ "text" @@ -144803,55 +153634,55 @@ }, "open_weights": false, "limit": { - "context": 4095, - "output": 4096 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 1, - "output": 2 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "openai/gpt-5-image-mini": { - "id": "openai/gpt-5-image-mini", - "name": "GPT-5 Image Mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "xai/grok-build-0.1": { + "id": "xai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-16", - "last_updated": "2025-10-16", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "pdf" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 2.5, + "input": 1, "output": 2, - "cache_read": 0.25 + "cache_read": 0.2 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", - "attachment": true, + "fireworks/gpt-oss-20b": { + "id": "fireworks/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -144865,79 +153696,63 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.07, + "output": 0.3, + "cache_read": 0.035 } }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "fireworks/glm-5.2": { + "id": "fireworks/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.4, + "output": 4.4, + "cache_read": 0.14 } }, - "openai/gpt-oss-120b:free": { - "id": "openai/gpt-oss-120b:free", - "name": "gpt-oss-120b (free)", + "fireworks/gpt-oss-120b": { + "id": "fireworks/gpt-oss-120b", + "name": "GPT OSS 120B", "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", "family": "gpt-oss", "attachment": false, @@ -144953,7 +153768,7 @@ } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, "release_date": "2025-08-05", "last_updated": "2025-08-05", @@ -144968,74 +153783,102 @@ "open_weights": true, "limit": { "context": 131072, - "output": 131072 + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "openai/gpt-4o-2024-08-06": { - "id": "openai/gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-08-06", - "last_updated": "2024-08-06", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.003 + } + }, + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -145043,45 +153886,41 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "qwen/qwen3.8-max-preview": { + "id": "qwen/qwen3.8-max-preview", + "name": "Qwen3.8 Max Preview", + "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-07-19", + "last_updated": "2026-07-19", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "video" ], "output": [ "text" @@ -145089,45 +153928,36 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 2.5, + "output": 7.5 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -145135,53 +153965,90 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.25, + "output": 1.5, + "cache_read": 0.05, + "tiers": [ + { + "input": 1, + "output": 4, + "cache_read": 0.2, + "tier": { + "type": "context", + "size": 262144 + } + } + ], + "context_over_200k": { + "input": 1, + "output": 4, + "cache_read": 0.2 + } } }, - "openai/gpt-audio-mini": { - "id": "openai/gpt-audio-mini", - "name": "GPT Audio Mini", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "o-mini", + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", - "audio" + "image", + "video" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 2.4 + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "tiers": [ + { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24, + "tier": { + "type": "context", + "size": 262144 + } + } + ], + "context_over_200k": { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24 + } } }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-nano", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -145192,17 +154059,16 @@ "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -145220,35 +154086,50 @@ "output": 128000 }, "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -145260,42 +154141,86 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 100000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 30, + "output": 180, + "cache_read": 3, + "tiers": [ + { + "input": 60, + "output": 270, + "cache_read": 3, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270, + "cache_read": 3 + } } }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-mini", + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16385, + "output": 4096 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -145313,31 +154238,56 @@ "output": 128000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125 + "input": 30, + "output": 180, + "cache_read": 3, + "tiers": [ + { + "input": 60, + "output": 270, + "cache_read": 3, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270, + "cache_read": 3 + } } }, - "openai/o4-mini-deep-research": { - "id": "openai/o4-mini-deep-research", - "name": "o4-mini-deep-research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "family": "o-mini", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-05", - "release_date": "2024-06-26", - "last_updated": "2024-06-26", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "pdf", - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -145345,32 +154295,44 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "image", "text", + "image", "pdf" ], "output": [ @@ -145379,21 +154341,22 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -145407,31 +154370,35 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.036, - "output": 0.18 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openai/gpt-4o-2024-11-20": { - "id": "openai/gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", "family": "gpt", "attachment": true, "reasoning": false, @@ -145439,8 +154406,8 @@ "structured_output": true, "temperature": true, "knowledge": "2023-09", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", @@ -145462,12 +154429,12 @@ "cache_read": 1.25 } }, - "openai/o1": { - "id": "openai/o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -145482,49 +154449,13 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 - } - }, - "openai/o1-pro": { - "id": "openai/o1-pro", - "name": "o1-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2025-03-19", - "last_updated": "2025-03-19", + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -145532,31 +154463,44 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 150, - "output": 600 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openai/gpt-chat-latest": { - "id": "openai/gpt-chat-latest", - "name": "GPT Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": false, - "release_date": "2026-05-05", - "last_updated": "2026-05-05", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -145565,80 +154509,84 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "openai/gpt-5-image": { - "id": "openai/gpt-5-image", - "name": "GPT-5 Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-10-01", - "release_date": "2025-10-14", - "last_updated": "2025-10-14", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 10, - "output": 10, - "cache_read": 1.25 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -145646,21 +154594,20 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -145677,15 +154624,15 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -145698,16 +154645,16 @@ "output": 128000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, "reasoning": false, "tool_call": true, @@ -145719,8 +154666,7 @@ "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -145732,30 +154678,39 @@ "output": 32768 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "openai/o3-deep-research": { - "id": "openai/o3-deep-research", - "name": "o3-deep-research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "family": "o", + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-05", - "release_date": "2024-06-26", - "last_updated": "2024-06-26", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text" @@ -145763,31 +154718,44 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 10, - "output": 40, - "cache_read": 2.5 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "openai/gpt-4-turbo-preview": { - "id": "openai/gpt-4-turbo-preview", - "name": "GPT-4 Turbo Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, + "openai/o1": { + "id": "openai/o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -145795,28 +154763,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 200000, + "output": 100000 }, "cost": { - "input": 10, - "output": 30 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", "high" ] } @@ -145824,14 +154790,13 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -145841,32 +154806,41 @@ "limit": { "context": 400000, "input": 272000, - "output": 128000 + "output": 272000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 15, + "output": 120 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -145874,32 +154848,34 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -145908,42 +154884,43 @@ "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 16384 }, "cost": { - "input": 10, - "output": 30 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -145956,22 +154933,24 @@ "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", "medium", "high", "xhigh" @@ -145981,9 +154960,9 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -146001,35 +154980,59 @@ "output": 128000 }, "cost": { - "input": 30, - "output": 180 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } } }, - "openai/o3-mini-high": { - "id": "openai/o3-mini-high", - "name": "o3 Mini High", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high" + "none", + "low", + "medium", + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2023-10-31", - "release_date": "2025-02-12", - "last_updated": "2025-02-12", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" @@ -146037,19 +155040,20 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai/gpt-5.4-image-2": { - "id": "openai/gpt-5.4-image-2", - "name": "GPT-5.4 Image 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", "family": "gpt", "attachment": true, "reasoning": true, @@ -146065,49 +155069,102 @@ ] } ], - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "image", "text", + "image", "pdf" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { - "context": 272000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 8, - "output": 15, - "cache_read": 2 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "openai/gpt-4o-search-preview": { - "id": "openai/gpt-4o-search-preview", - "name": "GPT-4o Search Preview", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", "family": "gpt", - "attachment": false, + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2023-10-31", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ + "text", + "image", + "pdf" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" ], "output": [ "text" @@ -146115,25 +155172,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", "medium", "high", "xhigh" @@ -146143,9 +155203,9 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -146163,28 +155223,47 @@ "output": 128000 }, "cost": { - "input": 30, - "output": 180 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -146193,20 +155272,19 @@ "open_weights": false, "limit": { "context": 128000, - "output": 16384 + "output": 4096 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 10, + "output": 30 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -146220,53 +155298,54 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.029, - "output": 0.14 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", @@ -146279,28 +155358,27 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -146311,8 +155389,8 @@ "structured_output": true, "temperature": false, "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -146334,17 +155412,18 @@ "cache_read": 0.125 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -146355,13 +155434,14 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -146369,22 +155449,53 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5 + } } - }, - "openai/gpt-oss-20b:free": { - "id": "openai/gpt-oss-20b:free", - "name": "gpt-oss-20b (free)", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + } + } + }, + "xpersona": { + "id": "xpersona", + "env": [ + "XPERSONA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://www.xpersona.co/v1", + "name": "Xpersona", + "doc": "https://www.xpersona.co/docs", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -146392,62 +155503,73 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.75, + "output": 6, + "reasoning": 6, + "cache_read": 0.075 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text" @@ -146455,46 +155577,44 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 200000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.6, + "output": 3.7, + "reasoning": 3.7, + "cache_read": 0.06 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "xpersona-frieren-coder": { + "id": "xpersona-frieren-coder", + "name": "Xpersona Frieren 1", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "knowledge": "2025-12-30", + "release_date": "2026-05-01", + "last_updated": "2026-05-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -146502,251 +155622,291 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 1.5, + "output": 6, + "reasoning": 6, + "cache_read": 0.15 } }, - "thedrummer/cydonia-24b-v4.1": { - "id": "thedrummer/cydonia-24b-v4.1", - "name": "Cydonia 24B V4.1", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gpt-5.6": { + "id": "gpt-5.6", + "name": "GPT-5.6", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2025-09-27", - "last_updated": "2025-09-27", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 372000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 0.5, + "input": 1.5, + "output": 12, + "reasoning": 12, "cache_read": 0.15 } }, - "thedrummer/skyfall-36b-v2": { - "id": "thedrummer/skyfall-36b-v2", - "name": "Skyfall 36B V2", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "xpersona-gpt-5.5": { + "id": "xpersona-gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2025-03-10", - "last_updated": "2025-03-10", + "temperature": false, + "knowledge": "2025-12-30", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.55, - "output": 0.8, - "cache_read": 0.25 + "input": 3, + "output": 18, + "reasoning": 18, + "cache_read": 0.3 } }, - "thedrummer/unslopnemo-12b": { - "id": "thedrummer/unslopnemo-12b", - "name": "UnslopNemo 12B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2024-11-08", - "last_updated": "2024-11-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.4, - "output": 0.4 - } - }, - "thedrummer/rocinante-12b": { - "id": "thedrummer/rocinante-12b", - "name": "Rocinante 12B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2024-09-30", - "last_updated": "2024-09-30", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 0.5 + "input": 1.55, + "output": 12.2, + "reasoning": 12.2, + "cache_read": 0.155 } }, - "bytedance/ui-tars-1.5-7b": { - "id": "bytedance/ui-tars-1.5-7b", - "name": "UI-TARS 7B ", - "description": "Multimodal model for analyzing text, images, documents, and rich media", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01-31", - "release_date": "2025-07-22", - "last_updated": "2025-07-22", - "modalities": { - "input": [ - "image", - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 2048 - }, - "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.1 - } - }, - "rekaai/reka-flash-3": { - "id": "rekaai/reka-flash-3", - "name": "Reka Flash 3", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "reka", - "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01-31", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.2 + "input": 3, + "output": 18.5, + "reasoning": 18.5, + "cache_read": 0.3 } }, - "rekaai/reka-edge": { - "id": "rekaai/reka-edge", - "name": "Reka Edge", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "reka", + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "image", "text", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 272000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 0.375, + "output": 4, + "reasoning": 4, + "cache_read": 0.0375 } }, - "mistralai/mistral-large-2407": { - "id": "mistralai/mistral-large-2407", - "name": "Mistral Large 2407", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-03-31", - "release_date": "2024-11-19", - "last_updated": "2024-11-19", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" @@ -146754,104 +155914,96 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "output": 128000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 1.5, + "output": 9.25, + "reasoning": 9.25, + "cache_read": 0.15 } }, - "mistralai/mistral-small-3.2-24b-instruct": { - "id": "mistralai/mistral-small-3.2-24b-instruct", - "name": "Mistral Small 3.2 24B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-10-31", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", - "modalities": { - "input": [ - "image", - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.075, - "output": 0.2 - } - }, - "mistralai/mistral-nemo": { - "id": "mistralai/mistral-nemo", - "name": "Mistral Nemo", - "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0.02, - "output": 0.03 + "input": 1.5, + "output": 12, + "reasoning": 12, + "cache_read": 0.15 } }, - "mistralai/mistral-medium-3-5": { - "id": "mistralai/mistral-medium-3-5", - "name": "Mistral Medium 3.5", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "high" + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -146859,26 +156011,44 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 128000 }, "cost": { - "input": 1.5, - "output": 7.5 + "input": 0.9, + "output": 5.55, + "reasoning": 5.55, + "cache_read": 0.09 } }, - "mistralai/ministral-8b-2512": { - "id": "mistralai/ministral-8b-2512", - "name": "Ministral 3 8B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -146888,30 +156058,46 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 372000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.015 + "input": 1.5, + "output": 2, + "reasoning": 2, + "cache_read": 0.15 } }, - "mistralai/mistral-small-3.1-24b-instruct": { - "id": "mistralai/mistral-small-3.1-24b-instruct", - "name": "Mistral Small 3.1 24B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2023-10-31", - "release_date": "2025-03-17", - "last_updated": "2025-03-17", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -146921,33 +156107,49 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 372000, "output": 128000 }, "cost": { - "input": 0.351, - "output": 0.555 + "input": 1.5, + "output": 12, + "reasoning": 12, + "cache_read": 0.15 } - }, - "mistralai/mistral-saba": { - "id": "mistralai/mistral-saba", - "name": "Saba", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": true, - "reasoning": false, + } + } + }, + "qiniu-ai": { + "id": "qiniu-ai", + "env": [ + "QINIU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.qnaigc.com/v1", + "name": "Qiniu", + "doc": "https://developer.qiniu.com/aitokenapi", + "models": { + "glm-4.5": { + "id": "glm-4.5", + "name": "GLM 4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-02-17", - "last_updated": "2025-02-17", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -146955,31 +156157,28 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.2, - "output": 0.6, - "cache_read": 0.02 + "context": 131072, + "output": 98304 } }, - "mistralai/mistral-large": { - "id": "mistralai/mistral-large", - "name": "Mistral Large", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", + "gemini-3.0-flash-preview": { + "id": "gemini-3.0-flash-preview", + "name": "Gemini 3.0 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-11-30", - "release_date": "2024-02-26", - "last_updated": "2024-02-26", + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ "text", + "image", + "audio", + "video", "pdf" ], "output": [ @@ -146988,33 +156187,31 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "context": 1000000, + "output": 64000 } }, - "mistralai/mistral-medium-3.1": { - "id": "mistralai/mistral-medium-3.1", - "name": "Mistral Medium 3.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "doubao-seed-2.0-mini": { + "id": "doubao-seed-2.0-mini", + "name": "Doubao Seed 2.0 Mini", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-08-13", - "last_updated": "2025-08-13", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -147022,28 +156219,23 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 262144 - }, - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.04 + "context": 256000, + "output": 32000 } }, - "mistralai/mistral-small-24b-instruct-2501": { - "id": "mistralai/mistral-small-24b-instruct-2501", - "name": "Mistral Small 3", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "mimo-v2-flash": { + "id": "mimo-v2-flash", + "name": "Mimo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2023-10-31", - "release_date": "2025-01-30", - "last_updated": "2025-01-30", + "knowledge": "2024-12-01", + "release_date": "2025-12-16", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -147054,205 +156246,168 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 16384 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.05, - "output": 0.08 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01 } }, - "mistralai/ministral-3b-2512": { - "id": "mistralai/ministral-3b-2512", - "name": "Ministral 3 3B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "doubao-seed-2.0-lite": { + "id": "doubao-seed-2.0-lite", + "name": "Doubao Seed 2.0 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.01 + "context": 256000, + "output": 32000 } }, - "mistralai/mistral-small-2603": { - "id": "mistralai/mistral-small-2603", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", + "doubao-seed-1.6-thinking": { + "id": "doubao-seed-1.6-thinking", + "name": "Doubao-Seed 1.6 Thinking", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "release_date": "2025-08-15", + "last_updated": "2025-08-15", "modalities": { "input": [ + "image", "text", - "image" + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "context": 256000, + "output": 32000 } }, - "mistralai/ministral-14b-2512": { - "id": "mistralai/ministral-14b-2512", - "name": "Ministral 3 14B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "gemini-2.5-flash-image": { + "id": "gemini-2.5-flash-image", + "name": "Gemini 2.5 Flash Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "release_date": "2025-10-22", + "last_updated": "2025-10-22", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.02 + "context": 32768, + "output": 8192 } }, - "mistralai/devstral-2512": { - "id": "mistralai/devstral-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", + "gemini-3.0-pro-image-preview": { + "id": "gemini-3.0-pro-image-preview", + "name": "Gemini 3.0 Pro Image Preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.04 + "context": 32768, + "output": 8192 } }, - "mistralai/mixtral-8x22b-instruct": { - "id": "mistralai/mixtral-8x22b-instruct", - "name": "Mixtral 8x22B Instruct", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", + "claude-3.5-haiku": { + "id": "claude-3.5-haiku", + "name": "Claude 3.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-01-31", - "release_date": "2024-04-17", - "last_updated": "2024-04-17", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "context": 200000, + "output": 8192 } }, - "mistralai/mistral-medium-3": { - "id": "mistralai/mistral-medium-3", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235b A22B Instruct 2507", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "release_date": "2025-08-12", + "last_updated": "2025-08-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -147260,99 +156415,79 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.04 + "context": 262144, + "output": 64000 } }, - "mistralai/voxtral-small-24b-2507": { - "id": "mistralai/voxtral-small-24b-2507", - "name": "Voxtral Small 24B 2507", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral", - "attachment": true, - "reasoning": false, + "qwen3-next-80b-a3b-thinking": { + "id": "qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B Thinking", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-10-30", - "last_updated": "2025-10-30", + "release_date": "2025-09-12", + "last_updated": "2025-09-12", "modalities": { "input": [ - "text", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01 + "context": 131072, + "output": 32768 } }, - "mistralai/mistral-large-2512": { - "id": "mistralai/mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", - "attachment": true, + "qwen-max-2025-01-25": { + "id": "qwen-max-2025-01-25", + "name": "Qwen2.5-Max-2025-01-25", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.05 + "context": 128000, + "output": 4096 } }, - "mistralai/codestral-2508": { - "id": "mistralai/codestral-2508", - "name": "Codestral 2508", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "codestral", - "attachment": true, + "deepseek-v3-0324": { + "id": "deepseek-v3-0324", + "name": "DeepSeek-V3-0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-01", - "last_updated": "2025-08-01", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -147360,27 +156495,26 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.03 + "context": 128000, + "output": 16000 } }, - "morph/morph-v3-fast": { - "id": "morph/morph-v3-fast", - "name": "Morph V3 Fast", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "morph", + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-07-07", - "last_updated": "2025-07-07", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -147391,26 +156525,21 @@ }, "open_weights": false, "limit": { - "context": 81920, - "output": 38000 - }, - "cost": { - "input": 0.8, - "output": 1.2 + "context": 40000, + "output": 4096 } }, - "morph/morph-v3-large": { - "id": "morph/morph-v3-large", - "name": "Morph V3 Large", + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3 Max", "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "morph", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-07-07", - "last_updated": "2025-07-07", + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ "text" @@ -147422,31 +156551,25 @@ "open_weights": false, "limit": { "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.9, - "output": 1.9 + "output": 65536 } }, - "bytedance-seed/seed-1.6-flash": { - "id": "bytedance-seed/seed-1.6-flash", - "name": "Seed 1.6 Flash", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "claude-4.5-sonnet": { + "id": "claude-4.5-sonnet", + "name": "Claude 4.5 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "image", "text", - "video" + "image" ], "output": [ "text" @@ -147454,32 +156577,24 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.075, - "output": 0.3 + "context": 200000, + "output": 64000 } }, - "bytedance-seed/seed-1.6": { - "id": "bytedance-seed/seed-1.6", - "name": "Seed 1.6", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "deepseek-v3": { + "id": "deepseek-v3", + "name": "DeepSeek-V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2025-08-13", + "last_updated": "2025-08-13", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" @@ -147487,41 +156602,81 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.25, - "output": 2 - } + "context": 128000, + "output": 16000 + } }, - "bytedance-seed/seed-2.0-mini": { - "id": "bytedance-seed/seed-2.0-mini", - "name": "Seed-2.0-Mini", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, + "qwen3-coder-480b-a35b-instruct": { + "id": "qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-14", + "last_updated": "2025-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262000, + "output": 4096 + } + }, + "qwen-turbo": { + "id": "qwen-turbo", + "name": "Qwen-Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 4096 + } + }, + "gemini-2.0-flash": { + "id": "gemini-2.0-flash", + "name": "Gemini 2.0 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ @@ -147530,41 +156685,27 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.1, - "output": 0.4 + "context": 1048576, + "output": 8192 } }, - "bytedance-seed/seed-2.0-lite": { - "id": "bytedance-seed/seed-2.0-lite", - "name": "Seed-2.0-Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "gemini-2.0-flash-lite": { + "id": "gemini-2.0-flash-lite", + "name": "Gemini 2.0 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-10", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ @@ -147573,36 +156714,22 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.25, - "output": 2 + "context": 1048576, + "output": 8192 } }, - "sakana/fugu-ultra": { - "id": "sakana/fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", + "claude-4.0-sonnet": { + "id": "claude-4.0-sonnet", + "name": "Claude 4.0 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "max", - "xhigh", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-24", - "last_updated": "2026-06-24", + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -147614,58 +156741,54 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "context": 200000, + "output": 64000 } }, - "anthracite-org/magnum-v4-72b": { - "id": "anthracite-org/magnum-v4-72b", - "name": "Magnum v4 72B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, + "qwen2.5-vl-7b-instruct": { + "id": "qwen2.5-vl-7b-instruct", + "name": "Qwen 2.5 VL 7B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 2048 - }, - "cost": { - "input": 3, - "output": 5 + "context": 128000, + "output": 8192 } }, - "nvidia/nemotron-3-nano-30b-a3b:free": { - "id": "nvidia/nemotron-3-nano-30b-a3b:free", - "name": "Nemotron 3 Nano 30B A3B (free)", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", + "qwen3-235b-a22b-thinking-2507": { + "id": "qwen3-235b-a22b-thinking-2507", + "name": "Qwen3 235B A22B Thinking 2507", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "release_date": "2025-08-12", + "last_updated": "2025-08-12", "modalities": { "input": [ "text" @@ -147674,29 +156797,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 262144, + "output": 4096 } }, - "nvidia/nemotron-nano-9b-v2:free": { - "id": "nvidia/nemotron-nano-9b-v2:free", - "name": "Nemotron Nano 9B V2 (free)", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", + "qwen3-235b-a22b": { + "id": "qwen3-235b-a22b", + "name": "Qwen 3 235B A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-08-18", - "last_updated": "2025-08-18", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -147705,29 +156822,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 + "output": 32000 } }, - "nvidia/nemotron-nano-12b-v2-vl:free": { - "id": "nvidia/nemotron-nano-12b-v2-vl:free", - "name": "Nemotron Nano 12B 2 VL (free)", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", + "doubao-seed-2.0-pro": { + "id": "doubao-seed-2.0-pro", + "name": "Doubao Seed 2.0 Pro", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", @@ -147738,78 +156854,55 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 256000, "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 } }, - "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free": { - "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free", - "name": "Nemotron 3 Nano Omni (free)", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", + "doubao-seed-1.6-flash": { + "id": "doubao-seed-1.6-flash", + "name": "Doubao-Seed 1.6 Flash", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "release_date": "2025-08-15", + "last_updated": "2025-08-15", "modalities": { "input": [ "text", "image", - "video", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 256000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 + "output": 32000 } }, - "nvidia/nemotron-3-ultra-550b-a55b:free": { - "id": "nvidia/nemotron-3-ultra-550b-a55b:free", - "name": "Nemotron 3 Ultra (free)", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", + "qwen3-max-preview": { + "id": "qwen3-max-preview", + "name": "Qwen3 Max Preview", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "release_date": "2025-09-06", + "last_updated": "2025-09-06", "modalities": { "input": [ "text" @@ -147818,40 +156911,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 + "context": 256000, + "output": 64000 } }, - "nvidia/nemotron-3-ultra-550b-a55b": { - "id": "nvidia/nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct", + "description": "Tool-capable chat model for instruction following and agentic application workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "release_date": "2025-09-12", + "last_updated": "2025-09-12", "modalities": { "input": [ "text" @@ -147860,62 +156936,50 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 16384 - }, - "cost": { - "input": 0.5, - "output": 2.2, - "cache_read": 0.1 + "context": 131072, + "output": 32768 } }, - "nvidia/nemotron-3.5-content-safety:free": { - "id": "nvidia/nemotron-3.5-content-safety:free", - "name": "Nemotron 3.5 Content Safety (free)", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "nemotron", + "qwen3-vl-30b-a3b-thinking": { + "id": "qwen3-vl-30b-a3b-thinking", + "name": "Qwen3-Vl 30b A3b Thinking", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning": false, + "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "release_date": "2026-02-09", + "last_updated": "2026-02-09", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 + "output": 32000 } }, - "nvidia/nemotron-3-nano-30b-a3b": { - "id": "nvidia/nemotron-3-nano-30b-a3b", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", + "doubao-1.5-pro-32k": { + "id": "doubao-1.5-pro-32k", + "name": "Doubao 1.5 Pro 32k", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -147924,71 +156988,51 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 228000 - }, - "cost": { - "input": 0.05, - "output": 0.2 + "context": 128000, + "output": 12000 } }, - "nvidia/llama-3.3-nemotron-super-49b-v1.5": { - "id": "nvidia/llama-3.3-nemotron-super-49b-v1.5", - "name": "Llama 3.3 Nemotron Super 49B v1.5", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, + "claude-3.5-sonnet": { + "id": "claude-3.5-sonnet", + "name": "Claude 3.5 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "release_date": "2025-09-09", + "last_updated": "2025-09-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.4, - "output": 0.4 + "context": 200000, + "output": 8200 } }, - "nvidia/nemotron-3-super-120b-a12b:free": { - "id": "nvidia/nemotron-3-super-120b-a12b:free", - "name": "Nemotron 3 Super (free)", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "gpt-oss-20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ "text" @@ -147997,110 +157041,81 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 + "context": 128000, + "output": 4096 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 16384 - }, - "cost": { - "input": 0.08, - "output": 0.45 + "context": 1048576, + "output": 65536 } }, - "cognitivecomputations/dolphin-mistral-24b-venice-edition:free": { - "id": "cognitivecomputations/dolphin-mistral-24b-venice-edition:free", - "name": "Uncensored (free)", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 + "context": 1048576, + "output": 64000 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "qwen-vl-max-2025-01-25": { + "id": "qwen-vl-max-2025-01-25", + "name": "Qwen VL-MAX-2025-01-25", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -148112,38 +157127,24 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.105, - "output": 0.28, - "cache_read": 0.028 + "context": 128000, + "output": 4096 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "doubao-1.5-thinking-pro": { + "id": "doubao-1.5-thinking-pro", + "name": "Doubao 1.5 Thinking Pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -148152,43 +157153,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "context": 128000, + "output": 16000 } }, - "inception/mercury-2": { - "id": "inception/mercury-2", - "name": "Mercury 2", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", - "attachment": false, + "claude-4.5-haiku": { + "id": "claude-4.5-haiku", + "name": "Claude 4.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-03-04", - "last_updated": "2026-03-04", + "release_date": "2025-10-16", + "last_updated": "2025-10-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -148196,43 +157182,53 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 50000 + "context": 200000, + "output": 64000 + } + }, + "kling-v2-6": { + "id": "kling-v2-6", + "name": "Kling-V2 6", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-01-13", + "last_updated": "2026-01-13", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "video" + ] }, - "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.025 + "open_weights": false, + "limit": { + "context": 99999999, + "output": 99999999 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "claude-4.1-opus": { + "id": "claude-4.1-opus", + "name": "Claude 4.1 Opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -148240,62 +157236,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "context": 200000, + "output": 32000 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude Sonnet 4", + "claude-3.7-sonnet": { + "id": "claude-3.7-sonnet", + "name": "Claude 3.7 Sonnet", "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "knowledge": "2025-01-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text" @@ -148303,62 +157263,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "context": 200000, + "output": 128000 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, + "MiniMax-M1": { + "id": "MiniMax-M1", + "name": "MiniMax M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -148366,48 +157293,31 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "context": 1000000, + "output": 80000 } }, - "anthropic/claude-opus-4.7-fast": { - "id": "anthropic/claude-opus-4.7-fast", - "name": "Claude Opus 4.7 (Fast)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "doubao-seed-1.6": { + "id": "doubao-seed-1.6", + "name": "Doubao-Seed 1.6", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "structured_output": false, + "temperature": true, + "release_date": "2025-08-15", + "last_updated": "2025-08-15", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -148415,46 +157325,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 + "context": 256000, + "output": 32000 } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "qwen3-30b-a3b": { + "id": "qwen3-30b-a3b", + "name": "Qwen3 30B A3B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -148462,49 +157355,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "context": 40000, + "output": 4096 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-22", + "last_updated": "2026-02-22", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -148512,67 +157386,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "context": 256000, + "output": 64000 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "doubao-seed-2.0-code": { + "id": "doubao-seed-2.0-code", + "name": "Doubao Seed 2.0 Code", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -148580,46 +157418,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 256000, "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "claude-4.0-opus": { + "id": "claude-4.0-opus", + "name": "Claude 4.0 Opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -148627,44 +157445,24 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "context": 200000, + "output": 32000 } }, - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - } - ], + "kimi-k2": { + "id": "kimi-k2", + "name": "Kimi K2", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", "release_date": "2025-08-05", "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -148672,52 +157470,24 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "context": 128000, + "output": 128000 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "qwen3-30b-a3b-instruct-2507": { + "id": "qwen3-30b-a3b-instruct-2507", + "name": "Qwen3 30b A3b Instruct 2507", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2026-02-04", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -148725,53 +157495,27 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "context": 128000, + "output": 32000 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "qwen2.5-vl-72b-instruct": { + "id": "qwen2.5-vl-72b-instruct", + "name": "Qwen 2.5 VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -148779,51 +157523,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "context": 128000, + "output": 8192 } }, - "anthropic/claude-3-haiku": { - "id": "anthropic/claude-3-haiku", - "name": "Claude 3 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude", + "doubao-1.5-vision-pro": { + "id": "doubao-1.5-vision-pro", + "name": "Doubao 1.5 Vision Pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": false, "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-03-13", - "last_updated": "2024-03-13", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -148831,44 +157550,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 4096 - }, - "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.03, - "cache_write": 0.3 + "context": 128000, + "output": 16000 } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Claude Opus 4", + "claude-4.5-opus": { + "id": "claude-4.5-opus", + "name": "Claude 4.5 Opus", "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "knowledge": "2025-01-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text" @@ -148877,47 +157578,24 @@ "open_weights": false, "limit": { "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "output": 200000 } }, - "anthropic/claude-opus-4.8-fast": { - "id": "anthropic/claude-opus-4.8-fast", - "name": "Claude Opus 4.8 (Fast)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek-v3.1": { + "id": "deepseek-v3.1", + "name": "DeepSeek-V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-05-27", - "last_updated": "2026-05-27", + "structured_output": false, + "temperature": true, + "release_date": "2025-08-19", + "last_updated": "2025-08-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -148925,53 +157603,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "context": 128000, + "output": 32000 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -148979,56 +157629,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "context": 128000, + "output": 32000 } }, - "tencent/hy3:free": { - "id": "tencent/hy3:free", - "name": "Hy3 (free)", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hy3", + "qwen3-30b-a3b-thinking-2507": { + "id": "qwen3-30b-a3b-thinking-2507", + "name": "Qwen3 30b A3b Thinking 2507", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "release_date": "2026-02-04", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -149037,71 +157657,52 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 + "context": 126000, + "output": 32000 } }, - "tencent/hy3": { - "id": "tencent/hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hy3", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 0.58, - "cache_read": 0.035 + "context": 1048576, + "output": 64000 } }, - "tencent/hunyuan-a13b-instruct": { - "id": "tencent/hunyuan-a13b-instruct", - "name": "Hunyuan A13B Instruct", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", + "deepseek-r1-0528": { + "id": "deepseek-r1-0528", + "name": "DeepSeek-R1-0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": true, + "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-07-08", - "last_updated": "2025-07-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -149110,38 +157711,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 0.57 + "context": 128000, + "output": 32000 } }, - "tencent/hy3-preview": { - "id": "tencent/hy3-preview", - "name": "Hy3 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "GLM 4.5 Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -149150,30 +157741,24 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.063, - "output": 0.21, - "cache_read": 0.021 + "context": 131000, + "output": 4096 } }, - "deepcogito/cogito-v2.1-671b": { - "id": "deepcogito/cogito-v2.1-671b", - "name": "Cogito v2.1 671B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "cogito", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": true, + "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ "text" @@ -149185,57 +157770,51 @@ "open_weights": false, "limit": { "context": 128000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 1.25 + "output": 4096 } }, - "cohere/command-a": { - "id": "cohere/command-a", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "gemini-3.0-pro-preview": { + "id": "gemini-3.0-pro-preview", + "name": "Gemini 3.0 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 2.5, - "output": 10 + "context": 1000000, + "output": 64000 } }, - "cohere/command-r-08-2024": { - "id": "cohere/command-r-08-2024", - "name": "Command R", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", + "deepseek/deepseek-v3.2-exp-thinking": { + "id": "deepseek/deepseek-v3.2-exp-thinking", + "name": "DeepSeek/DeepSeek-V3.2-Exp-Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text" @@ -149244,29 +157823,24 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 4000 - }, - "cost": { - "input": 0.15, - "output": 0.6 + "output": 32000 } }, - "cohere/command-r7b-12-2024": { - "id": "cohere/command-r7b-12-2024", - "name": "Command R7B", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", + "deepseek/deepseek-math-v2": { + "id": "deepseek/deepseek-math-v2", + "name": "Deepseek/Deepseek-Math-V2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-12-02", - "last_updated": "2024-12-02", + "release_date": "2025-12-04", + "last_updated": "2025-12-04", "modalities": { "input": [ "text" @@ -149275,29 +157849,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 0.0375, - "output": 0.15 + "context": 160000, + "output": 160000 } }, - "cohere/command-r-plus-08-2024": { - "id": "cohere/command-r-plus-08-2024", - "name": "Command R+", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", + "deepseek/deepseek-v3.1-terminus": { + "id": "deepseek/deepseek-v3.1-terminus", + "name": "DeepSeek/DeepSeek-V3.1-Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ "text" @@ -149306,29 +157874,24 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 4000 - }, - "cost": { - "input": 2.5, - "output": 10 + "output": 32000 } }, - "cohere/north-mini-code:free": { - "id": "cohere/north-mini-code:free", - "name": "North Mini Code (free)", - "description": "Cohere coding model for practical software engineering and agentic edits", - "family": "north", + "deepseek/deepseek-v3.2-251201": { + "id": "deepseek/deepseek-v3.2-251201", + "name": "Deepseek/DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -149337,44 +157900,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 128000, + "output": 32000 } }, - "~x-ai/grok-latest": { - "id": "~x-ai/grok-latest", - "name": "Grok Latest", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "deepseek/deepseek-v3.1-terminus-thinking": { + "id": "deepseek/deepseek-v3.1-terminus-thinking", + "name": "DeepSeek/DeepSeek-V3.1-Terminus-Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -149382,27 +157928,21 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 1000000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "context": 128000, + "output": 32000 } }, - "gryphe/mythomax-l2-13b": { - "id": "gryphe/mythomax-l2-13b", - "name": "MythoMax 13B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "deepseek/deepseek-v3.2-exp": { + "id": "deepseek/deepseek-v3.2-exp", + "name": "DeepSeek/DeepSeek-V3.2-Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-06-30", - "release_date": "2023-07-02", - "last_updated": "2023-07-02", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text" @@ -149411,73 +157951,53 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 4096, - "output": 4096 - }, - "cost": { - "input": 0.06, - "output": 0.06 + "context": 128000, + "output": 32000 } }, - "stepfun/step-3.7-flash": { - "id": "stepfun/step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "Z-AI/GLM 4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2026-01-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2025-10-11", + "last_updated": "2025-10-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 - }, - "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "context": 200000, + "output": 200000 } }, - "stepfun/step-3.5-flash": { - "id": "stepfun/step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "Z-Ai/GLM 5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -149486,62 +158006,53 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.3 + "context": 200000, + "output": 128000 } }, - "nex-agi/nex-n2-mini": { - "id": "nex-agi/nex-n2-mini", - "name": "Nex-N2-Mini", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "agi", - "attachment": true, + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "Z-Ai/GLM 4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-24", - "last_updated": "2026-06-24", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.025, - "output": 0.1, - "cache_read": 0.0025 + "context": 200000, + "output": 200000 } }, - "nex-agi/nex-n2-pro": { - "id": "nex-agi/nex-n2-pro", - "name": "Nex-N2-Pro", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "agi", + "z-ai/autoglm-phone-9b": { + "id": "z-ai/autoglm-phone-9b", + "name": "Z-Ai/Autoglm Phone 9b", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-06-08", - "last_updated": "2026-06-08", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text", @@ -149551,29 +158062,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.25, - "output": 1, - "cache_read": 0.025 + "context": 12800, + "output": 4096 } }, - "undi95/remm-slerp-l2-13b": { - "id": "undi95/remm-slerp-l2-13b", - "name": "ReMM SLERP 13B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "meituan/longcat-flash-chat": { + "id": "meituan/longcat-flash-chat", + "name": "Meituan/Longcat-Flash-Chat", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-06-30", - "release_date": "2023-07-22", - "last_updated": "2023-07-22", + "release_date": "2025-11-05", + "last_updated": "2025-11-05", "modalities": { "input": [ "text" @@ -149582,45 +158087,25 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 6144, - "output": 4096 - }, - "cost": { - "input": 0.45, - "output": 0.65 + "context": 131072, + "output": 131072 } }, - "~openai/gpt-mini-latest": { - "id": "~openai/gpt-mini-latest", - "name": "OpenAI GPT Mini Latest", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "meituan/longcat-flash-lite": { + "id": "meituan/longcat-flash-lite", + "name": "Meituan/Longcat-Flash-Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-06", + "last_updated": "2026-02-06", "modalities": { "input": [ - "pdf", - "image", "text" ], "output": [ @@ -149629,45 +158114,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "context": 256000, + "output": 320000 } }, - "~openai/gpt-latest": { - "id": "~openai/gpt-latest", - "name": "OpenAI GPT Latest", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "x-ai/grok-4-fast-reasoning": { + "id": "x-ai/grok-4-fast-reasoning", + "name": "X-Ai/Grok-4-Fast-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "audio", + "video" ], "output": [ "text" @@ -149675,33 +158143,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25 + "context": 2000000, + "output": 2000000 } }, - "~moonshotai/kimi-latest": { - "id": "~moonshotai/kimi-latest", - "name": "MoonshotAI Kimi Latest", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", + "x-ai/grok-4.1-fast-reasoning": { + "id": "x-ai/grok-4.1-fast-reasoning", + "name": "X-Ai/Grok 4.1 Fast Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2025-12-19", + "last_updated": "2025-12-19", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -149709,58 +158172,27 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.66, - "output": 3.41, - "cache_read": 0.15 + "context": 20000000, + "output": 2000000 } }, - "relace/relace-search": { - "id": "relace/relace-search", - "name": "Relace Search", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, + "x-ai/grok-4-fast-non-reasoning": { + "id": "x-ai/grok-4-fast-non-reasoning", + "name": "X-Ai/Grok-4-Fast-Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 1, - "output": 3 - } - }, - "relace/relace-apply-3": { - "id": "relace/relace-apply-3", - "name": "Relace Apply 3", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "release_date": "2025-09-26", - "last_updated": "2025-09-26", + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -149768,27 +158200,22 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.85, - "output": 1.25 + "context": 2000000, + "output": 2000000 } }, - "ai21/jamba-large-1.7": { - "id": "ai21/jamba-large-1.7", - "name": "Jamba Large 1.7", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "jamba", + "x-ai/grok-code-fast-1": { + "id": "x-ai/grok-code-fast-1", + "name": "x-AI/Grok-Code-Fast 1", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "release_date": "2025-09-02", + "last_updated": "2025-09-02", "modalities": { "input": [ "text" @@ -149797,28 +158224,24 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 256000, - "output": 4096 - }, - "cost": { - "input": 2, - "output": 8 + "output": 10000 } }, - "arcee-ai/coder-large": { - "id": "arcee-ai/coder-large", - "name": "Coder Large", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "x-ai/grok-4.1-fast": { + "id": "x-ai/grok-4.1-fast", + "name": "x-AI/Grok-4.1-Fast", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-05", - "last_updated": "2025-05-05", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ "text" @@ -149829,29 +158252,27 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 0.8 + "context": 2000000, + "output": 2000000 } }, - "arcee-ai/virtuoso-large": { - "id": "arcee-ai/virtuoso-large", - "name": "Virtuoso Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, + "x-ai/grok-4.1-fast-non-reasoning": { + "id": "x-ai/grok-4.1-fast-non-reasoning", + "name": "X-Ai/Grok 4.1 Fast Non Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-05", - "last_updated": "2025-05-05", + "release_date": "2025-12-19", + "last_updated": "2025-12-19", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -149859,125 +158280,79 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 64000 - }, - "cost": { - "input": 0.75, - "output": 1.2 + "context": 2000000, + "output": 2000000 } }, - "arcee-ai/trinity-large-thinking": { - "id": "arcee-ai/trinity-large-thinking", - "name": "Trinity Large Thinking", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "trinity", - "attachment": false, + "x-ai/grok-4-fast": { + "id": "x-ai/grok-4-fast", + "name": "x-AI/Grok-4-Fast", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "release_date": "2025-09-20", + "last_updated": "2025-09-20", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 80000 - }, - "cost": { - "input": 0.25, - "output": 0.8, - "cache_read": 0.06 + "context": 2000000, + "output": 2000000 } }, - "arcee-ai/trinity-mini": { - "id": "arcee-ai/trinity-mini", - "name": "Trinity Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "trinity-mini", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "stepfun-ai/gelab-zero-4b-preview": { + "id": "stepfun-ai/gelab-zero-4b-preview", + "name": "Stepfun-Ai/Gelab Zero 4b Preview", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.045, - "output": 0.15 + "context": 8192, + "output": 4096 } }, - "mancer/weaver": { - "id": "mancer/weaver", - "name": "Weaver (alpha)", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "alpha", + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2023-06-30", - "release_date": "2023-08-02", - "last_updated": "2023-08-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8000, - "output": 2000 - }, - "cost": { - "input": 0.75, - "output": 1 - } - }, - "perplexity/sonar-reasoning-pro": { - "id": "perplexity/sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded reasoning model for multi-step research and cited answers", - "family": "sonar-reasoning", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-03-07", - "last_updated": "2025-03-07", + "release_date": "2025-11-07", + "last_updated": "2025-11-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -149985,30 +158360,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 8 + "context": 256000, + "output": 100000 } }, - "perplexity/sonar": { - "id": "perplexity/sonar", - "name": "Sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Moonshotai/Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-01-27", - "last_updated": "2025-01-27", + "release_date": "2026-01-28", + "last_updated": "2026-01-28", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -150016,30 +158387,24 @@ }, "open_weights": false, "limit": { - "context": 127072, - "output": 127072 - }, - "cost": { - "input": 1, - "output": 1 + "context": 256000, + "output": 256000 } }, - "perplexity/sonar-pro": { - "id": "perplexity/sonar-pro", - "name": "Sonar Pro", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", - "attachment": true, + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-03-07", - "last_updated": "2025-03-07", + "release_date": "2025-09-08", + "last_updated": "2025-09-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -150047,59 +158412,58 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8000 - }, - "cost": { - "input": 3, - "output": 15 + "context": 256000, + "output": 100000 } }, - "perplexity/sonar-pro-search": { - "id": "perplexity/sonar-pro-search", - "name": "Sonar Pro Search", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", - "attachment": true, + "xiaomi/mimo-v2-flash": { + "id": "xiaomi/mimo-v2-flash", + "name": "Xiaomi/Mimo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "release_date": "2025-10-30", - "last_updated": "2025-10-30", + "knowledge": "2024-12-01", + "release_date": "2025-12-16", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 3, - "output": 15 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01 } }, - "perplexity/sonar-deep-research": { - "id": "perplexity/sonar-deep-research", - "name": "Sonar Deep Research", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar-deep-research", + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "Minimax/Minimax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-03-07", - "last_updated": "2025-03-07", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -150110,26 +158474,26 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 204800, "output": 128000 - }, - "cost": { - "input": 2, - "output": 8, - "reasoning": 3 } }, - "openrouter/bodybuilder": { - "id": "openrouter/bodybuilder", - "name": "Body Builder (beta)", - "description": "Preview model for early access evaluation, prototyping, and compatibility testing", + "minimax/minimax-m2.5-highspeed": { + "id": "minimax/minimax-m2.5-highspeed", + "name": "Minimax/Minimax-M2.5 Highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "structured_output": false, - "temperature": false, - "release_date": "2025-12-05", - "last_updated": "2025-12-05", + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text" @@ -150140,26 +158504,29 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 204800, "output": 128000 } }, - "openrouter/free": { - "id": "openrouter/free", - "name": "Free Models Router", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "Minimax/Minimax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -150167,26 +158534,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 8000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 204800, + "output": 128000 } }, - "openrouter/fusion": { - "id": "openrouter/fusion", - "name": "Fusion", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "Minimax/Minimax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "structured_output": false, - "temperature": false, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "temperature": true, + "release_date": "2025-10-28", + "last_updated": "2025-10-28", "modalities": { "input": [ "text" @@ -150197,24 +158564,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 200000, "output": 128000 } }, - "openrouter/pareto-code": { - "id": "openrouter/pareto-code", - "name": "Pareto Code Router", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, + "stepfun/step-3.5-flash": { + "id": "stepfun/step-3.5-flash", + "name": "Stepfun/Step-3.5 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": true, "reasoning": false, "tool_call": false, "structured_output": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "temperature": true, + "release_date": "2026-02-02", + "last_updated": "2026-02-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -150222,69 +158590,26 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 200000 + "context": 64000, + "output": 4096 } }, - "openrouter/auto": { - "id": "openrouter/auto", - "name": "Auto Router", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "auto", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "OpenAI/GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2023-11-08", - "last_updated": "2023-11-08", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", - "image", - "audio", - "pdf", - "video" - ], - "output": [ "text", "image" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 2000000 - } - }, - "qwen/qwen3.5-plus-20260420": { - "id": "qwen/qwen3.5-plus-20260420", - "name": "Qwen3.5 Plus 2026-04-20", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.5", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 81920 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", - "modalities": { - "input": [ - "text", - "image", - "video" ], "output": [ "text" @@ -150292,28 +158617,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 1.8, - "cache_write": 0.375 + "context": 400000, + "output": 128000 } }, - "qwen/qwen3-next-80b-a3b-instruct:free": { - "id": "qwen/qwen3-next-80b-a3b-instruct:free", - "name": "Qwen3 Next 80B A3B Instruct (free)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "OpenAI/GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2025-09-19", + "last_updated": "2025-09-19", "modalities": { "input": [ "text" @@ -150322,34 +158640,44 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 + "context": 400000, + "output": 128000 } - }, - "qwen/qwen3-vl-235b-a22b-thinking": { - "id": "qwen/qwen3-vl-235b-a22b-thinking", - "name": "Qwen3 VL 235B A22B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + } + } + }, + "modelscope": { + "id": "modelscope", + "env": [ + "MODELSCOPE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api-inference.modelscope.cn/v1", + "name": "ModelScope", + "doc": "https://modelscope.cn/docs/model-service/API-Inference/intro", + "models": { + "ZhipuAI/GLM-4.6": { + "id": "ZhipuAI/GLM-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2025-07", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -150357,32 +158685,34 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 202752, + "output": 98304 }, "cost": { - "input": 0.26, - "output": 2.6 + "input": 0, + "output": 0 } }, - "qwen/qwen3-vl-30b-a3b-thinking": { - "id": "qwen/qwen3-vl-30b-a3b-thinking", - "name": "Qwen3 VL 30B A3B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "ZhipuAI/GLM-4.5": { + "id": "ZhipuAI/GLM-4.5", + "name": "GLM-4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -150391,26 +158721,25 @@ "open_weights": true, "limit": { "context": 131072, - "output": 32768 + "output": 98304 }, "cost": { - "input": 0.13, - "output": 1.56 + "input": 0, + "output": 0 } }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "Qwen/Qwen3-30B-A3B-Instruct-2507": { + "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen3 30B A3B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ "text" @@ -150419,31 +158748,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.65, - "output": 3.25, - "cache_read": 0.13, - "cache_write": 0.8125 + "input": 0, + "output": 0 } }, - "qwen/qwen-plus": { - "id": "qwen/qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen3-235B-A22B-Thinking-2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", + "knowledge": "2025-04", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ "text" @@ -150452,31 +158779,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.26, - "output": 0.78, - "cache_read": 0.052, - "cache_write": 0.325 + "input": 0, + "output": 0 } }, - "qwen/qwen3-coder-30b-a3b-instruct": { - "id": "qwen/qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-07-21", "modalities": { "input": [ "text" @@ -150487,32 +158811,27 @@ }, "open_weights": true, "limit": { - "context": 160000, - "output": 32768 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.07, - "output": 0.27 + "input": 0, + "output": 0 } }, - "qwen/qwen3-32b": { - "id": "qwen/qwen3-32b", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "Qwen/Qwen3-30B-A3B-Thinking-2507": { + "id": "Qwen/Qwen3-30B-A3B-Thinking-2507", + "name": "Qwen3 30B A3B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ "text" @@ -150523,27 +158842,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.08, - "output": 0.28 + "input": 0, + "output": 0 } }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "Qwen/Qwen3-Coder-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "name": "Qwen3 Coder 30B A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2025-07-31", + "last_updated": "2025-07-31", "modalities": { "input": [ "text" @@ -150555,67 +158873,90 @@ "open_weights": true, "limit": { "context": 262144, - "output": 16384 + "output": 65536 }, "cost": { - "input": 0.09, - "output": 1.1 + "input": 0, + "output": 0 } - }, - "qwen/qwen3-vl-8b-instruct": { - "id": "qwen/qwen3-vl-8b-instruct", - "name": "Qwen3 VL 8B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + } + } + }, + "google": { + "id": "google", + "env": [ + "GOOGLE_API_KEY", + "GOOGLE_GENERATIVE_AI_API_KEY", + "GEMINI_API_KEY" + ], + "npm": "@ai-sdk/google", + "name": "Google", + "doc": "https://ai.google.dev/gemini-api/docs/models", + "models": { + "gemini-flash-latest": { + "id": "gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-14", - "last_updated": "2025-10-14", + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ + "text", "image", - "text" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.117, - "output": 0.455 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "input_audio": 0.75 } }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "gemma-4-31b-it": { + "id": "gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -150625,35 +158966,24 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.32, - "output": 1.28, - "cache_read": 0.064, - "cache_write": 0.4 + "context": 262144, + "output": 32768 } }, - "qwen/qwen3.6-35b-a3b": { - "id": "qwen/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "gemini-omni-flash-preview": { + "id": "gemini-omni-flash-preview", + "name": "Gemini Omni Flash Preview", + "description": "Video generation and editing model for fast, conversational text- and image-to-video workflows", + "family": "gemini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", @@ -150661,345 +158991,335 @@ "video" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 1 + "input": 1.5, + "output": 17.5 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "gemini-3.1-flash-image-preview": { + "id": "gemini-3.1-flash-image-preview", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 + "type": "effort", + "values": [ + "minimal", + "high" + ] } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2025-01", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, + "context": 65536, "output": 65536 }, "cost": { - "input": 1.25, - "output": 3.75, - "cache_read": 0.25, - "cache_write": 1.5625 + "input": 0.5, + "output": 60 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, + "veo-3.1-lite-generate-preview": { + "id": "veo-3.1-lite-generate-preview", + "name": "Veo 3.1 lite", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "tool_call": false, + "temperature": false, + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.78, - "output": 3.9, - "cache_read": 0.156, - "cache_write": 0.975 + "context": 480, + "output": 8192 } }, - "qwen/qwen3-8b": { - "id": "qwen/qwen3-8b", - "name": "Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "lyria-3-pro-preview": { + "id": "lyria-3-pro-preview", + "name": "Lyria 3 Pro Preview", + "description": "Music generation model for full-length songs from text or images with vocals and structure", + "family": "lyria", + "attachment": true, + "reasoning": false, + "tool_call": false, "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "release_date": "2026-03-25", + "last_updated": "2026-03-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.117, - "output": 0.455 + "input": 0, + "output": 0 } }, - "qwen/qwen-plus-2025-07-28": { - "id": "qwen/qwen-plus-2025-07-28", - "name": "Qwen Plus 0728", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "gemini-3.1-flash-tts-preview": { + "id": "gemini-3.1-flash-tts-preview", + "name": "Gemini 3.1 Flash TTS Preview", + "description": "Low-latency speech generation with steerable prompts and expressive audio tags", + "family": "gemini-flash", "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-08", - "last_updated": "2025-09-08", + "knowledge": "2025-01", + "release_date": "2026-04-15", + "last_updated": "2026-04-15", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 8192, + "output": 16384 }, "cost": { - "input": 0.26, - "output": 0.78 + "input": 1, + "output": 20 } }, - "qwen/qwen3.5-flash-02-23": { - "id": "qwen/qwen3.5-flash-02-23", - "name": "Qwen3.5-Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.5-flash-image": { + "id": "gemini-2.5-flash-image", + "name": "Nano Banana", + "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 81920 - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-02-25", - "last_updated": "2026-02-25", + "knowledge": "2024-06", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.065, - "output": 0.26 - } - }, - "qwen/qwen3-coder:free": { - "id": "qwen/qwen3-coder:free", - "name": "Qwen3 Coder 480B A35B (free)", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 262000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 30, + "cache_read": 0.075 } }, - "qwen/qwen3-30b-a3b-instruct-2507": { - "id": "qwen/qwen3-30b-a3b-instruct-2507", - "name": "Qwen3 30B A3B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gemini-flash-lite-latest": { + "id": "gemini-flash-lite-latest", + "name": "Gemini Flash-Lite Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.04815, - "output": 0.19305 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "qwen/qwen-2.5-coder-32b-instruct": { - "id": "qwen/qwen-2.5-coder-32b-instruct", - "name": "Qwen2.5 Coder 32B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "gemini-embedding-2": { + "id": "gemini-embedding-2", + "name": "Gemini Embedding 2", + "description": "Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space", + "family": "gemini", + "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2024-11-11", - "last_updated": "2024-11-11", + "temperature": false, + "knowledge": "2025-11", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 8192, + "output": 1 }, "cost": { - "input": 0.66, - "output": 1 + "input": 0.2, + "output": 0, + "input_audio": 6.5 } }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", - "attachment": false, + "gemini-3.6-flash": { + "id": "gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.0975, - "output": 0.78 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "input_audio": 0.75 } }, - "qwen/qwen3-235b-a22b-thinking-2507": { - "id": "qwen/qwen3-235b-a22b-thinking-2507", - "name": "Qwen3 235B A22B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "gemma-4-26b-a4b-it": { + "id": "gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -151008,102 +159328,103 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.1495, - "output": 1.495 + "output": 32768 } }, - "qwen/qwen3-vl-32b-instruct": { - "id": "qwen/qwen3-vl-32b-instruct", - "name": "Qwen3 VL 32B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-3-pro-image-preview": { + "id": "gemini-3-pro-image-preview", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2025-10-23", - "last_updated": "2025-10-23", + "knowledge": "2025-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 131072, "output": 32768 }, "cost": { - "input": 0.104, - "output": 0.416 + "input": 2, + "output": 120 } }, - "qwen/qwen3-coder": { - "id": "qwen/qwen3-coder", - "name": "Qwen3 Coder 480B A35B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "gemini-2.5-pro-preview-tts": { + "id": "gemini-2.5-pro-preview-tts", + "name": "Gemini 2.5 Pro Preview TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gemini-flash", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "knowledge": "2025-01", + "release_date": "2025-05-01", + "last_updated": "2025-05-01", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 8192, + "output": 16384 }, "cost": { - "input": 0.22, - "output": 1.8 + "input": 1, + "output": 20 } }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -151111,143 +159432,169 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.1875, - "output": 1.125, - "cache_write": 0.234375 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "qwen/qwen3.5-plus-02-15": { - "id": "qwen/qwen3.5-plus-02-15", - "name": "Qwen3.5 Plus 2026-02-15", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-3.1-flash-image": { + "id": "gemini-3.1-flash-image", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "effort", + "values": [ + "minimal", + "high" + ] } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", "image", - "video" + "video", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, + "context": 65536, "output": 65536 }, "cost": { - "input": 0.26, - "output": 1.56 + "input": 0.5, + "output": 60 } }, - "qwen/qwen-2.5-7b-instruct": { - "id": "qwen/qwen-2.5-7b-instruct", - "name": "Qwen2.5 7B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "lyria-3-clip-preview": { + "id": "lyria-3-clip-preview", + "name": "Lyria 3 Clip Preview", + "description": "Music generation model for short 30-second clips, loops, and previews from text or image prompts", + "family": "lyria", + "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2024-10-16", - "last_updated": "2024-10-16", + "release_date": "2026-03-25", + "last_updated": "2026-03-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.04, - "output": 0.1 + "input": 0, + "output": 0 } }, - "qwen/qwen3-vl-8b-thinking": { - "id": "qwen/qwen3-vl-8b-thinking", - "name": "Qwen3 VL 8B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-14", - "last_updated": "2025-10-14", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ + "text", "image", - "text" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.117, - "output": 1.365 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "qwen/qwen3-max-thinking": { - "id": "qwen/qwen3-max-thinking", - "name": "Qwen3 Max Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "gemini-2.5-computer-use-preview-10-2025": { + "id": "gemini-2.5-computer-use-preview-10-2025", + "name": "Gemini 2.5 Computer Use Preview 10-2025", + "description": "Specialized Gemini 2.5 model for browser-control agents that automate UI tasks", + "family": "gemini-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1, - "max": 81920 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-09", - "last_updated": "2026-02-09", + "knowledge": "2025-01", + "release_date": "2025-10-07", + "last_updated": "2025-10-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -151255,122 +159602,182 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.78, - "output": 3.9 + "input": 1.25, + "output": 10, + "tiers": [ + { + "input": 2.5, + "output": 15, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15 + } } }, - "qwen/qwen3-30b-a3b-thinking-2507": { - "id": "qwen/qwen3-30b-a3b-thinking-2507", - "name": "Qwen3 30B A3B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.13, - "output": 1.56 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 } }, - "qwen/qwen2.5-vl-72b-instruct": { - "id": "qwen/qwen2.5-vl-72b-instruct", - "name": "Qwen2.5 VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2025-02-01", - "last_updated": "2025-02-01", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.8, - "output": 1, - "cache_read": 0.4 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-3-pro-image": { + "id": "gemini-3-pro-image", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high" + ] } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.195, - "output": 1.56 + "input": 2, + "output": 120 } }, - "qwen/qwen3-235b-a22b": { - "id": "qwen/qwen3-235b-a22b", - "name": "Qwen3 235B-A22B", - "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", - "family": "qwen", - "attachment": false, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -151378,589 +159785,761 @@ }, { "type": "budget_tokens", - "min": 1, - "max": 38912 + "min": 0, + "max": 24576 } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.455, - "output": 1.82 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "qwen/qwen-2.5-72b-instruct": { - "id": "qwen/qwen-2.5-72b-instruct", - "name": "Qwen2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "gemini-2.5-flash-preview-tts": { + "id": "gemini-2.5-flash-preview-tts", + "name": "Gemini 2.5 Flash Preview TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gemini-flash", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "knowledge": "2025-01", + "release_date": "2025-05-01", + "last_updated": "2025-05-01", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 8192, "output": 16384 }, "cost": { - "input": 0.36, - "output": 0.4 + "input": 0.5, + "output": 10 } }, - "qwen/qwen-plus-2025-07-28:thinking": { - "id": "qwen/qwen-plus-2025-07-28:thinking", - "name": "Qwen Plus 0728 (thinking)", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "gemini-3.1-flash-lite-image": { + "id": "gemini-3.1-flash-lite-image", + "name": "Nano Banana 2 Lite", + "description": "Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "effort", + "values": [ + "minimal", + "high" + ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-08", - "last_updated": "2025-09-08", + "knowledge": "2025-01", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0.26, - "output": 0.78, - "cache_write": 0.325 + "input": 0.25, + "output": 30 } }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "deep-research-preview-04-2026": { + "id": "deep-research-preview-04-2026", + "name": "Deep Research Preview (Apr-21-2026)", + "description": "Agentic model for autonomous multi-step research, synthesis, and cited reports", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-04", - "last_updated": "2026-02-04", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.11, - "output": 0.8, - "cache_read": 0.07 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "qwen/qwen3.6-27b": { - "id": "qwen/qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "deep-research-max-preview-04-2026": { + "id": "deep-research-max-preview-04-2026", + "name": "Deep Research Max Preview (Apr-21-2026)", + "description": "Maximum-comprehensiveness agentic researcher for multi-step investigation, synthesis, and cited reports", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262140 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.285, - "output": 2.4, - "cache_read": 0.15 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "veo-3.1-fast-generate-preview": { + "id": "veo-3.1-fast-generate-preview", + "name": "Veo 3.1 fast", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-10-15", + "last_updated": "2026-01-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "video" + ] + }, + "open_weights": false, + "limit": { + "context": 480, + "output": 8192 + } + }, + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 81920 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 1, - "cache_read": 0.05 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "input_audio": 0.75 } }, - "qwen/qwen3.5-9b": { - "id": "qwen/qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "veo-3.1-generate-preview": { + "id": "veo-3.1-generate-preview", + "name": "Veo 3.1", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-10-15", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "video" + ] + }, + "open_weights": false, + "limit": { + "context": 480, + "output": 8192 + }, + "status": "beta" + }, + "gemini-3.1-pro-preview-customtools": { + "id": "gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.15 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.385, - "output": 2.45, - "cache_read": 0.111 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 } }, - "qwen/qwen3-vl-30b-a3b-instruct": { - "id": "qwen/qwen3-vl-30b-a3b-instruct", - "name": "Qwen3 VL 30B A3B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-3.8-flash": { + "id": "gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.13, - "output": 0.52 - } - }, - "qwen/qwen3-235b-a22b-2507": { - "id": "qwen/qwen3-235b-a22b-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-07-21", - "last_updated": "2025-07-21", - "modalities": { - "input": [ - "text" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.09, - "output": 0.1 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "input_audio": 0.75 } }, - "qwen/qwen3-coder-flash": { - "id": "qwen/qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "gemini-3.5-live-translate-preview": { + "id": "gemini-3.5-live-translate-preview", + "name": "Gemini 3.5 Live Translate Preview", + "description": "Low-latency audio-to-audio model for real-time speech translation across 70+ languages", + "family": "gemini-pro", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-01", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "audio" ], "output": [ + "audio", "text" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 16384, + "output": 32768 }, "cost": { - "input": 0.195, - "output": 0.975, - "cache_read": 0.039, - "cache_write": 0.24375 + "input": 3.5, + "output": 21, + "input_audio": 3.5, + "output_audio": 21 } }, - "qwen/qwen3-14b": { - "id": "qwen/qwen3-14b", - "name": "Qwen3 14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 512, + "max": 24576 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131702, - "output": 40960 + "context": 1048576, + "output": 65536 }, "cost": { "input": 0.1, - "output": 0.24 + "output": 0.4, + "cache_read": 0.01, + "input_audio": 0.3 } }, - "qwen/qwen3-vl-235b-a22b-instruct": { - "id": "qwen/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "gemini-embedding-001": { + "id": "gemini-embedding-001", + "name": "Gemini Embedding 001", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "gemini", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "tool_call": false, + "temperature": false, + "knowledge": "2025-05", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 2048, + "output": 1 }, "cost": { - "input": 0.2, - "output": 0.88, - "cache_read": 0.11 + "input": 0.15, + "output": 0 } }, - "qwen/qwen3-30b-a3b": { - "id": "qwen/qwen3-30b-a3b", - "name": "Qwen3 30B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "gemini-3.1-flash-lite-preview": { + "id": "gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1048576, + "output": 65536 }, + "status": "deprecated", "cost": { - "input": 0.12, - "output": 0.5 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "qwen/qwen3.6-max-preview": { - "id": "qwen/qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "gemini-3.1-flash-live-preview": { + "id": "gemini-3.1-flash-live-preview", + "name": "Gemini 3.1 Flash Live Preview", + "description": "High-quality, low-latency Live API model for real-time dialogue and voice-first AI applications", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 131072 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "knowledge": "2025-01", + "release_date": "2026-03-26", + "last_updated": "2026-03-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 262144, + "context": 131072, "output": 65536 }, "cost": { - "input": 1.04, - "output": 6.24, - "cache_write": 1.3 + "input": 0.75, + "output": 4.5, + "input_audio": 3, + "output_audio": 12 } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.26, - "output": 2.08 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } - }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", + } + } + }, + "vancine": { + "id": "vancine", + "env": [ + "VANCINE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://vancine.com/v1", + "name": "Vancine", + "doc": "https://vancine.com/docs", + "models": { + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 81920 } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", @@ -151971,65 +160550,99 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 1048576, + "output": 512000 }, "cost": { - "input": 0.325, - "output": 1.95, - "cache_write": 0.40625 + "input": 0.24, + "output": 0.96, + "cache_read": 0.048 } }, - "amazon/nova-lite-v1": { - "id": "amazon/nova-lite-v1", - "name": "Nova Lite 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 300000, - "output": 5120 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.06, - "output": 0.24 + "input": 0.12, + "output": 0.4, + "cache_read": 0.024 } }, - "amazon/nova-premier-v1": { - "id": "amazon/nova-premier-v1", - "name": "Nova Premier 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova", + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-10-31", - "last_updated": "2025-10-31", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" @@ -152038,96 +160651,140 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 32000 + "output": 131072 }, "cost": { - "input": 2.5, - "output": 12.5, - "cache_read": 0.625 + "input": 1.6, + "output": 4.8, + "cache_read": 0.2 } }, - "amazon/nova-pro-v1": { - "id": "amazon/nova-pro-v1", - "name": "Nova Pro 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 300000, - "output": 5120 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.8, - "output": 3.2 + "input": 2.4, + "output": 12, + "cache_read": 0.24 } }, - "amazon/nova-micro-v1": { - "id": "amazon/nova-micro-v1", - "name": "Nova Micro 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 5120 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.035, - "output": 0.14 + "input": 0.24, + "output": 0.96, + "cache_read": 0.0048 } }, - "amazon/nova-2-lite-v1": { - "id": "amazon/nova-2-lite-v1", - "name": "Nova 2 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova", + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", "image", - "video", - "pdf" + "video" ], "output": [ "text" @@ -152136,25 +160793,34 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 65535 + "output": 131072 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 0.12, + "output": 0.38, + "cache_read": 0.013 } }, - "aion-labs/aion-3.0-mini": { - "id": "aion-labs/aion-3.0-mini", - "name": "Aion-3.0-Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "hy4-preview": { + "id": "hy4-preview", + "name": "Hy4 preview", + "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", + "family": "Hy", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-07-07", - "last_updated": "2026-07-07", + "release_date": "2026-08-28", + "last_updated": "2026-08-28", "modalities": { "input": [ "text" @@ -152163,30 +160829,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1024000, + "output": 64000 }, "cost": { - "input": 0.7, - "output": 1.4, - "cache_read": 0.18 + "input": 0.67, + "output": 2, + "cache_read": 0.034 } }, - "aion-labs/aion-rp-llama-3.1-8b": { - "id": "aion-labs/aion-rp-llama-3.1-8b", - "name": "Aion-RP 1.0 (8B)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2025-02-04", - "last_updated": "2025-02-04", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -152195,59 +160870,102 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.8, - "output": 1.6 + "input": 1.12, + "output": 3.52, + "cache_read": 0.208 } - }, - "aion-labs/aion-3.0": { - "id": "aion-labs/aion-3.0", - "name": "Aion-3.0", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + } + } + }, + "zhipuai-coding-plan": { + "id": "zhipuai-coding-plan", + "env": [ + "ZHIPU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://open.bigmodel.cn/api/coding/paas/v4", + "name": "Zhipu AI Coding Plan", + "doc": "https://docs.bigmodel.cn/cn/coding-plan/overview", + "models": { + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-07-07", - "last_updated": "2026-07-07", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 3, - "output": 6, - "cache_read": 0.75 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "aion-labs/aion-2.0": { - "id": "aion-labs/aion-2.0", - "name": "Aion-2.0", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "glm-5.3-highspeed": { + "id": "glm-5.3-highspeed", + "name": "GLM-5.3 Highspeed", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -152256,29 +160974,43 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.8, - "output": 1.6, - "cache_read": 0.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "inflection/inflection-3-pi": { - "id": "inflection/inflection-3-pi", - "name": "Inflection 3 Pi", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2024-10-11", - "last_updated": "2024-10-11", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -152287,90 +161019,112 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8000, - "output": 1024 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "inflection/inflection-3-productivity": { - "id": "inflection/inflection-3-productivity", - "name": "Inflection 3 Productivity", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "glm-4.6v": { + "id": "glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2024-10-11", - "last_updated": "2024-10-11", + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8000, - "output": 1024 + "context": 128000, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.3, + "output": 0.9 } - }, - "sao10k/l3.1-euryale-70b": { - "id": "sao10k/l3.1-euryale-70b", - "name": "Llama 3.1 Euryale 70B v2.2", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + } + } + }, + "lucidquery": { + "id": "lucidquery", + "env": [ + "LUCIDQUERY_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.lucidquery.com/v1", + "name": "LucidQuery", + "doc": "https://lucidquery.com/docs", + "models": { + "lucidquery-agi-01-frontier": { + "id": "lucidquery-agi-01-frontier", + "name": "AGI-01 Frontier", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "agi", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-08-28", - "last_updated": "2024-08-28", + "knowledge": "2026-06-05", + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 300000, + "output": 120000 }, "cost": { - "input": 0.85, - "output": 0.85 + "input": 4.5, + "output": 22 } }, - "sao10k/l3.3-euryale-70b": { - "id": "sao10k/l3.3-euryale-70b", - "name": "Llama 3.3 Euryale 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-12-18", - "last_updated": "2024-12-18", + "lucidquery-nexus-coder": { + "id": "lucidquery-nexus-coder", + "name": "LucidQuery Nexus Coder", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "lucid", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": false, + "knowledge": "2025-08-01", + "release_date": "2025-09-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text" @@ -152379,29 +161133,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 250000, + "output": 60000 }, "cost": { - "input": 0.65, - "output": 0.75 + "input": 2, + "output": 5 } }, - "sao10k/l3-lunaris-8b": { - "id": "sao10k/l3-lunaris-8b", - "name": "Llama 3 8B Lunaris", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-08-13", - "last_updated": "2024-08-13", + "lucidnova-rf1-100b": { + "id": "lucidnova-rf1-100b", + "name": "LucidNova RF1 100B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "nova", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": false, + "knowledge": "2025-09-16", + "release_date": "2024-12-28", + "last_updated": "2025-09-10", "modalities": { "input": [ "text" @@ -152410,60 +161164,79 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 16384 + "context": 120000, + "output": 8000 }, "cost": { - "input": 0.04, - "output": 0.05 + "input": 2, + "output": 5 } }, - "sao10k/l3.1-70b-hanami-x1": { - "id": "sao10k/l3.1-70b-hanami-x1", - "name": "Llama 3.1 70B Hanami x1", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "lucidquery-agi-01-swift": { + "id": "lucidquery-agi-01-swift", + "name": "AGI-01 Swift", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "agi", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2025-01-08", - "last_updated": "2025-01-08", + "knowledge": "2026-06-05", + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16000, - "output": 16000 + "context": 300000, + "output": 120000 }, "cost": { - "input": 3, - "output": 3 + "input": 2.5, + "output": 15 } - }, - "upstage/solar-pro-3": { - "id": "upstage/solar-pro-3", - "name": "Solar Pro 3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "solar-pro", - "attachment": false, + } + } + }, + "gmicloud": { + "id": "gmicloud", + "env": [ + "GMICLOUD_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.gmi-serving.com/v1", + "name": "GMI Cloud", + "doc": "https://docs.gmicloud.ai/inference-engine/api-reference/llm-api-reference", + "models": { + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text" @@ -152474,28 +161247,39 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 409600, "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "allenai/olmo-3-32b-think": { - "id": "allenai/olmo-3-32b-think", - "name": "Olmo 3 32B Think", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "allenai", - "attachment": false, + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-11-21", - "last_updated": "2025-11-21", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -152504,114 +161288,101 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 65536 + "context": 409600, + "output": 128000 }, - "cost": { - "input": 0.15, - "output": 0.5 - } - }, - "deepseek/deepseek-r1-0528": { - "id": "deepseek/deepseek-r1-0528", - "name": "R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 32768 + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 0.5, - "output": 2.15, - "cache_read": 0.35 + "input": 4.5, + "output": 22.5, + "cache_read": 0.45 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", + "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.09, - "output": 0.18, - "cache_read": 0.018 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "deepseek/deepseek-v3.1-terminus": { - "id": "deepseek/deepseek-v3.1-terminus", - "name": "DeepSeek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text" @@ -152620,31 +161391,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 32768 + "context": 409600, + "output": 64000 }, "cost": { - "input": 0.27, - "output": 0.95, - "cache_read": 0.13 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "deepseek/deepseek-r1-distill-llama-70b": { - "id": "deepseek/deepseek-r1-distill-llama-70b", - "name": "R1 Distill Llama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "Qwen/Qwen3.7-Max": { + "id": "Qwen/Qwen3.7-Max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2025-01-23", - "last_updated": "2025-01-23", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -152653,35 +161422,26 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.8, - "output": 0.8 + "input": 2.5, + "output": 7.5, + "cache_read": 0.25, + "cache_write": 3.125 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" @@ -152701,29 +161461,32 @@ }, "open_weights": true, "limit": { - "context": 1048576, + "context": 1048575, "output": 384000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0.112, + "output": 0.224, + "cache_read": 0.022 } }, - "deepseek/deepseek-r1": { - "id": "deepseek/deepseek-r1", - "name": "DeepSeek-R1", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -152734,20 +161497,21 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 16000 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0.7, - "output": 2.5 + "input": 1.392, + "output": 2.784, + "cache_read": 0.116 } }, - "deepseek/deepseek-v3.2-exp": { - "id": "deepseek/deepseek-v3.2-exp", - "name": "DeepSeek V3.2 Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "MiniMaxAI/MiniMax-M3": { + "id": "MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -152755,14 +161519,14 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -152770,27 +161534,43 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 65536 + "context": 1048576, + "output": 512000 }, "cost": { - "input": 0.27, - "output": 0.41 + "input": 0.6, + "output": 2.4, + "cache_read": 0.12, + "tiers": [ + { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24, + "tier": { + "type": "context", + "size": 512000 + } + } + ], + "context_over_200k": { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24 + } } }, - "deepseek/deepseek-chat-v3-0324": { - "id": "deepseek/deepseek-chat-v3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "MiniMaxAI/MiniMax-M2.7": { + "id": "MiniMaxAI/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -152801,28 +161581,32 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 16384 + "context": 196608, + "output": 131072 }, "cost": { - "input": 0.24, - "output": 0.9, - "cache_read": 0.135 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "deepseek/deepseek-chat": { - "id": "deepseek/deepseek-chat", - "name": "DeepSeek Chat", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text" @@ -152833,32 +161617,32 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0.2002, - "output": 0.8001 + "input": 0.855, + "output": 3.6, + "cache_read": 0.144 } }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "moonshotai/kimi-k2.7-code-highspeed": { + "id": "moonshotai/kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text" @@ -152869,33 +161653,39 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.2288, - "output": 0.3432, - "cache_read": 0.02288 + "input": 1.9, + "output": 8, + "cache_read": 0.38 } }, - "deepseek/deepseek-chat-v3.1": { - "id": "deepseek/deepseek-chat-v3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "zai-org/GLM-5.2-FP8": { + "id": "zai-org/GLM-5.2-FP8", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -152906,28 +161696,27 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.21, - "output": 0.79, - "cache_read": 0.13 + "input": 0.979, + "output": 3.08, + "cache_read": 0.182 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "zai-org/GLM-5-FP8": { + "id": "zai-org/GLM-5-FP8", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "interleaved": { - "field": "reasoning_details" + "field": "reasoning_content" }, - "structured_output": true, "temperature": true, "release_date": "2026-02-12", "last_updated": "2026-02-12", @@ -152941,31 +161730,31 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 196608 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.9, - "cache_read": 0.05 + "input": 0.6, + "output": 1.92, + "cache_read": 0.12 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", + "zai-org/GLM-5.1-FP8": { + "id": "zai-org/GLM-5.1-FP8", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "interleaved": { - "field": "reasoning_details" + "field": "reasoning_content" }, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -152976,127 +161765,193 @@ }, "open_weights": true, "limit": { - "context": 204800, + "context": 202752, "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.98, + "output": 3.08, + "cache_read": 0.182 } }, - "minimax/minimax-01": { - "id": "minimax/minimax-01", - "name": "MiniMax-01", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-03-31", - "release_date": "2025-01-15", - "last_updated": "2025-01-15", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000192, - "output": 1000192 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.1 + "input": 5, + "output": 30, + "cache_read": 0.5 } - }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + } + } + }, + "oci": { + "id": "oci", + "env": [ + "OCI_GENAI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.generativeai.us-chicago-1.oci.oraclecloud.com/openai/v1", + "name": "OCI Generative AI", + "doc": "https://docs.oracle.com/en-us/iaas/Content/generative-ai/pretrained-models.htm", + "models": { + "xai.grok-4.3": { + "id": "xai.grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "xai.grok-4.20-non-reasoning": { + "id": "xai.grok-4.20-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, "structured_output": true, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.255, - "output": 1.02 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "minimax/minimax-m2-her": { - "id": "minimax/minimax-m2-her", - "name": "MiniMax M2-her", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "meta.llama-3.3-70b-instruct": { + "id": "meta.llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "release_date": "2026-01-23", - "last_updated": "2026-01-23", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -153105,33 +161960,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 2048 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.72, + "output": 0.72 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "meta.llama-4-scout-17b-16e-instruct": { + "id": "meta.llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout 17B Instruct", + "description": "Open Llama with long-context vision for efficient multimodal agents", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -153139,61 +161993,42 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 196608 + "context": 192000, + "output": 16384 }, "cost": { - "input": 0.24, - "output": 0.96 + "input": 0.72, + "output": 0.72 } }, - "minimax/minimax-m1": { - "id": "minimax/minimax-m1", - "name": "MiniMax M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "xai.grok-4.6": { + "id": "xai.grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 40000 - }, - "cost": { - "input": 0.4, - "output": 2.2 - } - }, - "kwaipilot/kat-coder-pro-v2": { - "id": "kwaipilot/kat-coder-pro-v2", - "name": "KAT-Coder-Pro V2", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "kat-coder", - "attachment": false, - "reasoning": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -153201,31 +162036,47 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 80000 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } }, - "nousresearch/hermes-3-llama-3.1-405b:free": { - "id": "nousresearch/hermes-3-llama-3.1-405b:free", - "name": "Hermes 3 405B Instruct (free)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "hermes", - "attachment": false, + "meta.llama-4-maverick-17b-128e-instruct-fp8": { + "id": "meta.llama-4-maverick-17b-128e-instruct-fp8", + "name": "Llama 4 Maverick 17B Instruct", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", + "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-08-16", - "last_updated": "2024-08-16", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -153233,63 +162084,39 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 512000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.72, + "output": 0.72 } }, - "nousresearch/hermes-4-405b": { - "id": "nousresearch/hermes-4-405b", - "name": "Hermes 4 405B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "hermes", + "openai.gpt-oss-120b": { + "id": "openai.gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 + "tool_call": true, + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 1, - "output": 3 - } - }, - "nousresearch/hermes-3-llama-3.1-70b": { - "id": "nousresearch/hermes-3-llama-3.1-70b", - "name": "Hermes 3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "nousresearch", - "attachment": false, - "reasoning": false, - "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-08-18", - "last_updated": "2024-08-18", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -153300,27 +162127,39 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 128000, "output": 16384 }, "cost": { - "input": 0.7, - "output": 0.7 + "input": 0.15, + "output": 0.6 } }, - "nousresearch/hermes-3-llama-3.1-405b": { - "id": "nousresearch/hermes-3-llama-3.1-405b", - "name": "Hermes 3 405B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "nousresearch", + "openai.gpt-oss-20b": { + "id": "openai.gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-08-16", - "last_updated": "2024-08-16", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -153331,77 +162170,108 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 128000, "output": 16384 }, "cost": { - "input": 1, - "output": 1 + "input": 0.07, + "output": 0.3 } }, - "nousresearch/hermes-4-70b": { - "id": "nousresearch/hermes-4-70b", - "name": "Hermes 4 70B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "hermes", - "attachment": false, + "xai.grok-4.20-reasoning": { + "id": "xai.grok-4.20-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": false, - "structured_output": false, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } } } }, - "jiekou": { - "id": "jiekou", + "cloudflare-ai-gateway": { + "id": "cloudflare-ai-gateway", "env": [ - "JIEKOU_API_KEY" + "CLOUDFLARE_API_TOKEN", + "CLOUDFLARE_ACCOUNT_ID", + "CLOUDFLARE_GATEWAY_ID" ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.jiekou.ai/openai", - "name": "Jiekou.AI", - "doc": "https://docs.jiekou.ai/docs/support/quickstart?utm_source=github_models.dev", + "npm": "ai-gateway-provider", + "name": "Cloudflare AI Gateway", + "doc": "https://developers.cloudflare.com/ai-gateway/", "models": { - "o3": { - "id": "o3", - "name": "o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -153409,30 +162279,52 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 10, - "output": 40 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "grok-code-fast-1": { - "id": "grok-code-fast-1", - "name": "grok-code-fast-1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -153440,31 +162332,49 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.18, - "output": 1.35 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gpt-5.2-pro": { - "id": "gpt-5.2-pro", - "name": "gpt-5.2-pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -153472,43 +162382,49 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, "cost": { - "input": 18.9, - "output": 151.2 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "gemini-2.5-pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -153516,30 +162432,43 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 1.125, - "output": 9 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "claude-haiku-4-5-20251001", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", "family": "claude-haiku", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -153547,31 +162476,49 @@ }, "open_weights": false, "limit": { - "context": 20000, + "context": 200000, "output": 64000 }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, "cost": { - "input": 0.9, - "output": 4.5 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "gpt-5-pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -153579,43 +162526,49 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 272000 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 13.5, - "output": 108 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "gemini-2.5-flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "anthropic/claude-fable-5.1": { + "id": "anthropic/claude-fable-5.1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -153623,30 +162576,43 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.27, - "output": 2.25 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -153654,43 +162620,49 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 1.1, - "output": 4.4 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "gemini-2.5-flash-lite-preview-09-2025": { - "id": "gemini-2.5-flash-lite-preview-09-2025", - "name": "gemini-2.5-flash-lite-preview-09-2025", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -153698,30 +162670,52 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.09, - "output": 0.36 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "claude-opus-4-1-20250805", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -153729,30 +162723,52 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 13.5, - "output": 67.5 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "grok-4-1-fast-non-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -153760,57 +162776,74 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.18, - "output": 0.45 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "gpt-5-chat-latest": { - "id": "gpt-5-chat-latest", - "name": "gpt-5-chat-latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": false, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 384000 }, "cost": { - "input": 1.125, - "output": 9 + "input": 1.74, + "output": 3.48, + "cache_read": 0.145 } }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "claude-opus-4-5-20251101", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "unbiased/pareto": { + "id": "unbiased/pareto", + "name": "Pareto", + "description": "Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-09-17", + "last_updated": "2026-09-17", "modalities": { "input": [ "text", @@ -153822,31 +162855,37 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 65536 + "context": 262144, + "output": 131072 }, "cost": { - "input": 4.5, - "output": 22.5 + "input": 2.5, + "output": 7.5, + "cache_read": 0.25 } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "gpt-5.1-codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "alibaba/qwen3.7-max": { + "id": "alibaba/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -153854,31 +162893,53 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.125, - "output": 9 + "input": 1.25, + "output": 3.75, + "cache_read": 0.25 } }, - "gpt-5.1-codex-max": { - "id": "gpt-5.1-codex-max", - "name": "gpt-5.1-codex-max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "alibaba/qwen3.8-max": { + "id": "alibaba/qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 + } + ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" @@ -153886,30 +162947,30 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.125, - "output": 9 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "claude-opus-4-20250514": { - "id": "claude-opus-4-20250514", - "name": "claude-opus-4-20250514", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "alibaba/qwen3-max": { + "id": "alibaba/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -153917,61 +162978,80 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 13.5, - "output": 67.5 + "input": 1.2, + "output": 6 } }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "alibaba/qwen3.5-397b-a17b": { + "id": "alibaba/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.1, - "output": 4.4 + "input": 0.6, + "output": 3.6 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "gpt-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "alibaba/qwen3.7-plus": { + "id": "alibaba/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -153979,30 +163059,44 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1.575, - "output": 12.6 + "input": 0.32, + "output": 1.28, + "cache_read": 0.064 } }, - "grok-4-0709": { - "id": "grok-4-0709", - "name": "grok-4-0709", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "xai/grok-4.7": { + "id": "xai/grok-4.7", + "name": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", "family": "grok", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -154010,19 +163104,36 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 500000, + "output": 500000 }, "cost": { - "input": 2.7, - "output": 13.5 + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "gemini-2.5-flash-lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "xai/grok-4.3": { + "id": "xai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -154039,14 +163150,13 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -154054,30 +163164,33 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.09, - "output": 0.36 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "claude-sonnet-4-20250514": { - "id": "claude-sonnet-4-20250514", - "name": "claude-sonnet-4-20250514", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "xai/grok-4.20-0309-reasoning": { + "id": "xai/grok-4.20-0309-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -154085,27 +163198,37 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 2.7, - "output": 13.5 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "gpt-5.1-codex-mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "xai/grok-4.5": { + "id": "xai/grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text", @@ -154117,30 +163240,48 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.225, - "output": 1.8 + "input": 2, + "output": 6, + "cache_read": 0.3, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 0.6 + } } }, - "grok-4-fast-reasoning": { - "id": "grok-4-fast-reasoning", - "name": "grok-4-fast-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "xai/grok-4.20-0309-non-reasoning": { + "id": "xai/grok-4.20-0309-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "family": "grok", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -154149,18 +163290,19 @@ "open_weights": false, "limit": { "context": 2000000, - "output": 2000000 + "output": 30000 }, "cost": { - "input": 0.18, - "output": 0.45 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "claude-opus-4-6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "xai/grok-4.6": { + "id": "xai/grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -154169,22 +163311,16 @@ "values": [ "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02", - "last_updated": "2026-02", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text", @@ -154196,94 +163332,107 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 500000, + "output": 500000 }, "cost": { - "input": 5, - "output": 25 - } - }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "claude-sonnet-4-5-20250929", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image" + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 2.7, - "output": 13.5 + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } }, - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "grok-4-1-fast-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.18, - "output": 0.45 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "gemini-3-pro-preview": { - "id": "gemini-3-pro-preview", - "name": "gemini-3-pro-preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -154291,27 +163440,42 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 1.8, - "output": 10.8 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "gpt-5-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -154323,30 +163487,46 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, + "input": 922000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai" + }, "cost": { - "input": 0.225, - "output": 1.8 + "input": 30, + "output": 180 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "gpt-5-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -154354,26 +163534,43 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, + "input": 922000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai" + }, "cost": { - "input": 0.045, - "output": 0.36 + "input": 30, + "output": 180 } }, - "grok-4-fast-non-reasoning": { - "id": "grok-4-fast-non-reasoning", - "name": "grok-4-fast-non-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -154385,32 +163582,37 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 128000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.18, - "output": 0.45 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "gemini-3-flash-preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -154418,26 +163620,30 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 16384 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.5, - "output": 3 + "input": 1.25, + "output": 5, + "cache_read": 0.625 } }, - "gemini-2.5-flash-preview-05-20": { - "id": "gemini-2.5-flash-preview-05-20", - "name": "gemini-2.5-flash-preview-05-20", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", "high" @@ -154446,15 +163652,14 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" @@ -154462,26 +163667,30 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 200000 + "context": 128000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.135, - "output": 3.15 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "gemini-2.5-flash-lite-preview-06-17": { - "id": "gemini-2.5-flash-lite-preview-06-17", - "name": "gemini-2.5-flash-lite-preview-06-17", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -154490,15 +163699,14 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "video", - "image", - "audio" + "image" ], "output": [ "text" @@ -154506,26 +163714,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 200000, + "output": 100000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.09, - "output": 0.36 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "gemini-2.5-pro-preview-06-05": { - "id": "gemini-2.5-pro-preview-06-05", - "name": "gemini-2.5-pro-preview-06-05", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -154534,15 +163745,13 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -154550,26 +163759,31 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 200000 + "context": 200000, + "output": 100000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 1.125, - "output": 9 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "gpt-5-codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -154581,27 +163795,42 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 32768 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 1.125, - "output": 9 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "gpt-5.2-codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -154613,27 +163842,110 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 128000, + "input": 272000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai" + }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "gpt-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "openai/gpt-6-astra": { + "id": "openai/gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } + } + }, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02", - "last_updated": "2026-02", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -154645,349 +163957,503 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 128000, + "input": 272000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai" + }, "cost": { - "input": 1.125, - "output": 9 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "moonshotai/kimi-k2-instruct": { - "id": "moonshotai/kimi-k2-instruct", - "name": "Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 16384 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.57, - "output": 2.3 + "input": 0.075, + "output": 0.3, + "cache_read": 0.0375 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 262143 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.6, - "output": 3 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.6, - "output": 2.5 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 } }, - "minimaxai/minimax-m1-80k": { - "id": "minimaxai/minimax-m1-80k", - "name": "MiniMax M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 40000 + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.55, - "output": 2.2 + "input": 5, + "output": 30 } }, - "baidu/ernie-4.5-vl-424b-a47b": { - "id": "baidu/ernie-4.5-vl-424b-a47b", - "name": "ERNIE 4.5 VL 424B A47B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ernie", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 123000, - "output": 16000 + "context": 1047576, + "output": 32768 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.42, - "output": 1.25 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "baidu/ernie-4.5-300b-a47b-paddle": { - "id": "baidu/ernie-4.5-300b-a47b-paddle", - "name": "ERNIE 4.5 300B A47B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "ernie", - "attachment": false, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 123000, - "output": 12000 + "context": 1047576, + "output": 32768 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.28, - "output": 1.1 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "xiaomimimo/mimo-v2-flash": { - "id": "xiaomimimo/mimo-v2-flash", - "name": "XiaomiMiMo/MiMo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "zai-org/glm-4.7": { - "id": "zai-org/glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "output": 100000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "zai-org/glm-4.5v": { - "id": "zai-org/glm-4.5v", - "name": "GLM 4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glmv", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 128000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.6, - "output": 1.8 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "zai-org/glm-4.5": { - "id": "zai-org/glm-4.5", - "name": "GLM-4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 2, + "output": 10, + "cache_read": 0.25, + "cache_write": 3.125 } - }, - "zai-org/glm-4.7-flash": { - "id": "zai-org/glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", + } + } + }, + "clarifai": { + "id": "clarifai", + "env": [ + "CLARIFAI_PAT" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.clarifai.com/v2/ext/openai/v1", + "name": "Clarifai", + "doc": "https://docs.clarifai.com/compute/inference/", + "models": { + "minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput": { + "id": "minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput", + "name": "MiniMax-M2.5 High Throughput", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-25", "modalities": { "input": [ "text" @@ -154998,59 +164464,60 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.07, - "output": 0.4 + "input": 0.3, + "output": 1.2 } }, - "qwen/qwen3-235b-a22b-instruct-2507": { - "id": "qwen/qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "clarifai/main/models/mm-poly-8b": { + "id": "clarifai/main/models/mm-poly-8b", + "name": "MM Poly 8B", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "mm-poly", + "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-06", + "last_updated": "2026-02-25", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 32768, + "output": 4096 }, "cost": { - "input": 0.15, - "output": 0.8 + "input": 0.658, + "output": 1.11 } }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "deepseek-ai/deepseek-ocr/models/DeepSeek-OCR": { + "id": "deepseek-ai/deepseek-ocr/models/DeepSeek-OCR", + "name": "DeepSeek OCR", + "description": "OCR model for extracting structured text from documents and screenshots", + "family": "deepseek", + "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-10-20", + "last_updated": "2026-02-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -155058,30 +164525,30 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 0.2, + "output": 0.7 } }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "mistralai/completion/models/Ministral-3-3B-Reasoning-2512": { + "id": "mistralai/completion/models/Ministral-3-3B-Reasoning-2512", + "name": "Ministral 3 3B Reasoning 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-12", + "last_updated": "2026-02-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -155089,30 +164556,31 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 1.039, + "output": 0.54825 } }, - "qwen/qwen3-235b-a22b-thinking-2507": { - "id": "qwen/qwen3-235b-a22b-thinking-2507", - "name": "Qwen3 235B A22b Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "mistralai/completion/models/Ministral-3-14B-Reasoning-2512": { + "id": "mistralai/completion/models/Ministral-3-14B-Reasoning-2512", + "name": "Ministral 3 14B Reasoning 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-12", + "release_date": "2025-12-01", + "last_updated": "2025-12-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -155120,30 +164588,35 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 3 + "input": 2.5, + "output": 1.7 } }, - "qwen/qwen3-32b-fp8": { - "id": "qwen/qwen3-32b-fp8", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "moonshotai/chat-completion/models/Kimi-K2_6": { + "id": "moonshotai/chat-completion/models/Kimi-K2_6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -155151,27 +164624,27 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 20000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.1, - "output": 0.45 + "input": 0.95, + "output": 4 } }, - "qwen/qwen3-30b-a3b-fp8": { - "id": "qwen/qwen3-30b-a3b-fp8", - "name": "Qwen3 30B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "arcee_ai/AFM/models/trinity-mini": { + "id": "arcee_ai/AFM/models/trinity-mini", + "name": "Trinity Mini", + "description": "Reasoning-tuned 26B MoE model with 3B active parameters for agents, tools, and multi-step workloads", + "family": "trinity", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2024-10", + "release_date": "2025-12-01", + "last_updated": "2026-02-25", "modalities": { "input": [ "text" @@ -155182,26 +164655,26 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 20000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.09, - "output": 0.45 + "input": 0.045, + "output": 0.15 } }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "qwen/qwen3-coder-next", + "qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct": { + "id": "qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct", + "name": "Qwen3 Coder 30B A3B Instruct", "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02", - "last_updated": "2026-02", + "knowledge": "2025-04", + "release_date": "2025-07-31", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -155216,22 +164689,22 @@ "output": 65536 }, "cost": { - "input": 0.2, - "output": 1.5 + "input": 0.11458, + "output": 0.74812 } }, - "qwen/qwen3-coder-480b-a35b-instruct": { - "id": "qwen/qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507": { + "id": "qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen3 30B A3B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-07-30", + "last_updated": "2026-02-25", "modalities": { "input": [ "text" @@ -155243,87 +164716,26 @@ "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "output": 262144 }, "cost": { - "input": 0.29, - "output": 1.2 + "input": 0.3, + "output": 0.5 } }, - "qwen/qwen3-235b-a22b-fp8": { - "id": "qwen/qwen3-235b-a22b-fp8", - "name": "Qwen3 235B A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507": { + "id": "qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507", + "name": "Qwen3 30B A3B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "output": 20000 - }, - "cost": { - "input": 0.2, - "output": 0.8 - } - }, - "deepseek/deepseek-r1-0528": { - "id": "deepseek/deepseek-r1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 32768 - }, - "cost": { - "input": 0.7, - "output": 2.5 - } - }, - "deepseek/deepseek-v3-0324": { - "id": "deepseek/deepseek-v3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-07-31", + "last_updated": "2026-02-25", "modalities": { "input": [ "text" @@ -155334,33 +164746,35 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.28, - "output": 1.14 + "input": 0.36, + "output": 1.3 } }, - "deepseek/deepseek-v3.1": { - "id": "deepseek/deepseek-v3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "openai/chat-completion/models/gpt-oss-120b-high-throughput": { + "id": "openai/chat-completion/models/gpt-oss-120b-high-throughput", + "name": "GPT OSS 120B High Throughput", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 32767 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-08-05", + "last_updated": "2026-02-25", "modalities": { "input": [ "text" @@ -155371,33 +164785,35 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 32768 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.27, - "output": 1 + "input": 0.09, + "output": 0.36 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "Minimax M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "openai/chat-completion/models/gpt-oss-20b": { + "id": "openai/chat-completion/models/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 131071 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-08-05", + "last_updated": "2025-12-12", "modalities": { "input": [ "text" @@ -155408,84 +164824,93 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.045, + "output": 0.18 } } } }, - "nova": { - "id": "nova", + "aiand": { + "id": "aiand", "env": [ - "NOVA_API_KEY" + "AIAND_API_KEY" ], "npm": "@ai-sdk/openai-compatible", - "api": "https://api.nova.amazon.com/v1", - "name": "Nova", - "doc": "https://nova.amazon.com/dev/documentation", + "api": "https://api.aiand.com/v1", + "name": "ai&", + "doc": "https://docs.aiand.com/", "models": { - "nova-2-pro-v1": { - "id": "nova-2-pro-v1", - "name": "Nova 2 Pro", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova-pro", - "attachment": true, + "motif-technologies/motif-3": { + "id": "motif-technologies/motif-3", + "name": "Motif 3", + "description": "Motif 3 is a large-scale, decoder-only Mixture-of-Experts (MoE) language model with 314 billion total parameters and 13.2 billion parameters activated per token.", + "family": "motif", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2026-01-03", + "release_date": "2026-08-07", + "last_updated": "2026-08-07", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "reasoning": 0 + "input": 0.5, + "output": 2, + "cache_read": 0.2 } }, - "nova-2-lite-v1": { - "id": "nova-2-lite-v1", - "name": "Nova 2 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova-lite", + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -155497,43 +164922,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "reasoning": 0 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } - } - } - }, - "alibaba-token-plan": { - "id": "alibaba-token-plan", - "env": [ - "ALIBABA_TOKEN_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://token-plan.ap-southeast-1.maas.aliyuncs.com/compatible-mode/v1", - "name": "Alibaba Token Plan", - "doc": "https://www.alibabacloud.com/help/en/model-studio/token-plan-overview", - "models": { - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + }, + "deepseek-ai/deepseek-v4-pro": { + "id": "deepseek-ai/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "high", "max" ] @@ -155558,99 +164969,108 @@ }, "open_weights": true, "limit": { - "context": 1000000, + "context": 1048576, "output": 384000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1, + "output": 2.5, + "cache_read": 0.25 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "deepseek-ai/deepseek-v4-flash": { + "id": "deepseek-ai/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 + "type": "effort", + "values": [ + "none", + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.15, + "output": 0.25, + "cache_read": 0.08 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 3, + "output": 12.5, + "cache_read": 0.5 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", "name": "Kimi K2.7 Code", "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", "family": "kimi-k2", @@ -155658,17 +165078,18 @@ "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "high" + ] } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "temperature": false, + "structured_output": true, + "temperature": true, "knowledge": "2025-01", "release_date": "2026-06-12", "last_updated": "2026-06-12", @@ -155676,7 +165097,7 @@ "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -155688,22 +165109,26 @@ "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.75, + "output": 3.5, + "cache_read": 0.2 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "zai-org/glm-5.2": { + "id": "zai-org/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high", + "max" + ] } ], "tool_call": true, @@ -155712,8 +165137,8 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -155724,30 +165149,27 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1, + "output": 4, + "cache_read": 0.3 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "zai-org/glm-5.3": { + "id": "zai-org/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", "high", "max" ] @@ -155759,9 +165181,8 @@ }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -155772,56 +165193,30 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "wan2.7-image-pro": { - "id": "wan2.7-image-pro", - "name": "Wan2.7 Image Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-05-29", - "last_updated": "2026-05-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 0 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 4, + "cache_read": 0.3 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "qwen/qwen3.8-27b": { + "id": "qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", - "max" + "none", + "low", + "medium", + "xhigh" ] } ], @@ -155831,11 +165226,14 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" @@ -155843,87 +165241,89 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.4, + "output": 3, + "cache_read": 0.2 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", + "qwen/qwen3.6-27b": { + "id": "qwen/qwen3.6-27b", + "name": "Qwen3.6 27B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "video" + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 262144, "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.32, + "output": 3.2, + "cache_read": 0.2 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -155931,205 +165331,298 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 131072, "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.15, + "output": 0.6, + "cache_read": 0.08 } - }, - "qwen-image-2.0": { - "id": "qwen-image-2.0", - "name": "Qwen Image 2.0", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, + } + } + }, + "frogbot": { + "id": "frogbot", + "env": [ + "FROGBOT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://app.frogbot.ai/api/v1", + "name": "FrogBot", + "doc": "https://docs.frogbot.ai", + "models": { + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "Grok 4.1 Fast (Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2025-11", + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 8192, - "output": 0 + "context": 2000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, + "gpt-5-4-mini": { + "id": "gpt-5-4-mini", + "name": "GPT-5.4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "input": 196601, - "output": 24576 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "qwen-image-2.0-pro": { - "id": "qwen-image-2.0-pro", - "name": "Qwen Image 2.0 Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", - "attachment": false, + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "Grok 4.1 Fast (Non-Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2025-11", + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } + }, + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", "image" + ], + "output": [ + "text" ] }, "open_weights": false, "limit": { - "context": 8192, - "output": 0 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "wan2.7-image": { - "id": "wan2.7-image", - "name": "Wan2.7 Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": false, + "minimax-m2-5": { + "id": "minimax-m2-5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "knowledge": "2024-09", + "release_date": "2025-01-15", + "last_updated": "2025-02-22", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 8192, - "output": 0 + "context": 192000, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "gpt-5-4-nano": { + "id": "gpt-5-4-nano", + "name": "GPT-5.4 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -156137,34 +165630,37 @@ }, "open_weights": false, "limit": { - "context": 202752, - "output": 16384 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-03", - "last_updated": "2025-12-05", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ "text" @@ -156176,39 +165672,39 @@ "open_weights": true, "limit": { "context": 131072, - "output": 65536 + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.07, + "output": 0.2 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": false, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "budget_tokens", - "max": 81920 + "min": 128, + "max": 32768 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-05", "modalities": { "input": [ "text", "image", - "video" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -156216,81 +165712,69 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.31 } - } - } - }, - "alibaba": { - "id": "alibaba", - "env": [ - "DASHSCOPE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1", - "name": "Alibaba", - "doc": "https://www.alibabacloud.com/help/en/model-studio/models", - "models": { - "qwen3-omni-flash": { - "id": "qwen3-omni-flash", - "name": "Qwen3-Omni Flash", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, + }, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 0, + "max": 24576 } ], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "knowledge": "2025-01", + "release_date": "2025-07-17", + "last_updated": "2025-07-17", "modalities": { "input": [ "text", "image", "audio", - "video" + "video", + "pdf" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.43, - "output": 1.66, - "input_audio": 3.81, - "output_audio": 15.11 + "input": 0.3, + "output": 2.5, + "cache_read": 0.075 } }, - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", - "attachment": false, - "reasoning": false, + "minimax-m2-7": { + "id": "minimax-m2-7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": true, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "knowledge": "2024-09", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -156299,39 +165783,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 192000, + "output": 8192 }, "cost": { - "input": 1, - "output": 5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "qwen-plus": { - "id": "qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "kimi-k2-6": { + "id": "kimi-k2-6", + "name": "Kimi-K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -156339,27 +165817,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 1.2, - "reasoning": 4 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", + "grok-code-fast-1": { + "id": "grok-code-fast-1", + "name": "Grok 4.1 Fast (Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "knowledge": "2023-10", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ "text" @@ -156368,141 +165847,81 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.45, - "output": 2.25 + "input": 0.2, + "output": 1.5, + "cache_read": 0.02 } }, - "qwen3-omni-flash-realtime": { - "id": "qwen3-omni-flash-realtime", - "name": "Qwen3-Omni Flash Realtime", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "qwen-3-6-plus": { + "id": "qwen-3-6-plus", + "name": "Qwen 3.6 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "release_date": "2026-04-02", + "last_updated": "2026-04-03", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.52, - "output": 1.99, - "input_audio": 4.57, - "output_audio": 18.13 + "input": 0.5, + "output": 3, + "cache_read": 0.1 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] }, { - "type": "budget_tokens" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.7, - "output": 2.8, - "reasoning": 8.4 - } - }, - "qwen-omni-turbo-realtime": { - "id": "qwen-omni-turbo-realtime", - "name": "Qwen-Omni Turbo Realtime", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-05-08", - "last_updated": "2025-05-08", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", "image", - "audio" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 2048 - }, - "cost": { - "input": 0.27, - "output": 1.07, - "input_audio": 4.44, - "output_audio": 8.89 - } - }, - "qwen-plus-character-ja": { - "id": "qwen-plus-character-ja", - "name": "Qwen Plus Character (Japanese)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01", - "last_updated": "2024-01", - "modalities": { - "input": [ - "text" + "pdf" ], "output": [ "text" @@ -156510,69 +165929,86 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 512 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 1.4 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "grok-4-3": { + "id": "grok-4-3", + "name": "Grok 4.3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "knowledge": "2024-11", + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 2 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": false, + "gpt-5-5": { + "id": "gpt-5-5", + "name": "GPT-5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-04", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -156580,97 +166016,91 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 272000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", "video", - "audio" + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.248, - "output": 1.485 + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "gemini-3-1-pro-preview": { + "id": "gemini-3-1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2026-01", + "release_date": "2026-02-18", + "last_updated": "2026-02-18", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -156679,27 +166109,26 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 65536 + "output": 64000 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek v4 Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2026-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -156710,135 +166139,127 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 1.2, - "output": 6 - } - }, - "qwen2-5-omni-7b": { - "id": "qwen2-5-omni-7b", - "name": "Qwen2.5-Omni 7B", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-12", - "last_updated": "2024-12", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 2048 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.4, - "input_audio": 6.76 + "input": 1.74, + "output": 3.48, + "cache_read": 0.14 } }, - "qwen3-8b": { - "id": "qwen3-8b", - "name": "Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] }, { - "type": "budget_tokens" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.18, - "output": 0.7, - "reasoning": 2.1 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "qwen2-5-14b-instruct": { - "id": "qwen2-5-14b-instruct", - "name": "Qwen2.5 14B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gpt-5-3-codex": { + "id": "gpt-5-3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.35, - "output": 1.4 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "qwen3-next-80b-a3b-thinking": { - "id": "qwen3-next-80b-a3b-thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ "text" @@ -156853,67 +166274,37 @@ "output": 32768 }, "cost": { - "input": 0.5, - "output": 6 - } - }, - "qvq-max": { - "id": "qvq-max", - "name": "QVQ Max", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qvq", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-03-25", - "last_updated": "2025-03-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 1.2, - "output": 4.8 + "input": 0.15, + "output": 0.6 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -156921,31 +166312,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.1875, - "output": 1.125, - "cache_write": 0.234375 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen2-5-vl-72b-instruct": { - "id": "qwen2-5-vl-72b-instruct", - "name": "Qwen2.5-VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "zai-glm-5-1": { + "id": "zai-glm-5-1", + "name": "Z.AI GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "knowledge": "2024-10", + "release_date": "2025-01-20", + "last_updated": "2025-02-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -156953,34 +166344,39 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 198000, "output": 8192 }, "cost": { - "input": 2.8, - "output": 8.4 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } - }, - "qwen3-vl-plus": { - "id": "qwen3-vl-plus", - "name": "Qwen3-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, + } + } + }, + "stackit": { + "id": "stackit", + "env": [ + "STACKIT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.openai-compat.model-serving.eu01.onstackit.cloud/v1", + "name": "STACKIT", + "doc": "https://docs.stackit.cloud/products/data-and-ai/ai-model-serving/basics/available-shared-models", + "models": { + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma 3 27B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2025-05-17", + "last_updated": "2025-05-17", "modalities": { "input": [ "text", @@ -156990,197 +166386,181 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 37000, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 1.6, - "reasoning": 4.8 + "input": 0.53, + "output": 0.76 } }, - "qwen-vl-ocr": { - "id": "qwen-vl-ocr", - "name": "Qwen-VL OCR", - "description": "OCR model for extracting structured text from documents and screenshots", - "family": "qwen", + "intfloat/e5-mistral-7b-instruct": { + "id": "intfloat/e5-mistral-7b-instruct", + "name": "E5 Mistral 7B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "mistral", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-10-28", - "last_updated": "2025-04-13", + "structured_output": false, + "temperature": false, + "release_date": "2023-12-11", + "last_updated": "2023-12-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 34096, + "context": 4096, "output": 4096 }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 0.02, + "output": 0.02 } }, - "qwen-mt-turbo": { - "id": "qwen-mt-turbo", - "name": "Qwen-MT Turbo", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "Qwen/Qwen3-VL-235B-A22B-Instruct-FP8": { + "id": "Qwen/Qwen3-VL-235B-A22B-Instruct-FP8", + "name": "Qwen3-VL 235B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01", - "last_updated": "2025-01", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "output": 8192 + "context": 218000, + "output": 16384 }, "cost": { - "input": 0.16, - "output": 0.49 + "input": 1.76, + "output": 2.05 } }, - "qwen-mt-plus": { - "id": "qwen-mt-plus", - "name": "Qwen-MT Plus", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "Qwen/Qwen3-VL-Embedding-8B": { + "id": "Qwen/Qwen3-VL-Embedding-8B", + "name": "Qwen3-VL Embedding 8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01", - "last_updated": "2025-01", + "structured_output": false, + "temperature": false, + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "output": 8192 + "context": 32000, + "output": 4096 }, "cost": { - "input": 2.46, - "output": 7.37 + "input": 0.09, + "output": 0.09 } }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.4, - "output": 2.4, - "reasoning": 2.4 + "input": 0.53, + "output": 0.76 } }, - "qwen-omni-turbo": { - "id": "qwen-omni-turbo", - "name": "Qwen-Omni Turbo", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", + "cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic": { + "id": "cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic", + "name": "Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01-19", - "last_updated": "2025-03-26", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 2048 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.07, - "output": 0.27, - "input_audio": 4.44, - "output_audio": 8.89 + "input": 0.53, + "output": 0.76 } }, - "qwen2-5-72b-instruct": { - "id": "qwen2-5-72b-instruct", - "name": "Qwen2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -157195,30 +166575,32 @@ "output": 8192 }, "cost": { - "input": 1.4, - "output": 5.6 + "input": 0.18, + "output": 0.29 } }, - "qwen-flash": { - "id": "qwen-flash", - "name": "Qwen Flash", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -157227,106 +166609,63 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 131000, + "output": 8192 }, "cost": { - "input": 0.05, - "output": 0.4 + "input": 0.53, + "output": 0.76 } - }, - "qwen3-vl-235b-a22b": { - "id": "qwen3-vl-235b-a22b", - "name": "Qwen3-VL 235B-A22B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + } + } + }, + "anyapi": { + "id": "anyapi", + "env": [ + "ANYAPI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.anyapi.ai/v1", + "name": "AnyAPI", + "doc": "https://docs.anyapi.ai", + "models": { + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.7, - "output": 2.8, - "reasoning": 8.4 - } - }, - "qwen3-vl-30b-a3b": { - "id": "qwen3-vl-30b-a3b", - "name": "Qwen3-VL 30B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.2, - "output": 0.8, - "reasoning": 2.4 - } - }, - "qwen-vl-max": { - "id": "qwen-vl-max", - "name": "Qwen-VL Max", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-08", - "last_updated": "2025-08-13", - "modalities": { - "input": [ - "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -157334,19 +166673,15 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.8, - "output": 3.2 + "context": 200000, + "output": 64000 } }, - "qwen3.5-27b": { - "id": "qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-sonnet-4-5": { + "id": "anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -157354,50 +166689,29 @@ "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.4 - } - }, - "qwen-max": { - "id": "qwen-max", - "name": "Qwen Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", - "modalities": { - "input": [ - "text" + "pdf" ], "output": [ "text" @@ -157405,94 +166719,81 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 1.6, - "output": 6.4 + "context": 200000, + "output": 64000 } }, - "qwen3-235b-a22b": { - "id": "qwen3-235b-a22b", - "name": "Qwen3 235B-A22B", - "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", - "family": "qwen", - "attachment": false, + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.7, - "output": 2.8, - "reasoning": 8.4 - } - }, - "qwen3-livetranslate-flash-realtime": { - "id": "qwen3-livetranslate-flash-realtime", - "name": "Qwen3-LiveTranslate Flash Realtime", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 53248, - "output": 4096 + "context": 1000000, + "output": 128000 }, - "cost": { - "input": 10, - "output": 10, - "input_audio": 10, - "output_audio": 38 + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } } }, - "qwen3.6-27b": { - "id": "qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -157500,40 +166801,45 @@ "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.6, - "output": 3.6 + "context": 1000000, + "output": 64000 } }, - "qwen3.5-35b-a3b": { - "id": "qwen3.5-35b-a3b", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -157541,47 +166847,67 @@ "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 128000 }, - "cost": { - "input": 0.25, - "output": 2 + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } } }, - "qwen3-coder-480b-a35b-instruct": { - "id": "qwen3-coder-480b-a35b-instruct", - "name": "Qwen3-Coder 480B-A35B Instruct", - "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", - "family": "qwen", + "cohere/command-r-plus-08-2024": { + "id": "cohere/command-r-plus-08-2024", + "name": "Command R+", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -157592,27 +166918,22 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 7.5 + "context": 128000, + "output": 4000 } }, - "qwq-plus": { - "id": "qwq-plus", - "name": "QwQ Plus", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "deepseek/deepseek-chat": { + "id": "deepseek/deepseek-chat", + "name": "DeepSeek Chat", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": true, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-03-05", - "last_updated": "2025-03-05", + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ "text" @@ -157621,28 +166942,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.8, - "output": 2.4 + "context": 1000000, + "output": 384000 } }, - "qwen2-5-32b-instruct": { - "id": "qwen2-5-32b-instruct", - "name": "Qwen2.5 32B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "deepseek/deepseek-r1": { + "id": "deepseek/deepseek-r1", + "name": "DeepSeek Reasoner", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ "text" @@ -157653,40 +166974,30 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.7, - "output": 2.8 + "context": 1000000, + "output": 384000 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -157694,26 +167005,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.6, - "output": 3.6 + "context": 1000000, + "output": 384000 } }, - "qwen3-coder-flash": { - "id": "qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -157722,70 +167034,56 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 1.5 + "output": 384000 } }, - "qwen3-14b": { - "id": "qwen3-14b", - "name": "Qwen3 14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, + "perplexity/sonar-pro": { + "id": "perplexity/sonar-pro", + "name": "Sonar Pro", + "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", + "family": "sonar-pro", + "attachment": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 200000, "output": 8192 - }, - "cost": { - "input": 0.35, - "output": 1.4, - "reasoning": 4.2 } }, - "qwen3-asr-flash": { - "id": "qwen3-asr-flash", - "name": "Qwen3-ASR Flash", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, + "perplexity/sonar-reasoning-pro": { + "id": "perplexity/sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", + "family": "sonar-reasoning", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "temperature": false, - "knowledge": "2024-04", - "release_date": "2025-09-08", - "last_updated": "2025-09-08", + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "audio" + "text", + "image" ], "output": [ "text" @@ -157793,37 +167091,31 @@ }, "open_weights": false, "limit": { - "context": 53248, + "context": 128000, "output": 4096 - }, - "cost": { - "input": 0.035, - "output": 0.035 } }, - "qwen-turbo": { - "id": "qwen-turbo", - "name": "Qwen Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, + "google/gemini-3-pro-preview": { + "id": "google/gemini-3-pro-preview", + "name": "Gemini 3 Pro Preview", + "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", + "family": "gemini-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11-01", - "last_updated": "2025-04-28", + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -157831,99 +167123,95 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 16384 - }, - "cost": { - "input": 0.05, - "output": 0.2, - "reasoning": 0.5 + "context": 1048576, + "output": 65536 } }, - "qwen2-5-7b-instruct": { - "id": "qwen2-5-7b-instruct", - "name": "Qwen2.5 7B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.175, - "output": 0.7 + "context": 1048576, + "output": 65536 } }, - "qwen2-5-vl-7b-instruct": { - "id": "qwen2-5-vl-7b-instruct", - "name": "Qwen2.5-VL 7B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.35, - "output": 1.05 + "context": 1048576, + "output": 65536 } }, - "qwen3.6-max-preview": { - "id": "qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -157931,73 +167219,60 @@ }, "open_weights": false, "limit": { - "context": 262144, + "context": 1048576, "output": 65536 - }, - "cost": { - "input": 1.3, - "output": 7.8, - "cache_read": 0.13, - "cache_write": 1.625 } }, - "qwen3.5-122b-a10b": { - "id": "qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", + "audio", "video", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1048576, "output": 65536 - }, - "cost": { - "input": 0.4, - "output": 3.2 } }, - "qwen-vl-plus": { - "id": "qwen-vl-plus", - "name": "Qwen-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "xai/grok-4.3": { + "id": "xai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-08-15", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -158005,149 +167280,66 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.21, - "output": 0.63 + "context": 1000000, + "output": 30000 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", + "mistralai/devstral-2512": { + "id": "mistralai/devstral-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "output": 262144 }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } - } - } - } - }, - "databricks": { - "id": "databricks", - "env": [ - "DATABRICKS_HOST", - "DATABRICKS_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://${DATABRICKS_HOST}/ai-gateway/mlflow/v1", - "name": "Databricks", - "doc": "https://docs.databricks.com/aws/en/machine-learning/foundation-models/", - "models": { - "databricks-claude-opus-4-7": { - "id": "databricks-claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "status": "deprecated" + }, + "mistralai/mistral-large-2512": { + "id": "mistralai/mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "knowledge": "2024-11", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "context": 262144, + "output": 262144 } }, - "databricks-gpt-5-4": { - "id": "databricks-gpt-5-4", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", "name": "GPT-5.4", "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", "family": "gpt", @@ -158165,7 +167357,7 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-08-31", "release_date": "2026-03-05", "last_updated": "2026-03-05", @@ -158200,34 +167392,13 @@ } } } - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } } }, - "databricks-gemini-3-flash": { - "id": "databricks-gemini-3-flash", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -158242,17 +167413,14 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -158260,39 +167428,38 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 + "context": 400000, + "input": 272000, + "output": 128000 } }, - "databricks-claude-opus-4-5": { - "id": "databricks-claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", + "modalities": { + "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -158301,27 +167468,20 @@ "open_weights": false, "limit": { "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "output": 100000 } }, - "databricks-gpt-5-nano": { - "id": "databricks-gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -158331,13 +167491,12 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -158345,28 +167504,21 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "context": 200000, + "output": 100000 } }, - "databricks-gpt-5-mini": { - "id": "databricks-gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -158375,10 +167527,10 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -158393,17 +167545,12 @@ "context": 400000, "input": 272000, "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 } }, - "databricks-gpt-5": { - "id": "databricks-gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", "family": "gpt", "attachment": true, "reasoning": true, @@ -158411,7 +167558,6 @@ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -158420,10 +167566,10 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -158438,39 +167584,25 @@ "context": 400000, "input": 272000, "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 } }, - "databricks-gemini-2-5-pro": { - "id": "databricks-gemini-2-5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -158479,36 +167611,44 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } + "context": 1047576, + "output": 32768 + } + }, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 } }, - "databricks-gemini-3-1-pro": { - "id": "databricks-gemini-3-1-pro", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -158523,16 +167663,14 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -158541,58 +167679,37 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "context": 200000, + "output": 100000 } }, - "databricks-gemini-2-5-flash": { - "id": "databricks-gemini-2-5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 0, - "max": 24576 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -158600,114 +167717,141 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 + } + } + } + }, + "crusoe": { + "id": "crusoe", + "env": [ + "CRUSOE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.inference.crusoecloud.com/v1", + "name": "Crusoe", + "doc": "https://docs.crusoecloud.com/managed-inference/overview", + "models": { + "zai/GLM-5.1": { + "id": "zai/GLM-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 + "input": 1.2, + "output": 4.4, + "cache_read": 0.25 } }, - "databricks-claude-sonnet-4": { - "id": "databricks-claude-sonnet-4", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "zai/GLM-5.2": { + "id": "zai/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "databricks-claude-haiku-4-5": { - "id": "databricks-claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.25, + "output": 0.75, + "cache_read": 0.13 } }, - "databricks-gpt-5-4-nano": { - "id": "databricks-gpt-5-4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -158717,209 +167861,233 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.14, + "output": 0.4, + "cache_read": 0.14 } }, - "databricks-claude-opus-4-6": { - "id": "databricks-claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B-A22B Instruct 2507", + "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-07-21", + "last_updated": "2025-07-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 16384 + }, + "cost": { + "input": 0.22, + "output": 0.8, + "cache_read": 0.11 + } + }, + "deepseek-ai/DeepSeek-V3-0324": { + "id": "deepseek-ai/DeepSeek-V3-0324", + "name": "DeepSeek V3 0324", + "description": "March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-03-24", + "last_updated": "2025-03-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 163840 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.25 + } + }, + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.7, + "output": 3.5, + "cache_read": 0.35 } }, - "databricks-gpt-5-1": { - "id": "databricks-gpt-5-1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B": { + "id": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.05, + "output": 0.2, + "cache_read": 0.03 } }, - "databricks-claude-sonnet-4-5": { - "id": "databricks-claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "nvidia/Nemotron-3-Nano-Omni-Reasoning-30B-A3B": { + "id": "nvidia/Nemotron-3-Nano-Omni-Reasoning-30B-A3B", + "name": "Nemotron 3 Nano Omni 30B A3B Reasoning", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, + "interleaved": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 256000, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, + "input": 0.3, + "output": 1.83, "cache_read": 0.3, - "cache_write": 3.75 + "input_audio": 0.5 } }, - "databricks-claude-opus-4-1": { - "id": "databricks-claude-opus-4-1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B": { + "id": "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B", + "name": "Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.3, + "output": 2.4, + "cache_read": 0.15 } }, - "databricks-gpt-5-5": { - "id": "databricks-gpt-5-5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -158932,95 +168100,72 @@ } ], "tool_call": true, + "interleaved": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.05, + "output": 0.2, + "cache_read": 0.05 } - }, - "databricks-gemini-3-1-flash-lite": { - "id": "databricks-gemini-3-1-flash-lite", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + } + } + }, + "volcengine-coding-plan": { + "id": "volcengine-coding-plan", + "env": [ + "ARK_CODING_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://ark.cn-beijing.volces.com/api/coding/v3", + "name": "Volcengine Ark Coding Plan", + "doc": "https://www.volcengine.com/docs/82379/1928261", + "models": { + "doubao-seed-2.0-lite": { + "id": "doubao-seed-2.0-lite", + "name": "Seed 2.0 Lite", + "description": "Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" @@ -159028,21 +168173,20 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "databricks-gemini-3-pro": { - "id": "databricks-gemini-3-pro", - "name": "Gemini 3 Pro Preview", - "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", - "family": "gemini-pro", + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -159050,161 +168194,160 @@ "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", "image", "video", - "audio", "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0, + "output": 0, + "cache_read": 0 } }, - "databricks-claude-sonnet-4-6": { - "id": "databricks-claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "databricks-gpt-oss-20b": { - "id": "databricks-gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "doubao-seed-2.1-turbo": { + "id": "doubao-seed-2.1-turbo", + "name": "Seed 2.1 Turbo", + "description": "Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows", + "family": "seed", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-06-23", + "last_updated": "2026-06-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.05, - "output": 0.2 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "databricks-gpt-oss-120b": { - "id": "databricks-gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -159212,19 +168355,20 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 512000 }, "cost": { - "input": 0.072, - "output": 0.28 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "databricks-gpt-5-2": { - "id": "databricks-gpt-5-2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "doubao-seed-evolving": { + "id": "doubao-seed-evolving", + "name": "Seed Evolving", + "description": "Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -159233,21 +168377,23 @@ "values": [ "none", "low", - "medium", "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-06-23", + "last_updated": "2026-06-23", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -159255,103 +168401,69 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "databricks-gpt-5-4-mini": { - "id": "databricks-gpt-5-4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0, + "output": 0, + "cache_read": 0 } - } - } - }, - "crof": { - "id": "crof", - "env": [ - "CROF_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://crof.ai/v1", - "name": "CrofAI", - "doc": "https://crof.ai/docs", - "models": { - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + }, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "medium", - "high" + "high", + "max" ] } ], @@ -159361,9 +168473,8 @@ }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -159377,98 +168488,34 @@ "context": 1000000, "output": 131072 }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, "cost": { - "input": 0.12, - "output": 0.21, - "cache_read": 0.003 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 0.11, - "output": 0.95, - "cache_read": 0.02, - "cache_write": 0.375 - } - }, - "greg-2-ultra": { - "id": "greg-2-ultra", - "name": "Greg 2 Ultra", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-06-14", - "last_updated": "2026-06-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 229376, - "output": 229376 - }, - "cost": { - "input": 3, - "output": 10, - "cache_read": 0.5 - } - }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -159476,35 +168523,25 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 202752 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.25, - "output": 1.1, - "cache_read": 0.05, - "cache_write": 0 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "deepseek-v4-pro-lightning": { - "id": "deepseek-v4-pro-lightning", - "name": "DeepSeek V4 Pro Lightning", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, @@ -159527,30 +168564,42 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "output": 384000 }, "cost": { - "input": 0.8, - "output": 1.6, - "cache_read": 0.02 + "input": 0, + "output": 0, + "cache_read": 0 } - }, - "greg-rp": { - "id": "greg-rp", - "name": "Greg (Roleplay)", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, + } + } + }, + "jiekou": { + "id": "jiekou", + "env": [ + "JIEKOU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.jiekou.ai/openai", + "name": "Jiekou.AI", + "doc": "https://docs.jiekou.ai/docs/support/quickstart?utm_source=github_models.dev", + "models": { + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "grok-4-1-fast-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -159558,38 +168607,26 @@ }, "open_weights": false, "limit": { - "context": 229376, - "output": 229376 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 0.18, + "output": 0.45 } }, - "gemma-4-31b-it": { - "id": "gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "grok-4-1-fast-non-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -159599,230 +168636,161 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 0.18, + "output": 0.45 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "gpt-5.2-codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.55, - "output": 2.25, - "cache_read": 0.05 + "input": 1.75, + "output": 14 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "gpt-5.1-codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 202752 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.45, - "output": 2.15, - "cache_read": 0.08, - "cache_write": 0 + "input": 1.125, + "output": 9 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "gemini-3-pro-preview": { + "id": "gemini-3-pro-preview", + "name": "gemini-3-pro-preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.35, - "output": 0.8, - "cache_read": 0.003 + "input": 1.8, + "output": 10.8 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "gpt-5-codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 2.2, - "cache_read": 0.08 + "input": 1.125, + "output": 9 } }, - "greg-2-super": { - "id": "greg-2-super", - "name": "Greg 2 Super", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "gpt-5-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-14", - "last_updated": "2026-06-14", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -159830,265 +168798,175 @@ }, "open_weights": false, "limit": { - "context": 229376, - "output": 229376 + "context": 400000, + "output": 128000 }, "cost": { - "input": 1.5, - "output": 5, - "cache_read": 0.25 + "input": 0.225, + "output": 1.8 } }, - "kimi-k2.5-lightning": { - "id": "kimi-k2.5-lightning", - "name": "Kimi K2.5 (Lightning)", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": false, + "gpt-5.2-pro": { + "id": "gpt-5.2-pro", + "name": "gpt-5.2-pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2 + "input": 18.9, + "output": 151.2 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", + "temperature": true, "release_date": "2026-01", "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.35, - "output": 1.7, - "cache_read": 0.07 + "input": 1.1, + "output": 4.4 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 1.99, - "cache_read": 0.05 + "input": 1.1, + "output": 4.4 } }, - "qwen3.6-27b": { - "id": "qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "grok-4-fast-reasoning": { + "id": "grok-4-fast-reasoning", + "name": "grok-4-fast-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.2, - "output": 1.5, - "cache_read": 0.04 + "input": 0.18, + "output": 0.45 } }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "gpt-5-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-03-13", - "last_updated": "2026-03-13", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.04, - "output": 0.15, - "cache_read": 0.008 + "input": 0.045, + "output": 0.36 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "gemini-2.5-flash-lite-preview-06-17": { + "id": "gemini-2.5-flash-lite-preview-06-17", + "name": "gemini-2.5-flash-lite-preview-06-17", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -160105,185 +168983,139 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", "video", + "image", "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 1048576, + "output": 65535 }, "cost": { - "input": 0.35, - "output": 1.75, - "cache_read": 0.07 + "input": 0.09, + "output": 0.36 } }, - "glm-4.7-flash": { - "id": "glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "claude-opus-4-20250514": { + "id": "claude-opus-4-20250514", + "name": "claude-opus-4-20250514", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.04, - "output": 0.3, - "cache_read": 0.008, - "cache_write": 0 + "input": 13.5, + "output": 67.5 } }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "grok-4-fast-non-reasoning": { + "id": "grok-4-fast-non-reasoning", + "name": "grok-4-fast-non-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.4, - "output": 0.8, - "cache_read": 0.003, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 0.18, + "output": 0.45 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "claude-opus-4-1-20250805", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 202752 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.48, - "output": 1.9, - "cache_read": 0.1, - "cache_write": 0 + "input": 13.5, + "output": 67.5 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "gpt-5-pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -160291,28 +169123,30 @@ }, "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 400000, + "output": 272000 }, "cost": { - "input": 0.18, - "output": 0.35, - "cache_read": 0.04 + "input": 13.5, + "output": 108 } }, - "greg-1-mini": { - "id": "greg-1-mini", - "name": "Greg 1 Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, + "gpt-5-chat-latest": { + "id": "gpt-5-chat-latest", + "name": "gpt-5-chat-latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -160320,212 +169154,150 @@ }, "open_weights": false, "limit": { - "context": 229376, - "output": 229376 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.07, - "output": 0.15, - "cache_read": 0.01 + "input": 1.125, + "output": 9 } - } - } - }, - "fastrouter": { - "id": "fastrouter", - "env": [ - "FASTROUTER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://go.fastrouter.ai/api/v1", - "name": "FastRouter", - "doc": "https://fastrouter.ai/models", - "models": { - "wanx/wan-v2-6": { - "id": "wanx/wan-v2-6", - "name": "Wan 2.6", - "description": "Video model for prompt-guided generation, editing, and motion workflows", + }, + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "claude-opus-4-5-20251101", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image" ], "output": [ - "video" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 400000, - "output": 0 + "context": 200000, + "output": 65536 + }, + "cost": { + "input": 4.5, + "output": 22.5 } }, - "moonshotai/kimi-k2": { - "id": "moonshotai/kimi-k2", - "name": "Kimi K2", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "gpt-5.1": { + "id": "gpt-5.1", + "name": "gpt-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-07-11", - "last_updated": "2025-07-11", + "release_date": "2026-02", + "last_updated": "2026-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.55, - "output": 2.2 + "input": 1.125, + "output": 9 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "gemini-2.5-pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.75, - "output": 3.5 - } - }, - "google/imagen-4.0-fast": { - "id": "google/imagen-4.0-fast", - "name": "Imagen 4 Fast", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 480, - "output": 0 - } - }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, + "open_weights": false, "limit": { "context": 1048576, - "output": 65536 + "output": 65535 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.31 + "input": 1.125, + "output": 9 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "gemini-2.5-flash", "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -160534,17 +169306,16 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65536 + "output": 65535 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.0375 + "input": 0.27, + "output": 2.25 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", + "gemini-2.5-flash-preview-05-20": { + "id": "gemini-2.5-flash-preview-05-20", + "name": "gemini-2.5-flash-preview-05-20", "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "family": "gemini-flash", "attachment": true, @@ -160553,6 +169324,7 @@ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -160562,16 +169334,14 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", "video", - "audio", - "pdf" + "audio" ], "output": [ "text" @@ -160580,52 +169350,57 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65536 + "output": 200000 }, "cost": { - "input": 1.5, - "output": 9 + "input": 0.135, + "output": 3.15 } }, - "google/veo3.1-lite": { - "id": "google/veo3.1-lite", - "name": "Veo 3.1 Lite", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", + "gpt-5.1-codex-max": { + "id": "gpt-5.1-codex-max", + "name": "gpt-5.1-codex-max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { "context": 400000, - "output": 0 + "output": 128000 + }, + "cost": { + "input": 1.125, + "output": 9 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "gpt-5.2": { + "id": "gpt-5.2", + "name": "gpt-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -160635,105 +169410,83 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.13, - "output": 0.38 + "input": 1.575, + "output": 12.6 } }, - "google/veo3.1": { - "id": "google/veo3.1", - "name": "Veo 3.1", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", + "grok-code-fast-1": { + "id": "grok-code-fast-1", + "name": "grok-code-fast-1", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image" ], "output": [ - "video" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 0 - } - }, - "google/imagen-4.0-ultra": { - "id": "google/imagen-4.0-ultra", - "name": "Imagen 4 Ultra", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", - "modalities": { - "input": [ "text" - ], - "output": [ - "image" ] }, "open_weights": false, "limit": { - "context": 480, - "output": 0 + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.18, + "output": 1.35 } }, - "google/gemini-3-pro-image-preview": { - "id": "google/gemini-3-pro-image-preview", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "claude-sonnet-4-5-20250929", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 2, - "output": 12 + "input": 2.7, + "output": 13.5 } }, - "google/gemini-3.1-flash-image-preview": { - "id": "google/gemini-3.1-flash-image-preview", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "claude-opus-4-6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -160742,66 +169495,59 @@ "values": [ "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 } ], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "knowledge": "2025-05-31", + "release_date": "2026-02", + "last_updated": "2026-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 3 + "input": 5, + "output": 25 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "gemini-3-flash-preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", "video", - "audio", - "pdf" + "audio" ], "output": [ "text" @@ -160813,47 +169559,53 @@ "output": 65536 }, "cost": { - "input": 2, - "output": 12 + "input": 0.5, + "output": 3 } }, - "google/veo3.1-fast": { - "id": "google/veo3.1-fast", - "name": "Veo 3.1 Fast", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "claude-haiku-4-5-20251001", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 0 + "context": 20000, + "output": 64000 + }, + "cost": { + "input": 0.9, + "output": 4.5 } }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "gemini-2.5-flash-lite-preview-09-2025": { + "id": "gemini-2.5-flash-lite-preview-09-2025", + "name": "gemini-2.5-flash-lite-preview-09-2025", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -160863,44 +169615,14 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 30000 - }, - "cost": { - "input": 1.25, - "output": 2.5 - } - }, - "x-ai/grok-4": { - "id": "x-ai/grok-4", - "name": "Grok 4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-09", - "last_updated": "2025-07-09", - "modalities": { - "input": [ - "text" + "video", + "audio" ], "output": [ "text" @@ -160908,27 +169630,26 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75, - "cache_write": 15 + "input": 0.09, + "output": 0.36 } }, - "x-ai/grok-build-0.1": { - "id": "x-ai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", + "gemini-2.5-pro-preview-06-05": { + "id": "gemini-2.5-pro-preview-06-05", + "name": "gemini-2.5-pro-preview-06-05", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -160938,13 +169659,14 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -160952,99 +169674,102 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1048576, + "output": 200000 }, "cost": { - "input": 1, - "output": 2 + "input": 1.125, + "output": 9 } }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "gemini-2.5-flash-lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 1.05, - "output": 3.5 + "input": 0.09, + "output": 0.36 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "claude-sonnet-4-20250514": { + "id": "claude-sonnet-4-20250514", + "name": "claude-sonnet-4-20250514", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.95, - "output": 3.15 + "input": 2.7, + "output": 13.5 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "gpt-5.1-codex-mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10-01", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -161060,34 +169785,21 @@ "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.225, + "output": 1.8 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "o3": { + "id": "o3", + "name": "o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -161099,43 +169811,30 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 1.25 + "input": 10, + "output": 40 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "grok-4-0709": { + "id": "grok-4-0709", + "name": "grok-4-0709", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -161143,68 +169842,63 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "output": 8192 }, "cost": { - "input": 1.75, - "output": 14 + "input": 2.7, + "output": 13.5 } }, - "openai/gpt-realtime-1.5": { - "id": "openai/gpt-realtime-1.5", - "name": "GPT Realtime 1.5", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", - "attachment": true, + "deepseek/deepseek-v3-0324": { + "id": "deepseek/deepseek-v3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-06-01", - "last_updated": "2025-06-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "audio", - "image" + "text" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "output": 4096 + "context": 163840, + "output": 163840 }, "cost": { - "input": 4, - "output": 16 + "input": 0.28, + "output": 1.14 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "deepseek/deepseek-v3.1": { + "id": "deepseek/deepseek-v3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024, + "max": 32767 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -161215,111 +169909,89 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 163840, "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.27, + "output": 1 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, + "deepseek/deepseek-r1-0528": { + "id": "deepseek/deepseek-r1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 163840, + "output": 32768 }, "cost": { - "input": 0.75, - "output": 4.5 + "input": 0.7, + "output": 2.5 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, + "xiaomimimo/mimo-v2-flash": { + "id": "xiaomimimo/mimo-v2-flash", + "name": "XiaomiMiMo/MiMo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "output": 131072 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0, + "output": 0 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "baidu/ernie-4.5-vl-424b-a47b": { + "id": "baidu/ernie-4.5-vl-424b-a47b", + "name": "ERNIE 4.5 VL 424B A47B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ernie", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-10-01", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -161329,124 +170001,128 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 123000, + "output": 16000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.42, + "output": 1.25 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, + "baidu/ernie-4.5-300b-a47b-paddle": { + "id": "baidu/ernie-4.5-300b-a47b-paddle", + "name": "ERNIE 4.5 300B A47B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "ernie", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 123000, + "output": 12000 + }, + "cost": { + "input": 0.28, + "output": 1.1 + } + }, + "minimaxai/minimax-m1-80k": { + "id": "minimaxai/minimax-m1-80k", + "name": "MiniMax M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-10-01", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 40000 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 0.55, + "output": 2.2 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024, + "max": 262143 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 30, - "output": 180 + "input": 0.6, + "output": 3 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "moonshotai/kimi-k2-instruct": { + "id": "moonshotai/kimi-k2-instruct", + "name": "Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -161458,281 +170134,253 @@ "open_weights": true, "limit": { "context": 131072, - "output": 65536 + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.2 + "input": 0.57, + "output": 2.3 } }, - "openai/gpt-image-2": { - "id": "openai/gpt-image-2", - "name": "GPT Image 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 0 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 2.5 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "zai-org/glm-4.5": { + "id": "zai-org/glm-4.5", + "name": "GLM-4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 98304 }, "cost": { - "input": 5, - "output": 30 + "input": 0.6, + "output": 2.2 } }, - "bytedance/seedance-2": { - "id": "bytedance/seedance-2", - "name": "Seedance 2", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", + "zai-org/glm-4.5v": { + "id": "zai-org/glm-4.5v", + "name": "GLM 4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glmv", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4096, - "output": 0 + "context": 65536, + "output": 16384 + }, + "cost": { + "input": 0.6, + "output": 1.8 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "zai-org/glm-4.7-flash": { + "id": "zai-org/glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 32000 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 200000, - "output": 64000 + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.07, + "output": 0.4 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "zai-org/glm-4.7": { + "id": "zai-org/glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 32000 - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 5, - "output": 25 + "input": 0.6, + "output": 2.2 } }, - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "Minimax M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "budget_tokens", "min": 1024, - "max": 32000 + "max": 131071 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.3, + "output": 1.2 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 32000 - } - ], + "qwen/qwen3-235b-a22b-instruct-2507": { + "id": "qwen/qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 3, - "output": 15 + "input": 0.15, + "output": 0.8 } }, - "sarvam/sarvam-105b": { - "id": "sarvam/sarvam-105b", - "name": "Sarvam 105B", - "description": "Flagship Indian-language reasoning model for enterprise multilingual applications", - "family": "sarvam", + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-09-01", - "last_updated": "2025-09-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -161743,26 +170391,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0.04, - "output": 0.16 + "input": 0.15, + "output": 1.5 } }, - "sarvam/sarvam-30b": { - "id": "sarvam/sarvam-30b", - "name": "Sarvam 30B", - "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", - "family": "sarvam", + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "qwen/qwen3-coder-next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-18", - "last_updated": "2026-02-18", + "release_date": "2026-02", + "last_updated": "2026-02", "modalities": { "input": [ "text" @@ -161773,27 +170421,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.02, - "output": 0.1 + "input": 0.2, + "output": 1.5 } }, - "deepseek-ai/deepseek-r1-distill-llama-70b": { - "id": "deepseek-ai/deepseek-r1-distill-llama-70b", - "name": "DeepSeek R1 Distill Llama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "qwen/qwen3-235b-a22b-fp8": { + "id": "qwen/qwen3-235b-a22b-fp8", + "name": "Qwen3 235B A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-01-23", - "last_updated": "2025-01-23", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -161804,26 +170452,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 40960, + "output": 20000 }, "cost": { - "input": 0.03, - "output": 0.14 + "input": 0.2, + "output": 0.8 } }, - "qwen/qwen3-coder": { - "id": "qwen/qwen3-coder", - "name": "Qwen3 Coder", + "qwen/qwen3-coder-480b-a35b-instruct": { + "id": "qwen/qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct", "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -161835,31 +170483,26 @@ "open_weights": true, "limit": { "context": 262144, - "output": 66536 + "output": 65536 }, "cost": { - "input": 0.3, + "input": 0.29, "output": 1.2 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "qwen/qwen3-235b-a22b-thinking-2507": { + "id": "qwen/qwen3-235b-a22b-thinking-2507", + "name": "Qwen3 235B A22b Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -161870,26 +170513,27 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.74, - "output": 3.48 + "input": 0.3, + "output": 3 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", + "qwen/qwen3-30b-a3b-fp8": { + "id": "qwen/qwen3-30b-a3b-fp8", + "name": "Qwen3 30B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -161900,26 +170544,26 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 40960, + "output": 20000 }, "cost": { - "input": 0.6, - "output": 2.4 + "input": 0.09, + "output": 0.45 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -161930,93 +170574,30 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "leonardo-ai/lucid-origin": { - "id": "leonardo-ai/lucid-origin", - "name": "Lucid Origin", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "lucid", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-06-01", - "last_updated": "2025-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 4096, - "output": 0 + "input": 0.15, + "output": 1.5 } }, - "leonardo-ai/lucid-realism": { - "id": "leonardo-ai/lucid-realism", - "name": "Lucid Realism", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "lucid", - "attachment": true, - "reasoning": false, + "qwen/qwen3-32b-fp8": { + "id": "qwen/qwen3-32b-fp8", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "temperature": false, - "release_date": "2025-06-01", - "last_updated": "2025-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 4096, - "output": 0 - } - } - } - }, - "abliteration-ai": { - "id": "abliteration-ai", - "env": [ - "ABLIT_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.abliteration.ai/v1", - "name": "abliteration.ai", - "doc": "https://docs.abliteration.ai/models", - "models": { - "abliterated-model": { - "id": "abliterated-model", - "name": "Abliterated Model", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, - "reasoning": false, - "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-01-06", - "last_updated": "2026-01-06", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -162024,42 +170605,39 @@ }, "open_weights": true, "limit": { - "context": 150000, - "input": 150000, - "output": 8192 + "context": 40960, + "output": 20000 }, "cost": { - "input": 3, - "output": 3 + "input": 0.1, + "output": 0.45 } } } }, - "xpersona": { - "id": "xpersona", + "ollama-cloud": { + "id": "ollama-cloud", "env": [ - "XPERSONA_API_KEY" + "OLLAMA_API_KEY" ], "npm": "@ai-sdk/openai-compatible", - "api": "https://www.xpersona.co/v1", - "name": "Xpersona", - "doc": "https://www.xpersona.co/docs", + "api": "https://ollama.com/v1", + "name": "Ollama Cloud", + "doc": "https://docs.ollama.com/cloud", "models": { - "xpersona-gpt-5.5": { - "id": "xpersona-gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": false, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "medium", "high", - "xhigh", "max" ] } @@ -162069,45 +170647,82 @@ "field": "reasoning_content" }, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-30", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 131072 }, "cost": { - "input": 3, - "output": 18, - "reasoning": 18, - "cache_read": 0.3 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "xpersona-frieren-coder": { - "id": "xpersona-frieren-coder", - "name": "Xpersona Frieren 1", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "nemotron-3-ultra": { + "id": "nemotron-3-ultra", + "name": "nemotron-3-ultra", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 128000 + }, + "cost": { + "input": 0.1, + "output": 3, + "cache_read": 0.1 + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "kimi-k3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", "high", - "xhigh", "max" ] } @@ -162115,9 +170730,8 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-12-30", - "release_date": "2026-05-01", - "last_updated": "2026-05-25", + "release_date": "2026-07-16", + "last_updated": "2026-07-27", "modalities": { "input": [ "text", @@ -162127,33 +170741,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 6, - "reasoning": 6, - "cache_read": 0.15 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", "high", - "xhigh", "max" ] } @@ -162162,10 +170776,11 @@ "interleaved": { "field": "reasoning_content" }, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -162175,180 +170790,160 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 3, - "output": 18, - "reasoning": 18, - "cache_read": 0.3 + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 } - } - } - }, - "azure-cognitive-services": { - "id": "azure-cognitive-services", - "env": [ - "AZURE_COGNITIVE_SERVICES_RESOURCE_NAME", - "AZURE_COGNITIVE_SERVICES_API_KEY" - ], - "npm": "@ai-sdk/azure", - "name": "Azure Cognitive Services", - "doc": "https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models", - "models": { - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + }, + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "minimax-m2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "knowledge": "2025-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + } + }, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "kimi-k2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-08-01", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "qwen3.5:397b": { + "id": "qwen3.5:397b", + "name": "qwen3.5:397b", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "interleaved": { + "field": "reasoning_details" + }, + "release_date": "2026-02-15", + "last_updated": "2026-02-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 262144, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.6, + "output": 3.6 } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 Nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, + "gpt-oss:20b": { + "id": "gpt-oss:20b", + "name": "gpt-oss:20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2025-08-05", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.07, + "output": 0.3, + "cache_read": 0.035 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "gpt-oss:120b": { + "id": "gpt-oss:120b", + "name": "gpt-oss:120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -162356,111 +170951,72 @@ "values": [ "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-12-31", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "release_date": "2025-08-05", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 0.15, + "output": 0.6, + "cache_read": 0.014 } }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "nemotron-3-nano:30b": { + "id": "nemotron-3-nano:30b", + "name": "nemotron-3-nano:30b", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2025-12-15", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 1048576, + "output": 131072 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.06, + "output": 0.24 } }, "kimi-k2.5": { "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "name": "kimi-k2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "family": "kimi-k2", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -162468,12 +171024,8 @@ } ], "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", @@ -162487,195 +171039,145 @@ "limit": { "context": 262144, "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" - }, - "cost": { - "input": 0.6, - "output": 3 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "mistral-large-3:675b": { + "id": "mistral-large-3:675b", + "name": "mistral-large-3:675b", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-31", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2025-12-02", + "last_updated": "2026-01-19", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 262144, + "output": 262144 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.5, + "output": 1.5 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "minimax-m2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 196608, + "output": 196608 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 Mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, + "deepseek-v4-pro:0813": { + "id": "deepseek-v4-pro:0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.66, + "output": 1.98, + "cache_read": 0.022 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": false, + "minimax-m3": { + "id": "minimax-m3", + "name": "minimax-m3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax-m3", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-05-31", + "last_updated": "2026-05-31", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -162683,113 +171185,75 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" + "context": 512000, + "output": 131072 }, "cost": { - "input": 0.95, - "output": 4 + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek-v4-flash:0731": { + "id": "deepseek-v4-flash:0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", "high", "max" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 } }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "gemma4:31b": { + "id": "gemma4:31b", + "name": "gemma4:31b", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2025-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-08", "modalities": { "input": [ "text", @@ -162799,153 +171263,163 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 0.14, + "output": 0.4, + "cache_read": 0.05 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 976000, + "output": 131072 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "glm-5.1": { + "id": "glm-5.1", + "name": "glm-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "release_date": "2026-03-27", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", - "attachment": true, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "deepseek-v4-pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ - "text", - "image", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 272000, - "output": 128000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.66, + "output": 1.98, + "cache_read": 0.022 } }, - "meta-llama-3-70b-instruct": { - "id": "meta-llama-3-70b-instruct", - "name": "Meta-Llama-3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-04-18", - "last_updated": "2024-04-18", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -162956,29 +171430,37 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 2048 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2.68, - "output": 3.54 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "phi-3-mini-4k-instruct": { - "id": "phi-3-mini-4k-instruct", - "name": "Phi-3-mini-instruct (4k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "kimi-k2.7-code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -162986,26 +171468,31 @@ }, "open_weights": true, "limit": { - "context": 4096, - "output": 1024 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.13, - "output": 0.52 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "phi-3-mini-128k-instruct": { - "id": "phi-3-mini-128k-instruct", - "name": "Phi-3-mini-instruct (128k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "nemotron-3-super": { + "id": "nemotron-3-super", + "name": "nemotron-3-super", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2026-03-11", + "last_updated": "2026-03-12", "modalities": { "input": [ "text" @@ -163016,27 +171503,37 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.13, - "output": 0.52 + "input": 0.015, + "output": 0.6, + "cache_read": 0.015 } }, - "phi-4-mini-reasoning": { - "id": "phi-4-mini-reasoning", - "name": "Phi-4-mini-reasoning", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "deepseek-v4-flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -163047,20 +171544,33 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 } - }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + } + } + }, + "agentrouter": { + "id": "agentrouter", + "env": [ + "AGENTROUTER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://agentrouter.org/v1", + "name": "AgentRouter", + "doc": "https://agentrouter.org/docs/opencode.html", + "models": { + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -163069,20 +171579,21 @@ "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -163090,59 +171601,80 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 128000 }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://agentrouter.org/v1" } }, - "llama-4-maverick-17b-128e-instruct-fp8": { - "id": "llama-4-maverick-17b-128e-instruct-fp8", - "name": "Llama 4 Maverick 17B 128E Instruct FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 128000 }, - "cost": { - "input": 0.25, - "output": 1 + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://agentrouter.org/v1" } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -163153,40 +171685,41 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.58, - "output": 1.68 + "context": 1000000, + "output": 131072 } }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -163194,93 +171727,133 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1050000, + "input": 922000, "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 } }, - "model-router": { - "id": "model-router", - "name": "Model Router", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "model-router", - "attachment": true, - "reasoning": false, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "release_date": "2025-05-19", - "last_updated": "2025-11-18", + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.14, - "output": 0 + "context": 1000000, + "output": 384000 } - }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + } + } + }, + "tencent-tokenhub": { + "id": "tencent-tokenhub", + "env": [ + "TENCENT_TOKENHUB_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://tokenhub.tencentmaas.com/v1", + "name": "Tencent TokenHub", + "doc": "https://cloud.tencent.com/document/product/1823/130050", + "models": { + "hy3-preview": { + "id": "hy3-preview", + "name": "Hy3 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "phi-4-multimodal": { - "id": "phi-4-multimodal", - "name": "Phi-4-multimodal", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "phi", - "attachment": true, - "reasoning": false, - "tool_call": false, + "hy3": { + "id": "hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -163288,28 +171861,40 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 256000, + "input": 192000, + "output": 128000 }, "cost": { - "input": 0.08, - "output": 0.32, - "input_audio": 4 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "phi-4-reasoning-plus": { - "id": "phi-4-reasoning-plus", - "name": "Phi-4-reasoning-plus", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", + "hy4-preview": { + "id": "hy4-preview", + "name": "Hy4 preview", + "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", + "family": "Hy", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2026-08-28", + "last_updated": "2026-08-28", "modalities": { "input": [ "text" @@ -163320,55 +171905,70 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 4096 + "context": 1024000, + "output": 64000 }, "cost": { - "input": 0.125, - "output": 0.5 + "input": 0.834, + "output": 2.501, + "cache_read": 0.042 } - }, - "codestral-2501": { - "id": "codestral-2501", - "name": "Codestral 25.01", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "codestral", - "attachment": false, + } + } + }, + "watsonx": { + "id": "watsonx", + "env": [ + "WATSONX_AI_APIKEY", + "WATSONX_AI_PROJECT_ID" + ], + "npm": "watsonx-ai-provider", + "name": "watsonx.ai", + "doc": "https://www.ibm.com/docs/en/watsonx/saas?topic=solutions-supported-foundation-models", + "models": { + "meta-llama/llama-4-maverick-17b-128e-instruct-fp8": { + "id": "meta-llama/llama-4-maverick-17b-128e-instruct-fp8", + "name": "Llama 4 Maverick 17B 128E Instruct FP8", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-03", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0.371, + "output": 1.484 } }, - "cohere-embed-v3-english": { - "id": "cohere-embed-v3-english", - "name": "Embed v3 English", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", + "meta-llama/llama-3-3-70b-instruct": { + "id": "meta-llama/llama-3-3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2023-11-07", - "last_updated": "2023-11-07", + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -163379,26 +171979,26 @@ }, "open_weights": true, "limit": { - "context": 512, - "output": 1024 + "context": 131072, + "output": 4096 }, "cost": { - "input": 0.1, - "output": 0 + "input": 0.7526, + "output": 0.7526 } }, - "phi-3-medium-4k-instruct": { - "id": "phi-3-medium-4k-instruct", - "name": "Phi-3-medium-instruct (4k)", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", + "ibm/granite-4-h-small": { + "id": "ibm/granite-4-h-small", + "name": "Granite-4.0-H-Small", + "description": "Open-weight hybrid model for enterprise chat, coding, retrieval-augmented generation, and tool-calling workloads", + "family": "granite", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2025-10-02", + "last_updated": "2025-10-02", "modalities": { "input": [ "text" @@ -163409,37 +172009,27 @@ }, "open_weights": true, "limit": { - "context": 4096, - "output": 1024 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.17, - "output": 0.68 + "input": 0.0636, + "output": 0.265 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "mistralai/mistral-small-3-1-24b-instruct-2503": { + "id": "mistralai/mistral-small-3-1-24b-instruct-2503", + "name": "Mistral Small 3.1 24B", + "description": "Efficient multimodal model for instruction following, coding, reasoning, and function calling", + "family": "mistral-small", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-17", "modalities": { "input": [ "text", @@ -163449,167 +172039,220 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 272000, - "output": 128000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01 + "input": 0.106, + "output": 0.318 } }, - "gpt-4-turbo": { - "id": "gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 131072, + "output": 32768 }, "cost": { - "input": 10, - "output": 30 + "input": 0.159, + "output": 0.636 } - }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + } + } + }, + "ambient": { + "id": "ambient", + "env": [ + "AMBIENT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.ambient.xyz/v1", + "name": "Ambient", + "doc": "https://ambient.xyz", + "models": { + "deepseek/deepseek-v4-flash-0731": { + "id": "deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.08, + "output": 0.18, + "cache_read": 0.016, + "cache_write": 0 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", - "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 272000, - "output": 128000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.03 - } + "input": 0.14, + "output": 0.28, + "cache_read": 0.028, + "cache_write": 0 + } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 202752, + "output": 202752 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.6, + "output": 2, + "cache_read": 0.15, + "cache_write": 0 } }, - "ministral-3b": { - "id": "ministral-3b", - "name": "Ministral 3B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "ambient/large": { + "id": "ambient/large", + "name": "Ambient Large", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -163620,26 +172263,33 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 202752, + "output": 202752 }, "cost": { - "input": 0.04, - "output": 0.04 + "input": 0.6, + "output": 2, + "cache_read": 0.15, + "cache_write": 0 } }, - "llama-3.2-90b-vision-instruct": { - "id": "llama-3.2-90b-vision-instruct", - "name": "Llama-3.2-90B-Vision-Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -163651,30 +172301,35 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2.04, - "output": 2.04 + "input": 0.95, + "output": 4, + "cache_read": 0.2, + "cache_write": 0 } }, - "deepseek-v3.1": { - "id": "deepseek-v3.1", - "name": "DeepSeek-V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -163682,136 +172337,161 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.56, - "output": 1.68 + "input": 0.69, + "output": 3.49, + "cache_read": 0.14, + "cache_write": 0 } }, - "o1": { - "id": "o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "zai-org/GLM-5.2-FP8": { + "id": "zai-org/GLM-5.2-FP8", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 202752, + "output": 202752 }, "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 1.2, + "output": 4.2, + "cache_read": 0.26, + "cache_write": 0 } }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, + "zai-org/GLM-5.1-FP8": { + "id": "zai-org/GLM-5.1-FP8", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 1.4, + "output": 4.4, + "cache_read": 0, + "cache_write": 0 } }, - "grok-4-fast-reasoning": { - "id": "grok-4-fast-reasoning", - "name": "Grok 4 Fast (Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo-V2.5", + "description": "MiMo omni model for text, image, video, audio, and agents", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-09-19", - "last_updated": "2025-09-19", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.4, + "output": 2, + "cache_read": 0.08, + "cache_write": 0 } }, - "phi-4": { - "id": "phi-4", - "name": "Phi-4", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", - "attachment": false, - "reasoning": false, - "tool_call": false, + "stepfun/step-3.7-flash": { + "id": "stepfun/step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -163819,25 +172499,81 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.125, - "output": 0.5 + "input": 0.19, + "output": 1.14, + "cache_read": 0.03, + "cache_write": 0 + } + } + } + }, + "model-oracle-ai": { + "id": "model-oracle-ai", + "env": [ + "MODEL_ORACLE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.modeloracle.com/api/v1", + "name": "Model Oracle AI", + "doc": "https://modeloracle.com/setup/", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 } }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -163846,10 +172582,10 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -163862,59 +172598,77 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 } }, - "meta-llama-3.1-8b-instruct": { - "id": "meta-llama-3.1-8b-instruct", - "name": "Meta-Llama-3.1-8B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 0.61 + "context": 200000, + "output": 100000 } }, - "gpt-3.5-turbo-0301": { - "id": "gpt-3.5-turbo-0301", - "name": "GPT-3.5 Turbo 0301", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-03-01", - "last_updated": "2023-03-01", + "claude-opus-4.8": { + "id": "claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -163922,58 +172676,79 @@ }, "open_weights": false, "limit": { - "context": 4096, - "output": 4096 - }, - "cost": { - "input": 1.5, - "output": 2 + "context": 1000000, + "output": 128000 } }, - "deepseek-v3.2-speciale": { - "id": "deepseek-v3.2-speciale", - "name": "DeepSeek-V3.2-Speciale", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "claude-haiku-4.5": { + "id": "claude-haiku-4.5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.58, - "output": 1.68 + "context": 200000, + "output": 64000 } }, - "text-embedding-3-small": { - "id": "text-embedding-3-small", - "name": "text-embedding-3-small", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -163981,39 +172756,34 @@ }, "open_weights": false, "limit": { - "context": 8191, - "output": 1536 - }, - "cost": { - "input": 0.02, - "output": 0 + "context": 1000000, + "output": 128000 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-08-31", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -164026,29 +172796,39 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 } }, - "mistral-large-2411": { - "id": "mistral-large-2411", - "name": "Mistral Large 24.11", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": false, - "reasoning": false, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-11-01", - "last_updated": "2024-11-01", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -164056,30 +172836,29 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 6 + "context": 1050000, + "input": 922000, + "output": 128000 } }, - "gpt-4-turbo-vision": { - "id": "gpt-4-turbo-vision", - "name": "GPT-4 Turbo Vision", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -164087,43 +172866,39 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 10, - "output": 30 + "context": 1047576, + "output": 32768 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -164131,37 +172906,35 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.125 } }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "none", + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -164170,30 +172943,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "context": 1000000, + "output": 131072 } }, - "text-embedding-ada-002": { - "id": "text-embedding-ada-002", - "name": "text-embedding-ada-002", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, "reasoning": false, - "tool_call": false, - "release_date": "2022-12-15", - "last_updated": "2022-12-15", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -164201,26 +172974,38 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 - }, - "cost": { - "input": 0.1, - "output": 0 + "context": 1047576, + "output": 32768 } }, - "meta-llama-3.1-70b-instruct": { - "id": "meta-llama-3.1-70b-instruct", - "name": "Meta-Llama-3.1-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -164231,26 +173016,34 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 2.68, - "output": 3.54 + "context": 1000000, + "output": 384000 } }, - "mistral-medium-2505": { - "id": "mistral-medium-2505", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -164262,51 +173055,69 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 400000, + "input": 272000, "output": 128000 - }, - "cost": { - "input": 0.4, - "output": 2 } }, - "phi-4-reasoning": { - "id": "phi-4-reasoning", - "name": "Phi-4-reasoning", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", - "attachment": false, + "auto": { + "id": "auto", + "name": "Auto", + "description": "Model Oracle AI decision engine that selects and routes among configured coding-agent models", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2026-06-29", + "last_updated": "2026-07-07", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 4096 - }, - "cost": { - "input": 0.125, - "output": 0.5 + "context": 1000000, + "output": 128000 } - }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt-codex", - "attachment": false, + } + } + }, + "xai": { + "id": "xai", + "env": [ + "XAI_API_KEY" + ], + "npm": "@ai-sdk/xai", + "name": "xAI", + "doc": "https://docs.x.ai/docs/models", + "models": { + "grok-4.7": { + "id": "grok-4.7", + "name": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -164314,342 +173125,462 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "temperature": true, + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ - "text", - "image", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 500000, + "output": 500000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } }, - "gpt-3.5-turbo-0613": { - "id": "gpt-3.5-turbo-0613", - "name": "GPT-3.5 Turbo 0613", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, + "grok-imagine-image": { + "id": "grok-imagine-image", + "name": "Grok Imagine Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-06-13", - "last_updated": "2023-06-13", + "temperature": false, + "release_date": "2026-01-28", + "last_updated": "2026-01-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "text" + "image", + "pdf" ] }, "open_weights": false, "limit": { - "context": 16384, - "output": 16384 - }, - "cost": { - "input": 3, - "output": 4 + "context": 16000, + "output": 0 } }, - "cohere-embed-v3-multilingual": { - "id": "cohere-embed-v3-multilingual", - "name": "Embed v3 Multilingual", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", - "attachment": false, + "grok-imagine-video": { + "id": "grok-imagine-video", + "name": "Grok Imagine Video", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": true, "reasoning": false, "tool_call": false, "temperature": false, - "release_date": "2023-11-07", - "last_updated": "2023-11-07", + "release_date": "2026-01-28", + "last_updated": "2026-01-28", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512, - "output": 1024 - }, - "cost": { - "input": 0.1, + "context": 1024, "output": 0 } }, - "gpt-3.5-turbo-0125": { - "id": "gpt-3.5-turbo-0125", - "name": "GPT-3.5 Turbo 0125", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2021-08", - "release_date": "2024-01-25", - "last_updated": "2024-01-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16384, - "output": 16384 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "phi-3-small-8k-instruct": { - "id": "phi-3-small-8k-instruct", - "name": "Phi-3-small-instruct (8k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, - "reasoning": false, - "tool_call": false, + "grok-4.3": { + "id": "grok-4.3", + "name": "Grok 4.3", + "description": "xAI's Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 2048 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "grok-4.20-0309-reasoning": { + "id": "grok-4.20-0309-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 1.35, - "output": 5.4 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-12-02", + "grok-imagine-video-1.5": { + "id": "grok-imagine-video-1.5", + "name": "Grok Imagine Video 1.5", + "description": "Video model for image-to-video generation, editing, and extension workflows", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-05-30", + "last_updated": "2026-05-30", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "pdf" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "context": 1024, + "output": 0 } }, - "gpt-5.1-chat": { - "id": "gpt-5.1-chat", - "name": "GPT-5.1 Chat", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt-codex", + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "medium" + "low", + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ - "text", - "image", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 500000, + "output": 500000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 2, + "output": 6, + "cache_read": 0.3, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 0.6 + } } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "grok-4.20-0309-non-reasoning": { + "id": "grok-4.20-0309-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.71, - "output": 0.71 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "llama-4-scout-17b-16e-instruct": { - "id": "llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout 17B 16E Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + "grok-imagine-image-quality": { + "id": "grok-imagine-image-quality", + "name": "Grok Imagine Image Quality", + "description": "Higher-fidelity Grok Imagine image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", "attachment": true, "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-03", + "last_updated": "2026-04-03", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "image", + "pdf" + ] + }, + "open_weights": false, + "limit": { + "context": 16000, + "output": 0 + } + }, + "grok-build-0.1": { + "id": "grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.2, - "output": 0.78 + "input": 1, + "output": 2, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 4, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 4, + "cache_read": 0.4 + } } }, - "gpt-3.5-turbo-1106": { - "id": "gpt-3.5-turbo-1106", - "name": "GPT-3.5 Turbo 1106", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, + "grok-4.20-multi-agent-0309": { + "id": "grok-4.20-multi-agent-0309", + "name": "Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-11-06", - "last_updated": "2023-11-06", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -164657,59 +173588,120 @@ }, "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 1, - "output": 2 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "phi-4-mini": { - "id": "phi-4-mini", - "name": "Phi-4-mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, - "reasoning": false, + "grok-4.6": { + "id": "grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } - }, - "cohere-command-r-plus-08-2024": { - "id": "cohere-command-r-plus-08-2024", - "name": "Command R+", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", - "attachment": false, + } + } + }, + "nebius": { + "id": "nebius", + "env": [ + "NEBIUS_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.tokenfactory.nebius.com/v1", + "name": "Nebius Token Factory", + "doc": "https://docs.tokenfactory.nebius.com/", + "models": { + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma-3-27b-it", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2025-10", + "release_date": "2026-01-20", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -164717,26 +173709,29 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 110000, + "input": 100000, + "output": 8192 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01, + "cache_write": 0.125 } }, - "meta-llama-3.1-405b-instruct": { - "id": "meta-llama-3.1-405b-instruct", - "name": "Meta-Llama-3.1-405B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "Qwen/Qwen3-30B-A3B-Instruct-2507": { + "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen3-30B-A3B-Instruct-2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-12", + "release_date": "2026-01-28", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -164747,26 +173742,29 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 262144, + "input": 262144, + "output": 8192 }, "cost": { - "input": 5.33, - "output": 16 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01, + "cache_write": 0.125 } }, - "gpt-4-32k": { - "id": "gpt-4-32k", - "name": "GPT-4 32K", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-03-14", - "last_updated": "2023-03-14", + "knowledge": "2025-07", + "release_date": "2025-07-25", + "last_updated": "2025-10-04", "modalities": { "input": [ "text" @@ -164777,26 +173775,27 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 262144, + "output": 8192 }, "cost": { - "input": 60, - "output": 120 + "input": 0.2, + "output": 0.6 } }, - "phi-3-medium-128k-instruct": { - "id": "phi-3-medium-128k-instruct", - "name": "Phi-3-medium-instruct (128k)", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", + "Qwen/Qwen3-Embedding-8B": { + "id": "Qwen/Qwen3-Embedding-8B", + "name": "Qwen3-Embedding-8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "structured_output": false, + "temperature": false, + "knowledge": "2025-10", + "release_date": "2026-01-10", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -164807,68 +173806,80 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 40960, + "input": 40960, + "output": 0 }, "cost": { - "input": 0.17, - "output": 0.68 + "input": 0.01, + "output": 0 } }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen3.5-397B-A17B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "structured_output": true, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-15", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "input": 250000, + "output": 8192 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 0.6, + "output": 3.6, + "cache_read": 0.06, + "cache_write": 0.75 } }, - "gpt-4": { - "id": "gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-03-14", - "last_updated": "2023-03-14", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -164877,63 +173888,93 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 1024000, + "output": 1024000 }, "cost": { - "input": 60, - "output": 120 + "input": 0.14, + "output": 0.28, + "cache_read": 0.14 } }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, - "reasoning": false, + "deepseek-ai/DeepSeek-V4-Pro-0813": { + "id": "deepseek-ai/DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 979000, + "output": 979000 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 1.32, + "output": 3.96, + "cache_read": 1.32 } }, - "cohere-command-r-08-2024": { - "id": "cohere-command-r-08-2024", - "name": "Command R", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, + "deepseek-ai/DeepSeek-V4.1-Flash": { + "id": "deepseek-ai/DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -164941,70 +173982,76 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 1048000, + "output": 1048000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.3, + "output": 1.2, + "cache_read": 0.3 } }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high" + "none", + "low", + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 272000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 15, - "output": 120 + "input": 1.75, + "output": 3.5, + "cache_read": 0.15 } }, - "llama-3.2-11b-vision-instruct": { - "id": "llama-3.2-11b-vision-instruct", - "name": "Llama-3.2-11B-Vision-Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", - "attachment": true, - "reasoning": false, + "MiniMaxAI/MiniMax-M3": { + "id": "MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -165012,27 +174059,41 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.37, - "output": 0.37 + "input": 0.3, + "output": 1.2 } }, - "cohere-command-a": { - "id": "cohere-command-a", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", + "moonshotai/Kimi-K3": { + "id": "moonshotai/Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text" @@ -165043,59 +174104,61 @@ }, "open_weights": true, "limit": { - "context": 256000, + "context": 1048576, "output": 8000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 3, + "output": 15, + "cache_read": 3 } }, - "gpt-5-chat": { - "id": "gpt-5-chat", - "name": "GPT-5 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": false, - "knowledge": "2024-10-24", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 8000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.95, + "output": 4 } }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -165103,45 +174166,59 @@ } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 272000, - "output": 128000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 1.4, + "output": 4.4 } }, - "cohere-embed-v-4-0": { - "id": "cohere-embed-v-4-0", - "name": "Embed v4", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "zai-org/GLM-5.3-Flash": { + "id": "zai-org/GLM-5.3-Flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -165149,26 +174226,40 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 1536 + "context": 1024000, + "output": 1024000 }, "cost": { - "input": 0.12, - "output": 0 + "input": 0.15, + "output": 0.5, + "cache_read": 0.15 } }, - "mistral-nemo": { - "id": "mistral-nemo", - "name": "Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + "zai-org/GLM-5.3": { + "id": "zai-org/GLM-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -165179,26 +174270,29 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 1024000, + "output": 1024000 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 1.4, + "output": 4.4, + "cache_read": 1.4 } }, - "phi-3.5-mini-instruct": { - "id": "phi-3.5-mini-instruct", - "name": "Phi-3.5-mini-instruct", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "Nemotron-3-Super-120B-A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-08-20", - "last_updated": "2024-08-20", + "knowledge": "2026-02", + "release_date": "2026-03-11", + "last_updated": "2026-03-12", "modalities": { "input": [ "text" @@ -165209,27 +174303,36 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.13, - "output": 0.52 + "input": 0.3, + "output": 0.9 } }, - "o1-mini": { - "id": "o1-mini", - "name": "o1-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "nvidia/Nemotron-3_5-Lightning": { + "id": "nvidia/Nemotron-3_5-Lightning", + "name": "Nemotron 3.5 Lightning 30B A3B", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-09-12", - "last_updated": "2024-09-12", + "structured_output": true, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ "text" @@ -165238,27 +174341,37 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 65536 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.06, + "output": 0.24, + "cache_read": 0.06 } }, - "text-embedding-3-large": { - "id": "text-embedding-3-large", - "name": "text-embedding-3-large", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "nvidia/Nemotron-3-Ultra-550b-a55b": { + "id": "nvidia/Nemotron-3-Ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -165267,59 +174380,80 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8191, - "output": 3072 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.13, - "output": 0 + "input": 1, + "output": 3, + "cache_read": 1 } }, - "mistral-small-2503": { - "id": "mistral-small-2503", - "name": "Mistral Small 3.1", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, - "reasoning": false, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2025-03-01", - "last_updated": "2025-03-01", + "knowledge": "2025-09", + "release_date": "2026-01-10", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 131072, + "input": 124000, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.15, + "output": 0.6, + "reasoning": 0.6, + "cache_read": 0.015, + "cache_write": 0.18 } }, - "meta-llama-3-8b-instruct": { - "id": "meta-llama-3-8b-instruct", - "name": "Meta-Llama-3-8B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "NousResearch/Hermes-4-405B": { + "id": "NousResearch/Hermes-4-405B", + "name": "Hermes-4-405B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-04-18", - "last_updated": "2024-04-18", + "knowledge": "2025-11", + "release_date": "2026-01-30", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -165330,29 +174464,51 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 2048 + "context": 131072, + "input": 120000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 0.61 + "input": 1, + "output": 3, + "reasoning": 3, + "cache_read": 0.1, + "cache_write": 1.25 } - }, - "phi-3-small-128k-instruct": { - "id": "phi-3-small-128k-instruct", - "name": "Phi-3-small-instruct (128k)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, - "reasoning": false, - "tool_call": false, + } + } + }, + "minimax-cn": { + "id": "minimax-cn", + "env": [ + "MINIMAX_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://api.minimax.cn/anthropic/v1", + "name": "MiniMax (minimax.cn)", + "doc": "https://platform.minimaxi.com/docs/guides/quickstart", + "models": { + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-04-23", - "last_updated": "2024-04-23", + "release_date": "2026-06-01", + "last_updated": "2026-06-25", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -165360,26 +174516,43 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 512000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "tiers": [ + { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12, + "tier": { + "type": "context", + "size": 512000 + } + } + ], + "context_over_200k": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 + } } }, - "deepseek-v3-0324": { - "id": "deepseek-v3-0324", - "name": "DeepSeek-V3-0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -165390,110 +174563,92 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 204800, "output": 131072 }, "cost": { - "input": 1.14, - "output": 4.56 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "o3": { - "id": "o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", - "attachment": true, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "gpt-5.2-chat": { - "id": "gpt-5.2-chat", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "deepseek-r1-0528": { - "id": "deepseek-r1-0528", - "name": "DeepSeek-R1-0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "MiniMax-M2.7-highspeed": { + "id": "MiniMax-M2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -165504,26 +174659,28 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.35, - "output": 5.4 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "gpt-3.5-turbo-instruct": { - "id": "gpt-3.5-turbo-instruct", - "name": "GPT-3.5 Turbo Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "MiniMax-M2": { + "id": "MiniMax-M2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-09-21", - "last_updated": "2023-09-21", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ "text" @@ -165532,28 +174689,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4096, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 2 + "input": 0.3, + "output": 1.2 } }, - "phi-3.5-moe-instruct": { - "id": "phi-3.5-moe-instruct", - "name": "Phi-3.5-MoE-instruct", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", + "MiniMax-M2.5-highspeed": { + "id": "MiniMax-M2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-08-20", - "last_updated": "2024-08-20", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -165564,36 +174721,53 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.16, - "output": 0.64 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } - }, - "codex-mini": { - "id": "codex-mini", - "name": "Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex-mini", - "attachment": true, + } + } + }, + "scaleway": { + "id": "scaleway", + "env": [ + "SCALEWAY_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.scaleway.ai/v1", + "name": "Scaleway", + "doc": "https://www.scaleway.com/en/docs/generative-apis/", + "models": { + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-04", - "release_date": "2025-05-16", - "last_updated": "2025-05-16", + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -165602,54 +174776,35 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 256000, + "output": 16384 }, + "status": "beta", "cost": { - "input": 1.5, - "output": 6, - "cache_read": 0.375 + "input": 0.468, + "output": 0.936, + "reasoning": 0.936, + "cache_read": 0.0936 } - } - } - }, - "baseten": { - "id": "baseten", - "env": [ - "BASETEN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.baseten.co/v1", - "name": "Baseten", - "doc": "https://docs.baseten.co/inference/model-apis/overview", - "models": { - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, + }, + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-04", + "release_date": "2025-07-01", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -165657,40 +174812,30 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 260000, + "output": 16384 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.75, + "output": 2.25, + "reasoning": 8.4 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-01-30", - "last_updated": "2026-02-12", + "whisper-large-v3": { + "id": "whisper-large-v3", + "name": "Whisper Large v3", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2023-09-01", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text", - "image" + "audio" ], "output": [ "text" @@ -165698,37 +174843,58 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 0, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.12 + "input": 0.003, + "output": 0 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "bge-multilingual-gemma2": { + "id": "bge-multilingual-gemma2", + "name": "BGE Multilingual Gemma2", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-07-26", + "last_updated": "2025-06-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8191, + "output": 3072 + }, + "cost": { + "input": 0.1, + "output": 0 + } + }, + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -165736,26 +174902,26 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.2, + "output": 0.8 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "OpenAI GPT 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "gemma-4-26b-a4b-it": { + "id": "gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -165765,12 +174931,13 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-04", + "release_date": "2026-04-01", + "last_updated": "2026-05-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -165778,35 +174945,27 @@ }, "open_weights": true, "limit": { - "context": 128072, - "output": 128072 + "context": 256000, + "output": 16384 }, + "status": "beta", "cost": { - "input": 0.1, + "input": 0.25, "output": 0.5 } }, - "nvidia/Nemotron-120B-A12B": { - "id": "nvidia/Nemotron-120B-A12B", - "name": "Nemotron Super", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2026-03-17", "modalities": { "input": [ "text" @@ -165817,38 +174976,42 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 202800 + "context": 100000, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 0.75, - "cache_read": 0.06 + "input": 0.9, + "output": 0.9 } }, - "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B": { - "id": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B", - "name": "Nemotron Ultra", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2025-04", + "release_date": "2026-05-01", + "last_updated": "2026-05-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -165856,39 +175019,31 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 202800 + "context": 128000, + "output": 16384 }, + "status": "beta", "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 + "input": 0.25, + "output": 1.5 } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM 5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "pixtral-12b-2409": { + "id": "pixtral-12b-2409", + "name": "Pixtral 12B 2409", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2024-09", + "release_date": "2024-09-25", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -165896,39 +175051,68 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 202800 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.95, - "output": 3.15, - "cache_read": 0.2 + "input": 0.2, + "output": 0.2 } }, - "zai-org/GLM-4.7": { - "id": "zai-org/GLM-4.7", - "name": "GLM 4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "qwen3-embedding-8b": { + "id": "qwen3-embedding-8b", + "name": "Qwen3 Embedding 8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-05", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 4096 + }, + "cost": { + "input": 0.1, + "output": 0 + } + }, + "mistral-medium-3.5-128b": { + "id": "mistral-medium-3.5-128b", + "name": "Mistral Medium 3.5 128B", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -165936,38 +175120,43 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 200000 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.12 + "input": 1.5, + "output": 7.5 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-04", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -165975,35 +175164,37 @@ }, "open_weights": true, "limit": { - "context": 202720, - "output": 202720 + "context": 256000, + "output": 16384 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.3 + "input": 0.6, + "output": 3.6 } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -166014,30 +175205,30 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 202800 + "context": 256000, + "output": 16384 }, "cost": { - "input": 1.3, - "output": 4.3, - "cache_read": 0.26 + "input": 1.8, + "output": 5.5 } }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek V3.1", - "description": "Legacy model retained for compatibility with older integrations", - "family": "deepseek", + "mistral-small-3.2-24b-instruct-2506": { + "id": "mistral-small-3.2-24b-instruct-2506", + "name": "Mistral Small 3.2 24B Instruct (2506)", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-08-25", - "last_updated": "2025-08-25", + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -166045,21 +175236,20 @@ }, "open_weights": true, "limit": { - "context": 164000, - "output": 131000 + "context": 128000, + "output": 32768 }, - "status": "deprecated", "cost": { - "input": 0.5, - "output": 1.5 + "input": 0.15, + "output": 0.35 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "Deepseek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT-OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -166067,55 +175257,14 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.145 - } - }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2024-01-01", + "last_updated": "2026-03-17", "modalities": { "input": [ "text" @@ -166126,38 +175275,36 @@ }, "open_weights": true, "limit": { - "context": 204000, - "output": 204000 + "context": 128000, + "output": 32768 }, - "status": "deprecated", "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.15, + "output": 0.6 } } } }, - "atomic-chat": { - "id": "atomic-chat", + "vercel": { + "id": "vercel", "env": [ - "ATOMIC_CHAT_API_KEY" + "AI_GATEWAY_API_KEY" ], - "npm": "@ai-sdk/openai-compatible", - "api": "http://127.0.0.1:1337/v1", - "name": "Atomic Chat", - "doc": "https://atomic.chat", + "npm": "@ai-sdk/gateway", + "name": "Vercel AI Gateway", + "doc": "https://github.com/vercel/ai/tree/5eb85cc45a259553501f535b8ac79a77d0e79223/packages/gateway", "models": { - "gemma-4-E4B-it-IQ4_XS": { - "id": "gemma-4-E4B-it-IQ4_XS", - "name": "Gemma 4 E4B Instruct (IQ4_XS)", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "voyage/voyage-law-2": { + "id": "voyage/voyage-law-2", + "name": "voyage-law-2", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "voyage", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2024-04-15", + "last_updated": "2024-03", "modalities": { "input": [ "text" @@ -166166,27 +175313,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 + "context": 8192, + "output": 1536 } }, - "Meta-Llama-3_1-8B-Instruct-GGUF": { - "id": "Meta-Llama-3_1-8B-Instruct-GGUF", - "name": "Meta Llama 3.1 8B Instruct (GGUF)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "voyage/rerank-2.5-lite": { + "id": "voyage/rerank-2.5-lite", + "name": "Voyage Rerank 2.5 Lite", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "voyage", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ "text" @@ -166195,57 +175338,48 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 + "context": 32000, + "output": 32000 } }, - "Qwen3_5-9B-MLX-4bit": { - "id": "Qwen3_5-9B-MLX-4bit", - "name": "Qwen 3.5 9B (MLX 4-bit)", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "voyage/voyage-3.5": { + "id": "voyage/voyage-3.5", + "name": "voyage-3.5", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "voyage", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2026-03-05", - "last_updated": "2026-04-04", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 + "context": 8192, + "output": 1536 } }, - "gemma-4-E4B-it-MLX-4bit": { - "id": "gemma-4-E4B-it-MLX-4bit", - "name": "Gemma 4 E4B Instruct (MLX 4-bit)", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "voyage/voyage-finance-2": { + "id": "voyage/voyage-finance-2", + "name": "voyage-finance-2", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "voyage", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2024-06-03", + "last_updated": "2024-03", "modalities": { "input": [ "text" @@ -166254,89 +175388,26 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 + "context": 8192, + "output": 1536 } }, - "Qwen3_5-9B-Q4_K_M": { - "id": "Qwen3_5-9B-Q4_K_M", - "name": "Qwen 3.5 9B (Q4_K_M)", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "voyage/voyage-4-large": { + "id": "voyage/voyage-4-large", + "name": "voyage-4-large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "voyage", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2026-03-05", - "last_updated": "2026-04-04", + "release_date": "2026-01-15", + "last_updated": "2026-03-06", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - } - } - }, - "meta": { - "id": "meta", - "env": [ - "META_MODEL_API_KEY" - ], - "npm": "@ai-sdk/openai", - "api": "https://api.meta.ai/v1", - "name": "Meta", - "doc": "https://dev.meta.ai/docs", - "models": { - "muse-spark-1.1": { - "id": "muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf", - "video" ], "output": [ "text" @@ -166344,81 +175415,46 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32000 - }, - "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "context": 32000, + "output": 0 } - } - } - }, - "routing-run": { - "id": "routing-run", - "env": [ - "ROUTING_RUN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.routing.run/v1", - "name": "routing.run", - "doc": "https://docs.routing.run/api-reference/models", - "models": { - "kimi-k2.6-nitro": { - "id": "kimi-k2.6-nitro", - "name": "Kimi K2.6 Nitro", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + }, + "voyage/voyage-3.5-lite": { + "id": "voyage/voyage-3.5-lite", + "name": "voyage-3.5-lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "voyage", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 0.275, - "output": 1.1 + "context": 8192, + "output": 1536 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "voyage/voyage-4": { + "id": "voyage/voyage-4", + "name": "voyage-4", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "voyage", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-01-15", + "last_updated": "2026-03-06", "modalities": { "input": [ "text" @@ -166427,70 +175463,48 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.112, - "output": 0.224 + "context": 32000, + "output": 0 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "voyage/voyage-3-large": { + "id": "voyage/voyage-3-large", + "name": "voyage-3-large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "voyage", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-01-07", + "last_updated": "2024-09", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 0.275, - "output": 1.1 + "context": 8192, + "output": 1536 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "voyage/voyage-code-2": { + "id": "voyage/voyage-code-2", + "name": "voyage-code-2", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "voyage", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2024-01-01", + "last_updated": "2024-01", "modalities": { "input": [ "text" @@ -166499,32 +175513,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.348, - "output": 0.696 + "context": 8192, + "output": 1536 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "voyage/voyage-code-3": { + "id": "voyage/voyage-code-3", + "name": "voyage-code-3", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "voyage", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2024-12-04", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -166533,37 +175538,26 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 0.8, - "output": 2.4 + "context": 8192, + "output": 1536 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "voyage/rerank-2.5": { + "id": "voyage/rerank-2.5", + "name": "Voyage Rerank 2.5", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "voyage", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -166571,30 +175565,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 32000, "output": 32000 - }, - "cost": { - "input": 5, - "output": 25 } }, - "glm-5.2-nitro": { - "id": "glm-5.2-nitro", - "name": "GLM 5.2 Nitro", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "voyage/voyage-4-lite": { + "id": "voyage/voyage-4-lite", + "name": "voyage-4-lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "voyage", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-01-15", + "last_updated": "2026-03-06", "modalities": { "input": [ "text" @@ -166603,69 +175588,70 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 0.8, - "output": 2.4 + "context": 32000, + "output": 0 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "interfaze/interfaze-beta": { + "id": "interfaze/interfaze-beta", + "name": "Interfaze Beta", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2025-10-07", + "last_updated": "2026-04-29", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, + "context": 1000000, "output": 32000 }, "cost": { - "input": 0.275, - "output": 1.1 + "input": 1.5, + "output": 3.5 } }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "zai/glm-4.5": { + "id": "zai/glm-4.5", + "name": "GLM 4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "interleaved": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-07", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -166676,30 +175662,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32000 + "context": 128000, + "output": 96000 }, "cost": { - "input": 0.16, - "output": 0.48 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "zai/glm-5v-turbo": { + "id": "zai/glm-5v-turbo", + "name": "GLM 5V Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", @@ -166712,32 +175699,41 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 200000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "nemotron-3-ultra": { - "id": "nemotron-3-ultra", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, + "zai/glm-5.3-flash": { + "id": "zai/glm-5.3-flash", + "name": "GLM 5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -166745,36 +175741,36 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32000 + "context": 1000000, + "output": 131000 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "kimi-k2.7-code-nitro": { - "id": "kimi-k2.7-code-nitro", - "name": "Kimi K2.7 Code Nitro", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "zai/glm-4.6": { + "id": "zai/glm-4.6", + "name": "GLM 4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "interleaved": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -166783,40 +175779,38 @@ "open_weights": true, "limit": { "context": 200000, - "output": 32000 + "output": 96000 }, "cost": { - "input": 0.275, - "output": 1.1 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } - } - } - }, - "aihubmix": { - "id": "aihubmix", - "env": [ - "AIHUBMIX_API_KEY" - ], - "npm": "@aihubmix/ai-sdk-provider", - "name": "AIHubMix", - "doc": "https://docs.aihubmix.com", - "models": { - "coding-minimax-m2.7": { - "id": "coding-minimax-m2.7", - "name": "Coding MiniMax M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + }, + "zai/glm-5.2-fast": { + "id": "zai/glm-5.2-fast", + "name": "GLM 5.2 Fast", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -166827,18 +175821,19 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 128100 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 2.1, + "output": 6.6, + "cache_read": 0.21 } }, - "alicloud-glm-5.1": { - "id": "alicloud-glm-5.1", - "name": "GLM-5.1 (Alibaba Cloud)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "zai/glm-5": { + "id": "zai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", "family": "glm", "attachment": false, "reasoning": true, @@ -166848,13 +175843,9 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -166865,143 +175856,98 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 202800, + "output": 131100 }, "cost": { - "input": 0.84, - "output": 3.38, - "cache_read": 0.169, - "cache_write": 1.05625 + "input": 1, + "output": 3.2 } }, - "claude-sonnet-4-6-think": { - "id": "claude-sonnet-4-6-think", - "name": "Claude Sonnet 4.6 Thinking", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "zai/glm-4.5v": { + "id": "zai/glm-4.5v", + "name": "GLM 4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-08", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 66000, + "output": 16000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 0.6, + "output": 1.8, + "cache_read": 0.11 } }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "zai/glm-5.3-flashx": { + "id": "zai/glm-5.3-flashx", + "name": "GLM 5.3 FlashX", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 1 + "input": 0.37, + "output": 1.25, + "cache_read": 0.075 } }, - "alicloud-deepseek-v4-flash": { - "id": "alicloud-deepseek-v4-flash", - "name": "DeepSeek V4 Flash (Alibaba Cloud)", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", + "zai/glm-4.7-flash": { + "id": "zai/glm-4.7-flash", + "name": "GLM 4.7 Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -167010,14 +175956,10 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -167028,20 +175970,19 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "output": 131000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0.07, + "output": 0.4 } }, - "xiaomi-mimo-v2.5-pro": { - "id": "xiaomi-mimo-v2.5-pro", - "name": "Xiaomi MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo-v2.5-pro", + "zai/glm-4.7": { + "id": "zai/glm-4.7", + "name": "GLM 4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -167050,13 +175991,11 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "interleaved": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -167067,37 +176006,21 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 120000 }, "cost": { - "input": 1.1, - "output": 3.3, - "cache_read": 0.22, - "tiers": [ - { - "input": 2.2, - "output": 6.6, - "cache_read": 0.44, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2.2, - "output": 6.6, - "cache_read": 0.44 - } + "input": 0.6, + "output": 2.2, + "cache_read": 0.12 } }, - "doubao-seed-2-0-code-preview": { - "id": "doubao-seed-2-0-code-preview", - "name": "Doubao Seed 2.0 Code Preview", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "seed", - "attachment": true, + "zai/glm-5.2": { + "id": "zai/glm-5.2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -167106,68 +176029,41 @@ { "type": "effort", "values": [ - "minimal", - "low", - "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.48, - "output": 2.41, - "cache_read": 0.09644, - "tiers": [ - { - "input": 0.72, - "output": 3.62, - "cache_read": 0.144656, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.45, - "output": 7.23, - "cache_read": 0.28932, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.8, + "output": 2.55, + "cache_read": 0.16 } }, - "coding-xiaomi-mimo-v2.5-pro": { - "id": "coding-xiaomi-mimo-v2.5-pro", - "name": "Coding Xiaomi MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo-v2.5-pro", - "attachment": false, + "zai/glm-5.1": { + "id": "zai/glm-5.1", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -167175,13 +176071,10 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -167192,106 +176085,57 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 202800, + "output": 64000 }, "cost": { - "input": 0.2, - "output": 0.6, - "cache_read": 0.04, - "tiers": [ - { - "input": 0.4, - "output": 1.2, - "cache_read": 0.08, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.2, - "cache_read": 0.08 - } + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "doubao-seed-2-0-pro": { - "id": "doubao-seed-2-0-pro", - "name": "Doubao Seed 2.0 Pro", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, + "zai/glm-4.7-flashx": { + "id": "zai/glm-4.7-flashx", + "name": "GLM 4.7 FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "interleaved": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-01", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, + "context": 200000, "output": 128000 }, "cost": { - "input": 0.48, - "output": 2.41, - "cache_read": 0.09644, - "tiers": [ - { - "input": 0.72, - "output": 3.62, - "cache_read": 0.144656, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.45, - "output": 7.23, - "cache_read": 0.28932, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.06, + "output": 0.4, + "cache_read": 0.01 } }, - "deep-deepseek-v4-flash": { - "id": "deep-deepseek-v4-flash", - "name": "DeepSeek V4 Flash (DeepSeek)", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", + "zai/glm-5-turbo": { + "id": "zai/glm-5-turbo", + "name": "GLM 5 Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -167300,14 +176144,10 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text" @@ -167316,42 +176156,34 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 202800, + "output": 131100 }, "cost": { - "input": 0.154, - "output": 0.308, - "cache_read": 0.0308 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "zai/glm-4.5-air": { + "id": "zai/glm-4.5-air", + "name": "GLM 4.5 Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -167360,107 +176192,115 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 991000, - "output": 64000 + "context": 128000, + "output": 96000 }, "cost": { - "input": 0.282, - "output": 1.128, - "cache_read": 0.0564, - "cache_write": 0.3525 + "input": 0.2, + "output": 1.1, + "cache_read": 0.03 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "zai/glm-5.3": { + "id": "zai/glm-5.3", + "name": "GLM 5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-05", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 1.4, + "output": 4.4, + "cache_read": 0.14 } }, - "grok-4.3": { - "id": "grok-4.3", - "name": "Grok 4.3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "zai/glm-5.3-fast": { + "id": "zai/glm-5.3-fast", + "name": "GLM 5.3 Fast", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 262144 + }, + "cost": { + "input": 2.1, + "output": 6.6, + "cache_read": 0.21 + } + }, + "spacexai/grok-4.20-multi-agent": { + "id": "spacexai/grok-4.20-multi-agent", + "name": "Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "release_date": "2026-03-10", + "last_updated": "2026-03-10", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -167468,8 +176308,8 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 2000000, + "output": 2000000 }, "cost": { "input": 1.25, @@ -167482,7 +176322,7 @@ "cache_read": 0.4, "tier": { "type": "context", - "size": 200000 + "size": 200001 } } ], @@ -167493,33 +176333,46 @@ } } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "spacexai/grok-tts": { + "id": "spacexai/grok-tts", + "name": "Grok TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "grok", "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "spacexai/grok-4.20-reasoning": { + "id": "spacexai/grok-4.20-reasoning", + "name": "Grok 4.20 Reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2026-03-10", + "last_updated": "2026-03-10", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -167527,27 +176380,39 @@ }, "open_weights": false, "limit": { - "context": 991000, - "output": 64000 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 1.69, - "output": 5.07, - "cache_read": 0.169, - "cache_write": 2.1125 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200001 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "spacexai/grok-4.7": { + "id": "spacexai/grok-4.7", + "name": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -167555,26 +176420,18 @@ "medium", "high" ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-05", + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -167582,89 +176439,67 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 + "input": 1.2, + "output": 3.6, + "cache_read": 0.3, + "tiers": [ + { + "input": 2.4, + "output": 7.2, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 200001 + } + } + ], + "context_over_200k": { + "input": 2.4, + "output": 7.2, + "cache_read": 0.6 + } } }, - "xiaomi-mimo-v2.5-free": { - "id": "xiaomi-mimo-v2.5-free", - "name": "Xiaomi MiMo-V2.5 (free)", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo-v2.5", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", + "spacexai/grok-voice-think-fast-1.0": { + "id": "spacexai/grok-voice-think-fast-1.0", + "name": "Grok Voice Think Fast 1.0", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "audio" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "context": 0, + "output": 0 } }, - "claude-opus-4-7-think": { - "id": "claude-opus-4-7-think", - "name": "Claude Opus 4.7 Thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "spacexai/grok-4.20-non-reasoning": { + "id": "spacexai/grok-4.20-non-reasoning", + "name": "Grok 4.20 Non-Reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "release_date": "2026-03-10", + "last_updated": "2026-03-10", "modalities": { "input": [ "text", @@ -167677,198 +176512,243 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, "tiers": [ { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, + "input": 2.5, + "output": 5, + "cache_read": 0.4, "tier": { "type": "context", - "size": 200000 + "size": 200001 } } ], "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 + "input": 2.5, + "output": 5, + "cache_read": 0.4 } } }, - "coding-glm-5.1-free": { - "id": "coding-glm-5.1-free", - "name": "Coding GLM 5.1 (free)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm-free", + "spacexai/grok-stt": { + "id": "spacexai/grok-stt", + "name": "Grok STT", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "grok", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-11", - "last_updated": "2026-04-11", + "reasoning": false, + "tool_call": false, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 0, + "output": 0 + } + }, + "spacexai/grok-imagine-image": { + "id": "spacexai/grok-imagine-image", + "name": "Grok Imagine Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-01-28", + "last_updated": "2026-01-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] }, - "cost": { - "input": 0, + "open_weights": false, + "limit": { + "context": 0, "output": 0 } }, - "xiaomi-mimo-v2.5-pro-free": { - "id": "xiaomi-mimo-v2.5-pro-free", - "name": "Xiaomi MiMo-V2.5-Pro (free)", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo-v2.5-pro", + "spacexai/grok-imagine-video": { + "id": "spacexai/grok-imagine-video", + "name": "Grok Imagine", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-01-28", + "last_updated": "2026-01-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "video" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "spacexai/grok-4.1-fast-reasoning": { + "id": "spacexai/grok-4.1-fast-reasoning", + "name": "Grok 4.1 Fast Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "xiaomi-mimo-v2.5": { - "id": "xiaomi-mimo-v2.5", - "name": "Xiaomi MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo-v2.5", + "spacexai/grok-4.3": { + "id": "spacexai/grok-4.3", + "name": "Grok 4.3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.44, - "output": 2.2, - "cache_read": 0.088, + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, "tiers": [ { - "input": 0.88, - "output": 4.4, - "cache_read": 0.176, + "input": 2.5, + "output": 5, + "cache_read": 0.4, "tier": { "type": "context", - "size": 256000 + "size": 200001 } } ], "context_over_200k": { - "input": 0.88, - "output": 4.4, - "cache_read": 0.176 + "input": 2.5, + "output": 5, + "cache_read": 0.4 } } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "spacexai/grok-imagine-video-1.5": { + "id": "spacexai/grok-imagine-video-1.5", + "name": "Grok Imagine Video 1.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-05-30", + "last_updated": "2026-05-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "video" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "spacexai/grok-4.5": { + "id": "spacexai/grok-4.5", + "name": "Grok 4.5", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text", @@ -167881,103 +176761,98 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 500000, + "output": 500000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, + "input": 2, + "output": 6, + "cache_read": 0.3, "tiers": [ { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, + "input": 4, + "output": 12, + "cache_read": 0.6, "tier": { "type": "context", - "size": 200000 + "size": 200001 } } ], "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 + "input": 4, + "output": 12, + "cache_read": 0.6 } } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "spacexai/grok-voice-think-fast-2.0": { + "id": "spacexai/grok-voice-think-fast-2.0", + "name": "Grok Voice Think Fast 2.0", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "grok", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "reasoning": false, + "tool_call": false, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text", + "audio" + ] }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "spacexai/grok-imagine-image-2.0": { + "id": "spacexai/grok-imagine-image-2.0", + "name": "Grok Imagine Image 2.0", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-08-07", + "last_updated": "2026-08-07", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 1.1268, - "output": 3.9438, - "cache_read": 0.2817 + "context": 0, + "output": 0 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.6", + "spacexai/grok-4.20-multi-agent-beta": { + "id": "spacexai/grok-4.20-multi-agent-beta", + "name": "Grok 4.20 Multi Agent Beta", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -167985,115 +176860,91 @@ }, "open_weights": false, "limit": { - "context": 991000, - "output": 64000 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.17, - "output": 1.01, - "cache_read": 0.0169, - "cache_write": 0.21125, + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, "tiers": [ { - "input": 0.68, - "output": 4.06, - "cache_read": 0.0676, - "cache_write": 0.845, + "input": 2.5, + "output": 5, + "cache_read": 0.4, "tier": { "type": "context", - "size": 256000 + "size": 200001 } } ], "context_over_200k": { - "input": 0.68, - "output": 4.06, - "cache_read": 0.0676, - "cache_write": 0.845 + "input": 2.5, + "output": 5, + "cache_read": 0.4 } } }, - "coding-xiaomi-mimo-v2.5": { - "id": "coding-xiaomi-mimo-v2.5", - "name": "Coding Xiaomi MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo-v2.5", + "spacexai/grok-4.20-non-reasoning-beta": { + "id": "spacexai/grok-4.20-non-reasoning-beta", + "name": "Grok 4.20 Beta Non-Reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.08, - "output": 0.4, - "cache_read": 0.016, + "input": 1.25, + "output": 2.5, + "cache_read": 0.4, "tiers": [ { - "input": 0.16, - "output": 0.8, - "cache_read": 0.032, + "input": 2.5, + "output": 5, + "cache_read": 0.4, "tier": { "type": "context", - "size": 256000 + "size": 200001 } } ], "context_over_200k": { - "input": 0.16, - "output": 0.8, - "cache_read": 0.032 + "input": 2.5, + "output": 5, + "cache_read": 0.4 } } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "spacexai/grok-build-0.1": { + "id": "spacexai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Grok coding model for agentic engineering, edits, and codebase workflows", + "family": "grok-build", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -168105,48 +176956,46 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1, + "output": 2, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 4, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200001 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 4, + "cache_read": 0.4 + } } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "spacexai/grok-4.1-fast-non-reasoning": { + "id": "spacexai/grok-4.1-fast-non-reasoning", + "name": "Grok 4.1 Fast Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -168154,45 +177003,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "gemini-3.1-pro-preview-customtools": { - "id": "gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "spacexai/grok-4.20-reasoning-beta": { + "id": "spacexai/grok-4.20-reasoning-beta", + "name": "Grok 4.20 Beta Reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -168201,46 +177035,42 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 2, - "output": 12, + "input": 1.25, + "output": 2.5, "cache_read": 0.2, "tiers": [ { - "input": 4, - "output": 18, + "input": 2.5, + "output": 5, "cache_read": 0.4, "tier": { "type": "context", - "size": 200000 + "size": 200001 } } ], "context_over_200k": { - "input": 4, - "output": 18, + "input": 2.5, + "output": 5, "cache_read": 0.4 } } }, - "doubao-seed-2-0-mini-260428": { - "id": "doubao-seed-2-0-mini-260428", - "name": "Doubao Seed 2.0 Mini 260428", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "spacexai/grok-4.6": { + "id": "spacexai/grok-4.6", + "name": "Grok 4.6", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -168248,18 +177078,15 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -168267,219 +177094,308 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.03, - "output": 0.28, - "cache_read": 0.00564, - "input_audio": 0.423, + "input": 2, + "output": 6, + "cache_read": 0.5, "tiers": [ { - "input": 0.06, - "output": 0.56, - "cache_read": 0.01128, - "input_audio": 0.846, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.11, - "output": 1.13, - "cache_read": 0.02256, - "input_audio": 1.692, + "input": 4, + "output": 12, + "cache_read": 1, "tier": { "type": "context", - "size": 128000 + "size": 200001 } } - ] + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } }, - "doubao-seed-2-0-lite-260428": { - "id": "doubao-seed-2-0-lite-260428", - "name": "Doubao Seed 2.0 Lite 260428", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "prodia/flux-fast-schnell": { + "id": "prodia/flux-fast-schnell", + "name": "Flux Schnell", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-08-02", + "last_updated": "2026-06-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] }, - "structured_output": true, + "open_weights": false, + "limit": { + "context": 512, + "output": 0 + } + }, + "recraft/recraft-v2": { + "id": "recraft/recraft-v2", + "name": "Recraft V2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "release_date": "2024-03-13", + "last_updated": "2024-03", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 512, + "output": 0 + } + }, + "recraft/recraft-v4-pro": { + "id": "recraft/recraft-v4-pro", + "name": "Recraft V4 Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-02-17", + "last_updated": "2026-02-17", + "modalities": { + "input": [ "text" + ], + "output": [ + "image" ] }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 + "context": 0, + "output": 0 + } + }, + "recraft/recraft-v4.1-utility": { + "id": "recraft/recraft-v4.1-utility", + "name": "Recraft V4.1 Utility", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-05-14", + "last_updated": "2026-05-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] }, - "cost": { - "input": 0.08, - "output": 0.51, - "cache_read": 0.01692, - "input_audio": 1.269, - "tiers": [ - { - "input": 0.13, - "output": 0.76, - "cache_read": 0.02536, - "input_audio": 1.902, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.25, - "output": 1.52, - "cache_read": 0.05072, - "input_audio": 3.804, - "tier": { - "type": "context", - "size": 128000 - } - } + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "recraft/recraft-v4.1-pro": { + "id": "recraft/recraft-v4.1-pro", + "name": "Recraft V4.1 Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-05-14", + "last_updated": "2026-05-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "recraft/recraft-v3": { + "id": "recraft/recraft-v3", + "name": "Recraft V3", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-10-30", + "last_updated": "2024-10", "modalities": { "input": [ - "text", + "text" + ], + "output": [ "image" + ] + }, + "open_weights": false, + "limit": { + "context": 512, + "output": 0 + } + }, + "recraft/recraft-v4.1": { + "id": "recraft/recraft-v4.1", + "name": "Recraft V4.1", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-05-14", + "last_updated": "2026-05-14", + "modalities": { + "input": [ + "text" ], "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "recraft/recraft-v4.1-utility-pro": { + "id": "recraft/recraft-v4.1-utility-pro", + "name": "Recraft V4.1 Utility Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-05-14", + "last_updated": "2026-05-14", + "modalities": { + "input": [ "text" + ], + "output": [ + "image" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 0, + "output": 0 + } + }, + "recraft/recraft-v4": { + "id": "recraft/recraft-v4", + "name": "Recraft V4", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-02-17", + "last_updated": "2026-02-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "open_weights": false, + "limit": { + "context": 0, + "output": 0 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "poolside/laguna-s-2.1-free": { + "id": "poolside/laguna-s-2.1-free", + "name": "Laguna S 2.1 Free", + "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", + "family": "laguna", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0, + "output": 0 } }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "poolside/laguna-s-2.1": { + "id": "poolside/laguna-s-2.1", + "name": "Laguna S 2.1", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text" @@ -168490,19 +177406,18 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.1, + "output": 0.2, + "cache_read": 0.01 } }, - "claude-opus-4-6-think": { - "id": "claude-opus-4-6-think", - "name": "Claude Opus 4.6 Thinking", + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude Opus 4.5", "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, @@ -168511,29 +177426,17 @@ { "type": "toggle" }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, { "type": "budget_tokens", "min": 1024 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "interleaved": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", @@ -168546,42 +177449,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { "input": 5, "output": 25, "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "cache_write": 6.25 } }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -168589,18 +177477,23 @@ "medium", "high" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "interleaved": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -168608,156 +177501,169 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": true, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "coding-glm-5.1": { - "id": "coding-glm-5.1", - "name": "Coding GLM 5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "anthropic/claude-3-haiku": { + "id": "anthropic/claude-3-haiku", + "name": "Claude Haiku 3", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-haiku", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-11", - "last_updated": "2026-04-11", + "knowledge": "2023-08-31", + "release_date": "2024-03-13", + "last_updated": "2024-03-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 200000, - "output": 128000 + "output": 4096 }, "cost": { - "input": 0.06, - "output": 0.22, - "cache_read": 0.013 + "input": 0.25, + "output": 1.25, + "cache_read": 0.03, + "cache_write": 0.3 } }, - "coding-minimax-m2.7-free": { - "id": "coding-minimax-m2.7-free", - "name": "Coding MiniMax M2.7 (Free)", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax-free", - "attachment": false, + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 128100 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "anthropic/claude-opus-5.5-fast": { + "id": "anthropic/claude-opus-5.5-fast", + "name": "Claude Opus 5.5 (Fast)", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", @@ -168770,77 +177676,43 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 8, + "output": 40, + "cache_read": 0.4, + "cache_write": 10 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "interleaved": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -168848,79 +177720,67 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "anthropic/claude-fable-5.1": { + "id": "anthropic/claude-fable-5.1", + "name": "Claude Fable 5.1", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -168929,22 +177789,21 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -168953,36 +177812,21 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -168995,21 +177839,15 @@ "low", "medium", "high", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", @@ -169026,95 +177864,78 @@ "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "deep-deepseek-v4-pro": { - "id": "deep-deepseek-v4-pro", - "name": "DeepSeek V4 Pro (DeepSeek)", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", - "attachment": false, + "anthropic/claude-opus-4.8-fast": { + "id": "anthropic/claude-opus-4.8-fast", + "name": "Claude Opus 4.8 (Fast)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 128000 }, "cost": { - "input": 0.478, - "output": 0.956, - "cache_read": 0.004302 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM 5 Vision Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glmv", + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-05-09", - "last_updated": "2026-05-09", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -169122,58 +177943,86 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.7042, - "output": 3.09848, - "cache_read": 0.169008 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200001 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "zai-glm-5.1": { - "id": "zai-glm-5.1", - "name": "GLM-5.1 (Z.ai)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.845, - "output": 3.38, - "cache_read": 0.183112 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", "family": "claude-sonnet", "attachment": true, "reasoning": true, @@ -169186,8 +178035,7 @@ "values": [ "low", "medium", - "high", - "max" + "high" ] }, { @@ -169197,7 +178045,6 @@ ], "tool_call": true, "interleaved": true, - "structured_output": true, "temperature": true, "knowledge": "2025-08-31", "release_date": "2026-02-17", @@ -169215,7 +178062,7 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 128000 }, "cost": { "input": 3, @@ -169242,36 +178089,34 @@ } } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "anthropic/claude-opus-5.5": { + "id": "anthropic/claude-opus-5.5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -169280,76 +178125,24 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "tiers": [ - { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 - } - } - }, - "coding-minimax-m2.7-highspeed": { - "id": "coding-minimax-m2.7-highspeed", - "name": "Coding MiniMax M2.7 Highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 128100 - }, - "cost": { - "input": 0.2, - "output": 0.2 + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 } }, - "claude-opus-4-8-think": { - "id": "claude-opus-4-8-think", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "anthropic/claude-opus-5-fast": { + "id": "anthropic/claude-opus-5-fast", + "name": "Claude Opus 5 (Fast)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -169362,17 +178155,15 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -169380,40 +178171,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "qwen3.6-max-preview": { - "id": "qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "qwen3.6", - "attachment": false, + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-05-09", - "last_updated": "2026-05-09", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -169421,52 +178206,35 @@ }, "open_weights": false, "limit": { - "context": 240000, - "output": 64000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1.27, - "output": 7.61, - "cache_read": 0.1268, - "cache_write": 1.585, - "tiers": [ - { - "input": 2.11, - "output": 12.67, - "cache_read": 0.2112, - "cache_write": 2.64, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", @@ -169479,42 +178247,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "cohere/rerank-v4-fast": { + "id": "cohere/rerank-v4-fast", + "name": "Cohere Rerank 4 Fast", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-05-09", - "last_updated": "2026-05-09", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -169522,55 +178277,20 @@ }, "open_weights": false, "limit": { - "context": 991000, - "output": 64000 - }, - "cost": { - "input": 0.28, - "output": 1.69, - "cache_read": 0.0282, - "cache_write": 0.3525, - "tiers": [ - { - "input": 1.13, - "output": 6.77, - "cache_read": 0.1128, - "cache_write": 1.41, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.13, - "output": 6.77, - "cache_read": 0.1128, - "cache_write": 1.41 - } + "context": 32000, + "output": 32000 } }, - "alicloud-deepseek-v4-pro": { - "id": "alicloud-deepseek-v4-pro", - "name": "DeepSeek V4 Pro (Alibaba Cloud)", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", + "cohere/rerank-v4-pro": { + "id": "cohere/rerank-v4-pro", + "name": "Cohere Rerank 4 Pro", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text" @@ -169579,45 +178299,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.69, - "output": 3.38, - "cache_read": 0.13 + "context": 32000, + "output": 32000 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "cohere/command-a": { + "id": "cohere/command-a", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -169625,46 +178327,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "output": 8000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 2.5, + "output": 10 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "cohere/embed-v4.0": { + "id": "cohere/embed-v4.0", + "name": "Embed v4.0", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -169672,347 +178356,252 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "context": 128000, + "output": 1536 } - } - } - }, - "google-vertex": { - "id": "google-vertex", - "env": [ - "GOOGLE_VERTEX_PROJECT", - "GOOGLE_VERTEX_LOCATION", - "GOOGLE_APPLICATION_CREDENTIALS" - ], - "npm": "@ai-sdk/google-vertex", - "name": "Vertex", - "doc": "https://cloud.google.com/vertex-ai/generative-ai/docs/models", - "models": { - "gemini-2.5-pro-tts": { - "id": "gemini-2.5-pro-tts", - "name": "Gemini 2.5 Pro TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gemini-pro", + }, + "cohere/rerank-v3.5": { + "id": "cohere/rerank-v3.5", + "name": "Cohere Rerank 3.5", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-09-30", - "last_updated": "2025-12-10", + "temperature": true, + "release_date": "2024-12-02", + "last_updated": "2024-12-02", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 1, - "output": 20 + "context": 4096, + "output": 4096 } }, - "claude-haiku-4-5@20251001": { - "id": "claude-haiku-4-5@20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "deepseek/deepseek-v4-flash-0731": { + "id": "deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.076, + "output": 0.153, + "cache_read": 0.014 } }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "deepseek/deepseek-v4.1-flash": { + "id": "deepseek/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", - "low", - "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 32768 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.007 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "deepseek/deepseek-v4-flash-vision-exp": { + "id": "deepseek/deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 1048576 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 } }, - "gemini-2.5-flash-tts": { - "id": "gemini-2.5-flash-tts", - "name": "Gemini 2.5 Flash TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gemini-flash", + "deepseek/deepseek-v3.1-terminus": { + "id": "deepseek/deepseek-v3.1-terminus", + "name": "DeepSeek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-09-30", - "last_updated": "2025-12-10", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 16384 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 10 + "input": 0.27, + "output": 1, + "cache_read": 0.135 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "deepseek/deepseek-v4-pro-0813": { + "id": "deepseek/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens", - "min": 0, - "max": 24576 + "type": "effort", + "values": [ + "high", + "xhigh" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075, - "cache_write": 0.383 + "input": 0.66, + "output": 1.98, + "cache_read": 0.066 } }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "deepseek/deepseek-v3.2-thinking": { + "id": "deepseek/deepseek-v3.2-thinking", + "name": "DeepSeek V3.2 Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -170020,116 +178609,106 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 8000 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 + "input": 0.62, + "output": 1.85 } }, - "claude-opus-4@20250514": { - "id": "claude-opus-4@20250514", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek/deepseek-v3.1": { + "id": "deepseek/deepseek-v3.1", + "name": "DeepSeek-V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" + "context": 163840, + "output": 128000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.25, + "output": 0.95, + "cache_read": 0.13 } }, - "claude-opus-4-1@20250805": { - "id": "claude-opus-4-1@20250805", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" + "context": 1000000, + "output": 384000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.66, + "output": 1.98, + "cache_read": 0.022 } }, - "gemini-embedding-001": { - "id": "gemini-embedding-001", - "name": "Gemini Embedding 001", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "gemini", + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "knowledge": "2025-05", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "structured_output": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -170138,455 +178717,648 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2048, - "output": 1 + "context": 128000, + "output": 8000 }, "cost": { - "input": 0.15, - "output": 0 + "input": 0.62, + "output": 1.85 } }, - "claude-opus-4-5@20251101": { - "id": "claude-opus-4-5@20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" + "context": 1000000, + "output": 384000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.13, + "output": 0.26, + "cache_read": 0.028 } }, - "claude-3-5-haiku@20241022": { - "id": "claude-3-5-haiku@20241022", - "name": "Claude Haiku 3.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": false, - "tool_call": true, + "deepseek/deepseek-r1": { + "id": "deepseek/deepseek-r1", + "name": "DeepSeek-R1", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "input": 1.35, + "output": 5.4 } }, - "gemini-3.1-pro-preview-customtools": { - "id": "gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "quiverai/arrow-1.1": { + "id": "quiverai/arrow-1.1", + "name": "Arrow 1.1", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "context": 131072, + "output": 131072 } }, - "gemini-flash-lite-latest": { - "id": "gemini-flash-lite-latest", - "name": "Gemini Flash-Lite Latest", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "quiverai/arrow-2": { + "id": "quiverai/arrow-2", + "name": "Arrow 2", + "description": "Fast SVG generation model for creation, vectorization, editing, and animation", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "release_date": "2026-09-16", + "last_updated": "2026-09-16", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5 } }, - "claude-sonnet-4@20250514": { - "id": "claude-sonnet-4@20250514", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "quiverai/arrow-2-telos": { + "id": "quiverai/arrow-2-telos", + "name": "Arrow 2 Telos", + "description": "High-fidelity SVG generation model for complex vector work and long-context refinement", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-09-16", + "last_updated": "2026-09-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" + "context": 131072, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 - } - ], - "tool_call": true, + "input": 6, + "output": 30, + "cache_read": 0.6, + "cache_write": 7.5 + } + }, + "bfl/flux-kontext-max": { + "id": "bfl/flux-kontext-max", + "name": "FLUX.1 Kontext Max", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2025-05-29", + "last_updated": "2025-06", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 512, + "output": 0 + } + }, + "bfl/flux-pro-1.1-ultra": { + "id": "bfl/flux-pro-1.1-ultra", + "name": "FLUX1.1 [pro] Ultra", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-11-01", + "last_updated": "2024-11", + "modalities": { + "input": [ + "text" ], "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 512, + "output": 0 + } + }, + "bfl/flux-3-video": { + "id": "bfl/flux-3-video", + "name": "Flux 3", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-08-04", + "last_updated": "2026-08-04", + "modalities": { + "input": [ "text" + ], + "output": [ + "video" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 0, + "output": 0 + } + }, + "bfl/flux-2-flex": { + "id": "bfl/flux-2-flex", + "name": "FLUX.2 [flex]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-11-25", + "last_updated": "2026-06-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "bfl/flux-2-klein-4b": { + "id": "bfl/flux-2-klein-4b", + "name": "FLUX.2 [klein] 4B", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-01-15", + "last_updated": "2026-06-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "bfl/flux-pro-1.0-fill": { + "id": "bfl/flux-pro-1.0-fill", + "name": "FLUX.1 Fill [pro]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-10-01", + "last_updated": "2024-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 512, + "output": 0 + } + }, + "bfl/flux-2-klein-9b": { + "id": "bfl/flux-2-klein-9b", + "name": "FLUX.2 [klein] 9B", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-01-15", + "last_updated": "2026-06-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "bfl/flux-2-pro": { + "id": "bfl/flux-2-pro", + "name": "FLUX.2 [pro]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-11-25", + "last_updated": "2026-06-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 67300, + "output": 67300 + } + }, + "bfl/flux-2-max": { + "id": "bfl/flux-2-max", + "name": "FLUX.2 [max]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-12-16", + "last_updated": "2026-06-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 67300, + "output": 67300 + } + }, + "bfl/flux-pro-1.1": { + "id": "bfl/flux-pro-1.1", + "name": "FLUX1.1 [pro]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-10-02", + "last_updated": "2024-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 512, + "output": 0 + } + }, + "bfl/flux-kontext-pro": { + "id": "bfl/flux-kontext-pro", + "name": "FLUX.1 Kontext Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-05-29", + "last_updated": "2025-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 512, + "output": 0 + } + }, + "tencent/hy-mt2-lite": { + "id": "tencent/hy-mt2-lite", + "name": "Tencent Hy-MT2-Lite", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8000, + "output": 4000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 + "input": 0.044, + "output": 0.177 } }, - "claude-opus-4-7@default": { - "id": "claude-opus-4-7@default", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "tencent/hy3": { + "id": "tencent/hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" + "cost": { + "input": 0.14, + "output": 0.58, + "cache_read": 0.035 + } + }, + "tencent/hy-mt2-plus": { + "id": "tencent/hy-mt2-plus", + "name": "Tencent Hy-MT2-Plus", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8000, + "output": 4000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } + "input": 0.074, + "output": 0.295 + } + }, + "tencent/hy-mt2-pro": { + "id": "tencent/hy-mt2-pro", + "name": "Tencent Hy-MT2-Pro", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8000, + "output": 4000 + }, + "cost": { + "input": 0.074, + "output": 0.295 } }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, + "tencent/hy4-preview": { + "id": "tencent/hy4-preview", + "name": "Tencent Hy4 Preview", + "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-08-28", + "last_updated": "2026-08-28", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 1024000, + "output": 64000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.834, + "output": 2.501, + "cache_read": 0.042 } }, - "claude-sonnet-4-5@20250929": { - "id": "claude-sonnet-4-5@20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "inference-net/schematron-v2-small": { + "id": "inference-net/schematron-v2-small", + "name": "Schematron V2 Small", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -170594,33 +179366,56 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 4096 }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" + "cost": { + "input": 0.05, + "output": 0.23, + "cache_read": 0.05 + } + }, + "inference-net/schematron-v2-turbo": { + "id": "inference-net/schematron-v2-turbo", + "name": "Schematron V2 Turbo", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.03, + "output": 0.15, + "cache_read": 0.03 } }, - "claude-sonnet-5@default": { - "id": "claude-sonnet-5@default", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "thinkingmachines/inkling-small": { + "id": "thinkingmachines/inkling-small", + "name": "Inkling Small", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high", @@ -170630,10 +179425,9 @@ } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", @@ -170644,101 +179438,102 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" + "output": 1000000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.5, + "output": 1.2, + "cache_read": 0.1 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "thinkingmachines/inkling": { + "id": "thinkingmachines/inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 + "input": 1, + "output": 4.05, + "cache_read": 0.17 } }, - "claude-opus-4-6@default": { - "id": "claude-opus-4-6@default", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "perplexity/pplx-embed-v1-4b": { + "id": "perplexity/pplx-embed-v1-4b", + "name": "Embed v1 4b", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-02-26", + "last_updated": "2026-02-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 0 + } + }, + "perplexity/sonar-pro": { + "id": "perplexity/sonar-pro", + "name": "Sonar Pro", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2025-09", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -170746,43 +179541,67 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } + "context": 200000, + "output": 8000 + } + }, + "perplexity/pplx-embed-v1-0.6b": { + "id": "perplexity/pplx-embed-v1-0.6b", + "name": "Embed v1 0.6b", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "v0", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-02-26", + "last_updated": "2026-02-26", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 0 } }, - "gemini-flash-latest": { - "id": "gemini-flash-latest", - "name": "Gemini Flash Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "perplexity/sonar": { + "id": "perplexity/sonar", + "name": "Sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar", "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-02", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 127000, + "output": 8000 + } + }, + "perplexity/sonar-reasoning-pro": { + "id": "perplexity/sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded reasoning model for multi-step research and cited answers", + "family": "sonar-reasoning", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -170795,18 +179614,15 @@ ] } ], - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "knowledge": "2025-09", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -170814,27 +179630,22 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075, - "cache_write": 0.383 + "context": 127000, + "output": 8000 } }, - "claude-opus-4-8@default": { - "id": "claude-opus-4-8@default", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "meta/muse-spark-1.3": { + "id": "meta/muse-spark-1.3", + "name": "Muse Spark 1.3", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high", @@ -170844,10 +179655,10 @@ } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", @@ -170860,42 +179671,20 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "claude-sonnet-4-6@default": { - "id": "claude-sonnet-4-6@default", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "meta/muse-glimmer-30b": { + "id": "meta/muse-glimmer-30b", + "name": "Muse Glimmer 30B", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -170905,68 +179694,42 @@ "low", "medium", "high", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" + "context": 131072, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 0.35, + "output": 1.5, + "cache_read": 0.04 } }, - "gemini-3.1-flash-lite-preview": { - "id": "gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "meta/muse-spark-1.1": { + "id": "meta/muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "muse", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -170975,22 +179738,20 @@ "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2026-04-08", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -171000,33 +179761,26 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65536 + "output": 1048576 }, - "status": "deprecated", "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "moonshotai/kimi-k2-thinking-maas": { - "id": "moonshotai/kimi-k2-thinking-maas", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "meta/llama-3.1-8b": { + "id": "meta/llama-3.1-8b", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -171035,41 +179789,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 2.5 + "input": 0.22, + "output": 0.22 } }, - "openai/gpt-oss-120b-maas": { - "id": "openai/gpt-oss-120b-maas", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "meta/llama-3.1-70b": { + "id": "meta/llama-3.1-70b", + "name": "Llama 3.1 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -171078,227 +179819,173 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.09, - "output": 0.36 + "input": 0.72, + "output": 0.72 } }, - "openai/gpt-oss-20b-maas": { - "id": "openai/gpt-oss-20b-maas", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "meta/muse-spark-1.2": { + "id": "meta/muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "muse", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.07, - "output": 0.25 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "zai-org/glm-4.7-maas": { - "id": "zai-org/glm-4.7-maas", - "name": "GLM-4.7", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-06", - "last_updated": "2026-01-06", - "modalities": { - "input": [ + "meta/muse-image-1.0": { + "id": "meta/muse-image-1.0", + "name": "Muse Image 1.0", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "muse", + "attachment": true, + "reasoning": false, + "tool_call": false, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ "text", - "pdf" + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.6, - "output": 2.2 + "context": 0, + "output": 0 } }, - "zai-org/glm-5-maas": { - "id": "zai-org/glm-5-maas", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "meta/muse-spark-1.2-contributor": { + "id": "meta/muse-spark-1.2-contributor", + "name": "Muse Spark 1.2 Contributor", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.1 - } - }, - "deepseek-ai/deepseek-v3.1-maas": { - "id": "deepseek-ai/deepseek-v3.1-maas", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 32768 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.6, - "output": 1.7 + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 } }, - "deepseek-ai/deepseek-v3.2-maas": { - "id": "deepseek-ai/deepseek-v3.2-maas", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "meta/muse-spark-1.3-contributor": { + "id": "meta/muse-spark-1.3-contributor", + "name": "Muse Spark 1.3 Contributor", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "muse", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-17", - "last_updated": "2026-04-04", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.56, - "output": 1.68, - "cache_read": 0.056 + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 } }, - "qwen/qwen3-235b-a22b-instruct-2507-maas": { - "id": "qwen/qwen3-235b-a22b-instruct-2507-maas", - "name": "Qwen3 235B A22B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "meta/llama-3.3-70b": { + "id": "meta/llama-3.3-70b", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-13", - "last_updated": "2025-08-13", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -171309,34 +179996,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 16384 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.22, - "output": 0.88 + "input": 0, + "output": 0 } }, - "meta/llama-3.3-70b-instruct-maas": { - "id": "meta/llama-3.3-70b-instruct-maas", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "meta/llama-4-maverick": { + "id": "meta/llama-4-maverick", + "name": "Llama-4-Maverick-17B-128E-Instruct-FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", "family": "llama", - "attachment": false, + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -171345,30 +180028,25 @@ "open_weights": true, "limit": { "context": 128000, - "output": 8192 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + "output": 4096 }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 0, + "output": 0 } }, - "meta/llama-4-maverick-17b-128e-instruct-maas": { - "id": "meta/llama-4-maverick-17b-128e-instruct-maas", - "name": "Llama 4 Maverick 17B 128E Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", + "meta/llama-4-scout": { + "id": "meta/llama-4-scout", + "name": "Llama-4-Scout-17B-16E-Instruct-FP8", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", "family": "llama", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2024-08", - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -171380,82 +180058,71 @@ }, "open_weights": true, "limit": { - "context": 524288, - "output": 8192 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.35, - "output": 1.15 + "input": 0, + "output": 0 } - } - } - }, - "nano-gpt": { - "id": "nano-gpt", - "env": [ - "NANO_GPT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://nano-gpt.com/api/v1", - "name": "NanoGPT", - "doc": "https://docs.nano-gpt.com", - "models": { - "step-3": { - "id": "step-3", - "name": "Step-3", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-31", - "last_updated": "2025-07-31", + }, + "arcee-ai/trinity-large-thinking": { + "id": "arcee-ai/trinity-large-thinking", + "name": "Trinity Large Thinking", + "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", + "family": "trinity", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "input": 65536, - "output": 8192 + "context": 262100, + "output": 80000 }, "cost": { - "input": 0.2499, - "output": 0.6494 + "input": 0.25, + "output": 0.8999999999999999 } }, - "qwen3.5-35b-a3b:thinking": { - "id": "qwen3.5-35b-a3b:thinking", - "name": "Qwen3.5 35B A3B Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.5-flash": { + "id": "alibaba/qwen3.5-flash", + "name": "Qwen 3.5 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "budget_tokens", - "min": 1024, + "min": 1, "max": 81920 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -171463,59 +180130,41 @@ }, "open_weights": false, "limit": { - "context": 260096, - "input": 260096, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.225, - "output": 1.8 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0.125 } }, - "glm-4.1v-thinking-flashx": { - "id": "glm-4.1v-thinking-flashx", - "name": "GLM 4.1V Thinking FlashX", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "alibaba/qwen-3-32b": { + "id": "alibaba/qwen-3-32b", + "name": "Qwen 3.32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 38912 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 64000, - "input": 64000, - "output": 8192 - }, - "cost": { - "input": 0.3, - "output": 0.3 - } - }, - "ernie-x1.1-preview": { - "id": "ernie-x1.1-preview", - "name": "ERNIE X1.1", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-10", - "last_updated": "2025-09-10", - "modalities": { - "input": [ - "text", - "pdf" ], "output": [ "text" @@ -171523,55 +180172,51 @@ }, "open_weights": false, "limit": { - "context": 64000, - "input": 64000, + "context": 128000, "output": 8192 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.16, + "output": 0.64 } }, - "qwen25-vl-72b-instruct": { - "id": "qwen25-vl-72b-instruct", - "name": "Qwen25 VL 72b", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "alibaba/wan-v2.6-t2v": { + "id": "alibaba/wan-v2.6-t2v", + "name": "Wan v2.6 Text-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-05-10", - "last_updated": "2025-05-10", + "temperature": true, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 32768 - }, - "cost": { - "input": 0.69989, - "output": 0.69989 + "context": 0, + "output": 0 } }, - "gemini-2.0-pro-exp-02-05": { - "id": "gemini-2.0-pro-exp-02-05", - "name": "Gemini 2.0 Pro 0205", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3-vl-235b-a22b-instruct": { + "id": "alibaba/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-02-05", - "last_updated": "2025-02-05", + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", @@ -171581,50 +180226,21 @@ "text" ] }, - "open_weights": false, - "limit": { - "context": 2097152, - "input": 2097152, - "output": 8192 - }, - "cost": { - "input": 1.989, - "output": 7.956 - } - }, - "doubao-seed-2-0-lite-260215": { - "id": "doubao-seed-2-0-lite-260215", - "name": "Doubao Seed 2.0 Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 32000 + "context": 131072, + "output": 129024 }, "cost": { - "input": 0.1462, - "output": 0.8738 + "input": 0.4, + "output": 1.6 } }, - "Qwen3.5-27B-Writer-V2-Derestricted": { - "id": "Qwen3.5-27B-Writer-V2-Derestricted", - "name": "Qwen3.5 27B Writer V2 Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.7-max": { + "id": "alibaba/qwen3.7-max", + "name": "Qwen 3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -171633,19 +180249,17 @@ }, { "type": "budget_tokens", - "min": 1024, - "max": 81920 + "min": 1, + "max": 262144 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-06", - "last_updated": "2026-04-06", + "tool_call": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -171653,26 +180267,39 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 991000, + "output": 64000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "claude-opus-4-thinking:8192": { - "id": "claude-opus-4-thinking:8192", - "name": "Claude 4 Opus Thinking (8K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.8-27b": { + "id": "alibaba/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", @@ -171683,128 +180310,140 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 14.994, - "output": 75.004 + "input": 0.5, + "output": 3, + "cache_read": 0.1, + "cache_write": 0.625 } }, - "qwen3-vl-235b-a22b-thinking": { - "id": "qwen3-vl-235b-a22b-thinking", - "name": "Qwen3 VL 235B A22B Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3-vl-thinking": { + "id": "alibaba/qwen3-vl-thinking", + "name": "Qwen3 VL Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "budget_tokens", - "min": 1024, + "min": 1, "max": 81920 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "tool_call": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-09-23", + "last_updated": "2025-09-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "input": 32768, + "context": 131072, "output": 32768 }, "cost": { - "input": 0.5, - "output": 6 + "input": 0.4, + "output": 4 } }, - "glm-4-air-0111": { - "id": "glm-4-air-0111", - "name": "GLM 4 Air 0111", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.8-2.4t-a95b": { + "id": "alibaba/qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-11", - "last_updated": "2025-01-11", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 262144, + "output": 128000 }, "cost": { - "input": 0.1394, - "output": 0.1394 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "Qwen3.5-27B-Queen-Derestricted": { - "id": "Qwen3.5-27B-Queen-Derestricted", - "name": "Qwen3.5 27B Queen Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "alibaba/qwen3-next-80b-a3b-thinking": { + "id": "alibaba/qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B Thinking", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "budget_tokens", - "min": 1024, - "max": 81920 + "min": 1 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "input": 262144, - "output": 16384 + "output": 262144 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.15, + "output": 1.2 } }, - "gemini-2.5-pro-preview-03-25": { - "id": "gemini-2.5-pro-preview-03-25", - "name": "Gemini 2.5 Pro Preview 0325", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.8-max": { + "id": "alibaba/qwen3.8-max", + "name": "Qwen 3.8 Max", + "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -171813,14 +180452,19 @@ "values": [ "low", "medium", - "high" + "xhigh" ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "tool_call": true, + "temperature": true, + "release_date": "2026-07-19", + "last_updated": "2026-07-19", "modalities": { "input": [ "text", @@ -171832,25 +180476,29 @@ }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 262144, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "brave-research": { - "id": "brave-research", - "name": "Brave (Research)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3-coder-next": { + "id": "alibaba/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2023-03-02", - "last_updated": "2024-01-01", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ "text" @@ -171859,66 +180507,60 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 256000, + "output": 256000 }, "cost": { - "input": 5, - "output": 5 + "input": 0.5, + "output": 1.2 } }, - "qwen-plus": { - "id": "qwen-plus", - "name": "Qwen Plus", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 - } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "alibaba/qwen3-vl-instruct": { + "id": "alibaba/qwen3-vl-instruct", + "name": "Qwen3 VL Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 995904, - "input": 995904, - "output": 32768 + "context": 131072, + "output": 129024 }, "cost": { - "input": 0.3995, - "output": 1.2002 + "input": 0.4, + "output": 1.6 } }, - "doubao-seed-2-0-mini-260215": { - "id": "doubao-seed-2-0-mini-260215", - "name": "Doubao Seed 2.0 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3-coder": { + "id": "alibaba/qwen3-coder", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-22", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -171929,19 +180571,40 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.0493, - "output": 0.4845 + "input": 1.5, + "output": 7.5, + "cache_read": 0.3, + "tiers": [ + { + "input": 2.7, + "output": 13.5, + "cache_read": 0.54, + "tier": { + "type": "context", + "size": 32001 + } + }, + { + "input": 4.5, + "output": 22.5, + "cache_read": 0.9, + "tier": { + "type": "context", + "size": 128001 + } + } + ] } }, - "Qwen3.5-27B-BlueStar-v3-Derestricted-Lite": { - "id": "Qwen3.5-27B-BlueStar-v3-Derestricted-Lite", - "name": "Qwen3.5 27B BlueStar v3 Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.5-plus": { + "id": "alibaba/qwen3.5-plus", + "name": "Qwen 3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -171950,19 +180613,20 @@ }, { "type": "budget_tokens", - "min": 1024, + "min": 1, "max": 81920 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -171970,149 +180634,152 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.306, - "output": 0.306 - } - }, - "deepseek-v3-0324": { - "id": "deepseek-v3-0324", - "name": "DeepSeek Chat 0324", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2025-03-24", - "last_updated": "2025-03-24", - "modalities": { - "input": [ - "text" + "input": 0.4, + "output": 2.5, + "cache_read": 0.04, + "cache_write": 0.5, + "tiers": [ + { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tier": { + "type": "context", + "size": 256001 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "input": 128000, - "output": 8192 - }, - "cost": { - "input": 0.25, - "output": 0.7 + "context_over_200k": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625 + } } }, - "Gemma-4-31B-Musica-v1": { - "id": "Gemma-4-31B-Musica-v1", - "name": "Gemma 4 31B Musica v1", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen-3.6-max-preview": { + "id": "alibaba/qwen-3.6-max-preview", + "name": "Qwen 3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 131072 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "tool_call": true, + "temperature": true, + "release_date": "2026-04-20", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 240000, + "output": 64000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 1.3, + "output": 7.8, + "cache_read": 0.13, + "cache_write": 1.625, + "tiers": [ + { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "mirothinker-1-7-deepresearch-mini": { - "id": "mirothinker-1-7-deepresearch-mini", - "name": "MiroThinker 1.7 Deep Research Mini", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "alibaba/wan-v2.6-i2v-flash": { + "id": "alibaba/wan-v2.6-i2v-flash", + "name": "Wan v2.6 Image-to-Video Flash", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "temperature": true, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 1.25, - "output": 10 + "context": 0, + "output": 0 } }, - "Baichuan4-Turbo": { - "id": "Baichuan4-Turbo", - "name": "Baichuan 4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/wan-v2.6-r2v-flash": { + "id": "alibaba/wan-v2.6-r2v-flash", + "name": "Wan v2.6 Reference-to-Video Flash", + "description": "Video model for prompt-guided generation, editing, and motion workflows", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-08-19", - "last_updated": "2025-08-19", + "temperature": true, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 32768 - }, - "cost": { - "input": 2.42, - "output": 2.42 + "context": 0, + "output": 0 } }, - "doubao-seed-1-6-flash-250615": { - "id": "doubao-seed-1-6-flash-250615", - "name": "Doubao Seed 1.6 Flash", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen-3-14b": { + "id": "alibaba/qwen-3-14b", + "name": "Qwen3-14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-15", - "last_updated": "2025-06-15", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -172123,28 +180790,31 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, + "context": 40960, "output": 16384 }, "cost": { - "input": 0.0374, - "output": 0.374 + "input": 0.12, + "output": 0.24 } }, - "kimi-k2-instruct-fast": { - "id": "kimi-k2-instruct-fast", - "name": "Kimi K2 0711 Fast", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.7-flash": { + "id": "alibaba/qwen3.7-flash", + "name": "Qwen 3.7 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ @@ -172153,27 +180823,52 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 16384 + "context": 991000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 2 + "input": 0.03, + "output": 0.13, + "cache_read": 0.006, + "cache_write": 0.038, + "tiers": [ + { + "input": 0.1, + "output": 0.4, + "cache_read": 0.02, + "cache_write": 0.125, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.2, + "output": 0.8, + "cache_read": 0.04, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 256000 + } + } + ] } }, - "glm-4-plus": { - "id": "glm-4-plus", - "name": "GLM-4 Plus", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3-max": { + "id": "alibaba/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-08-01", - "last_updated": "2024-08-01", - "modalities": { - "input": [ + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ "text" ], "output": [ @@ -172182,25 +180877,46 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 262144, + "output": 32768 }, "cost": { - "input": 7.497, - "output": 7.497 + "input": 1.2, + "output": 6, + "cache_read": 0.24, + "tiers": [ + { + "input": 2.4, + "output": 12, + "cache_read": 0.48, + "tier": { + "type": "context", + "size": 32001 + } + }, + { + "input": 3, + "output": 15, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 128001 + } + } + ] } }, - "v0-1.0-md": { - "id": "v0-1.0-md", - "name": "v0 1.0 MD", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3-embedding-0.6b": { + "id": "alibaba/qwen3-embedding-0.6b", + "name": "Qwen3 Embedding 0.6B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-07-04", - "last_updated": "2025-07-04", + "temperature": true, + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text" @@ -172211,25 +180927,27 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15 + "context": 32768, + "output": 32768 } }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3 Coder 30B A3B Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen-3-30b": { + "id": "alibaba/qwen-3-30b", + "name": "Qwen3-30B-A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -172240,31 +180958,42 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 40960, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 0.12, + "output": 0.5 } }, - "gemini-2.5-flash-preview-09-2025-thinking": { - "id": "gemini-2.5-flash-preview-09-2025-thinking", - "name": "Gemini 2.5 Flash Preview (09/2025) – Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.8-omni-flash": { + "id": "alibaba/qwen3.8-omni-flash", + "name": "Qwen 3.8 Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "release_date": "2026-09-17", + "last_updated": "2026-09-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -172272,25 +181001,28 @@ }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 0.15, + "output": 0.47, + "cache_read": 0.016 } }, - "v0-1.5-lg": { - "id": "v0-1.5-lg", - "name": "v0 1.5 LG", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3-coder-30b-a3b": { + "id": "alibaba/qwen3-coder-30b-a3b", + "name": "Qwen 3 Coder 30B A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-04", - "last_updated": "2025-07-04", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-31", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -172301,37 +181033,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 64000 + "context": 262144, + "output": 8192 }, "cost": { - "input": 15, - "output": 75 + "input": 0.15, + "output": 0.6 } }, - "qwen3.5-flash:thinking": { - "id": "qwen3.5-flash:thinking", - "name": "Qwen3.5 Flash Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 - } - ], + "alibaba/qwen3-embedding-4b": { + "id": "alibaba/qwen3-embedding-4b", + "name": "Qwen3 Embedding 4B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "temperature": true, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -172339,25 +181062,22 @@ }, "open_weights": false, "limit": { - "context": 991808, - "input": 991808, - "output": 65536 - }, - "cost": { - "input": 0.09, - "output": 0.36 + "context": 32768, + "output": 32768 } }, - "jamba-large": { - "id": "jamba-large", - "name": "Jamba Large", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3-max-preview": { + "id": "alibaba/qwen3-max-preview", + "name": "Qwen3 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-05", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" @@ -172368,159 +181088,183 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 4096 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.989, - "output": 7.99 + "input": 1.2, + "output": 6, + "cache_read": 0.24, + "tiers": [ + { + "input": 2.4, + "output": 12, + "cache_read": 0.48, + "tier": { + "type": "context", + "size": 32001 + } + }, + { + "input": 3, + "output": 15, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 128001 + } + } + ] } }, - "kimi-thinking-preview": { - "id": "kimi-thinking-preview", - "name": "Kimi Thinking Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "alibaba/wan-v2.6-i2v": { + "id": "alibaba/wan-v2.6-i2v", + "name": "Wan v2.6 Image-to-Video", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "temperature": true, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 - }, - "cost": { - "input": 31.46, - "output": 31.46 + "context": 0, + "output": 0 } }, - "gemini-2.0-flash-thinking-exp-01-21": { - "id": "gemini-2.0-flash-thinking-exp-01-21", - "name": "Gemini 2.0 Flash Thinking 0121", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-21", - "last_updated": "2025-01-21", + "alibaba/qwen3-next-80b-a3b-instruct": { + "id": "alibaba/qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 8192 + "context": 262114, + "output": 262114 }, "cost": { - "input": 0.306, - "output": 1.003 + "input": 0.15, + "output": 1.2 } }, - "claude-opus-4-1-thinking:1024": { - "id": "claude-opus-4-1-thinking:1024", - "name": "Claude 4.1 Opus Thinking (1K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "alibaba/wan-v2.5-t2v-preview": { + "id": "alibaba/wan-v2.5-t2v-preview", + "name": "Wan v2.5 Text-to-Video Preview", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 - }, - "cost": { - "input": 14.994, - "output": 75.004 + "context": 0, + "output": 0 } }, - "Qwen3.5-27B-NaNovel-Derestricted-Lite": { - "id": "Qwen3.5-27B-NaNovel-Derestricted-Lite", - "name": "Qwen3.5 27B NaNovel Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "alibaba/qwen3-max-thinking": { + "id": "alibaba/qwen3-max-thinking", + "name": "Qwen 3 Max Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "budget_tokens", - "min": 1024, + "min": 1, "max": 81920 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-23", + "last_updated": "2025-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 1.2, + "output": 6, + "cache_read": 0.24, + "tiers": [ + { + "input": 2.4, + "output": 12, + "cache_read": 0.48, + "tier": { + "type": "context", + "size": 32001 + } + }, + { + "input": 3, + "output": 15, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 128001 + } + } + ] } }, - "doubao-seed-1-6-250615": { - "id": "doubao-seed-1-6-250615", - "name": "Doubao Seed 1.6", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3-coder-plus": { + "id": "alibaba/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-15", - "last_updated": "2025-06-15", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -172531,61 +181275,68 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.204, - "output": 0.51 - } - }, - "ernie-5.0-thinking-preview": { - "id": "ernie-5.0-thinking-preview", - "name": "Ernie 5.0 Thinking Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-11-18", - "last_updated": "2025-11-18", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" + "input": 1, + "output": 5, + "cache_read": 0.2, + "tiers": [ + { + "input": 1.8, + "output": 9, + "cache_read": 0.36, + "tier": { + "type": "context", + "size": 32001 + } + }, + { + "input": 3, + "output": 15, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 128001 + } + }, + { + "input": 6, + "output": 60, + "cache_read": 1.2, + "tier": { + "type": "context", + "size": 256001 + } + } ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "input": 128000, - "output": 16384 - }, - "cost": { - "input": 1.1, - "output": 2 } }, - "gemini-2.5-flash-preview-05-20:thinking": { - "id": "gemini-2.5-flash-preview-05-20:thinking", - "name": "Gemini 2.5 Flash 0520 Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3-235b-a22b-thinking": { + "id": "alibaba/qwen3-235b-a22b-thinking", + "name": "Qwen3 235B A22B Thinking 2507", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1, + "max": 81920 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-04", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -172593,19 +181344,19 @@ }, "open_weights": false, "limit": { - "context": 1048000, - "input": 1048000, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 3.5 + "input": 0.4, + "output": 4 } }, - "Qwen3.5-27B-Omega-Evolution-v2.2-Derestricted": { - "id": "Qwen3.5-27B-Omega-Evolution-v2.2-Derestricted", - "name": "Qwen3.5 27B Omega Evolution v2.2 Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.8-flash": { + "id": "alibaba/qwen3.8-flash", + "name": "Qwen 3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -172613,20 +181364,18 @@ "type": "toggle" }, { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "budget_tokens" } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-02", - "last_updated": "2026-05-02", + "tool_call": true, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -172634,25 +181383,27 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 991000, + "output": 128000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.15, + "output": 0.47, + "cache_read": 0.016, + "cache_write": 0.2 } }, - "azure-o1": { - "id": "azure-o1", - "name": "Azure o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "alibaba/qwen3-embedding-8b": { + "id": "alibaba/qwen3-embedding-8b", + "name": "Qwen3 Embedding 8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-12-17", - "last_updated": "2024-12-17", + "temperature": true, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ "text" @@ -172663,156 +181414,104 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 - }, - "cost": { - "input": 14.994, - "output": 59.993 + "context": 32768, + "output": 32768 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/wan-v3.0-video": { + "id": "alibaba/wan-v3.0-video", + "name": "Wan v3.0 Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2026-08-23", + "last_updated": "2026-08-23", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 991808, - "input": 991808, - "output": 65536 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.04 + "context": 0, + "output": 0 } }, - "glm-4-airx": { - "id": "glm-4-airx", - "name": "GLM-4 AirX", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/wan-v2.7-r2v": { + "id": "alibaba/wan-v2.7-r2v", + "name": "Wan v2.7 Reference-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-06-05", - "last_updated": "2024-06-05", + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 8000, - "input": 8000, - "output": 4096 - }, - "cost": { - "input": 2.006, - "output": 2.006 + "context": 0, + "output": 0 } }, - "step-r1-v-mini": { - "id": "step-r1-v-mini", - "name": "Step R1 V Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, + "alibaba/wan-v3.0-video-prime": { + "id": "alibaba/wan-v3.0-video-prime", + "name": "Wan v3.0 Video Prime", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-04-08", - "last_updated": "2025-04-08", + "release_date": "2026-08-28", + "last_updated": "2026-08-28", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 - }, - "cost": { - "input": 2.5, - "output": 11 + "context": 0, + "output": 0 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.6-27b": { + "id": "alibaba/qwen3.6-27b", + "name": "Qwen 3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 131072 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 - }, - "cost": { - "input": 2.5, - "output": 10 - } - }, - "claude-haiku-4-5-20251001-thinking": { - "id": "claude-haiku-4-5-20251001-thinking", - "name": "Claude Haiku 4.5 Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", @@ -172823,28 +181522,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1 + "input": 0.6, + "output": 3.6 } }, - "jamba-mini-1.6": { - "id": "jamba-mini-1.6", - "name": "Jamba Mini 1.6", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen-3-235b": { + "id": "alibaba/qwen-3-235b", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-01", - "last_updated": "2025-03-01", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -172855,25 +181555,36 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 4096 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.1989, - "output": 0.408 + "input": 0.22, + "output": 0.88 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.6-plus": { + "id": "alibaba/qwen3.6-plus", + "name": "Qwen 3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 131072 + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -172886,57 +181597,56 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, + "context": 1000000, "output": 64000 }, "cost": { - "input": 1, - "output": 5 - } - }, - "claude-3-5-haiku-20241022": { - "id": "claude-3-5-haiku-20241022", - "name": "Claude 3.5 Haiku", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2024-10-22", - "last_updated": "2024-10-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "input": 200000, - "output": 8192 - }, - "cost": { - "input": 0.8, - "output": 4 + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "claude-sonnet-4-thinking:32768": { - "id": "claude-sonnet-4-thinking:32768", - "name": "Claude 4 Sonnet Thinking (32K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.7-plus": { + "id": "alibaba/qwen3.7-plus", + "name": "Qwen 3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", @@ -172950,58 +181660,86 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 1000000, "output": 64000 }, "cost": { - "input": 2.992, - "output": 14.994 + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "cache_write": 0.5, + "tiers": [ + { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24, + "cache_write": 1.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24, + "cache_write": 1.5 + } } }, - "hermes-low": { - "id": "hermes-low", - "name": "Hermes Low", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "alibaba/wan-v2.6-r2v": { + "id": "alibaba/wan-v2.6-r2v", + "name": "Wan v2.6 Reference-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025 + "context": 0, + "output": 0 } }, - "claude-sonnet-4-thinking": { - "id": "claude-sonnet-4-thinking", - "name": "Claude 4 Sonnet Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3.8-max-0902": { + "id": "alibaba/qwen3.8-max-0902", + "name": "Qwen3.8 Max 0902", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-02-24", - "last_updated": "2025-02-24", + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", @@ -173014,55 +181752,49 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 64000 + "context": 991000, + "output": 128000 }, "cost": { - "input": 2.992, - "output": 14.994 + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/wan-v2.7-t2v": { + "id": "alibaba/wan-v2.7-t2v", + "name": "Wan v2.7 Text-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 65536 - }, - "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.25 + "context": 0, + "output": 0 } }, - "yi-medium-200k": { - "id": "yi-medium-200k", - "name": "Yi Medium 200k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "mixedbread/toast-1": { + "id": "mixedbread/toast-1", + "name": "Toast 1", + "description": "Specialized search model for knowledge-intensive questions, multi-step retrieval, and evidence synthesis", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-03-01", - "last_updated": "2024-03-01", + "tool_call": true, + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ "text" @@ -173073,353 +181805,295 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 4096 + "context": 131000, + "output": 4000 }, "cost": { - "input": 2.499, - "output": 2.499 + "input": 0.3, + "output": 0.72, + "cache_read": 0.036 } }, - "Qwen3.5-27B-BlueStar-v2-Derestricted-Lite": { - "id": "Qwen3.5-27B-BlueStar-v2-Derestricted-Lite", - "name": "Qwen3.5 27B BlueStar v2 Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 - } - ], + "klingai/kling-v2.6-t2v": { + "id": "klingai/kling-v2.6-t2v", + "name": "Kling v2.6 Text-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ling", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-04-06", - "last_updated": "2026-04-06", + "temperature": true, + "release_date": "2025-12-03", + "last_updated": "2025-12-21", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306 + "context": 0, + "output": 0 } }, - "Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled": { - "id": "Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled", - "name": "Gemma 4 31B Claude 4.6 Opus Reasoning Distilled", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "klingai/kling-v2.6-i2v": { + "id": "klingai/kling-v2.6-i2v", + "name": "Kling v2.6 Image-to-Video", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ling", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "temperature": true, + "release_date": "2025-12-03", + "last_updated": "2025-12-21", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306, - "cache_read": 0.0306 + "context": 0, + "output": 0 } }, - "doubao-1-5-thinking-vision-pro-250428": { - "id": "doubao-1-5-thinking-vision-pro-250428", - "name": "Doubao 1.5 Thinking Vision Pro", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "klingai/kling-v2.5-turbo-t2v": { + "id": "klingai/kling-v2.5-turbo-t2v", + "name": "Kling v2.5 Turbo Text-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ling", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-05-15", - "last_updated": "2025-05-15", + "temperature": true, + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 - }, - "cost": { - "input": 0.55, - "output": 1.43 + "context": 0, + "output": 0 } }, - "Qwen3.5-27B-BlueStar-Derestricted-Lite": { - "id": "Qwen3.5-27B-BlueStar-Derestricted-Lite", - "name": "Qwen3.5 27B BlueStar Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 - } - ], + "klingai/kling-v3.0-i2v": { + "id": "klingai/kling-v3.0-i2v", + "name": "Kling v3.0 Image-to-Video", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ling", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-04-06", - "last_updated": "2026-04-06", + "temperature": true, + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306 + "context": 0, + "output": 0 } }, - "Qwen2.5-32B-EVA-v0.2": { - "id": "Qwen2.5-32B-EVA-v0.2", - "name": "Qwen 2.5 32b EVA", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "klingai/kling-v3.0-motion-control": { + "id": "klingai/kling-v3.0-motion-control", + "name": "Kling v3.0 Motion Control", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ling", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-09-01", - "last_updated": "2024-09-01", + "temperature": true, + "release_date": "2026-03-04", + "last_updated": "2026-03-04", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 24576, - "input": 24576, - "output": 8192 - }, - "cost": { - "input": 0.493, - "output": 0.493 + "context": 0, + "output": 0 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "klingai/kling-v3.0-t2v": { + "id": "klingai/kling-v3.0-t2v", + "name": "Kling v3.0 Text-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ling", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "temperature": true, + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5 + "context": 0, + "output": 0 } }, - "qwen-long": { - "id": "qwen-long", - "name": "Qwen Long 10M", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "klingai/kling-v2.6-motion-control": { + "id": "klingai/kling-v2.6-motion-control", + "name": "Kling v2.6 Motion Control", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ling", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-01-25", - "last_updated": "2025-01-25", + "temperature": true, + "release_date": "2025-12-18", + "last_updated": "2025-12-21", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 10000000, - "input": 10000000, - "output": 8192 - }, - "cost": { - "input": 0.1003, - "output": 0.408 + "context": 0, + "output": 0 } }, - "Gemma-4-31B-DarkIdol": { - "id": "Gemma-4-31B-DarkIdol", - "name": "Gemma 4 31B DarkIdol", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "klingai/kling-v2.5-turbo-i2v": { + "id": "klingai/kling-v2.5-turbo-i2v", + "name": "Kling v2.5 Turbo Image-to-Video", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ling", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "temperature": true, + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306 + "context": 0, + "output": 0 } }, - "step-2-16k-exp": { - "id": "step-2-16k-exp", - "name": "Step-2 16k Exp", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bytedance/seedream-5.0-lite": { + "id": "bytedance/seedream-5.0-lite", + "name": "Seedream 5.0 Lite", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "seed", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-07-05", - "last_updated": "2024-07-05", + "temperature": true, + "release_date": "2026-02-13", + "last_updated": "2026-01-28", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 16000, - "input": 16000, - "output": 8192 + "context": 0, + "output": 0 + } + }, + "bytedance/seedream-4.5": { + "id": "bytedance/seedream-4.5", + "name": "Seedream 4.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "seed", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-12-03", + "last_updated": "2025-11-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] }, - "cost": { - "input": 7.004, - "output": 19.992 + "open_weights": false, + "limit": { + "context": 0, + "output": 0 } }, - "Qwen3.5-27B-Marvin-V2-Derestricted-Lite": { - "id": "Qwen3.5-27B-Marvin-V2-Derestricted-Lite", - "name": "Qwen3.5 27B Marvin V2 Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "bytedance/seed-1.8": { + "id": "bytedance/seed-1.8", + "name": "Seed 1.8", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-01", + "last_updated": "2025-10", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -173427,40 +182101,54 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.25, + "output": 2, + "cache_read": 0.05, + "tiers": [ + { + "input": 0.5, + "output": 4, + "cache_read": 0.05, + "tier": { + "type": "context", + "size": 128001 + } + } + ] } }, - "Qwen3.5-27B-NaNovel-Derestricted": { - "id": "Qwen3.5-27B-NaNovel-Derestricted", - "name": "Qwen3.5 27B NaNovel Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bytedance/seed-2.1-turbo": { + "id": "bytedance/seed-2.1-turbo", + "name": "Seed 2.1 Turbo", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "release_date": "2026-06-23", + "last_updated": "2026-06-23", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -173469,277 +182157,246 @@ "open_weights": false, "limit": { "context": 262144, - "input": 262144, - "output": 16384 + "output": 262144 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.5, + "output": 2.5, + "cache_read": 0.1 } }, - "ernie-4.5-8k-preview": { - "id": "ernie-4.5-8k-preview", - "name": "Ernie 4.5 8k Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bytedance/seedream-4.0": { + "id": "bytedance/seedream-4.0", + "name": "Seedream 4.0", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "seed", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "temperature": true, + "release_date": "2025-09-09", + "last_updated": "2025-08-28", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 8000, - "input": 8000, - "output": 16384 - }, - "cost": { - "input": 0.66, - "output": 2.6 + "context": 0, + "output": 0 } }, - "gemini-2.0-flash-exp-image-generation": { - "id": "gemini-2.0-flash-exp-image-generation", - "name": "Gemini Text + Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "bytedance/seedance-v1.0-pro": { + "id": "bytedance/seedance-v1.0-pro", + "name": "Seedance v1.0 Pro", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "temperature": true, + "release_date": "2025-06-11", + "last_updated": "2025-06-11", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 32767, - "input": 32767, - "output": 8192 - }, - "cost": { - "input": 0.2, - "output": 0.8 + "context": 0, + "output": 0 } }, - "gemini-2.5-flash-lite-preview-09-2025": { - "id": "gemini-2.5-flash-lite-preview-09-2025", - "name": "Gemini 2.5 Flash Lite Preview (09/2025)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bytedance/seedance-2.0": { + "id": "bytedance/seedance-2.0", + "name": "Seedance 2.0", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-04-14", + "last_updated": "2026-04-14", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.4 + "context": 0, + "output": 0 } }, - "deepseek-reasoner-cheaper": { - "id": "deepseek-reasoner-cheaper", - "name": "Deepseek R1 Cheaper", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, + "bytedance/seedance-2.0-fast": { + "id": "bytedance/seedance-2.0-fast", + "name": "Seedance 2.0 Fast", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", + "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "temperature": true, + "release_date": "2026-04-14", + "last_updated": "2026-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 - }, - "cost": { - "input": 0.4, - "output": 1.7 + "context": 0, + "output": 0 } }, - "venice-uncensored": { - "id": "venice-uncensored", - "name": "Venice Uncensored", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bytedance/seedance-v1.5-pro": { + "id": "bytedance/seedance-v1.5-pro", + "name": "Seedance v1.5 Pro", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-02-24", - "last_updated": "2025-02-24", + "temperature": true, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 - }, - "cost": { - "input": 0.4, - "output": 0.4 + "context": 0, + "output": 0 } }, - "gemini-2.0-flash-001": { - "id": "gemini-2.0-flash-001", - "name": "Gemini 2.0 Flash", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bytedance/seedance-2.0-mini": { + "id": "bytedance/seedance-2.0-mini", + "name": "Seedance 2.0 Mini", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "tool_call": false, + "release_date": "2026-06-22", + "last_updated": "2026-06-22", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 8192 - }, - "cost": { - "input": 0.1003, - "output": 0.408 + "context": 0, + "output": 0 } }, - "gemma-4-31B-Larkspur-v0.5": { - "id": "gemma-4-31B-Larkspur-v0.5", - "name": "Gemma 4 31B Larkspur v0.5", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "bytedance/seedance-2.5": { + "id": "bytedance/seedance-2.5", + "name": "Seedance 2.5", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-05-02", - "last_updated": "2026-05-02", + "release_date": "2026-08-07", + "last_updated": "2026-08-07", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306 + "context": 0, + "output": 0 } }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "Claude 4.1 Opus", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "bytedance/seedance-v1.0-pro-fast": { + "id": "bytedance/seedance-v1.0-pro-fast", + "name": "Seedance v1.0 Pro Fast", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "tool_call": false, + "temperature": true, + "release_date": "2025-10-24", + "last_updated": "2025-10-31", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 0, + "output": 0 + } + }, + "bytedance/seedream-5.0-pro": { + "id": "bytedance/seedream-5.0-pro", + "name": "Seedream 5.0 Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "seed", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-07-11", + "last_updated": "2026-07-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] }, - "cost": { - "input": 14.994, - "output": 75.004 + "open_weights": false, + "limit": { + "context": 0, + "output": 0 } }, - "holo3-35b-a3b": { - "id": "holo3-35b-a3b", - "name": "Holo3-35B-A3B", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "bytedance/seed-1.6": { + "id": "bytedance/seed-1.6", + "name": "Seed 1.6", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -173747,9 +182404,10 @@ } ], "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-01", + "last_updated": "2025-09", "modalities": { "input": [ "text", @@ -173761,67 +182419,75 @@ }, "open_weights": false, "limit": { - "context": 65536, - "input": 65536, - "output": 65536 + "context": 256000, + "output": 32000 }, "cost": { "input": 0.25, - "output": 1.8 + "output": 2, + "cache_read": 0.05, + "tiers": [ + { + "input": 0.5, + "output": 4, + "cache_read": 0.05, + "tier": { + "type": "context", + "size": 128001 + } + } + ] } }, - "qwen3.5-omni-plus": { - "id": "qwen3.5-omni-plus", - "name": "Qwen3.5 Omni Plus", - "description": "Omni-modal model for text, vision, audio, and multimodal agent tasks", + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 983616, - "input": 983616, - "output": 65536 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.14, + "output": 0.4 } }, - "Gemma-4-31B-it": { - "id": "Gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-omni-flash-preview": { + "id": "google/gemini-omni-flash-preview", + "name": "Gemini Omni Flash Preview", + "description": "Omni-modal model for text, vision, audio, and multimodal agent tasks", + "family": "gemini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": false, - "structured_output": false, - "release_date": "2026-04-09", - "last_updated": "2026-04-09", + "temperature": true, + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -173829,70 +182495,69 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1000000, + "output": 57920 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 1.5, + "output": 9 } }, - "Qwen3.5-27B-Derestricted": { - "id": "Qwen3.5-27B-Derestricted", - "name": "Qwen3.5 27B Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3.1-flash-image-preview": { + "id": "google/gemini-3.1-flash-image-preview", + "name": "Gemini 3.1 Flash Image Preview (Nano Banana 2)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "minimal", + "high" + ] } ], "tool_call": false, - "structured_output": false, - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "sonar-reasoning-pro": { - "id": "sonar-reasoning-pro", - "name": "Perplexity Reasoning Pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "google/gemini-3.5-transcribe-live": { + "id": "google/gemini-3.5-transcribe-live", + "name": "Gemini 3.5 Transcribe Live", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "gemini", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" @@ -173900,54 +182565,54 @@ }, "open_weights": false, "limit": { - "context": 127000, - "input": 127000, - "output": 128000 - }, - "cost": { - "input": 2.006, - "output": 7.9985 + "context": 0, + "output": 0 } }, - "venice-uncensored:web": { - "id": "venice-uncensored:web", - "name": "Venice Uncensored Web", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-2.5-flash-image": { + "id": "google/gemini-2.5-flash-image", + "name": "Nano Banana (Gemini 2.5 Flash Image)", + "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", + "family": "gemini-flash", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-05-01", - "last_updated": "2024-05-01", + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 80000, - "input": 80000, - "output": 16384 + "context": 32768, + "output": 65535 }, "cost": { - "input": 0.4, - "output": 0.4 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "doubao-1.5-pro-32k": { - "id": "doubao-1.5-pro-32k", - "name": "Doubao 1.5 Pro 32k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/text-multilingual-embedding-002": { + "id": "google/text-multilingual-embedding-002", + "name": "Text Multilingual Embedding 002", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-01-22", - "last_updated": "2025-01-22", + "temperature": true, + "release_date": "2024-03-01", + "last_updated": "2024-03", "modalities": { "input": [ "text" @@ -173958,25 +182623,22 @@ }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 8192 - }, - "cost": { - "input": 0.1343, - "output": 0.3349 + "context": 8192, + "output": 1536 } }, - "qwen3-30b-a3b-instruct-2507": { - "id": "qwen3-30b-a3b-instruct-2507", - "name": "Qwen3 30B A3B Instruct 2507", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-embedding-2": { + "id": "google/gemini-embedding-2", + "name": "Gemini Embedding 2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "gemini", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-02-20", - "last_updated": "2025-02-20", + "temperature": true, + "knowledge": "2025-11", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -173987,31 +182649,39 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 - }, - "cost": { - "input": 0.2, - "output": 0.5 + "context": 0, + "output": 0 } }, - "ernie-5.1:thinking": { - "id": "ernie-5.1:thinking", - "name": "ERNIE 5.1 Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3.6-flash": { + "id": "google/gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-10", - "last_updated": "2026-05-10", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -174019,57 +182689,52 @@ }, "open_weights": false, "limit": { - "context": 119000, - "input": 119000, + "context": 1000000, "output": 64000 }, "cost": { "input": 0.75, - "output": 3, - "cache_read": 0.75 + "output": 3.75, + "cache_read": 0.075 } }, - "mirothinker-1-7-deepresearch": { - "id": "mirothinker-1-7-deepresearch", - "name": "MiroThinker 1.7 Deep Research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "google/veo-3.1-lite-generate-001": { + "id": "google/veo-3.1-lite-generate-001", + "name": "Veo 3.1 Lite Generate", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 4, - "output": 25 + "context": 0, + "output": 0 } }, - "hermes-high": { - "id": "hermes-high", - "name": "Hermes High", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -174080,59 +182745,46 @@ "text" ] }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 - }, - "cost": { - "input": 4.998, - "output": 25.007 - } - }, - "jamba-large-1.6": { - "id": "jamba-large-1.6", - "name": "Jamba Large 1.6", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 4096 + "context": 262144, + "output": 131072 }, "cost": { - "input": 1.989, - "output": 7.99 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "glm-4-long": { - "id": "glm-4-long", - "name": "GLM-4 Long", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-08-01", - "last_updated": "2024-08-01", + "google/gemini-3.5-flash-lite": { + "id": "google/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -174141,57 +182793,79 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 1000000, - "output": 4096 + "output": 65000 }, "cost": { - "input": 0.2006, - "output": 0.2006 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "claude-opus-4-thinking:32768": { - "id": "claude-opus-4-thinking:32768", - "name": "Claude 4 Opus Thinking (32K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3.1-flash-image": { + "id": "google/gemini-3.1-flash-image", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 14.994, - "output": 75.004 + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "claude-sonnet-4-thinking:8192": { - "id": "claude-sonnet-4-thinking:8192", - "name": "Claude 4 Sonnet Thinking (8K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", @@ -174205,57 +182879,83 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 1000000, "output": 64000 }, "cost": { - "input": 2.992, - "output": 14.994 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200001 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "azure-o3-mini": { - "id": "azure-o3-mini", - "name": "Azure o3-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "google/veo-3.1-fast-generate-001": { + "id": "google/veo-3.1-fast-generate-001", + "name": "Veo 3.1 Fast Generate", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "temperature": true, + "release_date": "2025-10-15", + "last_updated": "2026-06-08", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 65536 - }, - "cost": { - "input": 1.088, - "output": 4.3996 + "context": 0, + "output": 0 } }, - "qvq-max": { - "id": "qvq-max", - "name": "Qwen: QvQ Max", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3-flash": { + "id": "google/gemini-3-flash", + "name": "Gemini 3 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-28", - "last_updated": "2025-03-28", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -174263,40 +182963,44 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 1000000, + "output": 65000 }, "cost": { - "input": 1.4, - "output": 5.3 + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "Qwen3.5-27B-BlueStar-v2-Derestricted": { - "id": "Qwen3.5-27B-BlueStar-v2-Derestricted", - "name": "Qwen3.5 27B BlueStar v2 Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-06", - "last_updated": "2026-04-06", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -174304,198 +183008,191 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 1.5, + "output": 9, + "cache_read": 0.15 } }, - "Qwen3.5-27B-Vivid-Durian": { - "id": "Qwen3.5-27B-Vivid-Durian", - "name": "Qwen3.5 27B Vivid Durian", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 - } - ], + "google/gemini-3-pro-image": { + "id": "google/gemini-3-pro-image", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 65536, + "output": 32768 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "jamba-large-1.7": { - "id": "jamba-large-1.7", - "name": "Jamba Large 1.7", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/veo-3.0-fast-generate-001": { + "id": "google/veo-3.0-fast-generate-001", + "name": "Veo 3.0 Fast Generate", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "temperature": true, + "release_date": "2025-07-31", + "last_updated": "2026-06-08", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 4096 - }, - "cost": { - "input": 1.989, - "output": 7.99 + "context": 0, + "output": 0 } }, - "Baichuan-M2": { - "id": "Baichuan-M2", - "name": "Baichuan M2 32B Medical", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3.8-live": { + "id": "google/gemini-3.8-live", + "name": "Gemini 3.8 Live", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gemini", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-08-19", - "last_updated": "2025-08-19", + "release_date": "2026-09-15", + "last_updated": "2026-09-15", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 0, + "output": 0 }, "cost": { - "input": 15.73, - "output": 15.73 + "input": 0.75, + "output": 4.5 } }, - "Magistral-Small-2506": { - "id": "Magistral-Small-2506", - "name": "Magistral Small 2506", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/veo-3.1-generate-001": { + "id": "google/veo-3.1-generate-001", + "name": "Veo 3.1", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "temperature": true, + "release_date": "2025-10-15", + "last_updated": "2026-06-08", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 1.4 + "context": 0, + "output": 0 } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek R1", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, + "google/gemini-3.1-flash-lite-image": { + "id": "google/gemini-3.1-flash-lite-image", + "name": "Gemini 3.1 Flash Lite Image (Nano Banana 2 Lite)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": false, "structured_output": false, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 65536, + "output": 4096 }, "cost": { - "input": 0.4, - "output": 1.7 + "input": 0.25, + "output": 1.5, + "cache_read": 0.03 } }, - "Qwen3.5-27B-earica-Derestricted-Lite": { - "id": "Qwen3.5-27B-earica-Derestricted-Lite", - "name": "Qwen3.5 27B earica Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3.7-flash": { + "id": "google/gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -174503,25 +183200,26 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1000000, + "output": 65535 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 } }, - "doubao-seed-2-0-pro-260215": { - "id": "doubao-seed-2-0-pro-260215", - "name": "Doubao Seed 2.0 Pro", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/text-embedding-005": { + "id": "google/text-embedding-005", + "name": "Text Embedding 005", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "temperature": true, + "release_date": "2024-08-01", + "last_updated": "2024-08", "modalities": { "input": [ "text" @@ -174532,58 +183230,23 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 128000 - }, - "cost": { - "input": 0.782, - "output": 3.876 + "context": 8192, + "output": 1536 } }, - "step-2-mini": { - "id": "step-2-mini", - "name": "Step-2 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3.5-transcribe": { + "id": "google/gemini-3.5-transcribe", + "name": "Gemini 3.5 Transcribe", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "gemini", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-07-05", - "last_updated": "2024-07-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8000, - "input": 8000, - "output": 4096 - }, - "cost": { - "input": 0.2006, - "output": 0.408 - } - }, - "gemini-exp-1206": { - "id": "gemini-exp-1206", - "name": "Gemini 2.0 Pro 1206", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text", - "image" + "audio" ], "output": [ "text" @@ -174591,26 +183254,27 @@ }, "open_weights": false, "limit": { - "context": 2097152, - "input": 2097152, - "output": 8192 + "context": 0, + "output": 0 }, "cost": { - "input": 1.258, - "output": 4.998 + "input": 2, + "output": 12 } }, - "claude-opus-4-5-20251101:thinking": { - "id": "claude-opus-4-5-20251101:thinking", - "name": "Claude 4.5 Opus Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3.8-flash": { + "id": "google/gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text", @@ -174623,19 +183287,20 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 1000000, + "output": 65535 }, "cost": { - "input": 4.998, - "output": 25.007 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 } }, - "Qwen3.5-27B-BlueStar-v3-Derestricted": { - "id": "Qwen3.5-27B-BlueStar-v3-Derestricted", - "name": "Qwen3.5 27B BlueStar v3 Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -174644,46 +183309,16 @@ }, { "type": "budget_tokens", - "min": 1024, - "max": 81920 + "min": 512, + "max": 24576 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306 - } - }, - "claude-opus-4-thinking": { - "id": "claude-opus-4-thinking", - "name": "Claude 4 Opus Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", @@ -174696,30 +183331,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 14.994, - "output": 75.004 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 } }, - "qwen3.5-flash": { - "id": "qwen3.5-flash", - "name": "Qwen3.5 Flash", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "google/gemini-embedding-001": { + "id": "google/gemini-embedding-001", + "name": "Gemini Embedding 001", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "gemini", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -174727,84 +183362,89 @@ }, "open_weights": false, "limit": { - "context": 991808, - "input": 991808, - "output": 65536 - }, - "cost": { - "input": 0.09, - "output": 0.36 + "context": 8192, + "output": 1536 } }, - "exa-research-pro": { - "id": "exa-research-pro", - "name": "Exa (Research Pro)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3.8-live-extended-thinking": { + "id": "google/gemini-3.8-live-extended-thinking", + "name": "Gemini 3.8 Live Extended Thinking", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gemini", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-06-04", - "last_updated": "2025-06-04", + "release_date": "2026-09-15", + "last_updated": "2026-09-15", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 0, + "output": 0 }, "cost": { - "input": 2.5, - "output": 2.5 + "input": 0.75, + "output": 4.5 } }, - "jamba-mini": { - "id": "jamba-mini", - "name": "Jamba Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/veo-3.0-generate-001": { + "id": "google/veo-3.0-generate-001", + "name": "Veo 3.0", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "temperature": true, + "release_date": "2025-05-20", + "last_updated": "2026-06-08", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 4096 - }, - "cost": { - "input": 0.1989, - "output": 0.408 + "context": 0, + "output": 0 } }, - "claude-opus-4-1-thinking:32000": { - "id": "claude-opus-4-1-thinking:32000", - "name": "Claude 4.1 Opus Thinking (32K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", @@ -174817,57 +183457,45 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 - }, - "cost": { - "input": 14.994, - "output": 75.004 - } - }, - "KAT-Coder-Air-V1": { - "id": "KAT-Coder-Air-V1", - "name": "KAT Coder Air V1", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "input": 128000, - "output": 32768 + "context": 1000000, + "output": 65000 }, "cost": { - "input": 0.1, - "output": 0.2 + "input": 0.25, + "output": 1.5, + "cache_read": 0.03 } }, - "exa-research": { - "id": "exa-research", - "name": "Exa (Research)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-04", - "last_updated": "2025-06-04", + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -174875,28 +183503,42 @@ }, "open_weights": false, "limit": { - "context": 8192, - "input": 8192, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 2.5 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "deepclaude": { - "id": "deepclaude", - "name": "DeepClaude", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-01", - "last_updated": "2025-02-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", + "image", + "audio", + "video", "pdf" ], "output": [ @@ -174905,44 +183547,55 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 15 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "gemini-2.5-flash-preview-09-2025": { - "id": "gemini-2.5-flash-preview-09-2025", - "name": "Gemini 2.5 Flash Preview (09/2025)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "sakana/fugu-max": { + "id": "sakana/fugu-max", + "name": "Fugu Max", + "description": "Multi-agent model for routing expert agents across complex analytical tasks", + "family": "fugu", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "none", - "low", - "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -174950,46 +183603,39 @@ }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "Claude 4.5 Opus", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "sakana/fugu-ultra-v2": { + "id": "sakana/fugu-ultra-v2", + "name": "Fugu Ultra v2", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", - "medium", - "high" + "high", + "xhigh", + "max" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, - "structured_output": true, - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -174997,28 +183643,46 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 4.998, - "output": 25.007 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272001 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "GLM-4.6-Derestricted-v5": { - "id": "GLM-4.6-Derestricted-v5", - "name": "GLM 4.6 Derestricted v5", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "sakana/namazu": { + "id": "sakana/namazu", + "name": "Sakana Namazu", + "description": "Multi-agent model for routing expert agents across complex analytical tasks", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -175026,28 +183690,32 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 8192 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.4, - "output": 1.5 + "input": 0.95, + "output": 4, + "cache_read": 0.15 } }, - "glm-z1-airx": { - "id": "glm-z1-airx", - "name": "GLM Z1 AirX", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, + "sakana/fugu-ultra": { + "id": "sakana/fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "temperature": true, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -175055,31 +183723,49 @@ }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 16384 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.7, - "output": 0.7 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272001 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "claude-sonnet-4-thinking:1024": { - "id": "claude-sonnet-4-thinking:1024", - "name": "Claude 4 Sonnet Thinking (1K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "inclusionai/ling-3.0-flash-fin-free": { + "id": "inclusionai/ling-3.0-flash-fin-free", + "name": "Ling 3.0 Flash Fin (Free)", + "description": "Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution", + "family": "ling", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -175087,26 +183773,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 64000 + "context": 256000, + "output": 32000 }, "cost": { - "input": 2.992, - "output": 14.994 + "input": 0, + "output": 0 } }, - "holo3-35b-a3b:thinking": { - "id": "holo3-35b-a3b:thinking", - "name": "Holo3-35B-A3B Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "inclusionai/ling-3.0-flash-vl-free": { + "id": "inclusionai/ling-3.0-flash-vl-free", + "name": "Ling 3.0 Flash VL (Free)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ling", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ "text", @@ -175118,25 +183807,25 @@ }, "open_weights": false, "limit": { - "context": 65536, - "input": 65536, - "output": 65536 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.25, - "output": 1.8 + "input": 0, + "output": 0 } }, - "owl": { - "id": "owl", - "name": "OWL", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "inclusionai/ling-3.0-flash": { + "id": "inclusionai/ling-3.0-flash", + "name": "Ling 3.0 Flash", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "release_date": "2026-08-06", + "last_updated": "2026-08-06", "modalities": { "input": [ "text" @@ -175147,25 +183836,26 @@ }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 262144 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.021, + "output": 0.063, + "cache_read": 0.0042 } }, - "gemini-2.0-pro-reasoner": { - "id": "gemini-2.0-pro-reasoner", - "name": "Gemini 2.0 Pro Reasoner", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "inclusionai/ling-3.0-flash-sante": { + "id": "inclusionai/ling-3.0-flash-sante", + "name": "Ling 3.0 Flash Sante", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-05", - "last_updated": "2025-02-05", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ "text" @@ -175176,19 +183866,19 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 256000, + "output": 32000 }, "cost": { - "input": 1.292, - "output": 4.998 + "input": 0, + "output": 0 } }, - "Gemma-4-31B-Cognitive-Unshackled": { - "id": "Gemma-4-31B-Cognitive-Unshackled", - "name": "Gemma 4 31B Cognitive Unshackled", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "inclusionai/ling-3.0-flash-vl": { + "id": "inclusionai/ling-3.0-flash-vl", + "name": "Ling 3.0 Flash VL", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -175196,15 +183886,13 @@ "type": "toggle" } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "tool_call": true, + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -175212,40 +183900,32 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0, + "output": 0 } }, - "Qwen3.5-27B-earica-Derestricted": { - "id": "Qwen3.5-27B-earica-Derestricted", - "name": "Qwen3.5 27B earica Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "inclusionai/ling-3.0-flash-fin": { + "id": "inclusionai/ling-3.0-flash-fin", + "name": "Ling 3.0 Flash Fin", + "description": "Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution", + "family": "ling", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -175253,39 +183933,28 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0, + "output": 0 } }, - "command-a-plus-05-2026": { - "id": "command-a-plus-05-2026", - "name": "Cohere Command A+ (05/2026)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "inclusionai/ling-3.0-flash-sante-free": { + "id": "inclusionai/ling-3.0-flash-sante-free", + "name": "Ling 3.0 Flash Sante (Free)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, - "release_date": "2026-05-22", - "last_updated": "2026-05-22", + "reasoning_options": [], + "tool_call": true, + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -175293,173 +183962,178 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 64000 + "context": 256000, + "output": 32000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0, + "output": 0 } }, - "auto-model-premium": { - "id": "auto-model-premium", - "name": "Auto model (Premium)", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "input": 1000000, - "output": 1000000 + "output": 131072 }, "cost": { - "input": 9.996, - "output": 19.992 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "learnlm-1.5-pro-experimental": { - "id": "learnlm-1.5-pro-experimental", - "name": "Gemini LearnLM Experimental", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-05-14", - "last_updated": "2024-05-14", + "moonshotai/kimi-k3-fast": { + "id": "moonshotai/kimi-k3-fast", + "name": "Kimi K3 Fast", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32767, - "input": 32767, - "output": 8192 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 3.502, - "output": 10.506 - } - }, - "deepseek-r1-sambanova": { - "id": "deepseek-r1-sambanova", - "name": "DeepSeek R1 Fast", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-20", - "last_updated": "2025-02-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "input": 128000, - "output": 4096 - }, - "cost": { - "input": 4.998, - "output": 6.987 + "input": 4.5, + "output": 22.5, + "cache_read": 0.45 } }, - "claw-medium": { - "id": "claw-medium", - "name": "Claw Medium", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, + "moonshotai/kimi-k2.7-code-highspeed": { + "id": "moonshotai/kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code High Speed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "input": 204800, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 1.9, + "output": 8, + "cache_read": 0.38 } }, - "claude-opus-4-20250514": { - "id": "claude-opus-4-20250514", - "name": "Claude 4 Opus", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-05-14", - "last_updated": "2025-05-14", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 14.994, - "output": 75.004 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "yi-large": { - "id": "yi-large", - "name": "Yi Large", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ "text" @@ -175468,56 +184142,66 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "input": 32000, - "output": 4096 + "context": 216144, + "output": 216144 }, "cost": { - "input": 3.196, - "output": 3.196 + "input": 0.47, + "output": 2, + "cache_read": 0.141 } }, - "qwen3-max-2026-01-23": { - "id": "qwen3-max-2026-01-23", - "name": "Qwen3 Max 2026-01-23", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-01-26", - "last_updated": "2026-01-26", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 256000, - "input": 256000, - "output": 32768 + "output": 256000 }, "cost": { - "input": 1.2002, - "output": 6.001 + "input": 0.6, + "output": 3 } }, - "phi-4-mini-instruct": { - "id": "phi-4-mini-instruct", - "name": "Phi 4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "moonshotai/kimi-k2": { + "id": "moonshotai/kimi-k2", + "name": "Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "tool_call": true, + "temperature": true, + "release_date": "2025-07-11", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -175528,57 +184212,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 - }, - "cost": { - "input": 0.17, - "output": 0.68 - } - }, - "ernie-x1-turbo-32k": { - "id": "ernie-x1-turbo-32k", - "name": "Ernie X1 Turbo 32k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-08", - "last_updated": "2025-05-08", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "input": 32000, - "output": 16384 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.165, - "output": 0.66 + "input": 0.57, + "output": 2.3 } }, - "claw-low": { - "id": "claw-low", - "name": "Claw Low", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", @@ -175589,124 +184244,143 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 256000, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "gemini-3-pro-image-preview": { - "id": "gemini-3-pro-image-preview", - "name": "Gemini 3 Pro Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": false, + "nvidia/nemotron-3-nano-30b-a3b": { + "id": "nvidia/nemotron-3-nano-30b-a3b", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": false, - "structured_output": false, - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "temperature": true, + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2, - "output": 12 + "input": 0.05, + "output": 0.2, + "cache_read": 0.025 } }, - "gemma-4-31B-Garnet": { - "id": "gemma-4-31B-Garnet", - "name": "Gemma 4 31B Garnet", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "nvidia/nemotron-3.5-lightning": { + "id": "nvidia/nemotron-3.5-lightning", + "name": "Nemotron 3.5 Lightning 30B", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 32768 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-02", - "last_updated": "2026-05-02", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "input": 262144, - "output": 16384 + "output": 131072 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.05, + "output": 0.2, + "cache_read": 0.01 } }, - "doubao-1.5-vision-pro-32k": { - "id": "doubao-1.5-vision-pro-32k", - "name": "Doubao 1.5 Vision Pro 32k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-22", - "last_updated": "2025-01-22", + "nvidia/nemotron-nano-9b-v2": { + "id": "nvidia/nemotron-nano-9b-v2", + "name": "Nvidia Nemotron Nano 9B V2", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-18", + "last_updated": "2025-08-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "input": 32000, - "output": 8192 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.459, - "output": 1.377 + "input": 0.06, + "output": 0.23 } }, - "auto-model-standard": { - "id": "auto-model-standard", - "name": "Auto model (Standard)", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "NVIDIA Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": false, - "structured_output": false, - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text" @@ -175715,300 +184389,339 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 + "context": 256000, + "output": 32000 }, "cost": { - "input": 9.996, - "output": 19.992 + "input": 0.15, + "output": 0.65 } }, - "Qwen3.5-27B-Marvin-DPO-V2-Derestricted-Lite": { - "id": "Qwen3.5-27B-Marvin-DPO-V2-Derestricted-Lite", - "name": "Qwen3.5 27B Marvin DPO V2 Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "nvidia/nemotron-3-ultra-550b-a55b": { + "id": "nvidia/nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1000000, + "output": 65000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 } }, - "glm-4": { - "id": "glm-4", - "name": "GLM-4", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-16", - "last_updated": "2024-01-16", + "nvidia/nemotron-nano-12b-v2-vl": { + "id": "nvidia/nemotron-nano-12b-v2-vl", + "name": "Nvidia Nemotron Nano 12B V2 VL", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-10-28", + "last_updated": "2025-10-28", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 131072, + "output": 131072 }, "cost": { - "input": 14.994, - "output": 14.994 + "input": 0.2, + "output": 0.6 } }, - "Qwen3.5-27B-Writer-Derestricted-Lite": { - "id": "Qwen3.5-27B-Writer-Derestricted-Lite", - "name": "Qwen3.5 27B Writer Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "xiaomi/mimo-v2.6-pro": { + "id": "xiaomi/mimo-v2.6-pro", + "name": "MiMo V2.6 Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-06", - "last_updated": "2026-04-06", + "tool_call": true, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } }, - "qwen-3.6-plus": { - "id": "qwen-3.6-plus", - "name": "Qwen 3.6 Plus", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "qwen3.6", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo M2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 991800, - "output": 65536 + "context": 1050000, + "output": 131100 }, "cost": { - "input": 0.45, - "output": 2.7 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "brave-pro": { - "id": "brave-pro", - "name": "Brave (Pro)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2023-03-02", - "last_updated": "2024-01-01", + "xiaomi/mimo-v2.6-pro-ultraspeed": { + "id": "xiaomi/mimo-v2.6-pro-ultraspeed", + "name": "MiMo V2.6 Pro UltraSpeed", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-09-21", + "last_updated": "2026-09-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "input": 8192, - "output": 8192 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 5, - "output": 5 + "input": 4.35, + "output": 8.7, + "cache_read": 0.036 } }, - "deepseek-chat-cheaper": { - "id": "deepseek-chat-cheaper", - "name": "DeepSeek V3/Chat Cheaper", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo V2.5 Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 1050000, + "output": 131000 }, "cost": { - "input": 0.25, - "output": 0.7 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } }, - "Gemma-4-31B-Queen": { - "id": "Gemma-4-31B-Queen", - "name": "Gemma 4 31B Queen", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "xiaomi/mimo-v2.6-flash": { + "id": "xiaomi/mimo-v2.6-flash", + "name": "MiMo V2.6 Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "tool_call": true, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "Gemma-4-31B-Gemopus": { - "id": "Gemma-4-31B-Gemopus", - "name": "Gemma 4 31B Gemopus", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 204800, + "output": 131000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "mistral-code-latest": { - "id": "mistral-code-latest", - "name": "Mistral Code Latest", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "minimax/minimax-m2.1-lightning": { + "id": "minimax/minimax-m2.1-lightning", + "name": "MiniMax M2.1 Lightning", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-23", + "last_updated": "2025-10-27", "modalities": { "input": [ "text" @@ -176019,69 +184732,53 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { "input": 0.3, - "output": 0.9 + "output": 2.4, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "minimax/minimax-h3": { + "id": "minimax/minimax-h3", + "name": "MiniMax H3", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "minimax", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.4 + "context": 0, + "output": 0 } }, - "jamba-mini-1.7": { - "id": "jamba-mini-1.7", - "name": "Jamba Mini 1.7", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "Minimax M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -176090,128 +184787,124 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 4096 + "context": 204800, + "output": 131000 }, "cost": { - "input": 0.1989, - "output": 0.408 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "universal-summarizer": { - "id": "universal-summarizer", - "name": "Universal Summarizer", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, + "minimax/minimax-h3-max": { + "id": "minimax/minimax-h3-max", + "name": "MiniMax H3 Max", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "minimax", + "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2023-05-01", - "last_updated": "2024-01-01", + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 - }, - "cost": { - "input": 30, - "output": 30 + "context": 0, + "output": 0 } }, - "Qwen3.5-27B-Anko": { - "id": "Qwen3.5-27B-Anko", - "name": "Qwen3.5 27B Anko", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 512000, + "output": 512000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "claude-sonnet-4-20250514": { - "id": "claude-sonnet-4-20250514", - "name": "Claude 4 Sonnet", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax M2.7 High Speed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 + "context": 204800, + "output": 131100 }, "cost": { - "input": 2.992, - "output": 14.994 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "ernie-x1-32k-preview": { - "id": "ernie-x1-32k-preview", - "name": "Ernie X1 32k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "minimax/minimax-m2.5-highspeed": { + "id": "minimax/minimax-m2.5-highspeed", + "name": "MiniMax M2.5 High Speed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-03", - "last_updated": "2025-04-03", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -176220,27 +184913,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "input": 32000, - "output": 16384 + "context": 204800, + "output": 131000 }, "cost": { - "input": 0.33, - "output": 1.32 + "input": 0.6, + "output": 2.4, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "azure-gpt-4-turbo": { - "id": "azure-gpt-4-turbo", - "name": "Azure gpt-4-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2023-11-06", - "last_updated": "2024-01-01", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -176249,27 +184946,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 9.996, - "output": 30.005 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "Meta-Llama-3-1-8B-Instruct-FP8": { - "id": "Meta-Llama-3-1-8B-Instruct-FP8", - "name": "Llama 3.1 8B (decentralized)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ "text" @@ -176278,33 +184980,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 205000, + "output": 205000 }, "cost": { - "input": 0.02, - "output": 0.03 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "claude-sonnet-4-5-20250929-thinking": { - "id": "claude-sonnet-4-5-20250929-thinking", - "name": "Claude Sonnet 4.5 Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "stepfun/step-5-preview": { + "id": "stepfun/step-5-preview", + "name": "Step 5 Preview", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-09-16", + "last_updated": "2026-09-20", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -176314,131 +185015,111 @@ "limit": { "context": 1000000, "input": 1000000, - "output": 64000 + "output": 1000000 }, "cost": { - "input": 2.992, - "output": 14.994 + "input": 1, + "output": 2.7, + "cache_read": 0.05 } }, - "asi1-mini": { - "id": "asi1-mini", - "name": "ASI1 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "stepfun/step-3.5-flash": { + "id": "stepfun/step-3.5-flash", + "name": "StepFun 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 - }, - "cost": { - "input": 1, - "output": 1 - } - }, - "qwen3.5-27b": { - "id": "qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 260096, - "input": 260096, - "output": 65536 + "context": 262114, + "output": 262114 }, "cost": { - "input": 0.27, - "output": 2.16 + "input": 0.09, + "output": 0.3, + "cache_read": 0.02 } }, - "Qwen3.5-27B-Omega-Evolution-v2.2-Derestricted-Lite": { - "id": "Qwen3.5-27B-Omega-Evolution-v2.2-Derestricted-Lite", - "name": "Qwen3.5 27B Omega Evolution v2.2 Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "stepfun/step-3.7-flash": { + "id": "stepfun/step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-02", - "last_updated": "2026-05-02", + "tool_call": true, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 } }, - "claude-opus-4-1-thinking:32768": { - "id": "claude-opus-4-1-thinking:32768", - "name": "Claude 4.1 Opus Thinking (32K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "inception/mercury-coder-small": { + "id": "inception/mercury-coder-small", + "name": "Mercury Coder Small Beta", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "mercury", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "temperature": true, + "release_date": "2025-02-26", + "last_updated": "2025-02-26", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -176446,61 +185127,35 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 32000, + "output": 16384 }, "cost": { - "input": 14.994, - "output": 75.004 + "input": 0.25, + "output": 1 } }, - "gemma-4-31B-K1-v5": { - "id": "gemma-4-31B-K1-v5", - "name": "Gemma 4 31B K1 v5", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "inception/mercury-2.5": { + "id": "inception/mercury-2.5", + "name": "Mercury 2.5", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-02", - "last_updated": "2026-05-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306 - } - }, - "sonar": { - "id": "sonar", - "name": "Perplexity Simple", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "tool_call": true, + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ "text" @@ -176511,26 +185166,36 @@ }, "open_weights": false, "limit": { - "context": 127000, - "input": 127000, - "output": 128000 + "context": 260000, + "output": 65536 }, "cost": { - "input": 1.003, - "output": 1.003 + "input": 0.04, + "output": 0.15, + "cache_read": 0.004 } }, - "MiniMax-M2": { - "id": "MiniMax-M2", - "name": "MiniMax M2", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "inception/mercury-2": { + "id": "inception/mercury-2", + "name": "Mercury 2", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-10-25", - "last_updated": "2025-10-25", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-24", + "last_updated": "2026-03-06", "modalities": { "input": [ "text" @@ -176541,26 +185206,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 131072 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.17, - "output": 1.53 + "input": 0.25, + "output": 0.75, + "cache_read": 0.024999999999999998 } }, - "sarvam-105b": { - "id": "sarvam-105b", - "name": "Sarvam 105B", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "amazon/titan-embed-text-v2": { + "id": "amazon/titan-embed-text-v2", + "name": "Titan Text Embeddings V2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "titan-embed", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-04-30", + "last_updated": "2024-04", "modalities": { "input": [ "text" @@ -176571,20 +185236,15 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 4096 - }, - "cost": { - "input": 0.045, - "output": 0.177, - "cache_read": 0.028 + "context": 8192, + "output": 1536 } }, - "Qwen3.5-27B-Writer-V2-Derestricted-Lite": { - "id": "Qwen3.5-27B-Writer-V2-Derestricted-Lite", - "name": "Qwen3.5 27B Writer V2 Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "amazon/nova-2-lite": { + "id": "amazon/nova-2-lite", + "name": "Nova 2 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -176592,20 +185252,24 @@ "type": "toggle" }, { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": false, - "structured_output": false, - "release_date": "2026-04-06", - "last_updated": "2026-04-06", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-02", + "last_updated": "2024-12-01", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -176613,40 +185277,33 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.3, + "output": 2.5, + "cache_read": 0.075 } }, - "Qwen3.5-27B-Marvin-V2-Derestricted": { - "id": "Qwen3.5-27B-Marvin-V2-Derestricted", - "name": "Qwen3.5 27B Marvin V2 Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "amazon/nova-lite": { + "id": "amazon/nova-lite", + "name": "Nova Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 - } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text", "image", - "video" + "video", + "pdf" ], "output": [ "text" @@ -176654,30 +185311,33 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 300000, + "output": 10000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.06, + "output": 0.24, + "cache_read": 0.015, + "cache_write": 0.06 } }, - "claw-high": { - "id": "claw-high", - "name": "Claw High", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "amazon/nova-pro": { + "id": "amazon/nova-pro", + "name": "Nova Pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ @@ -176686,25 +185346,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 300000, + "output": 10000 }, "cost": { - "input": 4.998, - "output": 25.007 + "input": 0.8, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0.8 } }, - "qwen-max": { - "id": "qwen-max", - "name": "Qwen 2.5 Max", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "amazon/nova-micro": { + "id": "amazon/nova-micro", + "name": "Nova Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-04-03", - "last_updated": "2024-04-03", + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text" @@ -176715,61 +185378,27 @@ }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 8192 - }, - "cost": { - "input": 1.5997, - "output": 6.392 - } - }, - "gemma-4-31B-Fabled": { - "id": "gemma-4-31B-Fabled", - "name": "Gemma 4 31B Fabled", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-02", - "last_updated": "2026-05-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 128000, + "output": 10000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.035, + "output": 0.14, + "cache_read": 0.00875, + "cache_write": 0.035 } }, - "v0-1.5-md": { - "id": "v0-1.5-md", - "name": "v0 1.5 MD", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "mistral/codestral-embed": { + "id": "mistral/codestral-embed", + "name": "Codestral Embed", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "codestral-embed", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-07-04", - "last_updated": "2025-07-04", + "temperature": true, + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -176780,31 +185409,25 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15 + "context": 8192, + "output": 1536 } }, - "claude-opus-4-thinking:1024": { - "id": "claude-opus-4-thinking:1024", - "name": "Claude 4 Opus Thinking (1K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistral/mistral-nemo": { + "id": "mistral/mistral-nemo", + "name": "Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-07-18", + "last_updated": "2024-07-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -176812,39 +185435,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 60288, + "output": 16000 }, "cost": { - "input": 14.994, - "output": 75.004 + "input": 0.04, + "output": 0.17 } }, - "gemini-2.5-flash-preview-04-17": { - "id": "gemini-2.5-flash-preview-04-17", - "name": "Gemini 2.5 Flash Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "mistral/mistral-medium-3.5": { + "id": "mistral/mistral-medium-3.5", + "name": "Mistral Medium Latest", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "none", - "low", - "medium", "high" ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-17", - "last_updated": "2025-04-17", + "tool_call": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-05-21", "modalities": { "input": [ "text", @@ -176856,35 +185474,27 @@ }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 262144, + "output": 256000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 1.5, + "output": 7.5, + "cache_read": 0.15 } }, - "gemini-2.5-pro-exp-03-25": { - "id": "gemini-2.5-pro-exp-03-25", - "name": "Gemini 2.5 Pro Experimental 0325", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "mistral/mistral-small": { + "id": "mistral/mistral-small", + "name": "Mistral Small (latest)", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2024-09-17", + "last_updated": "2026-03-16", "modalities": { "input": [ "text", @@ -176894,31 +185504,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 262144, + "output": 4000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "hermes-medium": { - "id": "hermes-medium", - "name": "Hermes Medium", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "mistral/mistral-large-3": { + "id": "mistral/mistral-large-3", + "name": "Mistral Large 3", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -176926,25 +185538,26 @@ }, "open_weights": false, "limit": { - "context": 204800, - "input": 204800, - "output": 131072 + "context": 262144, + "output": 256000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.5, + "output": 1.5, + "cache_read": 0.05 } }, - "sonar-pro": { - "id": "sonar-pro", - "name": "Perplexity Pro", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "mistral/mistral-embed": { + "id": "mistral/mistral-embed", + "name": "Mistral Embed", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "mistral-embed", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "temperature": true, + "release_date": "2023-12-11", + "last_updated": "2023-12-11", "modalities": { "input": [ "text" @@ -176955,28 +185568,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 128000 - }, - "cost": { - "input": 2.992, - "output": 14.994 + "context": 8192, + "output": 1536 } }, - "doubao-1-5-thinking-pro-250415": { - "id": "doubao-1-5-thinking-pro-250415", - "name": "Doubao 1.5 Thinking Pro", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "mistral/ministral-14b": { + "id": "mistral/ministral-14b", + "name": "Ministral 14B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-04-17", - "last_updated": "2025-04-17", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-02", + "last_updated": "2025-12-01", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ @@ -176985,32 +185596,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 262144, + "output": 256000 }, "cost": { - "input": 0.6, - "output": 2.4 + "input": 0.2, + "output": 0.2, + "cache_read": 0.02 } }, - "qwen3.7-max:thinking": { - "id": "qwen3.7-max:thinking", - "name": "Qwen3.7 Max Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "mistral/codestral": { + "id": "mistral/codestral", + "name": "Codestral (latest)", + "description": "Mistral code model for completions, refactors, and developer IDE workflows", + "family": "codestral", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 262144 - } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-05-29", + "last_updated": "2025-01-04", "modalities": { "input": [ "text" @@ -177019,163 +185625,134 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 65536 + "context": 256000, + "output": 4096 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.25 + "input": 0.3, + "output": 0.9 } }, - "ernie-4.5-turbo-vl-32k": { - "id": "ernie-4.5-turbo-vl-32k", - "name": "Ernie 4.5 Turbo VL 32k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "mistral/ministral-8b": { + "id": "mistral/ministral-8b", + "name": "Ministral 8B (latest)", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-08", - "last_updated": "2025-05-08", + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "input": 32000, - "output": 16384 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.495, - "output": 1.43 + "input": 0.1, + "output": 0.1 } }, - "gemini-2.5-pro-preview-05-06": { - "id": "gemini-2.5-pro-preview-05-06", - "name": "Gemini 2.5 Pro Preview 0506", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-06", - "last_updated": "2025-05-06", + "mistral/ministral-3b": { + "id": "mistral/ministral-3b", + "name": "Ministral 3B (latest)", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 128000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.04, + "output": 0.04 } }, - "hunyuan-turbos-20250226": { - "id": "hunyuan-turbos-20250226", - "name": "Hunyuan Turbo S", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "fish-audio/s2.1-pro": { + "id": "fish-audio/s2.1-pro", + "name": "S2.1 Pro", + "description": "Speech generation model for controllable voice, narration, and audio delivery", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "release_date": "2026-07-28", + "last_updated": "2026-07-28", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 24000, - "input": 24000, - "output": 8192 - }, - "cost": { - "input": 0.187, - "output": 0.374 + "context": 0, + "output": 0 } }, - "claude-sonnet-4-thinking:64000": { - "id": "claude-sonnet-4-thinking:64000", - "name": "Claude 4 Sonnet Thinking (64K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "fish-audio/s1": { + "id": "fish-audio/s1", + "name": "S1", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2025-10-20", + "last_updated": "2025-10-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 64000 - }, - "cost": { - "input": 2.992, - "output": 14.994 + "context": 0, + "output": 0 } }, - "ernie-x1-32k": { - "id": "ernie-x1-32k", - "name": "Ernie X1 32k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "fish-audio/transcribe-1": { + "id": "fish-audio/transcribe-1", + "name": "Transcribe-1", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-05-08", - "last_updated": "2025-05-08", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ - "text", - "image" + "audio" ], "output": [ "text" @@ -177183,54 +185760,44 @@ }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 16384 - }, - "cost": { - "input": 0.33, - "output": 1.32 + "context": 0, + "output": 0 } }, - "command-a-reasoning-08-2025": { - "id": "command-a-reasoning-08-2025", - "name": "Cohere Command A (08/2025)", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "fish-audio/s2-pro": { + "id": "fish-audio/s2-pro", + "name": "S2 Pro", + "description": "Speech generation model for controllable voice, narration, and audio delivery", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-08-22", - "last_updated": "2025-08-22", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 8192 - }, - "cost": { - "input": 2.5, - "output": 10 + "context": 0, + "output": 0 } }, - "doubao-seed-2-0-code-preview-260215": { - "id": "doubao-seed-2-0-code-preview-260215", - "name": "Doubao Seed 2.0 Code Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "morph/morph-v3-fast": { + "id": "morph/morph-v3-fast", + "name": "Morph v3 Fast", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "morph", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "temperature": false, + "release_date": "2024-08-15", + "last_updated": "2024-08-15", "modalities": { "input": [ "text" @@ -177241,25 +185808,25 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 128000 + "context": 16000, + "output": 16000 }, "cost": { - "input": 0.782, - "output": 3.893 + "input": 0.8, + "output": 1.2 } }, - "doubao-1.5-pro-256k": { - "id": "doubao-1.5-pro-256k", - "name": "Doubao 1.5 Pro 256k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "morph/morph-v3-large": { + "id": "morph/morph-v3-large", + "name": "Morph v3 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "morph", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "temperature": false, + "release_date": "2024-08-15", + "last_updated": "2024-08-15", "modalities": { "input": [ "text" @@ -177270,30 +185837,32 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 16384 + "context": 32000, + "output": 32000 }, "cost": { - "input": 0.799, - "output": 1.445 + "input": 0.9, + "output": 1.9 } }, - "qwen3.5-35b-a3b": { - "id": "qwen3.5-35b-a3b", - "name": "Qwen3.5 35B A3B", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-4.1-mini-fast": { + "id": "openai/gpt-4.1-mini-fast", + "name": "GPT-4.1 mini (Fast)", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -177301,66 +185870,44 @@ }, "open_weights": false, "limit": { - "context": 260096, - "input": 260096, - "output": 65536 + "context": 1047576, + "input": 1014808, + "output": 32768 }, "cost": { - "input": 0.225, - "output": 1.8 + "input": 0.7, + "output": 2.8, + "cache_read": 0.175 } }, - "qwen3.5-122b-a10b:thinking": { - "id": "qwen3.5-122b-a10b:thinking", - "name": "Qwen3.5 122B A10B Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/o3-fast": { + "id": "openai/o3-fast", + "name": "o3 (Fast)", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 260096, - "input": 260096, - "output": 65536 - }, - "cost": { - "input": 0.36, - "output": 2.88 - } - }, - "yi-lightning": { - "id": "yi-lightning", - "name": "Yi Lightning", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-10-16", - "last_updated": "2024-10-16", - "modalities": { - "input": [ - "text" + "pdf" ], "output": [ "text" @@ -177368,25 +185915,41 @@ }, "open_weights": false, "limit": { - "context": 12000, - "input": 12000, - "output": 4096 + "context": 200000, + "input": 100000, + "output": 100000 }, "cost": { - "input": 0.2006, - "output": 0.2006 + "input": 3.5, + "output": 14, + "cache_read": 0.875 } }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT 5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -177399,55 +185962,55 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 64000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 2.992, - "output": 14.994 - } - }, - "deepseek-math-v2": { - "id": "deepseek-math-v2", - "name": "DeepSeek Math V2", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", - "modalities": { - "input": [ - "text" + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "input": 128000, - "output": 65536 - }, - "cost": { - "input": 0.6, - "output": 2.2 + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "claude-opus-4-1-thinking:8192": { - "id": "claude-opus-4-1-thinking:8192", - "name": "Claude 4.1 Opus Thinking (8K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT 5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "structured_output": false, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -177460,25 +186023,51 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 14.994, - "output": 75.004 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "deepseek-reasoner": { - "id": "deepseek-reasoner", - "name": "DeepSeek Reasoner", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-oss-safeguard-120b": { + "id": "openai/gpt-oss-safeguard-120b", + "name": "GPT OSS Safeguard 120B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", "modalities": { "input": [ "text" @@ -177487,32 +186076,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 64000, - "input": 64000, - "output": 65536 + "context": 128000, + "input": 112000, + "output": 16000 }, "cost": { - "input": 0.4, - "output": 1.7 + "input": 0.15, + "output": 0.6 } }, - "gemini-2.5-flash-nothinking": { - "id": "gemini-2.5-flash-nothinking", - "name": "Gemini 2.5 Flash (No Thinking)", + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5 Turbo", "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "family": "gpt", + "attachment": false, "reasoning": false, "tool_call": false, "structured_output": false, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "temperature": true, + "knowledge": "2021-09", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -177520,40 +186110,45 @@ }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 16385, + "input": 12289, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 0.5, + "output": 1.5 } }, - "Qwen3.5-27B-Omega-Evolution-v2.0-Derestricted-Lite": { - "id": "Qwen3.5-27B-Omega-Evolution-v2.0-Derestricted-Lite", - "name": "Qwen3.5 27B Omega Evolution v2.0 Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.2-fast": { + "id": "openai/gpt-5.2-fast", + "name": "GPT 5.2 (Fast)", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-06", - "last_updated": "2026-04-06", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -177561,28 +186156,30 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 3.5, + "output": 28, + "cache_read": 0.35 } }, - "fastgpt": { - "id": "fastgpt", - "name": "Web Answer", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-4o-transcribe": { + "id": "openai/gpt-4o-transcribe", + "name": "GPT-4o Transcribe", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "gpt", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2023-08-01", - "last_updated": "2024-01-01", + "temperature": true, + "release_date": "2024-03-13", + "last_updated": "2024-03-13", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" @@ -177590,40 +186187,42 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 0, + "output": 0 }, "cost": { - "input": 7.5, - "output": 7.5 + "input": 2.5, + "output": 10 } }, - "Qwen3.5-27B-Infracelestial": { - "id": "Qwen3.5-27B-Infracelestial", - "name": "Qwen3.5 27B Infracelestial", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT 5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -177631,25 +186230,40 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1000000, + "input": 872000, + "output": 128000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "glm-4-flash": { - "id": "glm-4-flash", - "name": "GLM-4 Flash", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/text-embedding-3-small": { + "id": "openai/text-embedding-3-small", + "name": "text-embedding-3-small", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-08-01", - "last_updated": "2024-08-01", + "temperature": true, + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -177660,58 +186274,41 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 - }, - "cost": { - "input": 0.1003, - "output": 0.1003 + "context": 8192, + "input": 6656, + "output": 1536 } }, - "azure-gpt-4o-mini": { - "id": "azure-gpt-4o-mini", - "name": "Azure gpt-4o-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT 5.4 Nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "input": 128000, - "output": 16384 - }, - "cost": { - "input": 0.1496, - "output": 0.595 - } - }, - "sonar-deep-research": { - "id": "sonar-deep-research", - "name": "Perplexity Deep Research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-25", - "last_updated": "2025-02-25", - "modalities": { - "input": [ - "text" + "image", + "pdf" ], "output": [ "text" @@ -177719,87 +186316,92 @@ }, "open_weights": false, "limit": { - "context": 60000, - "input": 60000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 3.4, - "output": 13.6 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "qwq-32b": { - "id": "qwq-32b", - "name": "Qwen: QwQ 32B", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-realtime-2.1": { + "id": "openai/gpt-realtime-2.1", + "name": "gpt-realtime-2.1", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": false, "structured_output": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { "context": 128000, - "input": 128000, - "output": 32768 + "input": 96000, + "output": 32000 }, "cost": { - "input": 0.25599999, - "output": 0.30499999 + "input": 4, + "output": 24, + "cache_read": 0.4 } }, - "mistral-code-agent-latest": { - "id": "mistral-code-agent-latest", - "name": "Mistral Code Agent Latest", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2-Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "input": 262144, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 2 - } - }, - "glm-4.1v-thinking-flash": { - "id": "glm-4.1v-thinking-flash", - "name": "GLM 4.1V Thinking Flash", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -177807,31 +186409,34 @@ }, "open_weights": false, "limit": { - "context": 64000, - "input": 64000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 0.3 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "qwen3.5-omni-flash": { - "id": "qwen3.5-omni-flash", - "name": "Qwen3.5 Omni Flash", - "description": "Omni-modal model for text, vision, audio, and multimodal agent tasks", + "openai/gpt-4o-fast": { + "id": "openai/gpt-4o-fast", + "name": "GPT-4o (Fast)", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -177839,66 +186444,44 @@ }, "open_weights": false, "limit": { - "context": 49152, - "input": 49152, + "context": 128000, + "input": 111616, "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 4.25, + "output": 17, + "cache_read": 2.125 } }, - "gemma-4-31B-MeroMero": { - "id": "gemma-4-31B-MeroMero", - "name": "Gemma 4 31B MeroMero", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1-Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-02", - "last_updated": "2026-05-02", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306 - } - }, - "gemini-2.5-flash-preview-04-17:thinking": { - "id": "gemini-2.5-flash-preview-04-17:thinking", - "name": "Gemini 2.5 Flash Preview Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-17", - "last_updated": "2025-04-17", - "modalities": { - "input": [ - "text", - "image" + "pdf" ], "output": [ "text" @@ -177906,86 +186489,92 @@ }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 3.5 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "glm-4-air": { - "id": "glm-4-air", - "name": "GLM-4 Air", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-live-1": { + "id": "openai/gpt-live-1", + "name": "GPT-Live 1", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-06-05", - "last_updated": "2024-06-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 - }, - "cost": { - "input": 0.2006, - "output": 0.2006 + "context": 0, + "output": 0 } }, - "doubao-seed-1-6-thinking-250615": { - "id": "doubao-seed-1-6-thinking-250615", - "name": "Doubao Seed 1.6 Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-realtime-1.5": { + "id": "openai/gpt-realtime-1.5", + "name": "GPT-Realtime-1.5", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-06-15", - "last_updated": "2025-06-15", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 16384 + "context": 0, + "output": 0 }, "cost": { - "input": 0.204, - "output": 2.04 + "input": 4, + "output": 16, + "cache_read": 0.4 } }, - "auto-model-basic": { - "id": "auto-model-basic", - "name": "Auto model (Basic)", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "attachment": false, + "openai/gpt-4.1-nano-fast": { + "id": "openai/gpt-4.1-nano-fast", + "name": "GPT-4.1 nano (Fast)", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -177993,40 +186582,44 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 + "context": 1047576, + "input": 1014808, + "output": 32768 }, "cost": { - "input": 9.996, - "output": 19.992 + "input": 0.2, + "output": 0.8, + "cache_read": 0.05 } }, - "Qwen3.5-27B-RpRMax-v1": { - "id": "Qwen3.5-27B-RpRMax-v1", - "name": "Qwen3.5 27B RpRMax v1", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -178034,72 +186627,77 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "ernie-5.1": { - "id": "ernie-5.1", - "name": "ERNIE 5.1", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "openai/gpt-image-2": { + "id": "openai/gpt-image-2", + "name": "GPT Image 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-05-10", - "last_updated": "2026-05-10", + "temperature": true, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 119000, - "input": 119000, - "output": 64000 + "context": 0, + "output": 0 }, "cost": { - "input": 0.75, - "output": 3, - "cache_read": 0.75 + "input": 5, + "output": 30, + "cache_read": 1.25 } }, - "Qwen3.5-27B-Marvin-DPO-V2-Derestricted": { - "id": "Qwen3.5-27B-Marvin-DPO-V2-Derestricted", - "name": "Qwen3.5 27B Marvin DPO V2 Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.6-luna-fast": { + "id": "openai/gpt-5.6-luna-fast", + "name": "GPT 5.6 Luna (Fast)", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -178107,29 +186705,65 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.4, + "output": 2.4, + "cache_read": 0.04, + "cache_write": 0.5, + "tiers": [ + { + "input": 0.8, + "output": 3.6, + "cache_read": 0.08, + "cache_write": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.8, + "output": 3.6, + "cache_read": 0.08, + "cache_write": 1 + } } }, - "azure-gpt-4o": { - "id": "azure-gpt-4o", - "name": "Azure gpt-4o", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-6-astra-fast": { + "id": "openai/gpt-6-astra-fast", + "name": "GPT-6 Astra (Fast)", + "description": "Fast variant of GPT-6 Astra for low-latency assistance and high-volume workloads.", + "family": "gpt-astra", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "temperature": false, + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -178137,28 +186771,62 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 2.499, - "output": 9.996 + "input": 20, + "output": 100, + "cache_read": 2, + "cache_write": 25, + "tiers": [ + { + "input": 40, + "output": 150, + "cache_read": 4, + "cache_write": 50, + "tier": { + "type": "context", + "size": 272001 + } + } + ], + "context_over_200k": { + "input": 40, + "output": 150, + "cache_read": 4, + "cache_write": 50 + } } }, - "deepseek-chat": { - "id": "deepseek-chat", - "name": "DeepSeek V3/Deepseek Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT 5.2 ", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "structured_output": false, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ @@ -178167,29 +186835,31 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 0.7 + "input": 21, + "output": 168 } }, - "gemini-2.5-flash-preview-05-20": { - "id": "gemini-2.5-flash-preview-05-20", - "name": "Gemini 2.5 Flash 0520", + "openai/gpt-5.1-thinking-fast": { + "id": "openai/gpt-5.1-thinking-fast", + "name": "GPT 5.1 Thinking (Fast)", "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "release_date": "2025-11-12", + "last_updated": "2025-11-12", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -178197,26 +186867,27 @@ }, "open_weights": false, "limit": { - "context": 1048000, - "input": 1048000, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 2.5, + "output": 20, + "cache_read": 0.25 } }, - "mercury-2": { - "id": "mercury-2", - "name": "Mercury 2", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/text-embedding-ada-002": { + "id": "openai/text-embedding-ada-002", + "name": "text-embedding-ada-002", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2022-12-15", + "last_updated": "2022-12-15", "modalities": { "input": [ "text" @@ -178227,38 +186898,40 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 50000 - }, - "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.025 + "context": 8192, + "input": 6656, + "output": 1536 } }, - "qwen3.7-plus:thinking": { - "id": "qwen3.7-plus:thinking", - "name": "Qwen3.7 Plus Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT 5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 262144 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -178266,29 +186939,30 @@ }, "open_weights": false, "limit": { - "context": 983616, - "input": 983616, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.04 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "gemini-2.0-flash-thinking-exp-1219": { - "id": "gemini-2.0-flash-thinking-exp-1219", - "name": "Gemini 2.0 Flash Thinking 1219", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/whisper-1": { + "id": "openai/whisper-1", + "name": "Whisper", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-12-19", - "last_updated": "2024-12-19", + "temperature": true, + "release_date": "2022-09-21", + "last_updated": "2022-09-21", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" @@ -178296,28 +186970,39 @@ }, "open_weights": false, "limit": { - "context": 32767, - "input": 32767, - "output": 8192 - }, - "cost": { - "input": 0.1003, - "output": 0.408 + "context": 0, + "output": 0 } }, - "glm-4-plus-0111": { - "id": "glm-4-plus-0111", - "name": "GLM 4 Plus 0111", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "openai/gpt-5.3-codex-fast": { + "id": "openai/gpt-5.3-codex-fast", + "name": "GPT 5.3 Codex (Fast)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -178325,28 +187010,46 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 9.996, - "output": 9.996 + "input": 3.5, + "output": 28, + "cache_read": 0.35 } }, - "brave": { - "id": "brave", - "name": "Brave (Answers)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2023-03-02", - "last_updated": "2024-01-01", + "openai/gpt-5.5-fast": { + "id": "openai/gpt-5.5-fast", + "name": "GPT 5.5 (Fast)", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -178354,72 +187057,76 @@ }, "open_weights": false, "limit": { - "context": 8192, - "input": 8192, - "output": 8192 + "context": 1000000, + "input": 872000, + "output": 128000 }, "cost": { - "input": 5, - "output": 5 + "input": 12.5, + "output": 75, + "cache_read": 1.25 } }, - "glm-zero-preview": { - "id": "glm-zero-preview", - "name": "GLM Zero Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-image-2.5-flare": { + "id": "openai/gpt-image-2.5-flare", + "name": "GPT Image 2.5 Flare", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "temperature": false, + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 8000, - "input": 8000, - "output": 4096 + "context": 0, + "output": 0 }, "cost": { - "input": 1.802, - "output": 1.802 + "input": 5, + "output": 30, + "cache_read": 1.25 } }, - "gemini-2.5-flash-lite-preview-06-17": { - "id": "gemini-2.5-flash-lite-preview-06-17", - "name": "Gemini 2.5 Flash Lite Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.4-fast": { + "id": "openai/gpt-5.4-fast", + "name": "GPT 5.4 (Fast)", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -178427,40 +187134,47 @@ }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "Qwen3.5-27B-Writer-Derestricted": { - "id": "Qwen3.5-27B-Writer-Derestricted", - "name": "Qwen3.5 27B Writer Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.6-sol-fast": { + "id": "openai/gpt-5.6-sol-fast", + "name": "GPT 5.6 Sol (Fast)", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-06", - "last_updated": "2026-04-06", + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -178468,28 +187182,63 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 8, + "output": 40, + "cache_read": 0.8, + "cache_write": 10, + "tiers": [ + { + "input": 16, + "output": 60, + "cache_read": 1.6, + "cache_write": 20, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 16, + "output": 60, + "cache_read": 1.6, + "cache_write": 20 + } } }, - "KAT-Coder-Exp-72B-1010": { - "id": "KAT-Coder-Exp-72B-1010", - "name": "KAT Coder Exp 72B 1010", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", + "openai/o4-mini-fast": { + "id": "openai/o4-mini-fast", + "name": "o4-mini (Fast)", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -178497,39 +187246,42 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 32768 + "context": 200000, + "input": 100000, + "output": 100000 }, "cost": { - "input": 0.1, - "output": 0.2 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "gemini-2.5-pro-preview-06-05": { - "id": "gemini-2.5-pro-preview-06-05", - "name": "Gemini 2.5 Pro Preview 0605", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5 pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", "high" ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -178537,66 +187289,72 @@ }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 400000, + "input": 128000, + "output": 272000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 15, + "output": 120 } }, - "Qwen3.5-27B-Musica-v1": { - "id": "Qwen3.5-27B-Musica-v1", - "name": "Qwen3.5 27B Musica v1", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.1-thinking": { + "id": "openai/gpt-5.1-thinking", + "name": "GPT 5.1 Thinking", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-11-12", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "MiniMax-M1": { - "id": "MiniMax-M1", - "name": "MiniMax M1", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/text-embedding-3-large": { + "id": "openai/text-embedding-3-large", + "name": "text-embedding-3-large", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-06-16", - "last_updated": "2025-06-16", + "temperature": true, + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -178607,29 +187365,42 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.1394, - "output": 1.3328 + "context": 8192, + "input": 6656, + "output": 1536 } }, - "doubao-1-5-thinking-pro-vision-250415": { - "id": "doubao-1-5-thinking-pro-vision-250415", - "name": "Doubao 1.5 Thinking Pro Vision", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-6-astra": { + "id": "openai/gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -178637,37 +187408,65 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.4 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272001 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } } }, - "qwen3.5-27b:thinking": { - "id": "qwen3.5-27b:thinking", - "name": "Qwen3.5 27B Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.4-mini-fast": { + "id": "openai/gpt-5.4-mini-fast", + "name": "GPT 5.4 Mini (Fast)", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -178675,40 +187474,77 @@ }, "open_weights": false, "limit": { - "context": 260096, - "input": 260096, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 2.16 + "input": 1.5, + "output": 9, + "cache_read": 0.15 } }, - "Qwen3.5-27B-Omega-Evolution-v2.0-Derestricted": { - "id": "Qwen3.5-27B-Omega-Evolution-v2.0-Derestricted", - "name": "Qwen3.5 27B Omega Evolution v2.0 Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-realtime-mini": { + "id": "openai/gpt-realtime-mini", + "name": "GPT-Realtime mini", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-10-10", + "last_updated": "2025-10-10", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text", + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.06 + } + }, + "openai/gpt-5-mini-fast": { + "id": "openai/gpt-5-mini-fast", + "name": "GPT-5 mini (Fast)", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-06", - "last_updated": "2026-04-06", + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -178716,35 +187552,44 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.45, + "output": 3.6, + "cache_read": 0.045 } }, - "Gemma-4-31B-GarnetV2": { - "id": "Gemma-4-31B-GarnetV2", - "name": "Gemma 4 31B Garnet V2", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/o3-pro": { + "id": "openai/o3-pro", + "name": "o3 Pro", + "description": "High-effort o3 tier for difficult technical reasoning and careful answers", + "family": "o-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -178752,25 +187597,37 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 200000, + "input": 100000, + "output": 100000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 20, + "output": 80 } }, - "qwen-turbo": { - "id": "qwen-turbo", - "name": "Qwen Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-11-01", - "last_updated": "2024-11-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -178779,27 +187636,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, + "context": 131072, + "input": 122880, "output": 8192 }, "cost": { - "input": 0.04998, - "output": 0.2006 + "input": 0.03, + "output": 0.14 } }, - "phi-4-multimodal-instruct": { - "id": "phi-4-multimodal-instruct", - "name": "Phi 4 Multimodal", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-oss-safeguard-20b": { + "id": "openai/gpt-oss-safeguard-20b", + "name": "gpt-oss-safeguard-20b", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", "modalities": { "input": [ "text" @@ -178808,61 +187677,78 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "input": 128000, - "output": 16384 + "input": 112000, + "output": 16000 }, "cost": { "input": 0.07, - "output": 0.11 + "output": 0.2 } }, - "mistral-small-31-24b-instruct": { - "id": "mistral-small-31-24b-instruct", - "name": "Mistral Small 31 24b Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "openai/gpt-image-2.5-sunburst": { + "id": "openai/gpt-image-2.5-sunburst", + "name": "GPT Image 2.5 Sunburst", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "temperature": false, + "release_date": "2026-09-08", + "last_updated": "2026-09-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 131072 + "context": 0, + "output": 0 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 5, + "output": 30, + "cache_read": 1.25 } }, - "ernie-4.5-turbo-128k": { - "id": "ernie-4.5-turbo-128k", - "name": "Ernie 4.5 Turbo 128k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.6-terra-fast": { + "id": "openai/gpt-5.6-terra-fast", + "name": "GPT 5.6 Terra (Fast)", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-08", - "last_updated": "2025-05-08", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -178870,67 +187756,90 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.132, - "output": 0.55 + "input": 4, + "output": 24, + "cache_read": 0.4, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 36, + "cache_read": 0.8, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 36, + "cache_read": 0.8, + "cache_write": 10 + } } }, - "Qwen3.5-27B-BlueStar-Derestricted": { - "id": "Qwen3.5-27B-BlueStar-Derestricted", - "name": "Qwen3.5 27B BlueStar Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 - } - ], + "openai/gpt-image-1": { + "id": "openai/gpt-image-1", + "name": "GPT Image 1", + "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", + "family": "gpt-image", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-04-06", - "last_updated": "2026-04-06", + "temperature": true, + "release_date": "2025-04-24", + "last_updated": "2025-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 0, + "output": 0 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 5, + "output": 40, + "cache_read": 1.25 } }, - "gemini-2.5-flash-lite-preview-09-2025-thinking": { - "id": "gemini-2.5-flash-lite-preview-09-2025-thinking", - "name": "Gemini 2.5 Flash Lite Preview (09/2025) – Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT 5.4 Mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -178943,28 +187852,47 @@ }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "doubao-seed-1-8-251215": { - "id": "doubao-seed-1-8-251215", - "name": "Doubao Seed 1.8", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT 5.6 Luna", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -178972,29 +187900,64 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.612, - "output": 6.12 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } } }, - "qwen3.6-max-preview": { - "id": "qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "qwen3.6", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-20", - "last_updated": "2026-04-21", + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT 5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179002,27 +187965,46 @@ }, "open_weights": false, "limit": { - "context": 245800, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.3, - "output": 7.8 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "exa-answer": { - "id": "exa-answer", - "name": "Exa (Answer)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-04", - "last_updated": "2025-06-04", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179030,28 +188012,46 @@ }, "open_weights": false, "limit": { - "context": 4096, - "input": 4096, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 2.5 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "Baichuan4-Air": { - "id": "Baichuan4-Air", - "name": "Baichuan 4 Air", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-19", - "last_updated": "2025-08-19", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT 5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179059,30 +188059,63 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 1000000, + "input": 872000, + "output": 128000 }, "cost": { - "input": 0.157, - "output": 0.157 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "qwen3.5-122b-a10b": { - "id": "qwen3.5-122b-a10b", - "name": "Qwen3.5 122B A10B", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-6-sol-fast": { + "id": "openai/gpt-6-sol-fast", + "name": "GPT-6 Sol (Fast)", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -179090,57 +188123,85 @@ }, "open_weights": false, "limit": { - "context": 260096, - "input": 260096, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.36, - "output": 2.88 + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272001 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10 + } } }, - "sarvam-30b": { - "id": "sarvam-30b", - "name": "Sarvam 30B", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/tts-1": { + "id": "openai/tts-1", + "name": "TTS-1", + "description": "Speech generation model for controllable voice, narration, and audio delivery", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 65536, - "input": 65536, - "output": 4096 - }, - "cost": { - "input": 0.028, - "output": 0.111, - "cache_read": 0.017 + "context": 0, + "output": 0 } }, - "claude-opus-4-1-thinking": { - "id": "claude-opus-4-1-thinking", - "name": "Claude 4.1 Opus Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-6-luna-fast": { + "id": "openai/gpt-6-luna-fast", + "name": "GPT-6 Luna (Fast)", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", @@ -179153,26 +188214,59 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 14.994, - "output": 75.004 + "input": 0.2, + "output": 1, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.5, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272001 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.5, + "cache_read": 0.04, + "cache_write": 0.5 + } } }, - "claude-opus-4-thinking:32000": { - "id": "claude-opus-4-thinking:32000", - "name": "Claude 4 Opus Thinking (32K)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5-fast": { + "id": "openai/gpt-5-fast", + "name": "GPT-5 (Fast)", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -179185,87 +188279,66 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 14.994, - "output": 75.004 + "input": 2.5, + "output": 20, + "cache_read": 0.25 } }, - "auto-model": { - "id": "auto-model", - "name": "Auto model", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "openai/gpt-realtime-2": { + "id": "openai/gpt-realtime-2", + "name": "gpt-realtime-2", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-06-01", - "last_updated": "2024-06-01", + "temperature": true, + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 + "context": 0, + "output": 0 }, "cost": { - "input": 0, - "output": 0 + "input": 4, + "output": 24, + "cache_read": 0.4 } }, - "qwen3-vl-235b-a22b-instruct-original": { - "id": "qwen3-vl-235b-a22b-instruct-original", - "name": "Qwen3 VL 235B A22B Instruct Original", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-4.1-fast": { + "id": "openai/gpt-4.1-fast", + "name": "GPT-4.1 (Fast)", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "input": 32768, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 1.2 - } - }, - "glm-z1-air": { - "id": "glm-z1-air", - "name": "GLM Z1 Air", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179273,40 +188346,47 @@ }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 16384 + "context": 1047576, + "input": 1014808, + "output": 32768 }, "cost": { - "input": 0.07, - "output": 0.07 + "input": 3.5, + "output": 14, + "cache_read": 0.875 } }, - "Qwen3.5-27B-Queen-Derestricted-Lite": { - "id": "Qwen3.5-27B-Queen-Derestricted-Lite", - "name": "Qwen3.5 27B Queen Derestricted Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-6-luna": { + "id": "openai/gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -179314,59 +188394,79 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.306, - "output": 0.306 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272001 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } } }, - "inclusionai/ling-2.6-1t": { - "id": "inclusionai/ling-2.6-1t", - "name": "Ling 2.6 1T", - "description": "Tool-capable chat model for instruction following and agentic application workflows", + "openai/gpt-image-1-mini": { + "id": "openai/gpt-image-1-mini", + "name": "GPT Image 1 Mini", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "tool_call": false, + "temperature": true, + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 0, + "output": 0 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.06 + "input": 2, + "output": 8, + "cache_read": 0.2 } }, - "inclusionai/ring-2.6-1t": { - "id": "inclusionai/ring-2.6-1t", - "name": "Ring 2.6 1T", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "openai/gpt-4o-mini-transcribe": { + "id": "openai/gpt-4o-mini-transcribe", + "name": "GPT-4o mini Transcribe", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "o-mini", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-08", - "last_updated": "2026-05-08", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-03-13", + "last_updated": "2024-03-13", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" @@ -179374,58 +188474,75 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 65536 + "context": 0, + "output": 0 }, "cost": { - "input": 1, - "output": 3 + "input": 1.25, + "output": 5 } }, - "inclusionai/ling-2.6-flash": { - "id": "inclusionai/ling-2.6-flash", - "name": "Ling 2.6 Flash", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "openai/gpt-image-1.5": { + "id": "openai/gpt-image-1.5", + "name": "GPT Image 1.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "tool_call": false, + "temperature": true, + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 0, + "output": 0 }, "cost": { - "input": 0.08, - "output": 0.24 + "input": 5, + "output": 32, + "cache_read": 1.25 } }, - "Alibaba-NLP/Tongyi-DeepResearch-30B-A3B": { - "id": "Alibaba-NLP/Tongyi-DeepResearch-30B-A3B", - "name": "Tongyi DeepResearch 30B A3B", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "yi", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT 5.6 Terra", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179433,25 +188550,57 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.08, - "output": 0.24000000000000002 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "ibm-granite/granite-4.1-8b": { - "id": "ibm-granite/granite-4.1-8b", - "name": "Granite 4.1 8B", - "description": "Tool-capable chat model for instruction following and agentic application workflows", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -179460,32 +188609,45 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, - "input": 131072, "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.1, - "cache_read": 0.05 + "input": 0.1, + "output": 0.5, + "cache_read": 0.1 } }, - "Salesforce/Llama-xLAM-2-70b-fc-r": { - "id": "Salesforce/Llama-xLAM-2-70b-fc-r", - "name": "Llama-xLAM-2 70B fc-r", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-13", - "last_updated": "2025-04-13", + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179493,29 +188655,47 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 2.5 + "input": 0.25, + "output": 2, + "cache_read": 0.03 } }, - "THUDM/GLM-Z1-32B-0414": { - "id": "THUDM/GLM-Z1-32B-0414", - "name": "GLM Z1 32B 0414", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm-z", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT 5.6 Sol", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179523,29 +188703,53 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10 + } } }, - "THUDM/GLM-4-32B-0414": { - "id": "THUDM/GLM-4-32B-0414", - "name": "GLM 4 32B 0414", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "openai/gpt-4o-mini-fast": { + "id": "openai/gpt-4o-mini-fast", + "name": "GPT-4o mini (Fast)", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179554,28 +188758,29 @@ "open_weights": false, "limit": { "context": 128000, - "input": 128000, - "output": 65536 + "input": 111616, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 0.25, + "output": 1, + "cache_read": 0.125 } }, - "THUDM/GLM-4-9B-0414": { - "id": "THUDM/GLM-4-9B-0414", - "name": "GLM 4 9B 0414", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "openai/gpt-realtime-whisper": { + "id": "openai/gpt-realtime-whisper", + "name": "gpt-realtime-whisper", + "description": "Streaming speech-to-text model for low-latency transcript deltas from live audio", + "family": "whisper", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "temperature": true, + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" @@ -179583,29 +188788,41 @@ }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 8000 - }, - "cost": { - "input": 0.2, - "output": 0.2 + "context": 0, + "output": 0 } }, - "THUDM/GLM-Z1-9B-0414": { - "id": "THUDM/GLM-Z1-9B-0414", - "name": "GLM Z1 9B 0414", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm-z", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "openai/gpt-6-sol": { + "id": "openai/gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179613,56 +188830,83 @@ }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 8000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272001 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "meta-llama/llama-3.1-8b-instruct": { - "id": "meta-llama/llama-3.1-8b-instruct", - "name": "Llama 3.1 8b Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "openai/tts-1-hd": { + "id": "openai/tts-1-hd", + "name": "TTS-1 HD", + "description": "Speech generation model for controllable voice, narration, and audio delivery", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "temperature": true, + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 16384 - }, - "cost": { - "input": 0.0544, - "output": 0.0544 + "context": 0, + "output": 0 } }, - "meta-llama/llama-4-maverick": { - "id": "meta-llama/llama-4-maverick", - "name": "Llama 4 Maverick", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -179674,29 +188918,44 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.18000000000000002, - "output": 0.8 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "meta-llama/llama-3.3-70b-instruct": { - "id": "meta-llama/llama-3.3-70b-instruct", - "name": "Llama 3.3 70b Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179704,26 +188963,28 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 16384 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.05, - "output": 0.23 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "meta-llama/llama-4-scout": { - "id": "meta-llama/llama-4-scout", - "name": "Llama 4 Scout", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ "text", @@ -179735,29 +188996,32 @@ }, "open_weights": false, "limit": { - "context": 328000, - "input": 328000, - "output": 65536 + "context": 128000, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 0.085, - "output": 0.46 + "input": 10, + "output": 30 } }, - "meta-llama/llama-3.2-3b-instruct": { - "id": "meta-llama/llama-3.2-3b-instruct", - "name": "Llama 3.2 3b Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ @@ -179766,29 +189030,33 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 8192 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.0306, - "output": 0.0493 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "featherless-ai/Qwerky-72B": { - "id": "featherless-ai/Qwerky-72B", - "name": "Qwerky 72B", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "qwerky", - "attachment": false, + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-20", - "last_updated": "2025-03-20", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179796,29 +189064,33 @@ }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 8192 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 0.5 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "moonshotai/kimi-k2-instruct-0711": { - "id": "moonshotai/kimi-k2-instruct-0711", - "name": "Kimi K2 0711", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2025-07-11", - "last_updated": "2025-07-11", + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179827,28 +189099,42 @@ "open_weights": false, "limit": { "context": 128000, - "input": 128000, - "output": 8192 + "output": 16384 }, "cost": { - "input": 0.1, - "output": 2 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "moonshotai/Kimi-K2-Instruct-0905": { - "id": "moonshotai/Kimi-K2-Instruct-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "openai/o1": { + "id": "openai/o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179856,30 +189142,44 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 262144 + "context": 200000, + "output": 100000 }, + "status": "deprecated", "cost": { - "input": 0.4, - "output": 2 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "moonshotai/kimi-k2-thinking-original": { - "id": "moonshotai/kimi-k2-thinking-original", - "name": "Kimi K2 Thinking Original", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -179887,27 +189187,29 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.5 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "moonshotai/kimi-k2.5:thinking": { - "id": "moonshotai/kimi-k2.5:thinking", - "name": "Kimi K2.5 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-01-26", - "last_updated": "2026-01-26", + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -179919,26 +189221,39 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 65536 + "context": 1047576, + "output": 32768 }, + "status": "deprecated", "cost": { - "input": 0.3, - "output": 1.9 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "moonshotai/kimi-k2-instruct": { - "id": "moonshotai/kimi-k2-instruct", - "name": "Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ "text" @@ -179949,29 +189264,43 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 8192 + "context": 200000, + "output": 100000 }, + "status": "deprecated", "cost": { - "input": 0.1, - "output": 2 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, - "reasoning": false, + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -179979,27 +189308,40 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 262144 + "context": 200000, + "output": 100000 }, + "status": "deprecated", "cost": { - "input": 0.3, - "output": 1.2 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "moonshotai/kimi-k2.6:thinking": { - "id": "moonshotai/kimi-k2.6:thinking", - "name": "Kimi K2.6 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-04-16", - "last_updated": "2026-04-21", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -180009,31 +189351,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.53, - "output": 2.73 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-01-26", - "last_updated": "2026-01-26", + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -180041,30 +189388,58 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.9 - } - }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + } + } + }, + "zai-coding-plan": { + "id": "zai-coding-plan", + "env": [ + "ZHIPU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.z.ai/api/coding/paas/v4", + "name": "Z.AI Coding Plan", + "doc": "https://docs.z.ai/devpack/overview", + "models": { + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-04-16", - "last_updated": "2026-04-21", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" @@ -180072,58 +189447,85 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.53, - "output": 2.73 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "moonshotai/kimi-latest": { - "id": "moonshotai/kimi-latest", - "name": "Kimi Latest", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "attachment": true, + "glm-5.3-highspeed": { + "id": "glm-5.3-highspeed", + "name": "GLM-5.3 Highspeed", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 2.6, - "cache_read": 0.125 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "moonshotai/kimi-k2-thinking-turbo-original": { - "id": "moonshotai/kimi-k2-thinking-turbo-original", - "name": "Kimi K2 Thinking Turbo Original", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "glm-5.2-highspeed": { + "id": "glm-5.2-highspeed", + "name": "GLM-5.2 Highspeed", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -180132,91 +189534,122 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 16384 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.15, - "output": 8 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "baidu/ernie-4.5-vl-28b-a3b": { - "id": "baidu/ernie-4.5-vl-28b-a3b", - "name": "ERNIE 4.5 VL 28B", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "ernie", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.13999999999999999, - "output": 0.5599999999999999 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "perceptron/perceptron-mk1": { - "id": "perceptron/perceptron-mk1", - "name": "Perceptron Mk1", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5": { - "id": "failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5", - "name": "Llama 3 70B abliterated", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "glm-5-turbo": { + "id": "glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text" @@ -180227,26 +189660,41 @@ }, "open_weights": false, "limit": { - "context": 8192, - "input": 8192, - "output": 8192 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.7, - "output": 0.7 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "nanogpt/coding-router:max": { - "id": "nanogpt/coding-router:max", - "name": "Coding Router Max", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -180255,29 +189703,47 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "input": 1000000, - "output": 128000 + "output": 131072 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "nanogpt/coding-router:high": { - "id": "nanogpt/coding-router:high", - "name": "Coding Router High", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + } + } + }, + "ebcloud": { + "id": "ebcloud", + "env": [ + "EBCLOUD_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://maas-api.ebcloud.com/v1", + "name": "EBCloud", + "doc": "https://docs.ebtech.com/ai/model-api.html", + "models": { + "Kimi-K2.6": { + "id": "Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text" @@ -180286,29 +189752,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.1, - "output": 2.2, - "cache_read": 0.11 + "input": 0.9286, + "output": 3.8571 } }, - "nanogpt/coding-router:low": { - "id": "nanogpt/coding-router:low", - "name": "Coding Router Low", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "GLM-5.1": { + "id": "GLM-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -180317,29 +189786,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0.8571, + "output": 3.4286 } }, - "nanogpt/coding-router:medium": { - "id": "nanogpt/coding-router:medium", - "name": "Coding Router Medium", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "DeepSeek-V4-Flash": { + "id": "DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -180348,29 +189821,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "input": 1000000, - "output": 128000 + "output": 384000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0.143, + "output": 0.2857 } }, - "nanogpt/coding-router": { - "id": "nanogpt/coding-router", - "name": "Coding Router", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "DeepSeek-V4-Pro": { + "id": "DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -180379,29 +189856,52 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "input": 1000000, - "output": 128000 + "output": 384000 }, "cost": { - "input": 1.1, - "output": 2.2, - "cache_read": 0.11 + "input": 0.4286, + "output": 0.8571 } - }, - "GalrionSoftworks/MN-LooseCannon-12B-v1": { - "id": "GalrionSoftworks/MN-LooseCannon-12B-v1", - "name": "MN-LooseCannon-12B-v1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + } + } + }, + "greenpt": { + "id": "greenpt", + "env": [ + "GREENPT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.greenpt.ai/v1", + "name": "GreenPT", + "doc": "https://docs.greenpt.ai", + "models": { + "glm-5.2-honey-ultra": { + "id": "glm-5.2-honey-ultra", + "name": "GLM-5.2 Honey Ultra", + "description": "glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -180410,27 +189910,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 1.254, + "output": 5.016, + "cache_read": 0.3135 } }, - "TheDrummer/Cydonia-24B-v4.3": { - "id": "TheDrummer/Cydonia-24B-v4.3", - "name": "The Drummer Cydonia 24B v4.3", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "glm-5.2-ponytail-lite": { + "id": "glm-5.2-ponytail-lite", + "name": "GLM-5.2 Ponytail Lite", + "description": "glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-25", - "last_updated": "2025-12-25", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -180439,56 +189953,73 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.1003, - "output": 0.1207 + "input": 1.254, + "output": 5.016, + "cache_read": 0.3135 } }, - "TheDrummer/Cydonia-24B-v4.1": { - "id": "TheDrummer/Cydonia-24B-v4.1", - "name": "The Drummer Cydonia 24B v4.1", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, + "gemma-3-27b-it": { + "id": "gemma-3-27b-it", + "name": "Gemma 3 27B", + "description": "Google Gemma 3 multimodal model for chat, reasoning, and image understanding", + "family": "gemma", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-19", - "last_updated": "2025-08-19", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 32768 + "context": 40000, + "output": 8192 }, "cost": { - "input": 0.1003, - "output": 0.1207 + "input": 0.342, + "output": 0.684 } }, - "TheDrummer/Cydonia-24B-v2": { - "id": "TheDrummer/Cydonia-24B-v2", - "name": "The Drummer Cydonia 24B v2", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-17", - "last_updated": "2025-02-17", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text" @@ -180497,27 +190028,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.1003, - "output": 0.1207 + "input": 0.127754, + "output": 0.511016, + "cache_read": 0.0255508 } }, - "TheDrummer/Anubis-70B-v1.1": { - "id": "TheDrummer/Anubis-70B-v1.1", - "name": "Anubis 70B v1.1", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -180526,116 +190072,160 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.31, - "output": 0.31 + "input": 0.1596, + "output": 0.399, + "cache_read": 0.0456 } }, - "TheDrummer/UnslopNemo-12B-v4.1": { - "id": "TheDrummer/UnslopNemo-12B-v4.1", - "name": "UnslopNemo 12b v4", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen3 235B MoE instruct model for long-context multilingual chat and reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-21", + "last_updated": "2025-07-21", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.493, - "output": 0.493 + "input": 1.026, + "output": 3.078 } }, - "TheDrummer/skyfall-36b-v2": { - "id": "TheDrummer/skyfall-36b-v2", - "name": "TheDrummer Skyfall 36B V2", - "description": "Multimodal model for analyzing text, images, documents, and rich media", + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-10", - "last_updated": "2025-03-10", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 64000, - "input": 64000, - "output": 32768 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.493, - "output": 0.493 + "input": 3.762, + "output": 18.81, + "cache_read": 0.9405 } }, - "TheDrummer/Rocinante-12B-v1.1": { - "id": "TheDrummer/Rocinante-12B-v1.1", - "name": "Rocinante 12b", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.408, - "output": 0.595 + "input": 0.255552, + "output": 1.27776, + "cache_read": 0.0127776 } }, - "TheDrummer/Magidonia-24B-v4.3": { - "id": "TheDrummer/Magidonia-24B-v4.3", - "name": "The Drummer Magidonia 24B v4.3", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-25", - "last_updated": "2025-12-25", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -180644,60 +190234,75 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.1003, - "output": 0.1207 + "input": 0.1938, + "output": 1.129, + "cache_read": 0.0627 } }, - "TheDrummer/Cydonia-24B-v4": { - "id": "TheDrummer/Cydonia-24B-v4", - "name": "The Drummer Cydonia 24B v4", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.2006, - "output": 0.2414 + "input": 0.7524, + "output": 4.275, + "cache_read": 0.2508 } }, - "TheDrummer/Skyfall-31B-v4.2": { - "id": "TheDrummer/Skyfall-31B-v4.2", - "name": "TheDrummer Skyfall 31B v4.2", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "green-s": { + "id": "green-s", + "name": "Green S", + "description": "GreenPT speech-to-text model for pre-recorded and live transcription", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-03-26", - "last_updated": "2026-03-26", + "temperature": false, + "release_date": "2025-01", + "last_updated": "2025-01", "modalities": { "input": [ - "text", - "pdf" + "audio" ], "output": [ "text" @@ -180705,55 +190310,57 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 16384 + "context": 0, + "output": 8192 }, "cost": { - "input": 0.55, - "output": 0.8 + "input": 0.00437, + "output": 0 } }, - "TheDrummer/Anubis-70B-v1": { - "id": "TheDrummer/Anubis-70B-v1", - "name": "Anubis 70B v1", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, + "green-l": { + "id": "green-l", + "name": "Green L", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "tool_call": true, + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "input": 65536, - "output": 16384 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.31, - "output": 0.31 + "input": 0.285, + "output": 0.912 } }, - "huihui-ai/Qwen2.5-32B-Instruct-abliterated": { - "id": "huihui-ai/Qwen2.5-32B-Instruct-abliterated", - "name": "Qwen 2.5 32B Abliterated", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-06", - "last_updated": "2025-01-06", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -180762,28 +190369,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.7, - "output": 0.7 + "input": 0.285, + "output": 1.083 } }, - "huihui-ai/Llama-3.3-70B-Instruct-abliterated": { - "id": "huihui-ai/Llama-3.3-70B-Instruct-abliterated", - "name": "Llama 3.3 70B Instruct abliterated", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", "family": "llama", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -180792,121 +190399,158 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, + "context": 100000, "output": 16384 }, "cost": { - "input": 0.7, - "output": 0.7 + "input": 1.254, + "output": 1.254 } }, - "huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated": { - "id": "huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated", - "name": "DeepSeek R1 Qwen Abliterated", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "gemma4": { + "id": "gemma4", + "name": "gemma4", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.4, - "output": 1.4 + "input": 0.57, + "output": 1.71 } }, - "huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated": { - "id": "huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated", - "name": "DeepSeek R1 Llama 70B Abliterated", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek", - "attachment": false, + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.7, - "output": 0.7 + "input": 0.342, + "output": 2.052 } }, - "Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B": { - "id": "Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B", - "name": "Llama 3.05 Storybreaker Ministral 70b", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "llama", - "attachment": false, + "pixtral-12b-2409": { + "id": "pixtral-12b-2409", + "name": "Pixtral 12B", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "tool_call": true, + "temperature": true, + "knowledge": "2024-09", + "release_date": "2024-09-01", + "last_updated": "2024-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 0.285, + "output": 0.285 } }, - "Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B": { - "id": "Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B", - "name": "Nemotron Tenyxchat Storybreaker 70b", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", + "glm-5.2-honey": { + "id": "glm-5.2-honey", + "name": "GLM-5.2 Honey", + "description": "glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", - "modalities": { + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { "input": [ "text" ], @@ -180914,22 +190558,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 1.254, + "output": 5.016, + "cache_read": 0.3135 } }, - "stepfun-ai/step-3.5-flash-2603": { - "id": "stepfun-ai/step-3.5-flash-2603", - "name": "Step 3.5 Flash 2603", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, + "holo2-30b-a3b": { + "id": "holo2-30b-a3b", + "name": "Holo2 30B A3B", + "description": "H Company Holo2 vision model for GUI navigation and computer-use agents", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -180941,35 +190585,36 @@ ] } ], - "tool_call": false, + "tool_call": true, "structured_output": false, - "release_date": "2026-04-14", - "last_updated": "2026-04-14", + "temperature": true, + "release_date": "2025-11", + "last_updated": "2025-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 22016, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.399, + "output": 0.969 } }, - "stepfun-ai/step-3.5-flash": { - "id": "stepfun-ai/step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "family": "step", - "attachment": false, + "green-r-raw": { + "id": "green-r-raw", + "name": "Green R Raw", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -180981,13 +190626,43 @@ ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-02", - "last_updated": "2026-02-02", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.399, + "output": 1.083 + } + }, + "green-s-pro": { + "id": "green-s-pro", + "name": "Green S Pro", + "description": "GreenPT advanced speech-to-text model with multilingual transcription support", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-02", + "last_updated": "2025-02", + "modalities": { + "input": [ + "audio" ], "output": [ "text" @@ -180995,28 +190670,63 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 0, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 0.5 + "input": 0.00437, + "output": 0 } }, - "mistral/mistral-medium-3.5": { - "id": "mistral/mistral-medium-3.5", + "devstral-2-123b-instruct-2512": { + "id": "devstral-2-123b-instruct-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 16384 + }, + "cost": { + "input": 0.57, + "output": 2.736 + } + }, + "mistral-medium-3.5-128b": { + "id": "mistral-medium-3.5-128b", "name": "Mistral Medium 3.5", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, "structured_output": true, + "temperature": true, "release_date": "2026-04-29", "last_updated": "2026-04-29", "modalities": { @@ -181028,59 +190738,82 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 262144, + "output": 16384 }, "cost": { - "input": 1.5, - "output": 7.5 + "input": 2.052, + "output": 10.26 } }, - "mistral/mistral-medium-3.5:thinking": { - "id": "mistral/mistral-medium-3.5:thinking", - "name": "Mistral Medium 3.5 Thinking", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": true, + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 262144, + "output": 16384 }, "cost": { - "input": 1.5, - "output": 7.5 + "input": 0.798, + "output": 4.959 } }, - "Tongyi-Zhiwen/QwenLong-L1-32B": { - "id": "Tongyi-Zhiwen/QwenLong-L1-32B", - "name": "QwenLong L1 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-25", - "last_updated": "2025-01-25", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -181089,58 +190822,61 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 40960 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.13999999999999999, - "output": 0.6 + "input": 1.254, + "output": 5.016, + "cache_read": 0.3135 } }, - "google/gemini-flash-1.5": { - "id": "google/gemini-flash-1.5", - "name": "Gemini 1.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": false, + "voxtral-small-24b-2507": { + "id": "voxtral-small-24b-2507", + "name": "Voxtral Small 24B", + "description": "Mistral Voxtral audio-understanding model for speech and transcription tasks", + "family": "mistral", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": false, - "release_date": "2024-05-14", - "last_updated": "2024-05-14", + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-15", + "last_updated": "2025-07-15", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "input": 2000000, - "output": 8192 + "context": 32768, + "output": 16384 }, "cost": { - "input": 0.0748, - "output": 0.306 + "input": 0.228, + "output": 0.513 } }, - "google/gemini-3.1-pro-preview-high": { - "id": "google/gemini-3.1-pro-preview-high", - "name": "Gemini 3.1 Pro (Preview High)", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "green-l-raw": { + "id": "green-l-raw", + "name": "Green L Raw", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-02-21", - "last_updated": "2026-02-21", + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", "modalities": { "input": [ "text", @@ -181150,28 +190886,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 128000, + "output": 32768 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 0.285, + "output": 0.912 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "attachment": true, + "glm-5.2-ponytail-ultra": { + "id": "glm-5.2-ponytail-ultra", + "name": "GLM-5.2 Ponytail Ultra", + "description": "glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "minimal", "low", "medium", @@ -181181,42 +190917,50 @@ ], "tool_call": true, "structured_output": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025 + "input": 1.254, + "output": 5.016, + "cache_read": 0.3135 } }, - "google/gemini-3-flash-preview-thinking": { - "id": "google/gemini-3-flash-preview-thinking", - "name": "Gemini 3 Flash Thinking", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "green-r": { + "id": "green-r", + "name": "Green R", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -181226,27 +190970,91 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.399, + "output": 1.083 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 1.756, + "output": 5.518 + } + }, + "mistral-small-3.2-24b-instruct-2506": { + "id": "mistral-small-3.2-24b-instruct-2506", + "name": "Mistral Small 3.2", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.228, + "output": 0.456 + } + }, + "glm-5.2-caveman": { + "id": "glm-5.2-caveman", + "name": "GLM-5.2 Caveman", + "description": "glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "minimal", "low", "medium", @@ -181256,45 +191064,42 @@ ], "tool_call": true, "structured_output": true, - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 + "input": 1.254, + "output": 5.016, + "cache_read": 0.3135 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "kimi-k2.6-fast": { + "id": "kimi-k2.6-fast", + "name": "Kimi K2.6 Fast", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": false, + "reasoning_options": [], + "tool_call": true, "structured_output": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -181304,21 +191109,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "input": 262144, - "output": 131072 + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.35 + "input": 1.655, + "output": 8.778 } }, - "google/gemini-pro-latest": { - "id": "google/gemini-pro-latest", - "name": "Gemini Pro Latest", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -181333,8 +191139,9 @@ ], "tool_call": true, "structured_output": true, - "release_date": "2026-03-29", - "last_updated": "2026-03-29", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -181344,28 +191151,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 0.228, + "output": 0.798 } }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro (Preview Custom Tools)", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "attachment": true, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high" @@ -181374,39 +191182,40 @@ ], "tool_call": true, "structured_output": true, - "release_date": "2026-02-27", - "last_updated": "2026-02-27", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 1.27754, + "output": 5.11016, + "cache_read": 0.319385 } }, - "google/gemini-flash-lite-latest": { - "id": "google/gemini-flash-lite-latest", - "name": "Gemini Flash Lite Latest", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "attachment": true, + "glm-5.2-caveman-ultra": { + "id": "glm-5.2-caveman-ultra", + "name": "GLM-5.2 Caveman Ultra", + "description": "glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "minimal", "low", "medium", @@ -181416,71 +191225,85 @@ ], "tool_call": true, "structured_output": true, - "release_date": "2026-03-29", - "last_updated": "2026-03-29", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025 + "input": 1.254, + "output": 5.016, + "cache_read": 0.3135 } }, - "google/gemma-4-26b-a4b-it:thinking": { - "id": "google/gemma-4-26b-a4b-it:thinking", - "name": "Gemma 4 26B A4B Thinking", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": true, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "structured_output": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "input": 262144, - "output": 131072 + "output": 262144 }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 0.9006, + "output": 4.389, + "cache_read": 0.1881 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro (Preview)", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "attachment": true, + "glm-5.2-ponytail": { + "id": "glm-5.2-ponytail", + "name": "GLM-5.2 Ponytail", + "description": "glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high" @@ -181489,253 +191312,278 @@ ], "tool_call": true, "structured_output": true, - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 1.254, + "output": 5.016, + "cache_read": 0.3135 } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": true, + "glm-5.2-honey-lite": { + "id": "glm-5.2-honey-lite", + "name": "GLM-5.2 Honey Lite", + "description": "glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] } ], - "tool_call": false, + "tool_call": true, "structured_output": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, + "context": 1000000, "output": 131072 }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 1.254, + "output": 5.016, + "cache_read": 0.3135 } }, - "google/gemini-3.1-pro-preview-low": { - "id": "google/gemini-3.1-pro-preview-low", - "name": "Gemini 3.1 Pro (Preview Low)", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "attachment": true, + "glm-5.2-caveman-lite": { + "id": "glm-5.2-caveman-lite", + "name": "GLM-5.2 Caveman Lite", + "description": "glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2026-02-21", - "last_updated": "2026-02-21", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 1.254, + "output": 5.016, + "cache_read": 0.3135 } - }, - "google/gemma-4-31b-it:thinking": { - "id": "google/gemma-4-31b-it:thinking", - "name": "Gemma 4 31B Thinking", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": true, + } + } + }, + "mixlayer": { + "id": "mixlayer", + "env": [ + "MIXLAYER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://models.mixlayer.ai/v1", + "name": "Mixlayer", + "doc": "https://docs.mixlayer.com", + "models": { + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "input": 262144, - "output": 131072 + "output": 262144 }, "cost": { - "input": 0.1, - "output": 0.35 + "input": 0.3, + "output": 2.4 } }, - "google/gemini-3.5-flash-thinking": { - "id": "google/gemini-3.5-flash-thinking", - "name": "Gemini 3.5 Flash Thinking", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "attachment": true, + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B A10B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 + "input": 0.4, + "output": 3.2 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash (Preview)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "qwen/qwen3.5-9b": { + "id": "qwen/qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.1, + "output": 0.4 } }, - "google/gemini-flash-latest": { - "id": "google/gemini-flash-latest", - "name": "Gemini Flash Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "attachment": true, + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5 35B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-03-29", - "last_updated": "2026-03-29", + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 + "input": 0.25, + "output": 1.3 } }, - "liquid/lfm-2-24b-a2b": { - "id": "liquid/lfm-2-24b-a2b", - "name": "LFM2 24B A2B", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-20", - "last_updated": "2025-12-20", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -181744,27 +191592,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.03, - "output": 0.12 + "input": 0.6, + "output": 3.6 } - }, - "x-ai/grok-4.20": { - "id": "x-ai/grok-4.20", - "name": "Grok 4.20", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "attachment": true, + } + } + }, + "hyper": { + "id": "hyper", + "env": [ + "HYPER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://hyper.charm.land/v1", + "name": "Charm Hyper", + "doc": "https://hyper.charm.land", + "models": { + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high" @@ -181772,13 +191634,12 @@ } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "temperature": true, + "release_date": "2026-05-28", + "last_updated": "2026-07-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -181786,25 +191647,28 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "input": 2000000, - "output": 131072 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 2, - "output": 6 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5 } }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "Grok 4.3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "qwen3.8-27b": { + "id": "qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high" @@ -181813,8 +191677,9 @@ ], "tool_call": true, "structured_output": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", @@ -181824,28 +191689,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "input": 1000000, - "output": 1000000 + "output": 128000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.5, + "output": 3, + "cache_read": 0.1 } }, - "x-ai/grok-4.20-multi-agent": { - "id": "x-ai/grok-4.20-multi-agent", - "name": "Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "attachment": true, + "qwen3.8-2.4t-a95b": { + "id": "qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -181854,32 +191720,33 @@ ], "tool_call": true, "structured_output": true, - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "input": 2000000, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { "input": 2, - "output": 6 + "output": 6, + "cache_read": 0.25 } }, - "x-ai/grok-latest": { - "id": "x-ai/grok-latest", - "name": "Grok Latest", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -181887,15 +191754,16 @@ "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-31", "modalities": { "input": [ "text", @@ -181905,133 +191773,86 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.16332, + "output": 0.5444, + "cache_read": 0.031575 } }, - "x-ai/grok-build-0.1": { - "id": "x-ai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Grok coding model for agentic engineering, edits, and codebase workflows", - "attachment": true, + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-05-20", - "last_updated": "2026-05-20", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-08-02", + "last_updated": "2026-08-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 + "input": 0.44, + "output": 1.32, + "cache_read": 0.044 } }, - "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0": { - "id": "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0", - "name": "EVA Llama 3.33 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16384, - "input": 16384, - "output": 16384 - }, - "cost": { - "input": 2.006, - "output": 2.006 - } - }, - "EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2": { - "id": "EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2", - "name": "EVA-Qwen2.5-72B-v0.2", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max Preview", + "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16384, - "input": 16384, - "output": 8192 - }, - "cost": { - "input": 0.7989999999999999, - "output": 0.7989999999999999 - } - }, - "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1": { - "id": "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1", - "name": "EVA-LLaMA-3.33-70B-v0.1", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-07-19", + "last_updated": "2026-07-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -182039,90 +191860,128 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 2.006, - "output": 2.006 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2": { - "id": "EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2", - "name": "EVA-Qwen2.5-32B-v0.2", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-27", + "last_updated": "2026-07-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1048576, + "output": 16000 }, "cost": { - "input": 0.7989999999999999, - "output": 0.7989999999999999 + "input": 3.2664, + "output": 16.332, + "cache_read": 0.32664 } }, - "microsoft/wizardlm-2-8x22b": { - "id": "microsoft/wizardlm-2-8x22b", - "name": "WizardLM-2 8x22B", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "input": 65536, - "output": 8192 + "context": 1048576, + "output": 32768 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "Doctor-Shotgun/MS3.2-24B-Magnum-Diamond": { - "id": "Doctor-Shotgun/MS3.2-24B-Magnum-Diamond", - "name": "MS3.2 24B Magnum Diamond", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "qwen3.7-flash": { + "id": "qwen3.7-flash", + "name": "Qwen3.7 Flash", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-27", + "last_updated": "2026-07-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -182130,28 +191989,29 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 32768 + "context": 1000000, + "input": 991000, + "output": 64000 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 0.2, + "output": 0.8, + "cache_read": 0.04 } }, - "poolside/laguna-xs.2": { - "id": "poolside/laguna-xs.2", - "name": "Laguna XS.2", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-06-13", + "knowledge": "2024-08", + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ "text" @@ -182163,26 +192023,27 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 26214 }, "cost": { - "input": 0.2, - "output": 0.4 + "input": 0.6, + "output": 2.5, + "cache_read": 0.3 } }, - "poolside/laguna-m.1": { - "id": "poolside/laguna-m.1", - "name": "Laguna M.1", - "description": "Poolside's flagship agentic coding model for long-horizon work", - "family": "laguna", + "gemma-4-26b-a4b-it": { + "id": "gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-06-13", + "release_date": "2026-04-30", + "last_updated": "2026-07-22", "modalities": { "input": [ "text" @@ -182193,58 +192054,68 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 256000, + "output": 25600 }, "cost": { - "input": 0.2, - "output": 0.4 + "input": 0.098, + "output": 0.334, + "cache_read": 0.049 } }, - "z-ai/glm-4.5v:thinking": { - "id": "z-ai/glm-4.5v:thinking", - "name": "GLM 4.5V Thinking", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glmv", - "attachment": true, + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-11-22", - "last_updated": "2025-11-22", + "tool_call": true, + "temperature": true, + "release_date": "2026-06-05", + "last_updated": "2026-07-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 64000, - "input": 64000, - "output": 96000 + "context": 262100, + "output": 6553 }, "cost": { - "input": 0.6, - "output": 1.7999999999999998 + "input": 0.484, + "output": 1.852, + "cache_read": 0.242 } }, - "z-ai/glm-4.6:thinking": { - "id": "z-ai/glm-4.6:thinking", - "name": "GLM 4.6 Thinking", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -182253,33 +192124,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 65535 + "context": 1000000, + "output": 262144 }, "cost": { - "input": 0.4, - "output": 1.5 + "input": 1.437216, + "output": 4.311648, + "cache_read": 0.047907 } }, - "z-ai/glm-4.5v": { - "id": "z-ai/glm-4.5v", - "name": "GLM 4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glmv", + "inkling": { + "id": "inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-11-22", - "last_updated": "2025-11-22", + "tool_call": true, + "temperature": true, + "release_date": "2026-08-15", + "last_updated": "2026-09-05", "modalities": { "input": [ "text", @@ -182289,63 +192168,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 64000, - "input": 64000, - "output": 96000 + "context": 1048576, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 1.7999999999999998 + "input": 1.0888, + "output": 4.40964, + "cache_read": 0.185096 } }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "GLM 4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "release_date": "2025-09-30", - "last_updated": "2025-09-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "input": 200000, - "output": 65535 - }, - "cost": { - "input": 0.4, - "output": 1.5 - } - }, - "z-ai/glm-5v-turbo:thinking": { - "id": "z-ai/glm-5v-turbo:thinking", - "name": "GLM 5V Turbo Thinking", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": true, + "release_date": "2026-07-30", + "last_updated": "2026-07-31", "modalities": { "input": [ "text", @@ -182355,28 +192209,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202800, - "input": 202800, - "output": 131100 + "context": 512000, + "output": 512000 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 0.32664, + "output": 1.30656, + "cache_read": 0.064239 } }, - "z-ai/glm-5v-turbo": { - "id": "z-ai/glm-5v-turbo", - "name": "GLM 5V Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", @@ -182388,26 +192254,36 @@ }, "open_weights": false, "limit": { - "context": 202800, - "input": 202800, - "output": 131100 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 0.15, + "output": 0.47, + "cache_read": 0.016 } }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "GLM 5 Turbo", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2026-03-15", - "last_updated": "2026-03-15", + "temperature": true, + "release_date": "2026-06-30", + "last_updated": "2026-07-22", "modalities": { "input": [ "text" @@ -182416,30 +192292,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202800, - "input": 202800, - "output": 131072 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 1.52432, + "output": 4.79072, + "cache_read": 0.152432 } }, - "openai/o3-mini-low": { - "id": "openai/o3-mini-low", - "name": "OpenAI o3-mini (Low)", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-06", + "last_updated": "2026-07-22", "modalities": { "input": [ "text" @@ -182448,21 +192333,21 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 9.996, - "output": 19.992 + "input": 2.4, + "output": 4.8, + "cache_read": 0.2 } }, - "openai/gpt-oss-safeguard-20b": { - "id": "openai/gpt-oss-safeguard-20b", - "name": "GPT OSS Safeguard 20B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", "family": "gpt-oss", "attachment": false, "reasoning": true, @@ -182476,10 +192361,11 @@ ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-13", + "last_updated": "2026-07-22", "modalities": { "input": [ "text" @@ -182488,28 +192374,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 131072, + "output": 13107 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 0.178, + "output": 0.68, + "cache_read": 0.089 } }, - "openai/o3": { - "id": "openai/o3", - "name": "OpenAI o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-28", + "last_updated": "2026-08-31", "modalities": { "input": [ "text" @@ -182518,59 +192415,64 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 2, - "output": 8 + "input": 1.52432, + "output": 4.79072, + "cache_read": 0.283088 } }, - "openai/o4-mini-high": { - "id": "openai/o4-mini-high", - "name": "OpenAI o4-mini high", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-07-03", + "last_updated": "2026-07-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 256000, + "output": 16000 }, "cost": { - "input": 1.1, - "output": 4.4 + "input": 1.03436, + "output": 4.3552, + "cache_read": 0.206872 } }, - "openai/o3-pro-2025-06-10": { - "id": "openai/o3-pro-2025-06-10", - "name": "OpenAI o3-pro (2025-06-10)", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": false, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", "high" @@ -182578,12 +192480,14 @@ } ], "tool_call": true, - "structured_output": true, - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-15", + "last_updated": "2026-07-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -182591,29 +192495,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 9.996, - "output": 19.992 + "input": 1.2, + "output": 4.8, + "cache_read": 0.24 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT 5.2 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", - "attachment": true, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", - "medium", "high", "xhigh" ] @@ -182621,113 +192522,126 @@ ], "tool_call": true, "structured_output": true, - "release_date": "2026-01-01", - "last_updated": "2026-01-01", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-06", + "last_updated": "2026-07-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 21, - "output": 168 + "input": 0.2, + "output": 0.4, + "cache_read": 0.04 } - }, - "openai/gpt-4o-mini-search-preview": { - "id": "openai/gpt-4o-mini-search-preview", - "name": "GPT-4o mini Search Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": false, + } + } + }, + "jalapeno": { + "id": "jalapeno", + "env": [ + "JALAPENO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.jalapeno-cloud.ai/v1", + "name": "Jalapeno Cloud", + "doc": "https://www.jalapeno-cloud.ai/docs/", + "models": { + "Qwen3-VL-235B-A22B-Instruct": { + "id": "Qwen3-VL-235B-A22B-Instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 129024, + "output": 32768 }, "cost": { - "input": 0.088, - "output": 0.35 + "input": 0.3, + "output": 1.5 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT 5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 524288, + "output": 512000 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.3, + "output": 1.2 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "Qwen3-Next-80B-A3B-Instruct": { + "id": "Qwen3-Next-80B-A3B-Instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2022-11-30", - "last_updated": "2024-01-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ "text" @@ -182736,108 +192650,113 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16385, - "input": 16385, - "output": 4096 + "context": 129024, + "output": 32768 }, "cost": { - "input": 0.5, + "input": 0.15, "output": 1.5 } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT 5 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", - "attachment": true, + "GLM-5.1": { + "id": "GLM-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 15, - "output": 120 + "input": 1.38, + "output": 4.4 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "Hy3": { + "id": "Hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 202752, + "input": 192000, + "output": 128000 }, "cost": { - "input": 2.499, - "output": 9.996 + "input": 0.14, + "output": 0.58 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "OpenAI o4-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "GLM-5.2": { + "id": "GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -182846,197 +192765,189 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 1.1, + "input": 1.4, "output": 4.4 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT 5.4 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "Qwen3-VL-235B-A22B-Thinking": { + "id": "Qwen3-VL-235B-A22B-Thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.98, + "output": 3.95 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT 5.1 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "Qwen3.5-27B": { + "id": "Qwen3.5-27B", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.3, + "output": 2.4 } }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT 5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "Kimi-K3": { + "id": "Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 20 + "input": 3, + "output": 15 } }, - "openai/gpt-4o-2024-08-06": { - "id": "openai/gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-08-06", - "last_updated": "2024-08-06", + "Qwen3-Next-80B-A3B-Thinking": { + "id": "Qwen3-Next-80B-A3B-Thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 131072, + "output": 32768 }, "cost": { - "input": 2.499, - "output": 9.996 + "input": 0.15, + "output": 1.5 } }, - "openai/o1-preview": { - "id": "openai/o1-preview", - "name": "OpenAI o1-preview", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "DeepSeek-V4-Flash": { + "id": "DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, - "release_date": "2024-09-12", - "last_updated": "2024-09-12", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -183045,38 +192956,37 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 32768 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 14.993999999999998, - "output": 59.993 + "input": 0.14, + "output": 0.28 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "OpenAI o3-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "DeepSeek-V4-Pro": { + "id": "DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -183085,208 +192995,244 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 1.1, - "output": 4.4 + "input": 1.6, + "output": 3.38 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT 5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "Qwen3.5-35B-A3B": { + "id": "Qwen3.5-35B-A3B", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-01-01", - "last_updated": "2026-01-01", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.25, + "output": 2 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT 5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "Qwen3.5-122B-A10B": { + "id": "Qwen3.5-122B-A10B", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.4, + "output": 3.2 } }, - "openai/gpt-latest": { - "id": "openai/gpt-latest", - "name": "GPT Latest", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "Qwen3.5-397B-A17B": { + "id": "Qwen3.5-397B-A17B", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-03-29", - "last_updated": "2026-03-29", + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.6, + "output": 3.6 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT 5.1 Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex-mini", + "Kimi-K2.7-Code": { + "id": "Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 271360, + "output": 262144 }, "cost": { - "input": 0.25, - "output": 2 + "input": 0.95, + "output": 4 } }, - "openai/o4-mini-deep-research": { - "id": "openai/o4-mini-deep-research", - "name": "OpenAI o4-mini Deep Research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "family": "o-mini", - "attachment": false, + "Kimi-K2.5": { + "id": "Kimi-K2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 180224 + }, + "cost": { + "input": 0.6, + "output": 3 + } + } + } + }, + "dinference": { + "id": "dinference", + "env": [ + "DINFERENCE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.dinference.com/v1", + "name": "DInference", + "doc": "https://dinference.com", + "models": { + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -183295,70 +193241,62 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 200000, - "input": 200000, - "output": 100000 + "output": 128000 }, "cost": { - "input": 9.996, - "output": 19.992 + "input": 0.75, + "output": 2.4 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT 4.1 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "input": 1047576, - "output": 32768 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 0.22, + "output": 0.88 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -183367,69 +193305,66 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.25 + "input": 0.45, + "output": 1.65 } }, - "openai/gpt-4o-2024-11-20": { - "id": "openai/gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-11-20", - "last_updated": "2024-11-20", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 1.25, + "output": 3.89 } }, - "openai/o1": { - "id": "openai/o1", - "name": "OpenAI o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-17", - "last_updated": "2024-12-17", + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -183438,33 +193373,72 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 200000, - "input": 200000, - "output": 100000 + "output": 128000 }, "cost": { - "input": 14.993999999999998, - "output": 59.993 + "input": 1.25, + "output": 3.89 } }, - "openai/o1-pro": { - "id": "openai/o1-pro", - "name": "OpenAI o1 Pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": true, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-08", + "last_updated": "2025-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.0675, + "output": 0.27 + } + } + } + }, + "neosmith": { + "id": "neosmith", + "env": [ + "NEOSMITH_API_KEY" + ], + "npm": "@ai-sdk/openai", + "api": "https://router.neosmith.ai/v1", + "name": "NeoSmith", + "doc": "https://neosmith.ai/docs", + "models": { + "neosmith.neolite": { + "id": "neosmith.neolite", + "name": "NeoSmith NeoLite", + "description": "Sealed single-model budget tier. 512K context, text and images, tool use, and no escalation of any kind.", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": false, - "release_date": "2025-01-25", - "last_updated": "2025-01-25", + "temperature": true, + "release_date": "2026-06-20", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -183472,30 +193446,32 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 512000, + "output": 64000 }, "cost": { - "input": 150, - "output": 600 + "input": 0.6, + "output": 2.4, + "cache_read": 0.08, + "cache_write": 0 } }, - "openai/gpt-chat-latest": { - "id": "openai/gpt-chat-latest", - "name": "GPT Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "neosmith.intelligent-maestro": { + "id": "neosmith.intelligent-maestro", + "name": "NeoSmith Maestro", + "description": "Highest-accuracy coding tier. Hard, self-contained problems run NeoSmith's premium multi-model solver; everything else gets the strongest intelligence tier.", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "structured_output": false, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-08", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -183503,43 +193479,32 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 400000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 2.4, + "output": 12, + "cache_read": 0.35, + "cache_write": 0 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT 5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "neosmith.intelligent-pro": { + "id": "neosmith.intelligent-pro", + "name": "NeoSmith Pro", + "description": "Default production tier. Intelligent NeoSmith routing with a Claude Opus ceiling on escalation.", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": false, + "temperature": true, + "release_date": "2026-04-26", + "last_updated": "2026-07-18", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -183547,43 +193512,32 @@ }, "open_weights": false, "limit": { - "context": 922000, - "input": 922000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 1.81, + "output": 8.39, + "cache_read": 0.3, + "cache_write": 0 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT 5.4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "neosmith.intelligent-basic": { + "id": "neosmith.intelligent-basic", + "name": "NeoSmith Basic", + "description": "Cost-capped tier. Intelligent routing with a Claude Sonnet ceiling — Opus is never invoked.", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "structured_output": false, + "temperature": true, + "release_date": "2026-04-26", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -183591,67 +193545,100 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 400000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 1.17, + "output": 4.37, + "cache_read": 0.22, + "cache_write": 0 } - }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT 4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + } + } + }, + "fireworks-ai": { + "id": "fireworks-ai", + "env": [ + "FIREWORKS_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.fireworks.ai/inference/v1/", + "name": "Fireworks AI", + "doc": "https://fireworks.ai/docs/", + "models": { + "accounts/fireworks/models/kimi-k2p6": { + "id": "accounts/fireworks/models/kimi-k2p6", + "name": "Kimi K2.6", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-09-10", - "last_updated": "2025-09-10", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "input": 1047576, - "output": 32768 + "context": 262000, + "output": 262000 + }, + "status": "deprecated", + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 1.5, + "output": 6, + "cache_read": 0.22 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 2, - "output": 8 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "openai/o3-deep-research": { - "id": "openai/o3-deep-research", - "name": "OpenAI o3 Deep Research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "family": "o", + "accounts/fireworks/models/nemotron-3-ultra-nvfp4": { + "id": "accounts/fireworks/models/nemotron-3-ultra-nvfp4", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "tool_call": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -183660,28 +193647,46 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 262144, + "output": 128000 }, "cost": { - "input": 9.996, - "output": 19.992 + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 } }, - "openai/gpt-4-turbo-preview": { - "id": "openai/gpt-4-turbo-preview", - "name": "GPT-4 Turbo Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "accounts/fireworks/models/deepseek-v4-flash-0731": { + "id": "accounts/fireworks/models/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2023-11-06", - "last_updated": "2024-01-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -183690,39 +193695,58 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 1000000, + "output": 384000 + }, + "status": "deprecated", + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 0.275, + "output": 0.825, + "cache_read": 0.00875 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 9.996, - "output": 30.004999999999995 + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT 5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "accounts/fireworks/models/muse-glimmer-30b": { + "id": "accounts/fireworks/models/muse-glimmer-30b", + "name": "Muse Glimmer 30B", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", "modalities": { "input": [ "text", @@ -183732,28 +193756,51 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 131072, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.25, - "output": 2 + "input": 0.35, + "output": 1.5, + "cache_read": 0.04 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT 4.1 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "accounts/fireworks/models/kimi-k3": { + "id": "accounts/fireworks/models/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-27", + "last_updated": "2026-07-27", "modalities": { "input": [ "text", @@ -183763,70 +193810,105 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "input": 1047576, - "output": 32768 + "context": 1048576, + "output": 131072 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 3.75, + "output": 18.75, + "cache_read": 0.375 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.4, - "output": 1.6 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2023-11-06", - "last_updated": "2024-01-01", + "accounts/fireworks/models/qwen3p8-2p4t-a95b": { + "id": "accounts/fireworks/models/qwen3p8-2p4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 262144, + "output": 131072 }, "cost": { - "input": 10, - "output": 30 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT 5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "accounts/fireworks/models/deepseek-v4-flash-vision-exp": { + "id": "accounts/fireworks/models/deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", "low", - "medium", - "high" + "high", + "max" ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", @@ -183836,76 +193918,115 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 1000000, + "output": 384000 }, + "status": "deprecated", "cost": { - "input": 0.05, - "output": 0.4 + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT 5.4 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "accounts/fireworks/models/deepseek-v4p1-flash": { + "id": "accounts/fireworks/models/deepseek-v4p1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 922000, - "input": 922000, - "output": 128000 + "context": 1000000, + "output": 384000 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 0.275, + "output": 0.825, + "cache_read": 0.00875 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 30, - "output": 180, - "cache_read": 3 + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 } }, - "openai/o3-mini-high": { - "id": "openai/o3-mini-high", - "name": "OpenAI o3-mini (High)", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, + "accounts/fireworks/models/qwen3p7-plus": { + "id": "accounts/fireworks/models/qwen3p7-plus", + "name": "Qwen 3.7 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1 + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "temperature": true, + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -183913,57 +194034,99 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.64, - "output": 2.588 + "input": 0.4, + "output": 1.6, + "cache_read": 0.08 } }, - "openai/gpt-4o-search-preview": { - "id": "openai/gpt-4o-search-preview", - "name": "GPT-4o Search Preview", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "accounts/fireworks/models/glm-5p3": { + "id": "accounts/fireworks/models/glm-5p3", + "name": "GLM 5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-09-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 1048573, + "output": 262144 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 1.75, + "output": 5.5, + "cache_read": 0.325 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 1.47, - "output": 5.88 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai/gpt-5.1-2025-11-13": { - "id": "openai/gpt-5.1-2025-11-13", - "name": "GPT-5.1 (2025-11-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "accounts/fireworks/models/minimax-m2p7": { + "id": "accounts/fireworks/models/minimax-m2p7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -183972,69 +194135,86 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 32768 + "context": 196608, + "output": 131072 + }, + "status": "deprecated", + "provider": { + "body": { + "service_tier": "priority" + } }, "cost": { - "input": 1.25, - "output": 10 + "input": 1.2, + "output": 1.2, + "cache_read": 0.6 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "accounts/fireworks/models/nemotron-lightning-3p5-30b-a3b": { + "id": "accounts/fireworks/models/nemotron-lightning-3p5-30b-a3b", + "name": "Nemotron 3.5 Lightning 30B A3B", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.1496, - "output": 0.595 + "input": 0.05, + "output": 0.2, + "cache_read": 0.01 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "accounts/fireworks/models/deepseek-v4-pro-0813": { + "id": "accounts/fireworks/models/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "max" ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -184043,102 +194223,140 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 1000000, + "output": 384000 + }, + "status": "deprecated", + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 1.65, + "output": 4.95, + "cache_read": 0.055 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.04, - "output": 0.15 + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "GPT-5 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "accounts/fireworks/models/inkling": { + "id": "accounts/fireworks/models/inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 9.996, - "output": 19.992 + "input": 1, + "output": 4.05, + "cache_read": 0.17 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT 5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "accounts/fireworks/models/glm-5p2": { + "id": "accounts/fireworks/models/glm-5p2", + "name": "GLM 5.2", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 1048575, + "output": 131072 + }, + "status": "deprecated", + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 1.75, + "output": 5.5, + "cache_read": 0.175 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 1.75, - "output": 14 + "input": 1.4, + "output": 4.4, + "cache_read": 0.14 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT 5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "accounts/fireworks/models/minimax-m3": { + "id": "accounts/fireworks/models/minimax-m3", + "name": "MiniMax-M3", + "description": "Fireworks text-only MiniMax coding model for long-context reasoning and agent tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -184146,114 +194364,192 @@ } ], "tool_call": true, - "structured_output": true, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 512000, + "output": 512000 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 0.45, + "output": 1.8, + "cache_read": 0.09 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT 5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "accounts/fireworks/models/kimi-k2p7-code": { + "id": "accounts/fireworks/models/kimi-k2p7-code", + "name": "Kimi K2.7 Code", + "description": "Kimi coding model for software agents, refactors, and repository reasoning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-12", + "last_updated": "2026-06-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262000, + "output": 262000 + }, + "status": "deprecated", + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 1.425, + "output": 6, + "cache_read": 0.285 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "Gryphe/MythoMax-L2-13b": { - "id": "Gryphe/MythoMax-L2-13b", - "name": "MythoMax 13B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "accounts/fireworks/models/glm-5p3-flash": { + "id": "accounts/fireworks/models/glm-5p3-flash", + "name": "GLM 5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-09-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4000, - "input": 4000, - "output": 4096 + "context": 1048573, + "output": 131072 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 0.1875, + "output": 0.625, + "cache_read": 0.0375 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.1003, - "output": 0.1003 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "Unbabel/M-Prometheus-14B": { - "id": "Unbabel/M-Prometheus-14B", - "name": "M-Prometheus 14B", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "accounts/fireworks/models/deepseek-v4-pro": { + "id": "accounts/fireworks/models/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -184262,28 +194558,55 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 1000000, + "output": 384000 + }, + "status": "deprecated", + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 1.2, + "output": 1.2, + "cache_read": 0.6 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 1.2, + "output": 1.2, + "cache_read": 0.6 } }, - "LLM360/K2-Think": { - "id": "LLM360/K2-Think", - "name": "K2-Think", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "accounts/fireworks/models/gpt-oss-120b": { + "id": "accounts/fireworks/models/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2026-06-16", "modalities": { "input": [ "text" @@ -184292,30 +194615,53 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, + "context": 131072, "output": 32768 }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 0.18, + "output": 0.72, + "cache_read": 0.018 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, "cost": { - "input": 0.17, - "output": 0.68 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "NousResearch/hermes-4-405b": { - "id": "NousResearch/hermes-4-405b", - "name": "Hermes 4 Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "accounts/fireworks/models/qwen3p8-max": { + "id": "accounts/fireworks/models/qwen3p8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -184323,25 +194669,55 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 262144, + "output": 131072 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 3, + "output": 9, + "cache_read": 0.375 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "NousResearch/hermes-3-llama-3.1-70b": { - "id": "NousResearch/hermes-3-llama-3.1-70b", - "name": "Hermes 3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "accounts/fireworks/routers/minimax-latest": { + "id": "accounts/fireworks/routers/minimax-latest", + "name": "MiniMax Latest", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-01-07", - "last_updated": "2026-01-07", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-12", + "last_updated": "2026-09-15", "modalities": { "input": [ "text" @@ -184350,56 +194726,119 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "input": 65536, - "output": 8192 + "context": 512000, + "output": 512000 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 0.45, + "output": 1.8, + "cache_read": 0.09 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.408, - "output": 0.408 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "NousResearch/Hermes-4-70B:thinking": { - "id": "NousResearch/Hermes-4-70B:thinking", - "name": "Hermes 4 (Thinking)", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-17", - "last_updated": "2025-09-17", + "accounts/fireworks/routers/glm-flash-latest": { + "id": "accounts/fireworks/routers/glm-flash-latest", + "name": "GLM Flash Latest (GLM 5.3 Flash)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-09-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 1048573, + "output": 131072 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 0.1875, + "output": 0.625, + "cache_read": 0.0375 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.2006, - "output": 0.3995 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "NousResearch/hermes-4-70b": { - "id": "NousResearch/hermes-4-70b", - "name": "Hermes 4 Medium", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "accounts/fireworks/routers/glm-fast-latest": { + "id": "accounts/fireworks/routers/glm-fast-latest", + "name": "GLM 5.3 Fast (Latest)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-03", - "last_updated": "2025-07-03", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-28", + "last_updated": "2026-09-15", "modalities": { "input": [ "text" @@ -184408,56 +194847,95 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 1048572, + "output": 262144 }, "cost": { - "input": 0.2006, - "output": 0.3995 + "input": 2.1, + "output": 6.6, + "cache_read": 0.39 } }, - "NousResearch/DeepHermes-3-Mistral-24B-Preview": { - "id": "NousResearch/DeepHermes-3-Mistral-24B-Preview", - "name": "DeepHermes-3 Mistral 24B (Preview)", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-10", - "last_updated": "2025-05-10", + "accounts/fireworks/routers/kimi-k3-fast": { + "id": "accounts/fireworks/routers/kimi-k3-fast", + "name": "Kimi K3 Fast", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-27", + "last_updated": "2026-07-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 32768 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 0.3 + "input": 4.5, + "output": 22.5, + "cache_read": 0.45 } }, - "NousResearch/hermes-4-405b:thinking": { - "id": "NousResearch/hermes-4-405b:thinking", - "name": "Hermes 4 Large (Thinking)", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "accounts/fireworks/routers/glm-5p3-fast": { + "id": "accounts/fireworks/routers/glm-5p3-fast", + "name": "GLM 5.3 Fast", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "temperature": true, + "release_date": "2026-08-28", + "last_updated": "2026-09-07", "modalities": { "input": [ "text" @@ -184466,63 +194944,90 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 1048572, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 2.1, + "output": 6.6, + "cache_read": 0.39 } }, - "unsloth/gemma-3-12b-it": { - "id": "unsloth/gemma-3-12b-it", - "name": "Gemma 3 12B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "unsloth", + "accounts/fireworks/routers/kimi-fast-latest": { + "id": "accounts/fireworks/routers/kimi-fast-latest", + "name": "Kimi Fast Latest", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-10", - "last_updated": "2025-03-10", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-27", + "last_updated": "2026-09-15", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, + "context": 1048576, "output": 131072 }, "cost": { - "input": 0.272, - "output": 0.272 + "input": 4.5, + "output": 22.5, + "cache_read": 0.45 } }, - "unsloth/gemma-3-4b-it": { - "id": "unsloth/gemma-3-4b-it", - "name": "Gemma 3 4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "unsloth", + "accounts/fireworks/routers/qwen-max-latest": { + "id": "accounts/fireworks/routers/qwen-max-latest", + "name": "Qwen Max Latest (Qwen3.8 Max)", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-10", - "last_updated": "2025-03-10", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-09-16", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" @@ -184530,87 +195035,191 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 262144, + "output": 131072 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 3, + "output": 9, + "cache_read": 0.375 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.2006, - "output": 0.2006 + "input": 2, + "output": 6, + "cache_read": 0.25 } }, - "unsloth/gemma-3-27b-it": { - "id": "unsloth/gemma-3-27b-it", - "name": "Gemma 3 27B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "unsloth", + "accounts/fireworks/routers/kimi-latest": { + "id": "accounts/fireworks/routers/kimi-latest", + "name": "Kimi Latest", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-10", - "last_updated": "2025-03-10", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-27", + "last_updated": "2026-09-15", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 96000 + "context": 1048576, + "output": 131072 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 3.75, + "output": 18.75, + "cache_read": 0.375 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.2992, - "output": 0.2992 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "NeverSleep/Lumimaid-v0.2-70B": { - "id": "NeverSleep/Lumimaid-v0.2-70B", - "name": "Lumimaid v0.2", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "accounts/fireworks/routers/deepseek-flash-latest": { + "id": "accounts/fireworks/routers/deepseek-flash-latest", + "name": "DeepSeek Flash Latest", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1000000, + "output": 384000 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 0.275, + "output": 0.825, + "cache_read": 0.00875 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 1, - "output": 1.5 + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 } }, - "mistralai/mixtral-8x7b-instruct-v0.1": { - "id": "mistralai/mixtral-8x7b-instruct-v0.1", - "name": "Mixtral 8x7B", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mixtral", + "accounts/fireworks/routers/glm-5p2-fast": { + "id": "accounts/fireworks/routers/glm-5p2-fast", + "name": "GLM 5.2 Fast", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-26", + "last_updated": "2026-06-26", "modalities": { "input": [ "text" @@ -184619,59 +195228,104 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 1048575, + "output": 131072 }, "cost": { - "input": 0.27, - "output": 0.27 + "input": 2.1, + "output": 6.6, + "cache_read": 0.21 } }, - "mistralai/mistral-small-4-119b-2603:thinking": { - "id": "mistralai/mistral-small-4-119b-2603:thinking", - "name": "Mistral Small 4 119B Thinking", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "attachment": true, + "accounts/fireworks/routers/deepseek-pro-latest": { + "id": "accounts/fireworks/routers/deepseek-pro-latest", + "name": "DeepSeek Pro Latest", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1000000, + "output": 384000 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 1.65, + "output": 4.95, + "cache_read": 0.055 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.4, - "output": 1.4 + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 } }, - "mistralai/mixtral-8x22b-instruct-v0.1": { - "id": "mistralai/mixtral-8x22b-instruct-v0.1", - "name": "Mixtral 8x22B", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mixtral", + "accounts/fireworks/routers/glm-latest": { + "id": "accounts/fireworks/routers/glm-latest", + "name": "GLM Latest", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-09-15", "modalities": { "input": [ "text" @@ -184680,31 +195334,73 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "input": 65536, - "output": 32768 + "context": 1048573, + "output": 262144 + }, + "experimental": { + "modes": { + "priority": { + "cost": { + "input": 1.75, + "output": 5.5, + "cache_read": 0.325 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.8999999999999999, - "output": 0.8999999999999999 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } - }, - "mistralai/Devstral-Small-2505": { - "id": "mistralai/Devstral-Small-2505", - "name": "Mistral Devstral Small 2505", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-02", - "last_updated": "2025-08-02", + } + } + }, + "stepfun-ai": { + "id": "stepfun-ai", + "env": [ + "STEPFUN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.stepfun.ai/v1", + "name": "StepFun (Global)", + "doc": "https://platform.stepfun.ai/docs/en/overview/concept", + "models": { + "step-5-preview": { + "id": "step-5-preview", + "name": "Step 5 Preview", + "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "release_date": "2026-09-16", + "last_updated": "2026-09-20", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -184712,56 +195408,85 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 1000000, + "input": 1000000, + "output": 1000000 }, "cost": { - "input": 0.060000000000000005, - "output": 0.060000000000000005 + "input": 1, + "output": 2.7, + "cache_read": 0.05 } }, - "mistralai/ministral-8b-2512": { - "id": "mistralai/ministral-8b-2512", - "name": "Ministral 8B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-04", - "last_updated": "2025-12-04", + "step-3.7-flash": { + "id": "step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-06-29", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 0.185, + "output": 1.11, + "cache_read": 0.037 } }, - "mistralai/mistral-saba": { - "id": "mistralai/mistral-saba", - "name": "Mistral Saba", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", + "step-3.5-flash": { + "id": "step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-17", - "last_updated": "2025-02-17", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-06-15", "modalities": { "input": [ "text" @@ -184770,31 +195495,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "input": 32000, - "output": 32768 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.1989, - "output": 0.595 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "mistralai/mistral-large": { - "id": "mistralai/mistral-large", - "name": "Mistral Large 2411", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", + "stepaudio-2.5-asr": { + "id": "stepaudio-2.5-asr", + "name": "StepAudio 2.5 ASR", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "step", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-02-26", - "last_updated": "2024-02-26", + "temperature": false, + "release_date": "2026-04-24", + "last_updated": "2026-07-02", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" @@ -184802,56 +195528,46 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 256000 - }, - "cost": { - "input": 2.006, - "output": 6.001 + "context": 0, + "output": 0 } }, - "mistralai/mistral-medium-3.1": { - "id": "mistralai/mistral-medium-3.1", - "name": "Mistral Medium 3.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "stepaudio-2.5-tts": { + "id": "stepaudio-2.5-tts", + "name": "StepAudio 2.5 TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "step", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "temperature": false, + "release_date": "2026-04-16", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 2 + "context": 0, + "output": 0 } }, - "mistralai/ministral-3b-2512": { - "id": "mistralai/ministral-3b-2512", - "name": "Ministral 3B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "step-2-16k": { + "id": "step-2-16k", + "name": "Step 2 (16K)", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-04", - "last_updated": "2025-12-04", + "tool_call": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-01-01", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -184862,92 +195578,95 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 5.21, + "output": 16.44, + "cache_read": 1.04 } }, - "mistralai/ministral-14b-instruct-2512": { - "id": "mistralai/ministral-14b-instruct-2512", - "name": "Ministral 3 14B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "step-3.5-flash-2603": { + "id": "step-3.5-flash-2603", + "name": "Step 3.5 Flash 2603", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { "input": 0.1, - "output": 0.4 + "output": 0.3, + "cache_read": 0.02 } }, - "mistralai/mistral-small-4-119b-2603": { - "id": "mistralai/mistral-small-4-119b-2603", - "name": "Mistral Small 4 119B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "step-tts-2": { + "id": "step-tts-2", + "name": "Step TTS 2", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "step", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-03-01", + "last_updated": "2026-07-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.4, - "output": 1.4 + "context": 0, + "output": 0 } }, - "mistralai/ministral-14b-2512": { - "id": "mistralai/ministral-14b-2512", - "name": "Ministral 14B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "step-1-32k": { + "id": "step-1-32k", + "name": "Step 1 (32K)", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-04", - "last_updated": "2025-12-04", + "tool_call": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-01-01", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -184958,30 +195677,52 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, + "context": 32768, + "input": 32768, "output": 32768 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 2.05, + "output": 9.59, + "cache_read": 0.41 } - }, - "mistralai/mistral-large-3-675b-instruct-2512": { - "id": "mistralai/mistral-large-3-675b-instruct-2512", - "name": "Mistral Large 3 675B", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", + } + } + }, + "fastrouter": { + "id": "fastrouter", + "env": [ + "FASTROUTER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://go.fastrouter.ai/api/v1", + "name": "FastRouter", + "doc": "https://fastrouter.ai/models", + "models": { + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 32000 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -184989,30 +195730,40 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 256000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1, - "output": 3 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "mistralai/mistral-medium-3": { - "id": "mistralai/mistral-medium-3", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 32000 + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -185020,29 +195771,38 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 2 + "input": 5, + "output": 25 } }, - "mistralai/devstral-2-123b-instruct-2512": { - "id": "mistralai/devstral-2-123b-instruct-2512", - "name": "Devstral 2 123B", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 32000 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -185050,29 +195810,40 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.4, - "output": 1.4 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "mistralai/codestral-2508": { - "id": "mistralai/codestral-2508", - "name": "Codestral 2508", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "codestral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-01", - "last_updated": "2025-08-01", + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 32000 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -185080,26 +195851,32 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 0.8999999999999999 + "input": 3, + "output": 15 } }, - "mistralai/Mistral-Nemo-Instruct-2407": { - "id": "mistralai/Mistral-Nemo-Instruct-2407", - "name": "Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -185108,27 +195885,35 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.1003, - "output": 0.1207 + "input": 1.74, + "output": 3.48 } }, - "bytedance-seed/seed-2.0-lite": { - "id": "bytedance-seed/seed-2.0-lite", - "name": "ByteDance Seed 2.0 Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-10", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -185137,28 +195922,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, + "context": 204800, "output": 131072 }, "cost": { - "input": 0.25, - "output": 2 + "input": 0.95, + "output": 3.15 } }, - "anthracite-org/magnum-v2-72b": { - "id": "anthracite-org/magnum-v2-72b", - "name": "Magnum V2 72B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -185167,59 +195957,58 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 200000, + "output": 131072 }, "cost": { - "input": 2.006, - "output": 2.992 + "input": 1.05, + "output": 3.5 } }, - "anthracite-org/magnum-v4-72b": { - "id": "anthracite-org/magnum-v4-72b", - "name": "Magnum v4 72B", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "sarvam/sarvam-30b": { + "id": "sarvam/sarvam-30b", + "name": "Sarvam 30B", + "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", + "family": "sarvam", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-18", + "last_updated": "2026-02-18", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 128000, + "output": 128000 }, "cost": { - "input": 2.006, - "output": 2.992 + "input": 0.02, + "output": 0.1 } }, - "inflatebot/MN-12B-Mag-Mell-R1": { - "id": "inflatebot/MN-12B-Mag-Mell-R1", - "name": "Mag Mell R1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + "sarvam/sarvam-105b": { + "id": "sarvam/sarvam-105b", + "name": "Sarvam 105B", + "description": "Flagship Indian-language reasoning model for enterprise multilingual applications", + "family": "sarvam", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-09-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text" @@ -185228,34 +196017,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 0.04, + "output": 0.16 } }, - "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning": { - "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning", - "name": "Nvidia Nemotron 3 Nano Omni", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", - "attachment": true, + "x-ai/grok-4": { + "id": "x-ai/grok-4", + "name": "Grok 4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "knowledge": "2025-07", + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -185264,29 +196051,42 @@ "open_weights": false, "limit": { "context": 256000, - "input": 256000, - "output": 65536 + "output": 64000 }, "cost": { - "input": 0.105, - "output": 0.42 + "input": 3, + "output": 15, + "cache_read": 0.75, + "cache_write": 15 } }, - "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF": { - "id": "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF", - "name": "Nvidia Nemotron 70b", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -185294,30 +196094,41 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.357, - "output": 0.408 + "input": 1.25, + "output": 2.5 } }, - "nvidia/nvidia-nemotron-nano-9b-v2": { - "id": "nvidia/nvidia-nemotron-nano-9b-v2", - "name": "Nvidia Nemotron Nano 9B v2", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "x-ai/grok-build-0.1": { + "id": "x-ai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-18", - "last_updated": "2025-08-18", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -185325,260 +196136,255 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.17, - "output": 0.68 + "input": 1, + "output": 2 } }, - "nvidia/Llama-3_3-Nemotron-Super-49B-v1_5": { - "id": "nvidia/Llama-3_3-Nemotron-Super-49B-v1_5", - "name": "Nvidia Nemotron Super 49B v1.5", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, + "bytedance/seedance-2": { + "id": "bytedance/seedance-2", + "name": "Seedance 2", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", + "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "temperature": false, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 - }, - "cost": { - "input": 0.05, - "output": 0.25 + "context": 4096, + "output": 0 } }, - "nvidia/nemotron-3-nano-30b-a3b": { - "id": "nvidia/nemotron-3-nano-30b-a3b", - "name": "Nvidia Nemotron 3 Nano 30B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 262144 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.17, - "output": 0.68 + "input": 0.13, + "output": 0.38 } }, - "nvidia/Llama-3.3-Nemotron-Super-49B-v1": { - "id": "nvidia/Llama-3.3-Nemotron-Super-49B-v1", - "name": "Nvidia Nemotron Super 49B", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, + "google/gemini-3.1-flash-image-preview": { + "id": "google/gemini-3.1-flash-image-preview", + "name": "Nano Banana 2 Preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": false, - "structured_output": false, "temperature": true, - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "knowledge": "2025-01", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 0.5, + "output": 3 } }, - "nvidia/nemotron-3-super-120b-a12b:thinking": { - "id": "nvidia/nemotron-3-super-120b-a12b:thinking", - "name": "Nvidia Nemotron 3 Super 120B Thinking", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "google/veo3.1": { + "id": "google/veo3.1", + "name": "Veo 3.1", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.05, - "output": 0.25 + "context": 400000, + "output": 0 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "Nvidia Nemotron 3 Super 120B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "google/imagen-4.0-ultra": { + "id": "google/imagen-4.0-ultra", + "name": "Imagen 4 Ultra", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.05, - "output": 0.25 + "context": 480, + "output": 0 } }, - "cognitivecomputations/dolphin-2.9.2-qwen2-72b": { - "id": "cognitivecomputations/dolphin-2.9.2-qwen2-72b", - "name": "Dolphin 72b", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "google/veo3.1-lite": { + "id": "google/veo3.1-lite", + "name": "Veo 3.1 Lite", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "temperature": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 8192, - "input": 8192, - "output": 4096 - }, - "cost": { - "input": 0.306, - "output": 0.306 + "context": 400000, + "output": 0 } }, - "xiaomi/mimo-v2-flash-original": { - "id": "xiaomi/mimo-v2-flash-original", - "name": "MiMo V2 Flash Original", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, + "google/gemini-3-pro-image-preview": { + "id": "google/gemini-3-pro-image-preview", + "name": "Nano Banana Pro Preview", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": false, - "structured_output": false, - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, + "context": 65536, "output": 32768 }, "cost": { - "input": 0.102, - "output": 0.306 + "input": 2, + "output": 12 } }, - "xiaomi/mimo-v2-flash-thinking-original": { - "id": "xiaomi/mimo-v2-flash-thinking-original", - "name": "MiMo V2 Flash (Thinking) Original", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -185586,35 +196392,44 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.102, - "output": 0.306 + "input": 2, + "output": 12 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo V2.5", - "description": "MiMo omni model for text, image, video, audio, and agents", + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -185623,36 +196438,37 @@ "open_weights": false, "limit": { "context": 1048576, - "input": 1048576, - "output": 131072 + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 1.5, + "output": 9 } }, - "xiaomi/mimo-v2-omni": { - "id": "xiaomi/mimo-v2-omni", - "name": "MiMo V2 Omni", - "description": "MiMo omni model for text, image, video, audio, and agents", + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "structured_output": false, - "release_date": "2026-03-19", - "last_updated": "2026-03-19", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -185660,69 +196476,68 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08 + "input": 1.25, + "output": 10, + "cache_read": 0.31 } }, - "xiaomi/mimo-v2-flash": { - "id": "xiaomi/mimo-v2-flash", - "name": "MiMo V2 Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "google/veo3.1-fast": { + "id": "google/veo3.1-fast", + "name": "Veo 3.1 Fast", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "temperature": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 - }, - "cost": { - "input": 0.102, - "output": 0.306 + "context": 400000, + "output": 0 } }, - "xiaomi/mimo-v2-pro": { - "id": "xiaomi/mimo-v2-pro", - "name": "MiMo V2 Pro", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", - "attachment": false, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-19", - "last_updated": "2026-03-19", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -185731,66 +196546,52 @@ "open_weights": false, "limit": { "context": 1048576, - "input": 1048576, - "output": 131072 + "output": 65536 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2 + "input": 0.3, + "output": 2.5, + "cache_read": 0.0375 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo V2.5 Pro", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "google/imagen-4.0-fast": { + "id": "google/imagen-4.0-fast", + "name": "Imagen 4 Fast", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "context": 480, + "output": 0 } }, - "xiaomi/mimo-v2-flash-thinking": { - "id": "xiaomi/mimo-v2-flash-thinking", - "name": "MiMo V2 Flash (Thinking)", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", + "deepseek-ai/deepseek-r1-distill-llama-70b": { + "id": "deepseek-ai/deepseek-r1-distill-llama-70b", + "name": "DeepSeek R1 Distill Llama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": false, - "structured_output": false, - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-01-23", + "last_updated": "2025-01-23", "modalities": { "input": [ "text" @@ -185799,214 +196600,274 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.102, - "output": 0.306 + "input": 0.03, + "output": 0.14 } }, - "anthropic/claude-haiku-latest": { - "id": "anthropic/claude-haiku-latest", - "name": "Claude Haiku Latest", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, "structured_output": true, - "release_date": "2026-03-29", - "last_updated": "2026-03-29", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1 + "input": 0.75, + "output": 3.5 } }, - "anthropic/claude-opus-4.7:thinking": { - "id": "anthropic/claude-opus-4.7:thinking", - "name": "Claude 4.7 Opus Thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "moonshotai/kimi-k2": { + "id": "moonshotai/kimi-k2", + "name": "Kimi K2", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-07-11", + "last_updated": "2025-07-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 4.998, - "output": 25.007, - "cache_read": 0.4998 + "input": 0.55, + "output": 2.2 } }, - "anthropic/claude-opus-4.6:thinking:max": { - "id": "anthropic/claude-opus-4.6:thinking:max", - "name": "Claude 4.6 Opus Thinking Max", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 4.998, - "output": 25.007 + "input": 0.3, + "output": 1.2 } }, - "anthropic/claude-opus-4.6:thinking:low": { - "id": "anthropic/claude-opus-4.6:thinking:low", - "name": "Claude 4.6 Opus Thinking Low", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 4.998, - "output": 25.007 + "input": 0.6, + "output": 2.4 } }, - "anthropic/claude-sonnet-4.6:thinking": { - "id": "anthropic/claude-sonnet-4.6:thinking", - "name": "Claude Sonnet 4.6 Thinking", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "wanx/wan-v2-6": { + "id": "wanx/wan-v2-6", + "name": "Wan 2.6", + "description": "Video model for prompt-guided generation, editing, and motion workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" + ], + "output": [ + "video" + ] + }, + "open_weights": true, + "limit": { + "context": 400000, + "output": 0 + } + }, + "qwen/qwen3-coder": { + "id": "qwen/qwen3-coder", + "name": "Qwen3 Coder", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 66536 }, "cost": { - "input": 2.992, - "output": 14.993999999999998 + "input": 0.3, + "output": 1.2 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude 4.7 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "leonardo-ai/lucid-realism": { + "id": "leonardo-ai/lucid-realism", + "name": "Lucid Realism", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "lucid", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-01", + "last_updated": "2025-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 4096, + "output": 0 + } + }, + "leonardo-ai/lucid-origin": { + "id": "leonardo-ai/lucid-origin", + "name": "Lucid Origin", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "lucid", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-01", + "last_updated": "2025-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 4096, + "output": 0 + } + }, + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -186019,33 +196880,42 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 4.998, - "output": 25.007, - "cache_read": 0.4998 + "input": 30, + "output": 180 } }, - "anthropic/claude-opus-4.6:thinking:medium": { - "id": "anthropic/claude-opus-4.6:thinking:medium", - "name": "Claude 4.6 Opus Thinking Medium", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -186053,80 +196923,73 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 4.998, - "output": 25.007 + "input": 0.2, + "output": 1.25 } }, - "anthropic/claude-opus-4.8:thinking": { - "id": "anthropic/claude-opus-4.8:thinking", - "name": "Claude Opus 4.8 Thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "openai/gpt-realtime-1.5": { + "id": "openai/gpt-realtime-1.5", + "name": "GPT Realtime 1.5", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "release_date": "2025-06-01", + "last_updated": "2025-06-01", "modalities": { "input": [ "text", - "image", - "pdf" + "audio", + "image" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 32000, + "output": 4096 }, "cost": { - "input": 4.998, - "output": 25.007, - "cache_read": 0.4998 + "input": 4, + "output": 16 } }, - "anthropic/claude-sonnet-latest": { - "id": "anthropic/claude-sonnet-latest", - "name": "Claude Sonnet Latest", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "temperature": true, + "knowledge": "2024-10-01", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -186134,38 +196997,64 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, + "context": 400000, "output": 128000 }, "cost": { - "input": 2.992, - "output": 14.994, - "cache_read": 0.2992 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "openai/gpt-image-2": { + "id": "openai/gpt-image-2", + "name": "GPT Image 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 0 + } + }, + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", @@ -186178,43 +197067,41 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 4.998, - "output": 25.007, - "cache_read": 0.4998 + "input": 1.75, + "output": 14 } }, - "anthropic/claude-opus-latest": { - "id": "anthropic/claude-opus-latest", - "name": "Claude Opus Latest", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-29", - "last_updated": "2026-03-29", + "temperature": true, + "knowledge": "2024-10-01", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -186222,66 +197109,81 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, + "context": 400000, "output": 128000 }, "cost": { - "input": 4.998, - "output": 25.007, - "cache_read": 0.4998 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": false, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 2.992, - "output": 14.993999999999998 + "input": 0.05, + "output": 0.2 } }, - "anthropic/claude-opus-4.6:thinking": { - "id": "anthropic/claude-opus-4.6:thinking", - "name": "Claude 4.6 Opus Thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -186289,43 +197191,38 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 4.998, - "output": 25.007 + "input": 0.75, + "output": 4.5 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude 4.6 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 } ], "tool_call": true, "structured_output": true, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -186338,29 +197235,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 4.998, - "output": 25.007 + "input": 5, + "output": 30 } }, - "tencent/Hunyuan-MT-7B": { - "id": "tencent/Hunyuan-MT-7B", - "name": "Hunyuan MT 7B", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "family": "hunyuan", - "attachment": false, + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-18", - "last_updated": "2025-09-18", + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -186368,25 +197267,36 @@ }, "open_weights": false, "limit": { - "context": 8192, - "input": 8192, - "output": 8192 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 10, - "output": 20 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "tencent/hy3-preview": { - "id": "tencent/hy3-preview", - "name": "Tencent: Hy3 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -186395,32 +197305,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 262144 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.066, - "output": 0.26, - "cache_read": 0.029 + "input": 0.15, + "output": 0.6 } }, - "chutesai/Mistral-Small-3.2-24B-Instruct-2506": { - "id": "chutesai/Mistral-Small-3.2-24B-Instruct-2506", - "name": "Mistral Small 3.2 24b Instruct", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "chutesai", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-10-01", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -186428,29 +197339,55 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.4 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } - }, - "dmind/dmind-1-mini": { - "id": "dmind/dmind-1-mini", - "name": "DMind-1-Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-01", - "last_updated": "2025-06-01", + } + } + }, + "orcarouter": { + "id": "orcarouter", + "env": [ + "ORCAROUTER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.orcarouter.ai/v1", + "name": "OrcaRouter", + "doc": "https://docs.orcarouter.ai", + "models": { + "grok/grok-4.3": { + "id": "grok/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -186458,29 +197395,57 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.2, - "output": 0.4 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "dmind/dmind-1": { - "id": "dmind/dmind-1", - "name": "DMind-1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-01", - "last_updated": "2025-06-01", + "grok/grok-4.5": { + "id": "grok/grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -186488,29 +197453,43 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.3, - "output": 0.6 + "input": 2, + "output": 6, + "cache_read": 0.5 } }, - "MarinaraSpaghetti/NemoMix-Unleashed-12B": { - "id": "MarinaraSpaghetti/NemoMix-Unleashed-12B", - "name": "NemoMix 12B Unleashed", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "grok/grok-4.6": { + "id": "grok/grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -186518,29 +197497,42 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 2, + "output": 6, + "cache_read": 0.5 } }, - "deepcogito/cogito-v1-preview-qwen-32B": { - "id": "deepcogito/cogito-v1-preview-qwen-32B", - "name": "Cogito v1 Preview Qwen 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-10", - "last_updated": "2025-05-10", + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -186548,29 +197540,43 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.7999999999999998, - "output": 1.7999999999999998 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "cohere/command-r": { - "id": "cohere/command-r", - "name": "Cohere: Command R", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-03-11", - "last_updated": "2024-03-11", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -186578,29 +197584,63 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 0.476, - "output": 1.428 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "cohere/command-r-plus-08-2024": { - "id": "cohere/command-r-plus-08-2024", - "name": "Cohere: Command R+", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -186608,31 +197648,63 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 2.856, - "output": 14.246 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "stepfun/step-3.7-flash:thinking": { - "id": "stepfun/step-3.7-flash:thinking", - "name": "Step 3.7 Flash Thinking", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -186640,30 +197712,43 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "nex-agi/deepseek-v3.1-nex-n1": { - "id": "nex-agi/deepseek-v3.1-nex-n1", - "name": "DeepSeek V3.1 Nex N1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-10", - "last_updated": "2025-12-10", + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -186671,29 +197756,42 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.27999999999999997, - "output": 0.42000000000000004 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "undi95/remm-slerp-l2-13b": { - "id": "undi95/remm-slerp-l2-13b", - "name": "ReMM SLERP 13B", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ @@ -186702,29 +197800,43 @@ }, "open_weights": false, "limit": { - "context": 6144, - "input": 6144, - "output": 4096 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.7989999999999999, - "output": 1.2069999999999999 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 10 } }, - "nothingiisreal/L3.1-70B-Celeste-V0.1-BF16": { - "id": "nothingiisreal/L3.1-70B-Celeste-V0.1-BF16", - "name": "Llama 3.1 70B Celeste v0.1", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -186732,29 +197844,43 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "zai-org/glm-4.7-flash-original:thinking": { - "id": "zai-org/glm-4.7-flash-original:thinking", - "name": "GLM 4.7 Flash Original Thinking", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "attachment": false, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -186762,64 +197888,87 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.07, - "output": 0.4 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "zai-org/glm-4.7": { - "id": "zai-org/glm-4.7", - "name": "GLM 4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-01-29", - "last_updated": "2026-01-29", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "input": 200000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.8 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "zai-org/GLM-4.5-Air:thinking": { - "id": "zai-org/GLM-4.5-Air:thinking", - "name": "GLM 4.5 Air (Thinking)", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "attachment": false, + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -186827,26 +197976,45 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 98304 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.12, - "output": 0.8 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "zai-org/GLM-4.5:thinking": { - "id": "zai-org/GLM-4.5:thinking", - "name": "GLM 4.5 (Thinking)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "deepseek/deepseek-v4-flash-0731": { + "id": "deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -186855,61 +198023,77 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 1.3 + "input": 0.147, + "output": 0.295, + "cache_read": 0.02 } }, - "zai-org/glm-4.5": { - "id": "zai-org/glm-4.5", - "name": "GLM 4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "deepseek/deepseek-v4-flash-vision-exp": { + "id": "deepseek/deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 1.3 + "input": 0.147, + "output": 0.295, + "cache_read": 0.02 } }, - "zai-org/glm-5-original": { - "id": "zai-org/glm-5-original", - "name": "GLM 5 Original", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "deepseek/deepseek-chat": { + "id": "deepseek/deepseek-chat", + "name": "DeepSeek Chat", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ "text" @@ -186918,34 +198102,44 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.147, + "output": 0.295, + "cache_read": 0.02 } }, - "zai-org/glm-5.1": { - "id": "zai-org/glm-5.1", - "name": "GLM 5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "deepseek/deepseek-v4-pro-0813": { + "id": "deepseek/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -186956,30 +198150,44 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 2.55 + "input": 0.442, + "output": 0.884, + "cache_read": 0.06 } }, - "zai-org/glm-4.7-original": { - "id": "zai-org/glm-4.7-original", - "name": "GLM 4.7 Original", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "deepseek/deepseek-v4-flash-free": { + "id": "deepseek/deepseek-v4-flash-free", + "name": "DeepSeek V4 Flash (free)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -186988,29 +198196,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 65535 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0, + "output": 0 } }, - "zai-org/glm-4.7-original:thinking": { - "id": "zai-org/glm-4.7-original:thinking", - "name": "GLM 4.7 Original Thinking", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, + "deepseek/deepseek-reasoner": { + "id": "deepseek/deepseek-reasoner", + "name": "DeepSeek Reasoner", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ "text" @@ -187019,33 +198230,45 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 65535 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0.147, + "output": 0.295, + "cache_read": 0.028 } }, - "zai-org/glm-4.7-flash-original": { - "id": "zai-org/glm-4.7-flash-original", - "name": "GLM 4.7 Flash Original", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -187054,28 +198277,46 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.07, - "output": 0.4 + "input": 0.442, + "output": 0.884, + "cache_read": 0.06 } }, - "zai-org/glm-4.7:thinking": { - "id": "zai-org/glm-4.7:thinking", - "name": "GLM 4.7 Thinking", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -187084,29 +198325,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 65535 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0.147, + "output": 0.295, + "cache_read": 0.02 } }, - "zai-org/glm-5.1:thinking": { - "id": "zai-org/glm-5.1:thinking", - "name": "GLM 5.1 Thinking", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "tencent/hy3-free": { + "id": "tencent/hy3-free", + "name": "Hy3 (free)", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ "text" @@ -187117,27 +198367,36 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 131072 + "context": 256000, + "input": 192000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 2.55 + "input": 0, + "output": 0 } }, - "zai-org/glm-5:thinking": { - "id": "zai-org/glm-5:thinking", - "name": "GLM 5 Thinking", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "tencent/hy3": { + "id": "tencent/hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ "text" @@ -187148,90 +198407,116 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, + "context": 256000, + "input": 192000, "output": 128000 }, "cost": { - "input": 0.3, - "output": 2.55 + "input": 0.18, + "output": 0.59, + "cache_read": 0.059 } }, - "zai-org/glm-4.6v": { - "id": "zai-org/glm-4.6v", - "name": "GLM 4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "z-ai/glm-4.5": { + "id": "z-ai/glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 24000 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "zai-org/glm-4.6-original": { - "id": "zai-org/glm-4.6-original", - "name": "GLM 4.6 Original", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, + "z-ai/glm-5.3-flash": { + "id": "z-ai/glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "max" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 65535 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.35, - "output": 1.4 + "input": 0.075, + "output": 0.25 } }, - "zai-org/glm-5-original:thinking": { - "id": "zai-org/glm-5-original:thinking", - "name": "GLM 5 Original Thinking", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -187240,29 +198525,37 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "zai-org/glm-4.7-flash:thinking": { - "id": "zai-org/glm-4.7-flash:thinking", - "name": "GLM 4.7 Flash Thinking", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -187271,82 +198564,109 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.07, - "output": 0.4 + "input": 1, + "output": 3.2, + "cache_read": 0.26, + "cache_write": 0 } }, - "zai-org/glm-4.6v-flash-original": { - "id": "zai-org/glm-4.6v-flash-original", - "name": "GLM 4.6V Flash", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 24000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "zai-org/glm-4.6v-original": { - "id": "zai-org/glm-4.6v-original", - "name": "GLM 4.6V Original", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "z-ai/glm-5.3-flash-free": { + "id": "z-ai/glm-5.3-flash-free", + "name": "GLM-5.3-Flash (free)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 24000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 0.9 + "input": 0, + "output": 0 } }, - "zai-org/glm-latest": { - "id": "zai-org/glm-latest", - "name": "GLM Latest", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -187354,14 +198674,18 @@ "type": "effort", "values": [ "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -187370,23 +198694,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, + "context": 1000000, "output": 131072 }, "cost": { - "input": 0.75, - "output": 2.6, - "cache_read": 0.15 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "zai-org/glm-4.7-flash": { - "id": "zai-org/glm-4.7-flash", - "name": "GLM 4.7 Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -187395,9 +198718,13 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -187409,25 +198736,32 @@ "open_weights": true, "limit": { "context": 200000, - "input": 200000, - "output": 128000 + "output": 131072 }, "cost": { - "input": 0.07, - "output": 0.4 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "zai-org/GLM-4.6-turbo:thinking": { - "id": "zai-org/GLM-4.6-turbo:thinking", - "name": "GLM 4.6 Turbo (Thinking)", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "z-ai/glm-4.5-air": { + "id": "z-ai/glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-10-02", - "last_updated": "2025-10-02", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -187436,27 +198770,43 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 204800 + "context": 131072, + "output": 98304 }, "cost": { - "input": 1, - "output": 3 + "input": 0.2, + "output": 1.1, + "cache_read": 0.03, + "cache_write": 0 } }, - "zai-org/GLM-4.5-Air": { - "id": "zai-org/GLM-4.5-Air", - "name": "GLM 4.5 Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "z-ai/glm-5.3": { + "id": "z-ai/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -187465,30 +198815,47 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 98304 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.12, - "output": 0.8 + "input": 1.26, + "output": 3.96, + "cache_read": 0.234 } }, - "zai-org/GLM-4.6-turbo": { - "id": "zai-org/GLM-4.6-turbo", - "name": "GLM 4.6 Turbo", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-10-02", - "last_updated": "2025-10-02", + "meta/muse-spark-1.1": { + "id": "meta/muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-08", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf", + "video" ], "output": [ "text" @@ -187496,64 +198863,46 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 204800 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 1, - "output": 3 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "zai-org/glm-5": { - "id": "zai-org/glm-5", - "name": "GLM 5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "meta/muse-spark-1.2": { + "id": "meta/muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "input": 200000, - "output": 128000 - }, - "cost": { - "input": 0.3, - "output": 2.55 - } - }, - "Infermatic/MN-12B-Inferor-v0.0": { - "id": "Infermatic/MN-12B-Inferor-v0.0", - "name": "Mistral Nemo Inferor 12B", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "temperature": true, + "release_date": "2026-08-05", + "last_updated": "2026-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf", + "audio" ], "output": [ "text" @@ -187561,29 +198910,45 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.25499999999999995, - "output": 0.49299999999999994 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "shisa-ai/shisa-v2.1-llama3.3-70b": { - "id": "shisa-ai/shisa-v2.1-llama3.3-70b", - "name": "Shisa V2.1 Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "google/gemini-flash-latest": { + "id": "google/gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -187591,59 +198956,79 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 4096 + "context": 1048576, + "output": 65536 }, "cost": { "input": 0.5, - "output": 0.5 + "output": 3, + "cache_read": 0.1 } }, - "shisa-ai/shisa-v2-llama3.3-70b": { - "id": "shisa-ai/shisa-v2-llama3.3-70b", - "name": "Shisa V2 Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 0.5 + "input": 0.13, + "output": 0.38, + "cache_read": 0.02 } }, - "abacusai/Dracarys-72B-Instruct": { - "id": "abacusai/Dracarys-72B-Instruct", - "name": "Llama 3.1 70B Dracarys 2", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-02", - "last_updated": "2025-08-02", + "google/gemini-flash-lite-latest": { + "id": "google/gemini-flash-lite-latest", + "name": "Gemini Flash-Lite Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -187651,29 +199036,47 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "deepseek-ai/DeepSeek-V3.1:thinking": { - "id": "deepseek-ai/DeepSeek-V3.1:thinking", - "name": "DeepSeek V3.1 Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "google/gemini-3.6-flash": { + "id": "google/gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -187681,29 +199084,43 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.2, - "output": 0.7 + "input": 1.5, + "output": 7.5, + "cache_read": 0.15 } }, - "deepseek-ai/DeepSeek-V3.1-Terminus": { - "id": "deepseek-ai/DeepSeek-V3.1-Terminus", - "name": "DeepSeek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "google/gemini-robotics-er-1.6-preview": { + "id": "google/gemini-robotics-er-1.6-preview", + "name": "Gemini Robotics-ER 1.6 Preview", + "description": "Vision-language model for embodied reasoning: spatial understanding, task planning, and physical-world agentic robotics", + "family": "gemini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0 + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-08-02", - "last_updated": "2025-08-02", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-14", + "last_updated": "2026-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -187711,61 +199128,78 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, + "context": 131072, "output": 65536 }, "cost": { - "input": 0.25, - "output": 0.7 + "input": 1, + "output": 5 } }, - "deepseek-ai/deepseek-v3.2-exp-thinking": { - "id": "deepseek-ai/deepseek-v3.2-exp-thinking", - "name": "DeepSeek V3.2 Exp Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek-thinking", - "attachment": false, + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163840, - "input": 163840, - "output": 65536 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.27999999999999997, - "output": 0.42000000000000004 + "input": 0.06, + "output": 0.33, + "cache_read": 0.0075 } }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek", - "attachment": false, + "google/gemini-3.5-flash-lite": { + "id": "google/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -187773,29 +199207,44 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 163840 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 1.7 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", + "image", + "video", + "audio", "pdf" ], "output": [ @@ -187804,29 +199253,63 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.2, - "output": 0.7 + "input": 2, + "output": 12, + "cache_read": 0.2, + "input_audio": 2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "deepseek-ai/deepseek-v3.2-exp": { - "id": "deepseek-ai/deepseek-v3.2-exp", - "name": "DeepSeek V3.2 Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -187834,29 +199317,47 @@ }, "open_weights": false, "limit": { - "context": 163840, - "input": 163840, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.27999999999999997, - "output": 0.42000000000000004 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 0.08333, + "input_audio": 3 } }, - "deepseek-ai/DeepSeek-V3.1-Terminus:thinking": { - "id": "deepseek-ai/DeepSeek-V3.1-Terminus:thinking", - "name": "DeepSeek V3.1 Terminus (Thinking)", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": false, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -187864,29 +199365,61 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.25, - "output": 0.7 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "arcee-ai/trinity-large-thinking": { - "id": "arcee-ai/trinity-large-thinking", - "name": "Trinity Large Thinking", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -187894,29 +199427,46 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 80000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 0.9 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "arcee-ai/trinity-mini": { - "id": "arcee-ai/trinity-mini", - "name": "Trinity Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "trinity-mini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -187924,29 +199474,62 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.045000000000000005, - "output": 0.15 - } - }, - "soob3123/GrayLine-Qwen3-8B": { - "id": "soob3123/GrayLine-Qwen3-8B", - "name": "Grayline Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", - "modalities": { - "input": [ - "text" + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -187954,29 +199537,46 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 0.3 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 } }, - "soob3123/Veiled-Calla-12B": { - "id": "soob3123/Veiled-Calla-12B", - "name": "Veiled Calla 12B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-13", - "last_updated": "2025-04-13", + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -187984,29 +199584,95 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 0.3 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "input_audio": 0.3 } }, - "soob3123/amoral-gemma3-27B-v2": { - "id": "soob3123/amoral-gemma3-27B-v2", - "name": "Amoral Gemma3 27B v2", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-23", - "last_updated": "2025-05-23", + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 + } + }, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -188014,86 +199680,239 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 0.3 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "meganova-ai/manta-mini-1.0": { - "id": "meganova-ai/manta-mini-1.0", - "name": "Manta Mini 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-20", - "last_updated": "2025-12-20", + "kimi/kimi-k3": { + "id": "kimi/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ + "text", + "image", + "video" + ], + "output": [ "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33 + } + }, + "kimi/kimi-k2.6": { + "id": "kimi/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "input": 8192, - "output": 8192 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.02, - "output": 0.16 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "meganova-ai/manta-flash-1.0": { - "id": "meganova-ai/manta-flash-1.0", - "name": "Manta Flash 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-20", - "last_updated": "2025-12-20", + "kimi/kimi-k2.5": { + "id": "kimi/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ + "text", + "image", + "video" + ], + "output": [ "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.1, + "cache_write": 0 + } + }, + "kimi/kimi-k2.7-code": { + "id": "kimi/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.02, - "output": 0.16 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "meganova-ai/manta-pro-1.0": { - "id": "meganova-ai/manta-pro-1.0", - "name": "Manta Pro 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova", + "orcarouter/fusion-mini": { + "id": "orcarouter/fusion-mini", + "name": "OrcaRouter Fusion Mini", + "description": "Leaner two-model Fusion panel that runs Claude Opus 4.8 and GPT-5.5 in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Easy requests fall through to a cheaper default and bill as one call.", + "family": "model-router", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + } + }, + "orcarouter/free": { + "id": "orcarouter/free", + "name": "OrcaRouter Free", + "description": "Built-in router over the free tier that scores each request's difficulty and sends light work to the smaller free model and harder work to the stronger one. Priced at zero and never falls back to a paid model.", + "family": "auto", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-20", - "last_updated": "2025-12-20", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ "text" @@ -188104,31 +199923,41 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, + "context": 65536, "output": 32768 }, "cost": { - "input": 0.060000000000000005, - "output": 0.5 + "input": 0, + "output": 0 } }, - "qwen/Qwen3.6-35B-A3B:thinking": { - "id": "qwen/Qwen3.6-35B-A3B:thinking", - "name": "Qwen3.6 35B A3B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "orcarouter/fusion": { + "id": "orcarouter/fusion", + "name": "OrcaRouter Fusion", + "description": "Curated fan-out router that runs Claude Opus 4.8, GPT-5.5 and Gemini 3.1 Pro in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Easy requests fall through to a cheaper default and bill as one call.", + "family": "model-router", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-19", - "last_updated": "2026-04-19", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -188136,25 +199965,89 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, + "context": 1000000, + "output": 128000 + } + }, + "orcarouter/fusion-flash": { + "id": "orcarouter/fusion-flash", + "name": "OrcaRouter Fusion Flash", + "description": "Budget Fusion panel that runs Gemini 3.5 Flash, MiniMax M2.7 and GLM 5.1 in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Cost-sensitive fan-out over a 200K window.", + "family": "model-router", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 128000 + } + }, + "orcarouter/auto": { + "id": "orcarouter/auto", + "name": "OrcaRouter Auto", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2026-05-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, "output": 16384 }, "cost": { - "input": 0.112, - "output": 0.8 + "input": 0, + "output": 0 } }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-17", - "last_updated": "2025-09-17", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -188163,27 +200056,66 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1, - "output": 5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "qwen/Qwen3.6-35B-A3B": { - "id": "qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", @@ -188194,31 +200126,108 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1048576, + "output": 512000 }, "cost": { - "input": 0.112, - "output": 0.8 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "qwen/qwen3-32b": { - "id": "qwen/qwen3-32b", - "name": "Qwen 3 32b", + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "minimax/minimax-m2.5-highspeed": { + "id": "minimax/minimax-m2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-13", + "last_updated": "2026-02-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "qwen/qwen3.5-flash": { + "id": "qwen/qwen3.5-flash", + "name": "Qwen3.5 Flash", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 81920 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "pdf" + "image", + "video" ], "output": [ "text" @@ -188226,25 +200235,27 @@ }, "open_weights": false, "limit": { - "context": 41000, - "input": 41000, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { "input": 0.1, - "output": 0.3 + "output": 0.4 } }, - "qwen/Qwen3-VL-235B-A22B-Instruct": { - "id": "qwen/Qwen3-VL-235B-A22B-Instruct", + "qwen/qwen3-vl-235b-a22b-instruct": { + "id": "qwen/qwen3-vl-235b-a22b-instruct", "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", @@ -188254,27 +200265,35 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 262144 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.4, + "output": 1.6 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -188285,19 +200304,64 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 3.75, + "cache_read": 0.25, + "cache_write": 1.563 + } + }, + "qwen/qwen3.8-27b": { + "id": "qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, "output": 32768 }, "cost": { - "input": 1.08018, - "output": 5.4009 + "input": 0.33, + "output": 2.4 } }, - "qwen/qwen3.5-397b-a17b-thinking": { - "id": "qwen/qwen3.5-397b-a17b-thinking", - "name": "Qwen3.5 397B A17B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen/qwen3-vl-235b-a22b-thinking": { + "id": "qwen/qwen3-vl-235b-a22b-thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -188307,8 +200371,145 @@ "max": 81920 } ], - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 40960 + }, + "cost": { + "input": 0.4, + "output": 4 + } + }, + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.086, + "output": 0.688 + } + }, + "qwen/qwen3.8-max": { + "id": "qwen/qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 + } + }, + "qwen/qwen3.5-plus": { + "id": "qwen/qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", "release_date": "2026-02-16", "last_updated": "2026-02-16", "modalities": { @@ -188323,28 +200524,40 @@ }, "open_weights": false, "limit": { - "context": 258048, - "input": 258048, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.115, + "output": 0.688, + "reasoning": 2.4 } }, - "qwen/Qwen3-Next-80B-A3B-Instruct": { - "id": "qwen/Qwen3-Next-80B-A3B-Instruct", - "name": "Qwen3 Next 80B A3B (Instruct)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, + "qwen/qwen3.7-flash": { + "id": "qwen/qwen3.7-flash", + "name": "Qwen3.7 Flash", + "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-09-11", - "last_updated": "2025-09-11", + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -188352,83 +200565,116222 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 262144 + "context": 1000000, + "input": 991000, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.65 + "input": 0.03, + "output": 0.13, + "cache_read": 0.006, + "cache_write": 0.038 } }, - "qwen/Qwen3-235B-A22B-Instruct-2507-TEE": { - "id": "qwen/Qwen3-235B-A22B-Instruct-2507-TEE", - "name": "Qwen 3 235b A22B 2507 (TEE)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.359, + "output": 1.434 + } + }, + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.115, + "output": 0.917 + } + }, + "qwen/qwen3.6-35b-a3b": { + "id": "qwen/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.248, + "output": 1.485 + } + }, + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5 + } + }, + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.057, + "output": 0.459 + } + }, + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.172, + "output": 1.032 + } + }, + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } + } + }, + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.35, + "output": 1.42, + "cache_read": 0.071 + } + }, + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } + } + }, + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } + } + }, + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16385, + "output": 4096 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0 + } + }, + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 100000 + }, + "cost": { + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } + } + }, + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "openai/gpt-4o-2024-05-13": { + "id": "openai/gpt-4o-2024-05-13", + "name": "GPT-4o (2024-05-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 5, + "output": 15 + } + }, + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2 Pro", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 21, + "output": 168 + } + }, + "openai/gpt-4": { + "id": "openai/gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 30, + "output": 60 + } + }, + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 + } + }, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 + } + }, + "openai/gpt-5.2-chat-latest": { + "id": "openai/gpt-5.2-chat-latest", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "openai/gpt-5.1-chat-latest": { + "id": "openai/gpt-5.1-chat-latest", + "name": "GPT-5.1 Chat", + "description": "Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 272000 + }, + "cost": { + "input": 15, + "output": 120 + } + }, + "openai/gpt-5-chat-latest": { + "id": "openai/gpt-5-chat-latest", + "name": "GPT-5 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 100000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/gpt-4o-2024-08-06": { + "id": "openai/gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-08-06", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 + } + }, + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 + } + }, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "openai/gpt-4o-2024-11-20": { + "id": "openai/gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 + } + }, + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 10, + "output": 30 + } + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.03, + "output": 0.17 + } + }, + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5 + } + } + } + }, + "friendli": { + "id": "friendli", + "env": [ + "FRIENDLI_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.friendli.ai/serverless/v1", + "name": "Friendli", + "doc": "https://friendli.ai/docs/guides/serverless_endpoints/introduction", + "models": { + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.14, + "output": 0.4 + } + }, + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "DeepSeek V3.2", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 64000 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.25 + } + }, + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "zai-org/GLM-5.3-Flash": { + "id": "zai-org/GLM-5.3-Flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 + } + }, + "zai-org/GLM-5.3": { + "id": "zai-org/GLM-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.26, + "output": 3.96, + "cache_read": 0.234 + } + } + } + }, + "kimi-code-plan-cn": { + "id": "kimi-code-plan-cn", + "env": [ + "KIMI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.kimi.com/coding/v1", + "name": "Kimi For Coding (kimi.com)", + "doc": "https://www.kimi.com/code/docs/en/kimi-code/models.html", + "models": { + "kimi-for-coding-highspeed": { + "id": "kimi-for-coding-highspeed", + "name": "Kimi For Coding HighSpeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "kimi-for-coding": { + "id": "kimi-for-coding", + "name": "kimi-for-coding", + "description": "Kimi coding model with more efficient thinking and up to 1M context, available through Kimi Code", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "k3-256k": { + "id": "k3-256k", + "name": "Kimi K3-256K", + "description": "256K-context version of Kimi K3, reducing token consumption for shorter coding sessions", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "k3": { + "id": "k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + } + } + }, + "inco": { + "id": "inco", + "env": [ + "INCO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.inco.ai/v1", + "name": "Inco", + "doc": "https://platform.inco.ai/docs", + "models": { + "kimi-k3:fast": { + "id": "kimi-k3:fast", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 6, + "output": 30 + } + }, + "glm-5.3:fast": { + "id": "glm-5.3:fast", + "name": "GLM-5.3 Fast", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 2.8, + "output": 8.8 + } + }, + "glm-5.3-flash:fast": { + "id": "glm-5.3-flash:fast", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.5 + } + }, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 512000 + }, + "cost": { + "input": 0.3, + "output": 1.2 + } + }, + "deepseek-v4.1-flash:fast": { + "id": "deepseek-v4.1-flash:fast", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.6, + "output": 2.4 + } + }, + "minimax-m3:fast": { + "id": "minimax-m3:fast", + "name": "MiniMax M3 Fast", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 512000 + }, + "cost": { + "input": 0.6, + "output": 2.4 + } + }, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4 + } + } + } + }, + "sakana": { + "id": "sakana", + "env": [ + "SAKANA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.sakana.ai/v1", + "name": "Sakana AI", + "doc": "https://console.sakana.ai/models", + "models": { + "fugu-ultra": { + "id": "fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "provider": { + "shape": "responses" + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "fugu-ultra-20260615": { + "id": "fugu-ultra-20260615", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "provider": { + "shape": "responses" + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "sakana-namazu": { + "id": "sakana-namazu", + "name": "Sakana Namazu", + "description": "Japanese-specialized reasoning model based on Kimi K2.6 and tuned for Japanese language, culture, and business workflows", + "family": "sakana-namazu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.15 + } + }, + "fugu": { + "id": "fugu", + "name": "Fugu", + "description": "Multi-agent model for routing expert agents across complex analytical tasks", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "provider": { + "shape": "responses" + } + } + } + }, + "scx-ai": { + "id": "scx-ai", + "env": [ + "SCX_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.scx.ai/v1", + "name": "SCX.ai", + "doc": "https://platform.scx.ai/docs", + "models": { + "GLM-5.2": { + "id": "GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.55, + "output": 1.784, + "cache_read": 0.111 + } + }, + "Qwen3.8-Max": { + "id": "Qwen3.8-Max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 983616, + "output": 131072 + }, + "cost": { + "input": 1.815, + "output": 5.4461, + "cache_read": 0.17, + "cache_write": 2.5 + } + }, + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 196608 + }, + "cost": { + "input": 0.48, + "output": 1.79, + "cache_read": 0.05 + } + }, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.17, + "output": 0.55 + } + } + } + }, + "zenifra": { + "id": "zenifra", + "env": [ + "ZENIFRA_AI_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://ai.zenifra.com/v1", + "name": "Zenifra", + "doc": "https://docs.zenifra.com", + "models": { + "alibaba/qwen3.6-35b-a3b": { + "id": "alibaba/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "provider": { + "shape": "completions" + }, + "cost": { + "input": 0.19, + "output": 0.48 + } + } + } + }, + "tokenrouter": { + "id": "tokenrouter", + "env": [ + "TOKENROUTER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.tokenrouter.com/v1", + "name": "TokenRouter", + "doc": "https://www.tokenrouter.com/docs/tokenrouter-feature-guide/", + "models": { + "z-ai/glm-5.3-free": { + "id": "z-ai/glm-5.3-free", + "name": "GLM-5.3 (free)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + } + } + }, + "google-vertex-anthropic": { + "id": "google-vertex-anthropic", + "env": [ + "GOOGLE_VERTEX_PROJECT", + "GOOGLE_VERTEX_LOCATION", + "GOOGLE_APPLICATION_CREDENTIALS" + ], + "npm": "@ai-sdk/google-vertex/anthropic", + "name": "Vertex (Anthropic)", + "doc": "https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude", + "models": { + "claude-opus-4-8@default": { + "id": "claude-opus-4-8@default", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "claude-sonnet-4-5@20250929": { + "id": "claude-sonnet-4-5@20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "claude-opus-4-6@default": { + "id": "claude-opus-4-6@default", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "claude-opus-4@20250514": { + "id": "claude-opus-4@20250514", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "claude-opus-5-5@default": { + "id": "claude-opus-5-5@default", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 + } + }, + "claude-opus-5@default": { + "id": "claude-opus-5@default", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "claude-sonnet-4-6@default": { + "id": "claude-sonnet-4-6@default", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } + } + }, + "claude-haiku-4-5@20251001": { + "id": "claude-haiku-4-5@20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "claude-opus-4-1@20250805": { + "id": "claude-opus-4-1@20250805", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "claude-opus-4-5@20251101": { + "id": "claude-opus-4-5@20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "claude-opus-4-7@default": { + "id": "claude-opus-4-7@default", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "claude-fable-5-1@default": { + "id": "claude-fable-5-1@default", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "claude-fable-5@default": { + "id": "claude-fable-5@default", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "claude-sonnet-4@20250514": { + "id": "claude-sonnet-4@20250514", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "status": "deprecated", + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "claude-sonnet-5@default": { + "id": "claude-sonnet-5@default", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + } + } + }, + "moonshotai": { + "id": "moonshotai", + "env": [ + "MOONSHOT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.moonshot.ai/v1", + "name": "Moonshot AI", + "doc": "https://platform.moonshot.ai/docs/api/chat", + "models": { + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "kimi-k2.7-code-highspeed": { + "id": "kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code HighSpeed", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.9, + "output": 8, + "cache_read": 0.38 + } + }, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } + }, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + } + } + }, + "ainetcafe": { + "id": "ainetcafe", + "env": [ + "AINETCAFE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://microquickjs.com/v1", + "name": "ainetcafe", + "doc": "https://ainetcafe.com/k3/guides/", + "models": { + "Kimi-K3": { + "id": "Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 2.1, + "output": 10.5, + "cache_read": 0.3 + } + } + } + }, + "wallaby": { + "id": "wallaby", + "env": [ + "WALLABY_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.wallabytoken.com/v1", + "name": "Wallaby", + "doc": "https://wallabytoken.com/docs", + "models": { + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 2.7, + "output": 13.5, + "cache_read": 0.27 + } + } + } + }, + "scnet-token-plan": { + "id": "scnet-token-plan", + "env": [ + "SCNET_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.scnet.cn/api/llm/v1", + "name": "SCNet Token Plan", + "doc": "https://www.scnet.cn/ac/openapi/doc/2.0/moduleapi/plans/token-plan.html", + "models": { + "DeepSeek-V4-Flash-0731": { + "id": "DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "Kimi-K2.6": { + "id": "Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 512000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "GLM-5.1": { + "id": "GLM-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "GLM-5.2": { + "id": "GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "Qwen3.8-Max": { + "id": "Qwen3.8-Max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "Kimi-K3": { + "id": "Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "DeepSeek-V4-Pro-0813": { + "id": "DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "DeepSeek-V4.1-Flash": { + "id": "DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "DeepSeek-V4-Flash": { + "id": "DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "GLM-5": { + "id": "GLM-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "GLM-5.3-Flash": { + "id": "GLM-5.3-Flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "DeepSeek-V4-Pro": { + "id": "DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "Qwen3.8-Flash": { + "id": "Qwen3.8-Flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "GLM-5.3": { + "id": "GLM-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "Kimi-K2.7-Code": { + "id": "Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "Kimi-K2.5": { + "id": "Kimi-K2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + } + } + }, + "ofox": { + "id": "ofox", + "env": [ + "OFOX_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.ofox.ai/v1", + "name": "Ofox", + "doc": "https://ofox.ai/docs", + "models": { + "bailian/qwen-flash": { + "id": "bailian/qwen-flash", + "name": "Qwen Flash", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 32768 + }, + "cost": { + "input": 0.022, + "output": 0.22, + "cache_read": 0.0043, + "cache_write": 0.027 + } + }, + "bailian/qwen3.5-flash": { + "id": "bailian/qwen3.5-flash", + "name": "Qwen3.5 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0.125 + } + }, + "bailian/qwen3.7-max": { + "id": "bailian/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 + } + }, + "bailian/qwen3.8-27b": { + "id": "bailian/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1131072, + "output": 131072 + }, + "cost": { + "input": 0.45, + "output": 3.2, + "cache_read": 0.05, + "cache_write": 0.5625 + } + }, + "bailian/qwen3.5-27b": { + "id": "bailian/qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.29, + "output": 2.05, + "cache_read": 0.29 + } + }, + "bailian/qwen-max": { + "id": "bailian/qwen-max", + "name": "Qwen Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 8192 + }, + "cost": { + "input": 0.35, + "output": 1.38, + "cache_read": 0.069 + } + }, + "bailian/qwen3.8-max": { + "id": "bailian/qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 + } + }, + "bailian/qwen3-coder-next": { + "id": "bailian/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.2, + "output": 1.5 + } + }, + "bailian/qwen3.5-plus": { + "id": "bailian/qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.4, + "output": 2.4, + "cache_read": 0.04, + "cache_write": 0.4 + } + }, + "bailian/qwen-plus": { + "id": "bailian/qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 32768 + }, + "cost": { + "input": 0.12, + "output": 0.29, + "cache_read": 0.023 + } + }, + "bailian/qwen3-max": { + "id": "bailian/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.36, + "output": 1.43, + "cache_read": 0.072 + } + }, + "bailian/qwen-turbo": { + "id": "bailian/qwen-turbo", + "name": "Qwen Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-11-01", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.05, + "output": 0.09, + "cache_read": 0.0086 + } + }, + "bailian/qwen-vl-max": { + "id": "bailian/qwen-vl-max", + "name": "Qwen-VL Max", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-08", + "last_updated": "2025-08-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.23, + "output": 0.58, + "cache_read": 0.046 + } + }, + "bailian/qwen3.5-122b-a10b": { + "id": "bailian/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.29, + "output": 2.29, + "cache_read": 0.29 + } + }, + "bailian/qwen3.6-flash": { + "id": "bailian/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 0.31 + } + }, + "bailian/qwen3-coder-flash": { + "id": "bailian/qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 2.5, + "cache_read": 0.06, + "cache_write": 0.27 + } + }, + "bailian/qwen3-coder-plus": { + "id": "bailian/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.8, + "output": 9, + "cache_read": 0.2, + "cache_write": 1 + } + }, + "bailian/qwen3.8-flash": { + "id": "bailian/qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.47, + "cache_read": 0.016, + "cache_write": 0.2 + } + }, + "bailian/qwen3.5-35b-a3b": { + "id": "bailian/qwen3.5-35b-a3b", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.29, + "output": 1.83, + "cache_read": 0.29 + } + }, + "bailian/qwen3.6-max-preview": { + "id": "bailian/qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 2.15, + "output": 12.86, + "cache_read": 0.2, + "cache_write": 1.17 + } + }, + "bailian/qwen3.5-397b-a17b": { + "id": "bailian/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.55, + "output": 3.5, + "cache_read": 0.55 + } + }, + "bailian/qwen3.6-27b": { + "id": "bailian/qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.6, + "output": 3.6 + } + }, + "bailian/qwen3.6-plus": { + "id": "bailian/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625 + } + }, + "bailian/qwen3.7-plus": { + "id": "bailian/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "cache_write": 0.5 + } + }, + "bailian/qwen3.8-max-0902": { + "id": "bailian/qwen3.8-max-0902", + "name": "Qwen3.8 Max 0902", + "description": "2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 + } + }, + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-fable-5.1": { + "id": "anthropic/claude-fable-5.1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic/claude-opus-5.5": { + "id": "anthropic/claude-opus-5.5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://api.ofox.ai/anthropic/v1" + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 + } + }, + "deepseek/deepseek-v4-flash-0731": { + "id": "deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.308, + "output": 0.924, + "cache_read": 0.0098 + } + }, + "deepseek/deepseek-v4-flash-0423": { + "id": "deepseek/deepseek-v4-flash-0423", + "name": "DeepSeek V4 Flash 0423", + "description": "Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.19, + "output": 0.51, + "cache_read": 0.028 + } + }, + "deepseek/deepseek-v4.1-flash": { + "id": "deepseek/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.21, + "output": 0.84, + "cache_read": 0.0042 + } + }, + "deepseek/deepseek-v4-flash-vision-exp": { + "id": "deepseek/deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "status": "beta", + "cost": { + "input": 0.44, + "output": 1.32, + "cache_read": 0.014 + } + }, + "deepseek/deepseek-v4-pro-0813": { + "id": "deepseek/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.924, + "output": 2.772, + "cache_read": 0.0308 + } + }, + "deepseek/deepseek-v4-pro-0423": { + "id": "deepseek/deepseek-v4-pro-0423", + "name": "DeepSeek V4 Pro 0423", + "description": "DeepSeek V4 Pro initial snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 1.32, + "output": 3.96, + "cache_read": 0.15 + } + }, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 + } + }, + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32000 + }, + "cost": { + "input": 0.29, + "output": 0.43, + "cache_read": 0.06 + } + }, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.44, + "output": 1.32, + "cache_read": 0.014 + } + }, + "z-ai/glm-5v-turbo": { + "id": "z-ai/glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.24 + } + }, + "z-ai/glm-5.3-flash": { + "id": "z-ai/glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 + } + }, + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 + } + }, + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.2 + } + }, + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.4, + "output": 2.2, + "cache_read": 0.11 + } + }, + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "z-ai/glm-4.7-flashx": { + "id": "z-ai/glm-4.7-flashx", + "name": "GLM-4.7-FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.072, + "output": 0.4, + "cache_read": 0.01 + } + }, + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.24 + } + }, + "z-ai/glm-5.3": { + "id": "z-ai/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "x-ai/grok-4.7": { + "id": "x-ai/grok-4.7", + "name": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 30000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 + } + }, + "x-ai/grok-4.5": { + "id": "x-ai/grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.3 + } + }, + "x-ai/grok-4.20": { + "id": "x-ai/grok-4.20", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 12, + "cache_read": 0.4 + } + }, + "x-ai/grok-4.1-fast": { + "id": "x-ai/grok-4.1-fast", + "name": "Grok 4.1 Fast", + "description": "xAI's fast agentic tool-calling model with a 2M context window; non-reasoning variant for low-latency responses", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 30000 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } + }, + "x-ai/grok-4.6": { + "id": "x-ai/grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "volcengine/doubao-seed-2.0-mini": { + "id": "volcengine/doubao-seed-2.0-mini", + "name": "Seed 2.0 Mini", + "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 0.06, + "output": 0.56, + "cache_read": 0.02, + "cache_write": 0.0024 + } + }, + "volcengine/doubao-seed-1-6-flash": { + "id": "volcengine/doubao-seed-1-6-flash", + "name": "Seed 1.6 Flash", + "description": "Low-latency ByteDance Seed model for high-throughput chat, extraction, and lightweight tool use", + "family": "seed", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-28", + "last_updated": "2025-08-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 0.03, + "output": 0.22, + "cache_read": 0.0043 + } + }, + "volcengine/doubao-seed-2.0-lite": { + "id": "volcengine/doubao-seed-2.0-lite", + "name": "Seed 2.0 Lite", + "description": "Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 0.13, + "output": 0.76, + "cache_read": 0.03, + "cache_write": 0.0024 + } + }, + "volcengine/doubao-seed-1-8": { + "id": "volcengine/doubao-seed-1-8", + "name": "Seed 1.8", + "description": "ByteDance Seed model for multimodal reasoning, long-context analysis, and agent workflows", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-28", + "last_updated": "2025-12-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.12, + "output": 0.29, + "cache_read": 0.023 + } + }, + "volcengine/doubao-seed-1-6": { + "id": "volcengine/doubao-seed-1-6", + "name": "Seed 1.6", + "description": "ByteDance Seed model for long-context reasoning, instruction following, and tool-assisted tasks", + "family": "seed", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.12, + "output": 0.29, + "cache_read": 0.023 + } + }, + "volcengine/doubao-seed-2.0-pro": { + "id": "volcengine/doubao-seed-2.0-pro", + "name": "Seed 2.0 Pro", + "description": "Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.67, + "output": 3.36, + "cache_read": 0.14, + "cache_write": 0.0024 + } + }, + "volcengine/doubao-seed-1-6-vision": { + "id": "volcengine/doubao-seed-1-6-vision", + "name": "Seed 1.6 Vision", + "description": "ByteDance Seed multimodal model for image understanding, visual reasoning, and tool-assisted tasks", + "family": "seed", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-15", + "last_updated": "2025-08-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 0.12, + "output": 1.15, + "cache_read": 0.023 + } + }, + "volcengine/doubao-seed-2.1-turbo": { + "id": "volcengine/doubao-seed-2.1-turbo", + "name": "Seed 2.1 Turbo", + "description": "Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-23", + "last_updated": "2026-06-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.3536, + "output": 1.7696, + "cache_read": 0.068, + "cache_write": 0.0019 + } + }, + "volcengine/doubao-seed-2.1-pro": { + "id": "volcengine/doubao-seed-2.1-pro", + "name": "Seed 2.1 Pro", + "description": "Flagship ByteDance Seed 2.1 model for complex multimodal reasoning, coding, and agents", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-23", + "last_updated": "2026-06-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.7072, + "output": 3.536, + "cache_read": 0.1416, + "cache_write": 0.002 + } + }, + "volcengine/doubao-seed-character": { + "id": "volcengine/doubao-seed-character", + "name": "Seed Character", + "description": "ByteDance Seed model optimized for character-driven dialogue and consistent conversational behavior", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-23", + "last_updated": "2026-06-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.177, + "output": 0.884, + "cache_read": 0.024, + "cache_write": 0.0025 + } + }, + "volcengine/doubao-seed-evolving": { + "id": "volcengine/doubao-seed-evolving", + "name": "Seed Evolving", + "description": "Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-23", + "last_updated": "2026-06-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.884, + "output": 4.42, + "cache_read": 0.177, + "cache_write": 0.0025 + } + }, + "volcengine/doubao-seed-2.0-code": { + "id": "volcengine/doubao-seed-2.0-code", + "name": "Seed 2.0 Code", + "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.67, + "output": 3.36, + "cache_read": 0.14, + "cache_write": 0.0024 + } + }, + "google/gemini-3.6-flash": { + "id": "google/gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google", + "api": "https://api.ofox.ai/gemini/v1beta" + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.0415, + "input_audio": 1.5 + } + }, + "google/gemini-3.5-flash-lite": { + "id": "google/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google", + "api": "https://api.ofox.ai/gemini/v1beta" + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "cache_write": 0.083 + } + }, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google", + "api": "https://api.ofox.ai/gemini/v1beta" + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 4.5 + } + }, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google", + "api": "https://api.ofox.ai/gemini/v1beta" + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 0.083, + "input_audio": 3 + } + }, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "cache_write": 4.5 + } + }, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "cache_write": 1, + "input_audio": 1 + } + }, + "google/gemini-3.7-flash": { + "id": "google/gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google", + "api": "https://api.ofox.ai/gemini/v1beta" + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.0415, + "input_audio": 1.5 + } + }, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 1 + } + }, + "google/gemini-3.8-flash": { + "id": "google/gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google", + "api": "https://api.ofox.ai/gemini/v1beta" + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.0415, + "input_audio": 0.75 + } + }, + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 512, + "max": 24576 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 1, + "input_audio": 0.3 + } + }, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 1, + "input_audio": 0.5 + } + }, + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "moonshotai/kimi-k2.7-code-highspeed": { + "id": "moonshotai/kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.9, + "output": 8, + "cache_read": 0.38 + } + }, + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } + }, + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.1 + } + }, + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + }, + "minimax/minimax-m2.5-lightning": { + "id": "minimax/minimax-m2.5-lightning", + "name": "MiniMax-M2.5 Lightning", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-13", + "last_updated": "2026-02-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 2.4, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "minimax/minimax-m2.1-lightning": { + "id": "minimax/minimax-m2.1-lightning", + "name": "MiniMax-M2.1 Lightning", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 2.4, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 512000 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 + } + }, + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "minimax/m2-her": { + "id": "minimax/m2-her", + "name": "MiniMax-M2 Her", + "description": "MiniMax M2 variant tuned for conversational and character-driven agent interactions", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-01-23", + "last_updated": "2026-01-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 2048 + }, + "cost": { + "input": 0.3, + "output": 1.2 + } + }, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "qwen/qwen-flash": { + "id": "qwen/qwen-flash", + "name": "Qwen Flash", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 32000 + }, + "cost": { + "input": 0.022, + "output": 0.22, + "cache_read": 0.0043, + "cache_write": 0.027 + } + }, + "qwen/qwen3.5-flash": { + "id": "qwen/qwen3.5-flash", + "name": "Qwen3.5 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0.125 + } + }, + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1064000, + "output": 64000 + }, + "cost": { + "input": 1.71, + "output": 5.14, + "cache_read": 0.17, + "cache_write": 2.14 + } + }, + "qwen/qwen3.8-27b": { + "id": "qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1131072, + "output": 131072 + }, + "cost": { + "input": 0.5, + "output": 1.71, + "cache_read": 0.043, + "cache_write": 0.63 + } + }, + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.29, + "output": 2.05, + "cache_read": 0.29 + } + }, + "qwen/qwen-max": { + "id": "qwen/qwen-max", + "name": "Qwen Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 8000 + }, + "cost": { + "input": 0.35, + "output": 1.38, + "cache_read": 0.069 + } + }, + "qwen/qwen3.8-max": { + "id": "qwen/qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.71, + "output": 5.14, + "cache_read": 0.17, + "cache_write": 2.14 + } + }, + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.2, + "output": 1.5 + } + }, + "qwen/qwen3.5-plus": { + "id": "qwen/qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.4, + "output": 2.4, + "cache_read": 0.04, + "cache_write": 0.4 + } + }, + "qwen/qwen-plus": { + "id": "qwen/qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 32000 + }, + "cost": { + "input": 0.12, + "output": 0.29, + "cache_read": 0.023 + } + }, + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.36, + "output": 1.43, + "cache_read": 0.072 + } + }, + "qwen/qwen-turbo": { + "id": "qwen/qwen-turbo", + "name": "Qwen Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-11-01", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16000 + }, + "cost": { + "input": 0.043, + "output": 0.09, + "cache_read": 0.0086 + } + }, + "qwen/qwen-vl-max": { + "id": "qwen/qwen-vl-max", + "name": "Qwen-VL Max", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-08", + "last_updated": "2025-08-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 8000 + }, + "cost": { + "input": 0.23, + "output": 0.58, + "cache_read": 0.023 + } + }, + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.29, + "output": 2.29, + "cache_read": 0.29 + } + }, + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 0.31 + } + }, + "qwen/qwen3-coder-flash": { + "id": "qwen/qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.5, + "output": 2.5, + "cache_read": 0.06, + "cache_write": 0.27 + } + }, + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 1.8, + "output": 9, + "cache_read": 0.2, + "cache_write": 1 + } + }, + "qwen/qwen3.8-flash": { + "id": "qwen/qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.11, + "output": 0.39, + "cache_read": 0.011, + "cache_write": 0.14 + } + }, + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.29, + "output": 1.83, + "cache_read": 0.29 + } + }, + "qwen/qwen3.6-max-preview": { + "id": "qwen/qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 2.15, + "output": 12.86, + "cache_read": 0.2, + "cache_write": 1.17 + } + }, + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.55, + "output": 3.5, + "cache_read": 0.55 + } + }, + "qwen/qwen3.6-27b": { + "id": "qwen/qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0.43, + "output": 2.57 + } + }, + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625 + } + }, + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1064000, + "output": 64000 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "cache_write": 0.5 + } + }, + "qwen/qwen3.8-max-0902": { + "id": "qwen/qwen3.8-max-0902", + "name": "Qwen3.8 Max 0902", + "description": "2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 262144 + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.71, + "output": 5.14, + "cache_read": 0.17, + "cache_write": 2.14 + } + }, + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2 + } + }, + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 24, + "output": 144 + } + }, + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 0.16, + "output": 1, + "cache_read": 0.016 + } + }, + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 1.4, + "output": 11.2, + "cache_read": 0.144 + } + }, + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 1 + } + }, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 32768 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 0.2, + "output": 1.6, + "cache_read": 0.024 + } + }, + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 1.4, + "output": 11.2, + "cache_read": 0.144 + } + }, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 0.04, + "output": 0.32, + "cache_read": 0.008 + } + }, + "openai/gpt-6-astra": { + "id": "openai/gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 8, + "output": 40, + "cache_read": 0.8, + "cache_write": 10 + } + }, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 1, + "output": 8, + "cache_read": 0.104 + } + }, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 0.12, + "output": 0.48, + "cache_read": 0.06 + } + }, + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 0.6, + "output": 3.6, + "cache_read": 0.06 + } + }, + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 + } + }, + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 1, + "output": 8, + "cache_read": 0.104 + } + }, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 1.4, + "output": 11.2, + "cache_read": 0.144 + } + }, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 4, + "output": 24, + "cache_read": 0.4 + } + }, + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 1.6, + "output": 6.4, + "cache_read": 0.4 + } + }, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 0.32, + "output": 1.28, + "cache_read": 0.08 + } + }, + "openai/gpt-6-luna": { + "id": "openai/gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 0.08, + "output": 0.4, + "cache_read": 0.008, + "cache_write": 0.1 + } + }, + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 0.2, + "output": 1.6, + "cache_read": 0.024 + } + }, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 1, + "output": 8, + "cache_read": 0.104 + } + }, + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125 + } + }, + "openai/gpt-6-sol": { + "id": "openai/gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://api.ofox.ai/v1" + }, + "cost": { + "input": 1.6, + "output": 8, + "cache_read": 0.16, + "cache_write": 2 + } + } + } + }, + "neon": { + "id": "neon", + "env": [ + "NEON_AI_GATEWAY_BASE_URL", + "NEON_AI_GATEWAY_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "${NEON_AI_GATEWAY_BASE_URL}/v1", + "name": "Neon", + "doc": "https://neon.com/docs", + "models": { + "gpt-5-4-mini": { + "id": "gpt-5-4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "gpt-5-6-luna": { + "id": "gpt-5-6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } + } + }, + "glm-5-3-flash": { + "id": "glm-5-3-flash", + "name": "GLM-5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 + } + }, + "gpt-5-6-terra": { + "id": "gpt-5-6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "llama-4-maverick": { + "id": "llama-4-maverick", + "name": "Llama 4 Maverick 17B Instruct", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 8192 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 + } + }, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "claude-fable-5-1": { + "id": "claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "gpt-5-4-nano": { + "id": "gpt-5-4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "gpt-5-6-sol": { + "id": "gpt-5-6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "gpt-5-5-pro": { + "id": "gpt-5-5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } + } + }, + "gemini-3-1-pro": { + "id": "gemini-3-1-pro", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "glm-5-2": { + "id": "glm-5-2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } + } + }, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 10000 + }, + "cost": { + "input": 0.15, + "output": 1.2 + } + }, + "gemini-3-flash": { + "id": "gemini-3-flash", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 + } + }, + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 25000 + }, + "cost": { + "input": 0.07, + "output": 0.3 + } + }, + "gpt-5-4": { + "id": "gpt-5-4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } + } + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "inkling": { + "id": "inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1, + "output": 4.05, + "cache_read": 0.17 + } + }, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gemini-3-1-flash-lite": { + "id": "gemini-3-1-flash-lite", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 + } + }, + "gpt-5-5": { + "id": "gpt-5-5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "grok-4-6": { + "id": "grok-4-6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 524288 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "qwen35-122b-a10b": { + "id": "qwen35-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 25000 + }, + "cost": { + "input": 0.22, + "output": 2.2 + } + }, + "gemini-3-6-flash": { + "id": "gemini-3-6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 7.5, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "meta-llama-3-3-70b-instruct": { + "id": "meta-llama-3-3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "gpt-5-2": { + "id": "gpt-5-2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "gpt-5-3-codex": { + "id": "gpt-5-3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "gemma-3-12b": { + "id": "gemma-3-12b", + "name": "Gemma 3 12B", + "description": "Google's open-weight Gemma 3 vision-language model for text and image understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08-31", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.5 + } + }, + "gemini-3-5-flash-lite": { + "id": "gemini-3-5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 + } + }, + "gemini-3-5-flash": { + "id": "gemini-3-5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "gpt-5-1": { + "id": "gpt-5-1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 25000 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "meta-llama-3-1-8b-instruct": { + "id": "meta-llama-3-1-8b-instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Meta's compact open-weight Llama 3.1 model for fast, low-cost text generation", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12-31", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.45 + } + } + } + }, + "aihubmix": { + "id": "aihubmix", + "env": [ + "AIHUBMIX_API_KEY" + ], + "npm": "@aihubmix/ai-sdk-provider", + "name": "AIHubMix", + "doc": "https://docs.aihubmix.com", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } + } + }, + "coding-minimax-m2.7-free": { + "id": "coding-minimax-m2.7-free", + "name": "Coding MiniMax M2.7 (Free)", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 128100 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 991000, + "output": 64000 + }, + "cost": { + "input": 1.69, + "output": 5.07, + "cache_read": 0.169, + "cache_write": 2.1125 + } + }, + "muse-spark-1.3": { + "id": "muse-spark-1.3", + "name": "Muse Spark 1.3", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.375, + "output": 4.675, + "cache_read": 0.165 + } + }, + "deep-deepseek-v4-pro": { + "id": "deep-deepseek-v4-pro", + "name": "DeepSeek V4 Pro (DeepSeek)", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.478, + "output": 0.956, + "cache_read": 0.004302 + } + }, + "qwen3.8-2.4t-a95b": { + "id": "qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 262000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM 5 Vision Turbo", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glmv", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-09", + "last_updated": "2026-05-09", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.7042, + "output": 3.09848, + "cache_read": 0.169008 + } + }, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0.11268, + "output": 0.39438, + "cache_read": 0.02817 + } + }, + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.142, + "output": 0.284, + "cache_read": 0.0284 + } + }, + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 991000, + "output": 128000 + }, + "cost": { + "input": 1.69, + "output": 5.07, + "cache_read": 0.169, + "cache_write": 2.1125 + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "doubao-seed-2-0-lite-260428": { + "id": "doubao-seed-2-0-lite-260428", + "name": "Doubao Seed 2.0 Lite 260428", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-28", + "last_updated": "2026-04-28", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.08, + "output": 0.51, + "cache_read": 0.01692, + "input_audio": 1.269, + "tiers": [ + { + "input": 0.13, + "output": 0.76, + "cache_read": 0.02536, + "input_audio": 1.902, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.25, + "output": 1.52, + "cache_read": 0.05072, + "input_audio": 3.804, + "tier": { + "type": "context", + "size": 128000 + } + } + ] + } + }, + "claude-sonnet-4-6-think": { + "id": "claude-sonnet-4-6-think", + "name": "Claude Sonnet 4.6 Thinking", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } + } + }, + "deepseek-v4-flash-0731-fast": { + "id": "deepseek-v4-flash-0731-fast", + "name": "DeepSeek V4 Flash 0731 Fast", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.28, + "output": 1.4, + "cache_read": 0.07 + } + }, + "hy3-preview": { + "id": "hy3-preview", + "name": "Hy3 Preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-20", + "last_updated": "2026-04-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.17, + "output": 0.566661, + "cache_read": 0.051 + } + }, + "alicloud-deepseek-v4-pro": { + "id": "alicloud-deepseek-v4-pro", + "name": "DeepSeek V4 Pro (Alibaba Cloud)", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 1.69, + "output": 3.38, + "cache_read": 0.13 + } + }, + "zai-glm-5.1": { + "id": "zai-glm-5.1", + "name": "GLM-5.1 (Z.ai)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-27", + "last_updated": "2026-03-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.845, + "output": 3.38, + "cache_read": 0.183112 + } + }, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.155, + "output": 0.62, + "cache_read": 0.0031 + } + }, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "coding-minimax-m2.7-highspeed": { + "id": "coding-minimax-m2.7-highspeed", + "name": "Coding MiniMax M2.7 Highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 128100 + }, + "cost": { + "input": 0.2, + "output": 0.2 + } + }, + "muse-spark-1.1": { + "id": "muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-08", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.375, + "output": 4.675 + } + }, + "deepseek-v4-flash-vision-exp": { + "id": "deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.155, + "output": 0.62, + "cache_read": 0.0031 + } + }, + "kimi-k2.7-code-highspeed": { + "id": "kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 1.9, + "output": 7.999, + "cache_read": 0.32167 + } + }, + "qwen3.7-flash": { + "id": "qwen3.7-flash", + "name": "Qwen3.7 Flash", + "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 991000, + "input": 991000, + "output": 64000 + }, + "cost": { + "input": 0.0282, + "output": 0.1128, + "cache_read": 0.00564, + "cache_write": 0.03525 + } + }, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } + }, + "claude-opus-4-8-think": { + "id": "claude-opus-4-8-think", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "doubao-seed-2-0-pro": { + "id": "doubao-seed-2-0-pro", + "name": "Doubao Seed 2.0 Pro", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.48, + "output": 2.41, + "cache_read": 0.09644, + "tiers": [ + { + "input": 0.72, + "output": 3.62, + "cache_read": 0.144656, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 1.45, + "output": 7.23, + "cache_read": 0.28932, + "tier": { + "type": "context", + "size": 128000 + } + } + ] + } + }, + "xiaomi-mimo-v2.5": { + "id": "xiaomi-mimo-v2.5", + "name": "Xiaomi MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo-v2.5", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.44, + "output": 2.2, + "cache_read": 0.088, + "tiers": [ + { + "input": 0.88, + "output": 4.4, + "cache_read": 0.176, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.88, + "output": 4.4, + "cache_read": 0.176 + } + } + }, + "claude-fable-5-1": { + "id": "claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 11, + "output": 55, + "cache_read": 0.275, + "cache_write": 13.75 + } + }, + "gemini-3.6-flash": { + "id": "gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 7.5, + "cache_read": 0.15 + } + }, + "longcat-2.0": { + "id": "longcat-2.0", + "name": "LongCat-2.0", + "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", + "family": "longcat", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.7746, + "output": 3.0984, + "cache_read": 0.015492 + } + }, + "doubao-seed-2-0-code-preview": { + "id": "doubao-seed-2-0-code-preview", + "name": "Doubao Seed 2.0 Code Preview", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.48, + "output": 2.41, + "cache_read": 0.09644, + "tiers": [ + { + "input": 0.72, + "output": 3.62, + "cache_read": 0.144656, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 1.45, + "output": 7.23, + "cache_read": 0.28932, + "tier": { + "type": "context", + "size": 128000 + } + } + ] + } + }, + "grok-4.3": { + "id": "grok-4.3", + "name": "Grok 4.3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "qwen3.8-omni-flash": { + "id": "qwen3.8-omni-flash", + "name": "Qwen3.8 Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "release_date": "2026-09-17", + "last_updated": "2026-09-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.1126, + "output": 0.380025, + "cache_read": 0.014075, + "cache_write": 0.175937 + } + }, + "doubao-seed-2-0-mini-260428": { + "id": "doubao-seed-2-0-mini-260428", + "name": "Doubao Seed 2.0 Mini 260428", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-28", + "last_updated": "2026-04-28", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.03, + "output": 0.28, + "cache_read": 0.00564, + "input_audio": 0.423, + "tiers": [ + { + "input": 0.06, + "output": 0.56, + "cache_read": 0.01128, + "input_audio": 0.846, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.11, + "output": 1.13, + "cache_read": 0.02256, + "input_audio": 1.692, + "tier": { + "type": "context", + "size": 128000 + } + } + ] + } + }, + "xiaomi-mimo-v2.5-pro-free": { + "id": "xiaomi-mimo-v2.5-pro-free", + "name": "Xiaomi MiMo-V2.5-Pro (free)", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo-v2.5-pro", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.499999, + "cache_read": 0.03 + } + }, + "gemini-3.1-flash-image": { + "id": "gemini-3.1-flash-image", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], + "tool_call": false, + "interleaved": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 3 + } + }, + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } + } + }, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.1 + } + }, + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.13 + } + }, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "coding-minimax-m2.7": { + "id": "coding-minimax-m2.7", + "name": "Coding MiniMax M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 128100 + }, + "cost": { + "input": 0.2, + "output": 0.2 + } + }, + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 128000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 1.5 + } + }, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-05", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } + } + }, + "deep-deepseek-v4-flash": { + "id": "deep-deepseek-v4-flash", + "name": "DeepSeek V4 Flash (DeepSeek)", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.154, + "output": 0.308, + "cache_read": 0.0308 + } + }, + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 991000, + "output": 64000 + }, + "cost": { + "input": 0.17, + "output": 1.01, + "cache_read": 0.0169, + "cache_write": 0.21125, + "tiers": [ + { + "input": 0.68, + "output": 4.06, + "cache_read": 0.0676, + "cache_write": 0.845, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.68, + "output": 4.06, + "cache_read": 0.0676, + "cache_write": 0.845 + } + } + }, + "coding-glm-5.1": { + "id": "coding-glm-5.1", + "name": "Coding GLM 5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-11", + "last_updated": "2026-04-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.06, + "output": 0.22, + "cache_read": 0.013 + } + }, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-05", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 + } + }, + "xiaomi-mimo-v2.5-free": { + "id": "xiaomi-mimo-v2.5-free", + "name": "Xiaomi MiMo-V2.5 (free)", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo-v2.5", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 11, + "output": 55, + "cache_read": 1.1, + "cache_write": 13.75 + } + }, + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.6918, + "output": 2.0754, + "cache_read": 0.023058 + } + }, + "muse-spark-1.2": { + "id": "muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-05", + "last_updated": "2026-08-05", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.375, + "output": 4.675 + } + }, + "gemini-3.1-flash-lite-image": { + "id": "gemini-3.1-flash-lite-image", + "name": "Nano Banana 2 Lite", + "description": "Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 4096 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025 + } + }, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 512000 + }, + "cost": { + "input": 0.288, + "output": 1.152 + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 1, + "output": 6, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.1126, + "output": 0.380025, + "cache_read": 0.014075, + "cache_write": 0.175937 + } + }, + "claude-opus-4-6-think": { + "id": "claude-opus-4-6-think", + "name": "Claude Opus 4.6 Thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "coding-xiaomi-mimo-v2.5": { + "id": "coding-xiaomi-mimo-v2.5", + "name": "Coding Xiaomi MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo-v2.5", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.08, + "output": 0.4, + "cache_read": 0.016, + "tiers": [ + { + "input": 0.16, + "output": 0.8, + "cache_read": 0.032, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.16, + "output": 0.8, + "cache_read": 0.032 + } + } + }, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "ox-alpha": { + "id": "ox-alpha", + "name": "Ox Alpha", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen3.6-max-preview": { + "id": "qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "qwen3.6", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-05-09", + "last_updated": "2026-05-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 240000, + "output": 64000 + }, + "cost": { + "input": 1.27, + "output": 7.61, + "cache_read": 0.1268, + "cache_write": 1.585, + "tiers": [ + { + "input": 2.11, + "output": 12.67, + "cache_read": 0.2112, + "cache_write": 2.64, + "tier": { + "type": "context", + "size": 128000 + } + } + ] + } + }, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 + } + }, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 1.1268, + "output": 3.9438, + "cache_read": 0.2817 + } + }, + "hy3": { + "id": "hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 192000, + "output": 128000 + }, + "cost": { + "input": 0.1562, + "output": 0.6248, + "cache_read": 0.03905 + } + }, + "gemini-3.1-pro-preview-customtools": { + "id": "gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "tiers": [ + { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 0.5, + "output": 3, + "cache_read": 0.05 + } + } + }, + "grok-build-0.1": { + "id": "grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 1, + "output": 2, + "cache_read": 0.2 + } + }, + "claude-opus-4-7-think": { + "id": "claude-opus-4-7-think", + "name": "Claude Opus 4.7 Thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "alicloud-deepseek-v4-flash": { + "id": "alicloud-deepseek-v4-flash", + "name": "DeepSeek V4 Flash (Alibaba Cloud)", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 + } + }, + "coding-xiaomi-mimo-v2.5-pro": { + "id": "coding-xiaomi-mimo-v2.5-pro", + "name": "Coding Xiaomi MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo-v2.5-pro", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.2, + "output": 0.6, + "cache_read": 0.04, + "tiers": [ + { + "input": 0.4, + "output": 1.2, + "cache_read": 0.08, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.2, + "cache_read": 0.08 + } + } + }, + "gemini-3.8-flash": { + "id": "gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 + } + }, + "step-3.7-flash": { + "id": "step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 256000, + "output": 256000 + }, + "cost": { + "input": 0.22, + "output": 1.32, + "cache_read": 0.044 + } + }, + "alicloud-glm-5.1": { + "id": "alicloud-glm-5.1", + "name": "GLM-5.1 (Alibaba Cloud)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-27", + "last_updated": "2026-03-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.84, + "output": 3.38, + "cache_read": 0.169, + "cache_write": 1.05625 + } + }, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } + } + }, + "coding-glm-5.1-free": { + "id": "coding-glm-5.1-free", + "name": "Coding GLM 5.1 (free)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-11", + "last_updated": "2026-04-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-05-09", + "last_updated": "2026-05-09", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 991000, + "output": 64000 + }, + "cost": { + "input": 0.28, + "output": 1.69, + "cache_read": 0.0282, + "cache_write": 0.3525, + "tiers": [ + { + "input": 1.13, + "output": 6.77, + "cache_read": 0.1128, + "cache_write": 1.41, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.13, + "output": 6.77, + "cache_read": 0.1128, + "cache_write": 1.41 + } + } + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125 + } + }, + "hy4-preview": { + "id": "hy4-preview", + "name": "Hy4 preview", + "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-28", + "last_updated": "2026-08-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1024000, + "output": 64000 + }, + "cost": { + "input": 0.845, + "output": 2.535, + "cache_read": 0.04225 + } + }, + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 1.1268, + "output": 3.9438, + "cache_read": 0.2817 + } + }, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.95, + "output": 3.9995, + "cache_read": 0.160835 + } + }, + "grok-4.6": { + "id": "grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "xiaomi-mimo-v2.5-pro": { + "id": "xiaomi-mimo-v2.5-pro", + "name": "Xiaomi MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo-v2.5-pro", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.1, + "output": 3.3, + "cache_read": 0.22, + "tiers": [ + { + "input": 2.2, + "output": 6.6, + "cache_read": 0.44, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2.2, + "output": 6.6, + "cache_read": 0.44 + } + } + }, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 991000, + "output": 64000 + }, + "cost": { + "input": 0.282, + "output": 1.128, + "cache_read": 0.0564, + "cache_write": 0.3525 + } + }, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 1 + } + } + } + }, + "merge-gateway": { + "id": "merge-gateway", + "env": [ + "MERGE_GATEWAY_API_KEY" + ], + "npm": "merge-gateway-ai-sdk-provider", + "api": "https://api-gateway.merge.dev/v1/ai-sdk", + "name": "Merge Gateway", + "doc": "https://docs.merge.dev/merge-gateway", + "models": { + "zai/glm-4.5": { + "id": "zai/glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 98304 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 + } + }, + "zai/glm-5.3-flash": { + "id": "zai/glm-5.3-flash", + "name": "GLM-5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.015, + "output": 0.05, + "cache_read": 0.003 + } + }, + "zai/glm-4.6": { + "id": "zai/glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 + } + }, + "zai/glm-5": { + "id": "zai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0 + } + }, + "zai/glm-4.5v": { + "id": "zai/glm-4.5v", + "name": "Glm 4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32000 + }, + "cost": { + "input": 0.6, + "output": 1.8, + "cache_read": 0.11, + "cache_write": 0 + } + }, + "zai/glm-4.7-flash": { + "id": "zai/glm-4.7-flash", + "name": "GLM 4.7 Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.07, + "output": 0.4 + } + }, + "zai/glm-4.7": { + "id": "zai/glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 + } + }, + "zai/glm-5.2": { + "id": "zai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 50000 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.05, + "output": 3.3, + "cache_read": 0.195 + } + }, + "zai/glm-5.1": { + "id": "zai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 + } + }, + "zai/glm-4.7-flashx": { + "id": "zai/glm-4.7-flashx", + "name": "GLM-4.7 FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0.07, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0 + } + }, + "zai/glm-5-turbo": { + "id": "zai/glm-5-turbo", + "name": "GLM-5 Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.24, + "cache_write": 0 + } + }, + "zai/glm-4.5-air": { + "id": "zai/glm-4.5-air", + "name": "GLM-4.5 Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 98304 + }, + "cost": { + "input": 0.2, + "output": 1.1, + "cache_read": 0.03, + "cache_write": 0 + } + }, + "zai/glm-5.3": { + "id": "zai/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.7, + "output": 2.2, + "cache_read": 0.13 + } + }, + "anthropic/claude-3-7-sonnet-20250219": { + "id": "anthropic/claude-3-7-sonnet-20250219", + "name": "Claude 3.7 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10-31", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic/claude-opus-5-5": { + "id": "anthropic/claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 + } + }, + "anthropic/claude-opus-4-20250514": { + "id": "anthropic/claude-opus-4-20250514", + "name": "Claude Opus 4 (20250514)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "anthropic/claude-fable-5-1": { + "id": "anthropic/claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "anthropic/claude-opus-4-1-20250805": { + "id": "anthropic/claude-opus-4-1-20250805", + "name": "Claude Opus 4.1 (20250805)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "anthropic/claude-opus-4-5-20251101": { + "id": "anthropic/claude-opus-4-5-20251101", + "name": "Claude Opus 4.5 (20251101)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "anthropic/claude-opus-4-8": { + "id": "anthropic/claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 128000 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-sonnet-4-5-20250929": { + "id": "anthropic/claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5 (20250929)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-haiku-4-5-20251001": { + "id": "anthropic/claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5 (20251001)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic/claude-sonnet-4-20250514": { + "id": "anthropic/claude-sonnet-4-20250514", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "cohere/command-r-08-2024": { + "id": "cohere/command-r-08-2024", + "name": "Command R 08-2024", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "cohere/command-a-03-2025": { + "id": "cohere/command-a-03-2025", + "name": "Command A 03-2025", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8000 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "cohere/command-r7b-12-2024": { + "id": "cohere/command-r7b-12-2024", + "name": "Command R7B 12-2024", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2024-12-02", + "last_updated": "2024-12-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.0375, + "output": 0.15 + } + }, + "cohere/command-r-plus-08-2024": { + "id": "cohere/command-r-plus-08-2024", + "name": "Command R+ 08-2024", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "deepseek/deepseek-v4-flash-0731": { + "id": "deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.035, + "output": 0.07, + "cache_read": 0.007 + } + }, + "deepseek/deepseek-v4-flash-0423": { + "id": "deepseek/deepseek-v4-flash-0423", + "name": "DeepSeek V4 Flash 0423", + "description": "Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.139, + "output": 0.278 + } + }, + "deepseek/deepseek-v4-flash-0731-fast": { + "id": "deepseek/deepseek-v4-flash-0731-fast", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.28, + "output": 0.56, + "cache_read": 0.07 + } + }, + "deepseek/deepseek-v4.1-flash": { + "id": "deepseek/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 + } + }, + "deepseek/deepseek-v3": { + "id": "deepseek/deepseek-v3", + "name": "DeepSeek V3", + "description": "Open DeepSeek MoE chat model for coding, math, and general reasoning", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2024-12-26", + "last_updated": "2024-12-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 81920 + }, + "cost": { + "input": 0.58, + "output": 1.68 + } + }, + "deepseek/deepseek-v4-pro-0813": { + "id": "deepseek/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 1.74, + "output": 3.48, + "cache_read": 0.003625 + } + }, + "deepseek/deepseek-v4-pro-0423": { + "id": "deepseek/deepseek-v4-pro-0423", + "name": "DeepSeek V4 Pro 0423", + "description": "DeepSeek V4 Pro initial snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 393216 + }, + "cost": { + "input": 1.65, + "output": 3.3 + } + }, + "deepseek/deepseek-v3.1": { + "id": "deepseek/deepseek-v3.1", + "name": "DeepSeek V3.1", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-21", + "last_updated": "2025-08-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 164000, + "output": 41000 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "deepseek/deepseek-r1": { + "id": "deepseek/deepseek-r1", + "name": "DeepSeek R1", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 40960 + }, + "cost": { + "input": 1.35, + "output": 5.4 + } + }, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 262000 + }, + "cost": { + "input": 1.74, + "output": 3.48, + "cache_read": 0.145 + } + }, + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 40960 + }, + "cost": { + "input": 0.28, + "output": 0.45, + "cache_read": 0.14 + } + }, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.035, + "output": 0.07, + "cache_read": 0.007 + } + }, + "thinkingmachines/inkling": { + "id": "thinkingmachines/inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048000, + "output": 32000 + }, + "cost": { + "input": 1, + "output": 4.05, + "cache_read": 0.17 + } + }, + "meta/muse-spark-1.1": { + "id": "meta/muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-08", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 262144 + }, + "cost": { + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 + } + }, + "meta/llama-3.3-70b-instruct": { + "id": "meta/llama-3.3-70b-instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.22, + "output": 0.5, + "cache_read": 0.11 + } + }, + "meta/muse-spark-1.2": { + "id": "meta/muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-05", + "last_updated": "2026-08-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 262144 + }, + "cost": { + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 + } + }, + "meta/llama-3.1-8b-instruct": { + "id": "meta/llama-3.1-8b-instruct", + "name": "Llama 3.1 8B", + "description": "Compact open Llama model for lightweight chat, drafting, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 2048 + }, + "cost": { + "input": 0.22, + "output": 0.22 + } + }, + "meta/llama-3.1-70b-instruct": { + "id": "meta/llama-3.1-70b-instruct", + "name": "Llama 3.1 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 2048 + }, + "cost": { + "input": 0.99, + "output": 0.99 + } + }, + "bytedance/dola-seed-2.0-code-preview": { + "id": "bytedance/dola-seed-2.0-code-preview", + "name": "Dola Seed 2.0 Code (preview)", + "description": "Preview coding model for repository understanding, refactors, and engineering tasks", + "family": "seed", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-03-28", + "last_updated": "2026-03-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 3 + } + }, + "bytedance/dola-seed-2.0-code": { + "id": "bytedance/dola-seed-2.0-code", + "name": "Seed 2.0 Code", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "seed", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.4, + "output": 2.4 + } + }, + "bytedance/dola-seed-2.0-mini": { + "id": "bytedance/dola-seed-2.0-mini", + "name": "Seed 2.0 Mini", + "description": "Low-cost Seed model for general chat, extraction, and lightweight production tasks", + "family": "seed", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.4 + } + }, + "bytedance/dola-seed-2.0-lite": { + "id": "bytedance/dola-seed-2.0-lite", + "name": "Seed 2.0 Lite", + "description": "Efficient Seed model for general chat, analysis, and lightweight production tasks", + "family": "seed", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-28", + "last_updated": "2026-02-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.25, + "output": 2 + } + }, + "bytedance/dola-seed-2.0-pro": { + "id": "bytedance/dola-seed-2.0-pro", + "name": "Seed 2.0 Pro", + "description": "Higher-capability Seed model for complex chat, analysis, and production tasks", + "family": "seed", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-03-28", + "last_updated": "2026-03-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 3 + } + }, + "google/gemini-flash-latest": { + "id": "google/gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B It", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.14, + "output": 0.4 + } + }, + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma 3 27B IT", + "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.08, + "output": 0.45, + "cache_read": 0.04 + } + }, + "google/gemini-3-pro-preview": { + "id": "google/gemini-3-pro-preview", + "name": "Gemini 3 Pro Preview", + "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "google/gemini-2.5-flash-image": { + "id": "google/gemini-2.5-flash-image", + "name": "Gemini 2.5 Flash Image", + "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 2.5 + } + }, + "google/gemini-flash-lite-latest": { + "id": "google/gemini-flash-lite-latest", + "name": "Gemini Flash-Lite Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 + } + }, + "google/gemini-3.6-flash": { + "id": "google/gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 7.5, + "cache_read": 0.15 + } + }, + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.13, + "output": 0.4 + } + }, + "google/gemini-3.5-flash-lite": { + "id": "google/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash-Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 + } + }, + "google/gemini-3.1-flash-image": { + "id": "google/gemini-3.1-flash-image", + "name": "Gemini 3.1 Flash Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 3 + } + }, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "google/gemma-3-12b-it": { + "id": "google/gemma-3-12b-it", + "name": "Gemma 3 12B", + "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.09, + "output": 0.29 + } + }, + "google/gemini-2.5-computer-use-preview-10-2025": { + "id": "google/gemini-2.5-computer-use-preview-10-2025", + "name": "Gemini 2.5 Computer Use Preview (10-2025)", + "description": "Specialized Gemini 2.5 model for browser-control agents that automate UI tasks", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-10-07", + "last_updated": "2025-10-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 64000 + }, + "cost": { + "input": 1.25, + "output": 10 + } + }, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } + } + }, + "google/gemini-3-pro-image": { + "id": "google/gemini-3-pro-image", + "name": "Gemini 3 Pro Image", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12 + } + }, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 + } + }, + "google/gemini-3.7-flash": { + "id": "google/gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 + } + }, + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 + } + }, + "google/gemini-3.8-flash": { + "id": "google/gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 + } + }, + "google/gemma-3-4b-it": { + "id": "google/gemma-3-4b-it", + "name": "Gemma 3 4B", + "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.04, + "output": 0.08 + } + }, + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 512, + "max": 24576 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "input_audio": 0.3 + } + }, + "google/gemini-embedding-001": { + "id": "google/gemini-embedding-001", + "name": "Gemini Embedding 001", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "gemini", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": false, + "knowledge": "2025-05", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2048, + "output": 4096 + }, + "cost": { + "input": 0.15, + "output": 0 + } + }, + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 + } + }, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash-Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 + } + }, + "writer/palmyra-x4": { + "id": "writer/palmyra-x4", + "name": "Palmyra X4", + "description": "Enterprise language model for writing, analysis, and tool-assisted workflows", + "family": "palmyra", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2024-10-09", + "last_updated": "2024-10-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 32000 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "writer/palmyra-x5": { + "id": "writer/palmyra-x5", + "name": "Palmyra X5", + "description": "Enterprise multimodal model for writing, analysis, and tool-assisted workflows", + "family": "palmyra", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-04-28", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 250000 + }, + "cost": { + "input": 0.6, + "output": 6 + } + }, + "xai/grok-4.7": { + "id": "xai/grok-4.7", + "name": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "xai/grok-4.3": { + "id": "xai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "xai/grok-4.20-0309-reasoning": { + "id": "xai/grok-4.20-0309-reasoning", + "name": "Grok 4.20", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "xai/grok-4.5": { + "id": "xai/grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "xai/grok-4.20-0309-non-reasoning": { + "id": "xai/grok-4.20-0309-non-reasoning", + "name": "Grok 4.20 Non-Reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 + } + }, + "xai/grok-build-0.1": { + "id": "xai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 1, + "output": 2, + "cache_read": 0.2 + } + }, + "xai/grok-4.6": { + "id": "xai/grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "sakana/fugu-ultra": { + "id": "sakana/fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + } + }, + "sakana/sakana-namazu": { + "id": "sakana/sakana-namazu", + "name": "Sakana Namazu", + "description": "Japanese-specialized reasoning model based on Kimi K2.6 and tuned for Japanese language, culture, and business workflows", + "family": "sakana-namazu", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.15 + } + }, + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 32768 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.6, + "output": 2.5 + } + }, + "nvidia/nemotron-3.5-lightning-30b-a3b": { + "id": "nvidia/nemotron-3.5-lightning-30b-a3b", + "name": "Nemotron 3.5 Lightning 30B A3B", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nemotron-nano-9b-v2": { + "id": "nvidia/nemotron-nano-9b-v2", + "name": "Nemotron Nano 9B", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-18", + "last_updated": "2025-08-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.06, + "output": 0.23 + } + }, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 128000 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax M2.7 Highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 8192 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "minimax/minimax-m2.5-highspeed": { + "id": "minimax/minimax-m2.5-highspeed", + "name": "MiniMax M2.5 Highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-13", + "last_updated": "2026-02-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 8192 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "mistral/codestral-latest": { + "id": "mistral/codestral-latest", + "name": "Codestral (latest)", + "description": "Mistral code model for completions, refactors, and developer IDE workflows", + "family": "codestral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-05-29", + "last_updated": "2025-01-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 4096 + }, + "cost": { + "input": 0.3, + "output": 0.9 + } + }, + "mistral/mistral-large-2411": { + "id": "mistral/mistral-large-2411", + "name": "Mistral Large 2.1", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-18", + "last_updated": "2024-11-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "mistral/mistral-large-latest": { + "id": "mistral/mistral-large-latest", + "name": "Mistral Large (latest)", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "mistral/mistral-small-latest": { + "id": "mistral/mistral-small-latest", + "name": "Mistral Small (latest)", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "mistral/devstral-medium-latest": { + "id": "mistral/devstral-medium-latest", + "name": "Devstral 2 (latest)", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2 + } + }, + "mistral/devstral-2512": { + "id": "mistral/devstral-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.4, + "output": 2, + "cache_read": 0.04 + } + }, + "mistral/mistral-medium-2505": { + "id": "mistral/mistral-medium-2505", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.4, + "output": 2, + "cache_read": 0.04 + } + }, + "mistral/magistral-medium-latest": { + "id": "mistral/magistral-medium-latest", + "name": "Magistral Medium (latest)", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-medium", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2, + "output": 5 + } + }, + "mistral/devstral-medium-2507": { + "id": "mistral/devstral-medium-2507", + "name": "Devstral Medium", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-07-10", + "last_updated": "2025-07-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2 + } + }, + "mistral/mistral-large-2512": { + "id": "mistral/mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.05 + } + }, + "mistral/devstral-small-2507": { + "id": "mistral/devstral-small-2507", + "name": "Devstral Small", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-07-10", + "last_updated": "2025-07-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "mistral/pixtral-large-latest": { + "id": "mistral/pixtral-large-latest", + "name": "Pixtral Large (latest)", + "description": "Mistral's larger vision model for document-heavy image understanding and chat", + "family": "pixtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2024-11-04", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "mistral/mistral-medium-latest": { + "id": "mistral/mistral-medium-latest", + "name": "Mistral Medium (latest)", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.4, + "output": 2 + } + }, + "qwen/qwen-flash": { + "id": "qwen/qwen-flash", + "name": "Qwen Flash", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 250000 + }, + "cost": { + "input": 0.022, + "output": 0.216, + "cache_read": 0.0044 + } + }, + "qwen/qwen3.5-flash": { + "id": "qwen/qwen3.5-flash", + "name": "Qwen3.5 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 250000 + }, + "cost": { + "input": 0.029, + "output": 0.287, + "cache_read": 0.0058 + } + }, + "qwen/qwen3-vl-235b-a22b-instruct": { + "id": "qwen/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3-VL 235B A22B Instruct", + "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.287, + "output": 1.147, + "cache_read": 0.15785 + } + }, + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 250000 + }, + "cost": { + "input": 0.825, + "output": 2.4755, + "cache_read": 0.165 + } + }, + "qwen/qwen3-vl-235b-a22b-thinking": { + "id": "qwen/qwen3-vl-235b-a22b-thinking", + "name": "Qwen3-VL 235B A22B Thinking", + "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.287, + "output": 2.867, + "cache_read": 0.0574 + } + }, + "qwen/qwen3.8-2.4t-a95b": { + "id": "qwen/qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 1010000 + }, + "cost": { + "input": 2.5, + "output": 6.25, + "cache_read": 0.5 + } + }, + "qwen/qwen3-vl-plus": { + "id": "qwen/qwen3-vl-plus", + "name": "Qwen3-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.143, + "output": 1.434, + "cache_read": 0.0286 + } + }, + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.086, + "output": 0.688, + "cache_read": 0.0172 + } + }, + "qwen/qwen3-235b-a22b-instruct-2507": { + "id": "qwen/qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-07-21", + "last_updated": "2025-07-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.6, + "cache_read": 0.05 + } + }, + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 1.2 + } + }, + "qwen/qwen3.8-max": { + "id": "qwen/qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25 + } + }, + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.15, + "output": 0.8, + "cache_read": 0.075 + } + }, + "qwen/qwen3.5-plus": { + "id": "qwen/qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 250000 + }, + "cost": { + "input": 0.115, + "output": 0.688, + "cache_read": 0.023 + } + }, + "qwen/qwen3-32b": { + "id": "qwen/qwen3-32b", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "qwen/qwen-plus": { + "id": "qwen/qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 250000 + }, + "cost": { + "input": 0.115, + "output": 0.287, + "cache_read": 0.023 + } + }, + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.359, + "output": 1.434, + "cache_read": 0.0718 + } + }, + "qwen/qwen3-coder-480b-a35b-instruct": { + "id": "qwen/qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.22, + "output": 1.8 + } + }, + "qwen/qwen3-235b-a22b": { + "id": "qwen/qwen3-235b-a22b", + "name": "Qwen3 235B A22B", + "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.287, + "output": 1.147, + "cache_read": 0.0574 + } + }, + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.115, + "output": 0.917, + "cache_read": 0.023 + } + }, + "qwen/qwen3.6-35b-a3b": { + "id": "qwen/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.248, + "output": 1.485, + "cache_read": 0.0496 + } + }, + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.144, + "output": 0.574, + "cache_read": 0.0288 + } + }, + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 250000 + }, + "cost": { + "input": 0.165, + "output": 0.99, + "cache_read": 0.033 + } + }, + "qwen/qwen3-coder-flash": { + "id": "qwen/qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 250000 + }, + "cost": { + "input": 0.144, + "output": 0.574, + "cache_read": 0.0288 + } + }, + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 250000 + }, + "cost": { + "input": 0.574, + "output": 2.294, + "cache_read": 0.1148 + } + }, + "qwen/qwen3.5-9b": { + "id": "qwen/qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.09, + "output": 0.13 + } + }, + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5 35B A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.057, + "output": 0.459, + "cache_read": 0.020357 + } + }, + "qwen/qwen3.6-max-preview": { + "id": "qwen/qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 65536 + }, + "cost": { + "input": 1.31, + "output": 7.88 + } + }, + "qwen/qwen3-30b-a3b": { + "id": "qwen/qwen3-30b-a3b", + "name": "Qwen3 30B A3B", + "description": "Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-04-28", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.108, + "output": 1.076, + "cache_read": 0.0216 + } + }, + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.172, + "output": 1.032, + "cache_read": 0.0344 + } + }, + "qwen/qwen3.6-27b": { + "id": "qwen/qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.289, + "output": 2.4 + } + }, + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 250000 + }, + "cost": { + "input": 0.276, + "output": 1.651, + "cache_read": 0.0552 + } + }, + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.4, + "output": 1.6 + } + }, + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } + } + }, + "openai/gpt-oss-safeguard-120b": { + "id": "openai/gpt-oss-safeguard-120b", + "name": "GPT OSS Safeguard 120B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 4096, + "output": 4096 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16385, + "output": 4096 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 Nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "openai/gpt-4o-2024-05-13": { + "id": "openai/gpt-4o-2024-05-13", + "name": "GPT-4o (2024-05-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 5, + "output": 15 + } + }, + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4 Mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 + } + }, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3 Mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 + } + }, + "openai/gpt-4": { + "id": "openai/gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 30, + "output": 60 + } + }, + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 Nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 + } + }, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 + } + }, + "openai/gpt-5.2-chat-latest": { + "id": "openai/gpt-5.2-chat-latest", + "name": "GPT-5.2 Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "openai/gpt-5.1-chat-latest": { + "id": "openai/gpt-5.1-chat-latest", + "name": "GPT-5.1 Chat Latest", + "description": "Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations", + "family": "gpt-codex", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/o1": { + "id": "openai/o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 15, + "output": 60, + "cache_read": 7.5 + } + }, + "openai/gpt-5-chat-latest": { + "id": "openai/gpt-5-chat-latest", + "name": "GPT-5 Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/gpt-4o-2024-08-06": { + "id": "openai/gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-08-06", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "openai/gpt-6-astra": { + "id": "openai/gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o Mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 + } + }, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT-OSS 20B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.04, + "output": 0.2, + "cache_read": 0.02 + } + }, + "openai/gpt-oss-safeguard-20b": { + "id": "openai/gpt-oss-safeguard-20b", + "name": "GPT OSS Safeguard 20B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 4096, + "output": 4096 + }, + "cost": { + "input": 0.07, + "output": 0.2, + "cache_read": 0, + "cache_write": 0 + } + }, + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 Mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 + } + }, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "openai/gpt-4o-2024-11-20": { + "id": "openai/gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 Mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 + } + }, + "openai/gpt-6-luna": { + "id": "openai/gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125 + } + }, + "openai/gpt-5.3-chat-latest": { + "id": "openai/gpt-5.3-chat-latest", + "name": "GPT-5.3 Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 10, + "output": 30 + } + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT-OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.09, + "output": 0.36 + } + }, + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 5 + } + }, + "openai/gpt-6-sol": { + "id": "openai/gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "moonshot/kimi-k3": { + "id": "moonshot/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 2.9, + "output": 14, + "cache_read": 0.3 + } + }, + "moonshot/kimi-k2.7-code-highspeed": { + "id": "moonshot/kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 32768 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 1.9, + "output": 8, + "cache_read": 0.38 + } + }, + "moonshot/kimi-k2.6": { + "id": "moonshot/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } + }, + "moonshot/kimi-k2.5": { + "id": "moonshot/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.1 + } + }, + "moonshot/kimi-k2.7-code": { + "id": "moonshot/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 32768 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + } + } + }, + "opper": { + "id": "opper", + "env": [ + "OPPER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.opper.ai/v3/compat", + "name": "Opper", + "doc": "https://opper.ai/models", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } + } + }, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1.1, + "output": 5.5, + "cache_read": 0.11, + "cache_write": 1.375 + } + }, + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } + } + }, + "gemma-4-31b-it": { + "id": "gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.46488, + "output": 2.44062 + } + }, + "muse-spark-1.3": { + "id": "muse-spark-1.3", + "name": "Muse Spark 1.3", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 + } + }, + "qwen3.8-27b": { + "id": "qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.5811, + "output": 3 + } + }, + "gpt-5.5-pro": { + "id": "gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180 + } + }, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "qwen3.8-2.4t-a95b": { + "id": "qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 2.5, + "output": 6, + "cache_read": 0.63 + } + }, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.07 + } + }, + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 983616, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25 + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15 + } + }, + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.5811, + "output": 2.3244 + } + }, + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125, + "tiers": [ + { + "input": 6.6, + "output": 24.75, + "cache_read": 0.66, + "cache_write": 8.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6.6, + "output": 24.75, + "cache_read": 0.66, + "cache_write": 8.25 + } + } + }, + "claude-fable-5-1": { + "id": "claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "devstral-2512": { + "id": "devstral-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.4, + "output": 2 + } + }, + "sonar-pro": { + "id": "sonar-pro", + "name": "Sonar Pro", + "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", + "family": "sonar-pro", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8000 + }, + "cost": { + "input": 3, + "output": 15 + } + }, + "grok-4.3": { + "id": "grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 30000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 + } + }, + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 + } + }, + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } + } + }, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.248, + "output": 1.485 + } + }, + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.11622, + "output": 0.488124 + } + }, + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 131072 + }, + "cost": { + "input": 0.69732, + "output": 2.78928 + } + }, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + } + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "muse-spark-1.2": { + "id": "muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-05", + "last_updated": "2026-08-05", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 + } + }, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12, + "tiers": [ + { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24, + "tier": { + "type": "context", + "size": 524288 + } + } + ], + "context_over_200k": { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24 + } + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } + } + }, + "mistral-small-2603": { + "id": "mistral-small-2603", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.5811, + "output": 2.44062 + } + }, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 2.2, + "output": 11, + "cache_read": 0.22, + "cache_write": 2.75 + } + }, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 + } + }, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.62708, + "output": 5.811 + } + }, + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05 + } + }, + "grok-build-0.1": { + "id": "grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 1, + "output": 2, + "cache_read": 0.2 + } + }, + "sonar": { + "id": "sonar", + "name": "Sonar", + "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", + "family": "sonar", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 1, + "output": 1 + } + }, + "mistral-large-2512": { + "id": "mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "gemini-3.8-flash": { + "id": "gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.825, + "output": 4.125, + "cache_read": 0.0825 + } + }, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 1.78812, + "output": 3.57624 + } + }, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 + } + }, + "gpt-5.3-chat-latest": { + "id": "gpt-5.3-chat-latest", + "name": "GPT-5.3 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "status": "deprecated", + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "sonar-reasoning-pro": { + "id": "sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", + "family": "sonar-reasoning", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 2, + "output": 8 + } + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } + } + }, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 1.1622, + "output": 4.88124 + } + }, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.75, + "output": 4.6488, + "cache_read": 0.44 + } + }, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "grok-4.6": { + "id": "grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } + } + }, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.25, + "output": 0.66 + } + } + } + }, + "nvidia": { + "id": "nvidia", + "env": [ + "NVIDIA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://integrate.api.nvidia.com/v1", + "name": "Nvidia", + "doc": "https://docs.api.nvidia.com/nim/", + "models": { + "baai/bge-m3": { + "id": "baai/bge-m3", + "name": "BGE M3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "bge", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-01-30", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 1024 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "poolside/laguna-xs-2.1": { + "id": "poolside/laguna-xs-2.1", + "name": "Laguna XS 2.1", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-02", + "last_updated": "2026-07-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "abacusai/dracarys-llama-3.1-70b-instruct": { + "id": "abacusai/dracarys-llama-3.1-70b-instruct", + "name": "dracarys-llama-3.1-70b-instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-09-11", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "z-ai/glm-5.3-flash": { + "id": "z-ai/glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "z-ai/glm-5.3": { + "id": "z-ai/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "thinkingmachines/inkling": { + "id": "thinkingmachines/inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "black-forest-labs/flux_1-schnell": { + "id": "black-forest-labs/flux_1-schnell", + "name": "FLUX.1-schnell", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": false, + "knowledge": "2024-07", + "release_date": "2024-08-01", + "last_updated": "2026-02-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": true, + "limit": { + "context": 77, + "input": 77, + "output": 0 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "black-forest-labs/flux.1-dev": { + "id": "black-forest-labs/flux.1-dev", + "name": "FLUX.1-dev", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2024-08-01", + "last_updated": "2025-09-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 4096, + "output": 0 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "black-forest-labs/flux_1-kontext-dev": { + "id": "black-forest-labs/flux_1-kontext-dev", + "name": "FLUX.1-Kontext-dev", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-08-12", + "last_updated": "2025-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": true, + "limit": { + "context": 40960, + "output": 40960 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "black-forest-labs/flux_2-klein-4b": { + "id": "black-forest-labs/flux_2-klein-4b", + "name": "FLUX.2 Klein 4B", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-01-14", + "last_updated": "2026-01-31", + "modalities": { + "input": [ + "image", + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": true, + "limit": { + "context": 40960, + "output": 40960 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/llama-guard-4-12b": { + "id": "meta/llama-guard-4-12b", + "name": "Llama Guard 4 12B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-04-05", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/muse-glimmer-30b": { + "id": "meta/muse-glimmer-30b", + "name": "Muse Glimmer 30B", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/esmfold": { + "id": "meta/esmfold", + "name": "esmfold", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-03-15", + "last_updated": "2025-06-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/llama-3.3-70b-instruct": { + "id": "meta/llama-3.3-70b-instruct", + "name": "Llama 3.3 70b Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-11-26", + "last_updated": "2024-11-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/llama-3.2-90b-vision-instruct": { + "id": "meta/llama-3.2-90b-vision-instruct", + "name": "Llama-3.2-90B-Vision-Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/llama-4-maverick-17b-128e-instruct": { + "id": "meta/llama-4-maverick-17b-128e-instruct", + "name": "Llama 4 Maverick 17b 128e Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-02", + "release_date": "2025-04-01", + "last_updated": "2025-04-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/llama-3.1-8b-instruct": { + "id": "meta/llama-3.1-8b-instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 16000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/llama-3.2-11b-vision-instruct": { + "id": "meta/llama-3.2-11b-vision-instruct", + "name": "Llama 3.2 11b Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-18", + "last_updated": "2024-09-18", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/llama-3.2-1b-instruct": { + "id": "meta/llama-3.2-1b-instruct", + "name": "Llama 3.2 1b Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-18", + "last_updated": "2024-09-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/llama-3.2-3b-instruct": { + "id": "meta/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2024-09-18", + "last_updated": "2024-09-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/llama-3.1-70b-instruct": { + "id": "meta/llama-3.1-70b-instruct", + "name": "Llama 3.1 70b Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-07-16", + "last_updated": "2024-07-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/esm2-650m": { + "id": "meta/esm2-650m", + "name": "esm2-650m", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-08-29", + "last_updated": "2025-03-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "upstage/solar-10.7b-instruct": { + "id": "upstage/solar-10.7b-instruct", + "name": "solar-10.7b-instruct", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-06-05", + "last_updated": "2025-04-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "bytedance/seed-oss-36b-instruct": { + "id": "bytedance/seed-oss-36b-instruct", + "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "seed", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-09-04", + "last_updated": "2025-11-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262000, + "output": 262000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma-4-31B-IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "google/gemma-2-2b-it": { + "id": "google/gemma-2-2b-it", + "name": "Gemma 2 2b It", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-07-16", + "last_updated": "2024-07-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "google/gemma-3-12b-it": { + "id": "google/gemma-3-12b-it", + "name": "Gemma 3 12B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "google/gemma-3n-e2b-it": { + "id": "google/gemma-3n-e2b-it", + "name": "Gemma 3n E2b It", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-06-12", + "last_updated": "2025-06-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "google/google-paligemma": { + "id": "google/google-paligemma", + "name": "paligemma", + "description": "Gemini multimodal model for text, image, audio, video, and document tasks", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-05-14", + "last_updated": "2024-08-26", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "google/gemma-3n-e4b-it": { + "id": "google/gemma-3n-e4b-it", + "name": "Gemma 3n E4b It", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-06-03", + "last_updated": "2025-06-03", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "google/gemma-3-4b-it": { + "id": "google/gemma-3-4b-it", + "name": "Gemma 3 4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "minimaxai/minimax-m2.7": { + "id": "minimaxai/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-04-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "minimaxai/minimax-m3": { + "id": "minimaxai/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "deepseek-ai/deepseek-v4-flash-0731": { + "id": "deepseek-ai/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "deepseek-ai/deepseek-v4-pro-0813": { + "id": "deepseek-ai/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "deepseek-ai/deepseek-v4-pro": { + "id": "deepseek-ai/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 393216 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 + } + }, + "deepseek-ai/deepseek-v4-flash": { + "id": "deepseek-ai/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 393216 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 + } + }, + "stepfun-ai/step-3.5-flash": { + "id": "stepfun-ai/step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-02", + "last_updated": "2026-02-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "stepfun-ai/step-3.7-flash": { + "id": "stepfun-ai/step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistralai/mistral-nemotron": { + "id": "mistralai/mistral-nemotron", + "name": "mistral-nemotron", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-06-11", + "last_updated": "2025-06-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistralai/mistral-7b-instruct-v0.3": { + "id": "mistralai/mistral-7b-instruct-v0.3", + "name": "Mistral-7B-Instruct-v0.3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-04-01", + "last_updated": "2025-04-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 65536, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistralai/mistral-large-3-675b-instruct-2512": { + "id": "mistralai/mistral-large-3-675b-instruct-2512", + "name": "Mistral Large 3 675B Instruct 2512", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistralai/magistral-small-2506": { + "id": "mistralai/magistral-small-2506", + "name": "Magistral Small 2506", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "input": 32768, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistralai/mistral-small-4-119b-2603": { + "id": "mistralai/mistral-small-4-119b-2603", + "name": "mistral-small-4-119b-2603", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistralai/mixtral-8x7b-instruct": { + "id": "mistralai/mixtral-8x7b-instruct", + "name": "Mistral: Mixtral 8x7B Instruct", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2023-12-10", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistralai/mistral-medium-3.5-128b": { + "id": "mistralai/mistral-medium-3.5-128b", + "name": "Mistral Medium 3.5", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistralai/ministral-14b-instruct-2512": { + "id": "mistralai/ministral-14b-instruct-2512", + "name": "Ministral 3 14B Instruct 2512", + "description": "Compact Mistral VLM for chat and instruction-based workloads", + "family": "ministral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistralai/mixtral-8x22b-instruct": { + "id": "mistralai/mixtral-8x22b-instruct", + "name": "Mistral: Mixtral 8x22B Instruct", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-04-17", + "last_updated": "2024-04-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 65536, + "output": 13108 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistralai/mistral-medium-3-instruct": { + "id": "mistralai/mistral-medium-3-instruct", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "input": 131072, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0 + } + }, + "moonshotai/kimi-k2-instruct-0905": { + "id": "moonshotai/kimi-k2-instruct-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0 + } + }, + "sarvamai/sarvam-m": { + "id": "sarvamai/sarvam-m", + "name": "sarvam-m", + "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-07-25", + "last_updated": "2025-07-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nvidia-nemotron-nano-9b-v2": { + "id": "nvidia/nvidia-nemotron-nano-9b-v2", + "name": "nvidia-nemotron-nano-9b-v2", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-09", + "release_date": "2025-08-18", + "last_updated": "2025-08-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/llama-3.1-nemotron-ultra-253b-v1": { + "id": "nvidia/llama-3.1-nemotron-ultra-253b-v1", + "name": "Llama 3.1 Nemotron Ultra 253B", + "description": "Flagship Nemotron model for high-throughput reasoning and complex agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-04-07", + "last_updated": "2025-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/sparsedrive": { + "id": "nvidia/sparsedrive", + "name": "sparsedrive", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-03-18", + "last_updated": "2025-07-20", + "modalities": { + "input": [ + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nemotron-3-nano-30b-a3b": { + "id": "nvidia/nemotron-3-nano-30b-a3b", + "name": "nemotron-3-nano-30b-a3b", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-09", + "release_date": "2024-12", + "last_updated": "2024-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nemotron-3.5-lightning-30b-a3b": { + "id": "nvidia/nemotron-3.5-lightning-30b-a3b", + "name": "Nemotron 3.5 Lightning 30B A3B", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/llama-nemotron-embed-vl-1b-v2": { + "id": "nvidia/llama-nemotron-embed-vl-1b-v2", + "name": "llama-nemotron-embed-vl-1b-v2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "nemotron", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-02-10", + "last_updated": "2026-02-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 2048 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nv-embedcode-7b-v1": { + "id": "nvidia/nv-embedcode-7b-v1", + "name": "nv-embedcode-7b-v1", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-03-17", + "last_updated": "2025-05-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 2048 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nemotron-voicechat": { + "id": "nvidia/nemotron-voicechat", + "name": "nemotron-voicechat", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/bevformer": { + "id": "nvidia/bevformer", + "name": "bevformer", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-03-18", + "last_updated": "2025-07-20", + "modalities": { + "input": [ + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/cosmos-transfer1-7b": { + "id": "nvidia/cosmos-transfer1-7b", + "name": "cosmos-transfer1-7b", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-13", + "last_updated": "2025-06-30", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "video" + ] + }, + "open_weights": true, + "limit": { + "context": 0, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/cosmos-predict1-5b": { + "id": "nvidia/cosmos-predict1-5b", + "name": "cosmos-predict1-5b", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-03-18", + "last_updated": "2025-03-18", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "video" + ] + }, + "open_weights": true, + "limit": { + "context": 0, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/llama-3.1-nemotron-nano-vl-8b-v1": { + "id": "nvidia/llama-3.1-nemotron-nano-vl-8b-v1", + "name": "Llama 3.1 Nemotron Nano VL 8B v1", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-04-10", + "last_updated": "2025-04-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "Nemotron 3 Super", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.2, + "output": 0.8 + } + }, + "nvidia/studiovoice": { + "id": "nvidia/studiovoice", + "name": "studiovoice", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-10-03", + "last_updated": "2025-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nemotron-content-safety-reasoning-4b": { + "id": "nvidia/nemotron-content-safety-reasoning-4b", + "name": "nemotron-content-safety-reasoning-4b", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": false, + "release_date": "2026-01-22", + "last_updated": "2026-01-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/llama-3.1-nemotron-safety-guard-8b-v3": { + "id": "nvidia/llama-3.1-nemotron-safety-guard-8b-v3", + "name": "llama-3.1-nemotron-safety-guard-8b-v3", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-10-28", + "last_updated": "2025-10-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nv-embed-v1": { + "id": "nvidia/nv-embed-v1", + "name": "nv-embed-v1", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-06-07", + "last_updated": "2025-07-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 2048 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/synthetic-video-detector": { + "id": "nvidia/synthetic-video-detector", + "name": "synthetic-video-detector", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 0, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/cosmos-transfer2_5-2b": { + "id": "nvidia/cosmos-transfer2_5-2b", + "name": "cosmos-transfer2.5-2b", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-02-26", + "last_updated": "2026-02-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "video" + ] + }, + "open_weights": true, + "limit": { + "context": 0, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nemotron-3-ultra-550b-a55b": { + "id": "nvidia/nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 2.5, + "cache_read": 0.15 + } + }, + "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning": { + "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning", + "name": "Nemotron 3 Nano Omni", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": -1, + "max": 32768 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-28", + "last_updated": "2026-04-28", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/llama-3.1-nemotron-nano-8b-v1": { + "id": "nvidia/llama-3.1-nemotron-nano-8b-v1", + "name": "Llama 3.1 Nemotron Nano 8B v1", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-03-18", + "last_updated": "2025-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/active-speaker-detection": { + "id": "nvidia/active-speaker-detection", + "name": "Active Speaker Detection", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 0, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/streampetr": { + "id": "nvidia/streampetr", + "name": "streampetr", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/llama-nemotron-rerank-vl-1b-v2": { + "id": "nvidia/llama-nemotron-rerank-vl-1b-v2", + "name": "llama-nemotron-rerank-vl-1b-v2", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "nemotron", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-03-31", + "last_updated": "2026-03-31", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nemotron-mini-4b-instruct": { + "id": "nvidia/nemotron-mini-4b-instruct", + "name": "nemotron-mini-4b-instruct", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-08-21", + "last_updated": "2024-08-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/cosmos-reason2-8b": { + "id": "nvidia/cosmos-reason2-8b", + "name": "Cosmos Reason2 8B", + "description": "Vision language model for physical-world understanding with structured reasoning on video and images", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-12-01", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nemotron-3-content-safety": { + "id": "nvidia/nemotron-3-content-safety", + "name": "nemotron-3-content-safety", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/llama-3.3-nemotron-super-49b-v1": { + "id": "nvidia/llama-3.3-nemotron-super-49b-v1", + "name": "Llama 3.3 Nemotron Super 49B v1", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-04-07", + "last_updated": "2025-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/magpie-tts-zeroshot": { + "id": "nvidia/magpie-tts-zeroshot", + "name": "magpie-tts-zeroshot", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-05-22", + "last_updated": "2025-06-12", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "audio" + ] + }, + "open_weights": true, + "limit": { + "context": 0, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/riva-translate-4b-instruct-v1.1": { + "id": "nvidia/riva-translate-4b-instruct-v1.1", + "name": "riva-translate-4b-instruct-v1_1", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-12-12", + "last_updated": "2025-12-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nemotron-nano-12b-v2-vl": { + "id": "nvidia/nemotron-nano-12b-v2-vl", + "name": "Nemotron Nano 12B v2 VL", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-10-28", + "last_updated": "2025-10-28", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/llama-3.3-nemotron-super-49b-v1.5": { + "id": "nvidia/llama-3.3-nemotron-super-49b-v1.5", + "name": "Llama 3.3 Nemotron Super 49B v1.5", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-07-25", + "last_updated": "2025-07-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/llama-3_2-nemoretriever-300m-embed-v1": { + "id": "nvidia/llama-3_2-nemoretriever-300m-embed-v1", + "name": "llama-3_2-nemoretriever-300m-embed-v1", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-07-24", + "last_updated": "2025-07-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 2048 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/gliner-pii": { + "id": "nvidia/gliner-pii", + "name": "gliner-pii", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-03-03", + "last_updated": "2026-03-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/usdcode": { + "id": "nvidia/usdcode", + "name": "usdcode", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-01-01", + "last_updated": "2026-01-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/usdvalidate": { + "id": "nvidia/usdvalidate", + "name": "usdvalidate", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-07-24", + "last_updated": "2025-01-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 0, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/llama-3.1-nemotron-70b-instruct": { + "id": "nvidia/llama-3.1-nemotron-70b-instruct", + "name": "Llama 3.1 Nemotron 70B Instruct", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/rerank-qa-mistral-4b": { + "id": "nvidia/rerank-qa-mistral-4b", + "name": "rerank-qa-mistral-4b", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-03-17", + "last_updated": "2025-01-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen/qwen-image-edit": { + "id": "qwen/qwen-image-edit", + "name": "Qwen Image Edit", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-19", + "last_updated": "2025-08-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen/qwen3-coder-480b-a35b-instruct": { + "id": "qwen/qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 66536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next-80B-A3B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2024-12-01", + "last_updated": "2025-09-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5-397B-A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-01", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen/qwen2.5-coder-32b-instruct": { + "id": "qwen/qwen2.5-coder-32b-instruct", + "name": "Qwen2.5 Coder 32b Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-11-06", + "last_updated": "2024-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen/qwen-image": { + "id": "qwen/qwen-image", + "name": "Qwen Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "openai/whisper-large-v3": { + "id": "openai/whisper-large-v3", + "name": "Whisper Large v3", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2023-09-01", + "last_updated": "2025-09-05", + "modalities": { + "input": [ + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 0, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT-OSS-120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08", + "release_date": "2025-08-04", + "last_updated": "2025-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "microsoft/phi-4-mini-instruct": { + "id": "microsoft/phi-4-mini-instruct", + "name": "Phi-4-Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2024-12-01", + "last_updated": "2025-09-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "microsoft/phi-4-multimodal-instruct": { + "id": "microsoft/phi-4-multimodal-instruct", + "name": "Phi 4 Multimodal", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 128000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + } + } + }, + "pioneer": { + "id": "pioneer", + "env": [ + "PIONEER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.pioneer.ai/v1", + "name": "Pioneer", + "doc": "https://agent.pioneer.ai/llms.txt", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 2.5 + } + }, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 991000, + "output": 64000 + }, + "cost": { + "input": 1.25, + "output": 3.75, + "cache_read": 0.25, + "cache_write": 1.5625 + } + }, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02, + "cache_write": 0.2 + } + }, + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25, + "cache_write": 2.5 + } + }, + "ministral-3b": { + "id": "ministral-3b", + "name": "Ministral 3B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-03", + "release_date": "2024-10-16", + "last_updated": "2024-10-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.1, + "output": 0.1, + "cache_read": 0.1, + "cache_write": 0.1 + } + }, + "magistral-medium": { + "id": "magistral-medium", + "name": "Magistral Medium (latest)", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-medium", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 64000 + }, + "cost": { + "input": 2, + "output": 5, + "cache_read": 2, + "cache_write": 2 + } + }, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025, + "cache_write": 0.25 + } + }, + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "claude-3-7-sonnet-latest": { + "id": "claude-3-7-sonnet-latest", + "name": "Claude Sonnet 3.7", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-10-31", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175, + "cache_write": 1.75 + } + }, + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.05, + "cache_write": 0.1 + } + }, + "mistral-medium-3.5": { + "id": "mistral-medium-3.5", + "name": "Mistral Medium 3.5", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 131072 + }, + "cost": { + "input": 1.5, + "output": 7.5, + "cache_read": 1.5, + "cache_write": 1.5 + } + }, + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005, + "cache_write": 0.05 + } + }, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "gemini-3.6-flash": { + "id": "gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 1.5, + "output": 7.5, + "cache_read": 0.15, + "cache_write": 1.5 + } + }, + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65000 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "cache_write": 0.3 + } + }, + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "cache_write": 2 + } + }, + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 131072 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "cache_write": 1.25 + } + }, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075, + "cache_write": 0.15 + } + }, + "gemini-3-flash": { + "id": "gemini-3-flash", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65000 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.083333 + } + }, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 0.083333 + } + }, + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.1875, + "output": 1.125, + "cache_read": 0.0375, + "cache_write": 0.234375 + } + }, + "mistral-large-3": { + "id": "mistral-large-3", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 131072 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.5, + "cache_write": 0.5 + } + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 11, + "output": 55, + "cache_read": 1.1, + "cache_write": 13.75 + } + }, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075, + "cache_write": 0.75 + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 1, + "output": 6, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "devstral-small-2": { + "id": "devstral-small-2", + "name": "Devstral Small 2", + "description": "Compact multimodal coding model for repository exploration, file editing, and software agents", + "family": "devstral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.3, + "cache_read": 0.1, + "cache_write": 0.1 + } + }, + "devstral-2": { + "id": "devstral-2", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 131072 + }, + "cost": { + "input": 0.4, + "output": 2, + "cache_read": 0.4, + "cache_write": 0.4 + } + }, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 5 + } + }, + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 1, + "cache_write": 2 + } + }, + "mistral-medium": { + "id": "mistral-medium", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 64000 + }, + "cost": { + "input": 0.4, + "output": 2, + "cache_read": 0.4, + "cache_write": 0.4 + } + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "qwen3.6-max-preview": { + "id": "qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 240000, + "output": 64000 + }, + "cost": { + "input": 1.04, + "output": 6.24, + "cache_read": 0.208, + "cache_write": 1.3 + } + }, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.2, + "cache_write": 0.4 + } + }, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "claude-opus-5-fast": { + "id": "claude-opus-5-fast", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.325, + "output": 1.95, + "cache_read": 0.065, + "cache_write": 0.40625 + } + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125 + } + }, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "ministral-14b": { + "id": "ministral-14b", + "name": "Ministral 14B", + "description": "Compact multimodal Mistral model for local assistants, edge agents, and efficient tool use", + "family": "ministral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 131072 + }, + "cost": { + "input": 0.2, + "output": 0.2, + "cache_read": 0.2, + "cache_write": 0.2 + } + }, + "gemini-3.1-pro": { + "id": "gemini-3.1-pro", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 0.375 + } + }, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.32, + "output": 1.28, + "cache_read": 0.064, + "cache_write": 0.4 + } + }, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65000 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.03, + "cache_write": 0.25 + } + }, + "poolside/laguna-s-2.1": { + "id": "poolside/laguna-s-2.1", + "name": "Laguna S 2.1", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.2, + "cache_read": 0.01, + "cache_write": 0.1 + } + }, + "meta-llama/Llama-3.2-1B-Instruct": { + "id": "meta-llama/Llama-3.2-1B-Instruct", + "name": "Llama 3.2 1B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2023-12-31", + "release_date": "2024-08-31", + "last_updated": "2024-09-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 60000 + }, + "cost": { + "input": 0.1, + "output": 0.201, + "cache_read": 0.1, + "cache_write": 0.1 + } + }, + "meta-llama/Llama-3.2-3B-Instruct": { + "id": "meta-llama/Llama-3.2-3B-Instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2023-12-31", + "release_date": "2024-08-31", + "last_updated": "2024-09-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 80000 + }, + "cost": { + "input": 0.1, + "output": 0.335, + "cache_read": 0.1, + "cache_write": 0.1 + } + }, + "meta-llama/Llama-3.1-8B-Instruct": { + "id": "meta-llama/Llama-3.1-8B-Instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2023-12-31", + "release_date": "2024-06-30", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.2, + "output": 0.2, + "cache_read": 0.2, + "cache_write": 0.2 + } + }, + "meta-llama/Llama-3.2-3B": { + "id": "meta-llama/Llama-3.2-3B", + "name": "Llama-3.2-3B", + "description": "Small open Llama base model for lightweight text generation and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.1, + "cache_read": 0.1, + "cache_write": 0.1 + } + }, + "meta-llama/Llama-3.2-1B": { + "id": "meta-llama/Llama-3.2-1B", + "name": "Llama-3.2-1B", + "description": "Compact open Llama base model for lightweight and on-device use", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.1, + "cache_read": 0.1, + "cache_write": 0.1 + } + }, + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 16384, + "output": 16384 + }, + "cost": { + "input": 0.9, + "output": 0.9, + "cache_read": 0.9, + "cache_write": 0.9 + } + }, + "XiaomiMiMo/MiMo-V2.5-Pro": { + "id": "XiaomiMiMo/MiMo-V2.5-Pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1050000, + "output": 131000 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036, + "cache_write": 0.435 + } + }, + "XiaomiMiMo/MiMo-V2.5": { + "id": "XiaomiMiMo/MiMo-V2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1050000, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028, + "cache_write": 0.14 + } + }, + "thinkingmachines/inkling-small": { + "id": "thinkingmachines/inkling-small", + "name": "Inkling Small", + "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-07-30", + "last_updated": "2026-07-30", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.5, + "output": 1.2, + "cache_read": 0.1, + "cache_write": 0.5 + } + }, + "meta/muse-spark-1.1": { + "id": "meta/muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-08", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.25, + "output": 4.25, + "cache_read": 0.15, + "cache_write": 1.25 + } + }, + "HuggingFaceTB/SmolLM3-3B-Base": { + "id": "HuggingFaceTB/SmolLM3-3B-Base", + "name": "SmolLM3 3B Base", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-06-30", + "last_updated": "2025-06-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "google/diffusiongemma-26B-A4B-it": { + "id": "google/diffusiongemma-26B-A4B-it", + "name": "DiffusionGemma 26B-A4B IT", + "description": "Gemini model for general assistance, reasoning, and multimodal workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-05-31", + "last_updated": "2026-05-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.5, + "output": 0.5, + "cache_read": 0.5, + "cache_write": 0.5 + } + }, + "google/gemma-4-12B-it": { + "id": "google/gemma-4-12B-it", + "name": "Gemma 4 12B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-05-31", + "last_updated": "2026-05-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.25, + "output": 0.25, + "cache_read": 0.25, + "cache_write": 0.25 + } + }, + "google/gemma-4-E2B-it": { + "id": "google/gemma-4-E2B-it", + "name": "Gemma 4 E2B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.1, + "cache_read": 0.1, + "cache_write": 0.1 + } + }, + "google/gemma-3-4b-pt": { + "id": "google/gemma-3-4b-pt", + "name": "Gemma 3 4B (Pretrained)", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-02-28", + "last_updated": "2025-02-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 0.5, + "cache_read": 0.5, + "cache_write": 0.5 + } + }, + "google/gemma-4-E4B-it": { + "id": "google/gemma-4-E4B-it", + "name": "Gemma 4 E4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.2, + "output": 0.2, + "cache_read": 0.2, + "cache_write": 0.2 + } + }, + "Qwen/Qwen3-4B-Instruct-2507": { + "id": "Qwen/Qwen3-4B-Instruct-2507", + "name": "Qwen3 4B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-07-31", + "last_updated": "2025-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.2, + "output": 0.2, + "cache_read": 0.2, + "cache_write": 0.2 + } + }, + "Qwen/Qwen2.5-Coder-0.5B": { + "id": "Qwen/Qwen2.5-Coder-0.5B", + "name": "Qwen2.5-Coder-0.5B", + "description": "Tiny open Qwen code model for lightweight completion and on-device coding", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-11-12", + "last_updated": "2024-11-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.1, + "cache_read": 0.1, + "cache_write": 0.1 + } + }, + "Qwen/Qwen3-1.7B-Base": { + "id": "Qwen/Qwen3-1.7B-Base", + "name": "Qwen3 1.7B Base", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-03-31", + "last_updated": "2025-03-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.1, + "cache_read": 0.1, + "cache_write": 0.1 + } + }, + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 1, + "cache_read": 0.028, + "cache_write": 0.175 + } + }, + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B-A22B Instruct 2507", + "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-07-21", + "last_updated": "2025-07-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 1.2, + "cache_read": 1.2, + "cache_write": 1.2 + } + }, + "Qwen/Qwen3-8B": { + "id": "Qwen/Qwen3-8B", + "name": "Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-03-31", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 40960, + "output": 40960 + }, + "cost": { + "input": 0.2, + "output": 0.2, + "cache_read": 0.2, + "cache_write": 0.2 + } + }, + "Qwen/Qwen3-4B-Base": { + "id": "Qwen/Qwen3-4B-Base", + "name": "Qwen3 4B Base", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-03-31", + "last_updated": "2025-03-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.9, + "output": 0.9, + "cache_read": 0.9, + "cache_write": 0.9 + } + }, + "Qwen/Qwen3.5-9B": { + "id": "Qwen/Qwen3.5-9B", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 0.3, + "cache_read": 0.3, + "cache_write": 0.3 + } + }, + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.6, + "output": 0.6, + "cache_read": 0.6, + "cache_write": 0.6 + } + }, + "LiquidAI/LFM2-24B-A2B": { + "id": "LiquidAI/LFM2-24B-A2B", + "name": "LFM2 24B A2B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "liquid", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-01-31", + "last_updated": "2026-02-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.03, + "output": 0.12, + "cache_read": 0.03, + "cache_write": 0.03 + } + }, + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek-V3.1", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-08-21", + "last_updated": "2025-08-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 131072 + }, + "cost": { + "input": 0.56, + "output": 1.68, + "cache_read": 0.56, + "cache_write": 0.56 + } + }, + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.2, + "cache_read": 0.0197, + "cache_write": 0.1 + } + }, + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625, + "cache_write": 0.435 + } + }, + "deepseek-ai/DeepSeek-V3": { + "id": "deepseek-ai/DeepSeek-V3", + "name": "DeepSeek-V3", + "description": "Open DeepSeek MoE chat model for coding, math, and general reasoning", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-12-26", + "last_updated": "2024-12-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 8192 + }, + "cost": { + "input": 0.27, + "output": 1.12, + "cache_read": 0.135, + "cache_write": 0.27 + } + }, + "MiniMaxAI/MiniMax-M3": { + "id": "MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.3 + } + }, + "MiniMaxAI/MiniMax-M2.7": { + "id": "MiniMaxAI/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131000 + }, + "cost": { + "input": 0.279, + "output": 1.2, + "cache_read": 0.279, + "cache_write": 0.279 + } + }, + "mistralai/Mistral-Small-4-119B-2603": { + "id": "mistralai/Mistral-Small-4-119B-2603", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.015, + "cache_write": 0.15 + } + }, + "mistralai/Pixtral-12B-2409": { + "id": "mistralai/Pixtral-12B-2409", + "name": "Pixtral 12B", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-09", + "release_date": "2024-09-01", + "last_updated": "2024-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "mistralai/Mistral-Nemo-Instruct-2407": { + "id": "mistralai/Mistral-Nemo-Instruct-2407", + "name": "Mistral Nemo", + "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.02, + "output": 0.03, + "cache_read": 0.02, + "cache_write": 0.02 + } + }, + "mistralai/Magistral-Small-2506": { + "id": "mistralai/Magistral-Small-2506", + "name": "Magistral Small", + "description": "Open Mistral reasoning model for transparent step-by-step problem solving", + "family": "magistral", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-06-10", + "last_updated": "2025-06-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 64000 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.5, + "cache_write": 0.5 + } + }, + "mistralai/Ministral-8B-Instruct-2410": { + "id": "mistralai/Ministral-8B-Instruct-2410", + "name": "Ministral 8B Instruct", + "description": "Efficient open Mistral edge model for on-device chat and function calling", + "family": "ministral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-10-16", + "last_updated": "2024-10-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "mistralai/Codestral-22B-v0.1": { + "id": "mistralai/Codestral-22B-v0.1", + "name": "Codestral-22B-v0.1", + "description": "Open Mistral code model for fill-in-the-middle and 80+ programming languages", + "family": "codestral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-05-29", + "last_updated": "2024-05-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.3, + "output": 0.9, + "cache_read": 0.3, + "cache_write": 0.3 + } + }, + "mistralai/Mistral-7B-Instruct-v0.3": { + "id": "mistralai/Mistral-7B-Instruct-v0.3", + "name": "Mistral 7B Instruct v0.3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2023-04-30", + "last_updated": "2023-04-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.2, + "output": 0.2, + "cache_read": 0.2, + "cache_write": 0.2 + } + }, + "sakana/fugu-ultra": { + "id": "sakana/fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 5 + } + }, + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 131072 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.34, + "cache_write": 0.95 + } + }, + "moonshotai/Kimi-K3": { + "id": "moonshotai/Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3 + } + }, + "moonshotai/Kimi-K3-Fast": { + "id": "moonshotai/Kimi-K3-Fast", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 4.5, + "output": 22.5, + "cache_read": 0.45, + "cache_write": 4.5 + } + }, + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 32768 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19, + "cache_write": 0.95 + } + }, + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202000, + "output": 131072 + }, + "cost": { + "input": 0.98, + "output": 3.08, + "cache_read": 0.182, + "cache_write": 0.98 + } + }, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1040000, + "output": 128000 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 1.4 + } + }, + "zai-org/GLM-5.2-Fast": { + "id": "zai-org/GLM-5.2-Fast", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2.1, + "output": 6.6, + "cache_read": 0.21, + "cache_write": 2.1 + } + }, + "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16": { + "id": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-12-15", + "last_updated": "2025-12-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.05, + "output": 0.2, + "cache_read": 0.05, + "cache_write": 0.05 + } + }, + "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16": { + "id": "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16", + "name": "Nemotron 3.5 Lightning 30B A3B", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 4096 + }, + "cost": { + "input": 0.5, + "output": 0.5, + "cache_read": 0.5, + "cache_write": 0.5 + } + }, + "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16": { + "id": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 65000 + }, + "cost": { + "input": 0.5, + "output": 2.5, + "cache_read": 0.15, + "cache_write": 0.5 + } + }, + "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8": { + "id": "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8", + "name": "Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 0.09, + "output": 0.45, + "cache_read": 0.09, + "cache_write": 0.09 + } + }, + "pioneer/auto": { + "id": "pioneer/auto", + "name": "Pioneer Auto", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2024-01-01", + "last_updated": "2025-06-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 4096 + } + }, + "fastino/gliner2-multi-v1": { + "id": "fastino/gliner2-multi-v1", + "name": "GLiNER2 Multi", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-11-30", + "last_updated": "2025-11-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "fastino/gliner2-privacy-filter-PII-multi": { + "id": "fastino/gliner2-privacy-filter-PII-multi", + "name": "GLiNER2 Privacy Filter PII (Multi)", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "fastino/gliner2.5-multi-v1": { + "id": "fastino/gliner2.5-multi-v1", + "name": "GLiNER 2.5 Multi", + "description": "Multilingual boundary NER and span extraction; non-trainable encoder.", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-08-24", + "last_updated": "2026-08-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 4096, + "output": 4096 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "fastino/gliguard-LLMGuardrails-300M": { + "id": "fastino/gliguard-LLMGuardrails-300M", + "name": "GLiGuard LLM Guardrails 300M", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "fastino/gliner2-large-v1": { + "id": "fastino/gliner2-large-v1", + "name": "GLiNER2 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-06-30", + "last_updated": "2025-06-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "fastino/gliner2-base-v1": { + "id": "fastino/gliner2-base-v1", + "name": "GLiNER2 Base", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-06-30", + "last_updated": "2025-06-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "fastino/gliguard-PII-multi": { + "id": "fastino/gliguard-PII-multi", + "name": "GLiNER2-Guardrails-PII-Multi", + "description": "A 300M-parameter multilingual model that runs LLM safety moderation and PII detection in a single forward pass.", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "fastino/gliner2-multi-large-v1": { + "id": "fastino/gliner2-multi-large-v1", + "name": "GLiNER2 Multi Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-11-30", + "last_updated": "2025-11-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.15, + "cache_write": 0.15 + } + }, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.07, + "output": 0.3, + "cache_read": 0.035, + "cache_write": 0.07 + } + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.015, + "cache_write": 0.15 + } + } + } + }, + "xiaomi": { + "id": "xiaomi", + "env": [ + "XIAOMI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.xiaomimimo.com/v1", + "name": "Xiaomi", + "doc": "https://platform.xiaomimimo.com/#/docs", + "models": { + "mimo-v2.6-pro": { + "id": "mimo-v2.6-pro", + "name": "MiMo-V2.6-Pro", + "description": "MiMo Pro model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 + } + }, + "mimo-v2-flash": { + "id": "mimo-v2-flash", + "name": "MiMo-V2-Flash", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12-01", + "release_date": "2025-12-16", + "last_updated": "2026-06-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "status": "deprecated", + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 + } + }, + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-06-24", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 + } + }, + "mimo-v2-omni": { + "id": "mimo-v2-omni", + "name": "MiMo-V2-Omni", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-06-24", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 + } + }, + "mimo-v2.6-pro-ultraspeed": { + "id": "mimo-v2.6-pro-ultraspeed", + "name": "MiMo-V2.6-Pro-UltraSpeed", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-09-21", + "last_updated": "2026-09-21", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 4.35, + "output": 8.7, + "cache_read": 0.036 + } + }, + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-06-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 + } + }, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-06-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 + } + }, + "mimo-v2.6-flash": { + "id": "mimo-v2.6-flash", + "name": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 + } + }, + "mimo-v2.5-pro-ultraspeed": { + "id": "mimo-v2.5-pro-ultraspeed", + "name": "MiMo-V2.5-Pro-UltraSpeed", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-06-08", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "status": "beta", + "cost": { + "input": 1.305, + "output": 2.61, + "cache_read": 0.0108 + } + } + } + }, + "xiaomi-token-plan-sgp": { + "id": "xiaomi-token-plan-sgp", + "env": [ + "XIAOMI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://token-plan-sgp.xiaomimimo.com/v1", + "name": "Xiaomi Token Plan (Singapore)", + "doc": "https://platform.xiaomimimo.com/#/docs", + "models": { + "mimo-v2.6-pro": { + "id": "mimo-v2.6-pro", + "name": "MiMo-V2.6-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "mimo-v2.6-flash": { + "id": "mimo-v2.6-flash", + "name": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "mimo-v2.5-tts-voiceclone": { + "id": "mimo-v2.5-tts-voiceclone", + "name": "MiMo-V2.5-TTS-VoiceClone", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "mimo-v2-tts": { + "id": "mimo-v2-tts", + "name": "MiMo-V2-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "mimo-v2.5-tts": { + "id": "mimo-v2.5-tts", + "name": "MiMo-V2.5-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mimo-v2.5-tts-voicedesign": { + "id": "mimo-v2.5-tts-voicedesign", + "name": "MiMo-V2.5-TTS-VoiceDesign", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + } + } + }, + "minimax": { + "id": "minimax", + "env": [ + "MINIMAX_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://api.minimax.io/anthropic/v1", + "name": "MiniMax (minimax.io)", + "doc": "https://platform.minimax.io/docs/guides/quickstart", + "models": { + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-25", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 512000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "tiers": [ + { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12, + "tier": { + "type": "context", + "size": 512000 + } + } + ], + "context_over_200k": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 + } + } + }, + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "MiniMax-M2.7-highspeed": { + "id": "MiniMax-M2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "MiniMax-M2": { + "id": "MiniMax-M2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2 + } + }, + "MiniMax-M2.5-highspeed": { + "id": "MiniMax-M2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-13", + "last_updated": "2026-02-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 + } + } + } + }, + "github-copilot": { + "id": "github-copilot", + "env": [ + "GITHUB_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.githubcopilot.com", + "name": "GitHub Copilot", + "doc": "https://docs.github.com/en/copilot", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } + } + }, + "grok-4.7": { + "id": "grok-4.7", + "name": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "input": 372000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } + } + }, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "claude-opus-4.7": { + "id": "claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "input": 168000, + "output": 32000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 264000, + "input": 128000, + "output": 64000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "claude-opus-4.8": { + "id": "claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "input": 168000, + "output": 64000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "gemini-3.6-flash": { + "id": "gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 256, + "max": 32000 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 936000, + "output": 64000 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 + } + }, + "mai-code-1-flash-picker": { + "id": "mai-code-1-flash-picker", + "name": "MAI-Code-1-Flash", + "description": "Microsoft coding model built for fast, efficient assistance in everyday developer workflows", + "family": "mai", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2026-06-02", + "last_updated": "2026-06-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "input": 128000, + "output": 128000 + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "mai-code-1.1-flash": { + "id": "mai-code-1.1-flash", + "name": "MAI-Code-1.1-Flash", + "description": "Microsoft coding model with native vision support, optimized for fast and efficient software development", + "family": "mai", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "input": 128000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02 + } + }, + "claude-haiku-4.5": { + "id": "claude-haiku-4.5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 32000 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "input": 136000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } + } + }, + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "input": 372000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } + } + }, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 936000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 256, + "max": 24000 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "input": 128000, + "output": 64000 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "claude-fable-5.1": { + "id": "claude-fable-5.1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } + } + }, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 936000, + "output": 64000 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 + } + }, + "claude-sonnet-4.6": { + "id": "claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 32000 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "input": 168000, + "output": 32000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "gemini-3.8-flash": { + "id": "gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 936000, + "output": 64000 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 + } + }, + "gpt-6-luna": { + "id": "gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } + } + }, + "claude-opus-5.5": { + "id": "claude-opus-5.5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 + } + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } + } + }, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 224000, + "output": 32000 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + }, + "grok-4.6": { + "id": "grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "input": 372000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } + } + }, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10 + } + } + }, + "gpt-6-sol": { + "id": "gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } + } + } + } + }, + "inferx": { + "id": "inferx", + "env": [ + "INFERX_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://model.inferx.net/endpoints/v1", + "name": "InferX", + "doc": "https://model.inferx.net/endpoints", + "models": { + "gemma-4-31B-it-fp8": { + "id": "gemma-4-31B-it-fp8", + "name": "Gemma 4 31B IT FP8", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "Qwen3.6-35B-A3B-fp8-no-thinking": { + "id": "Qwen3.6-35B-A3B-fp8-no-thinking", + "name": "Qwen3.6-35B-A3B-fp8-no-thinking", + "description": "Qwen3.6-35B-A3B-fp8 disable thinking", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "Devstral-2-123B-Instruct-2512-int4-AutoRound": { + "id": "Devstral-2-123B-Instruct-2512-int4-AutoRound", + "name": "Devstral-2-123B-Instruct-2512-int4-AutoRound", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "Qwen3.6-27B-FP8": { + "id": "Qwen3.6-27B-FP8", + "name": "Qwen3.6 27B FP8", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "Qwen3-Coder-Next-FP8": { + "id": "Qwen3-Coder-Next-FP8", + "name": "Qwen3 Coder Next FP8", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256144, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "Qwen3-Coder-Next-FP8-no-thinking": { + "id": "Qwen3-Coder-Next-FP8-no-thinking", + "name": "Qwen3-Coder-Next-FP8-no-thinking", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 260000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mimo-v25": { + "id": "mimo-v25", + "name": "mimo-v25", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 100000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "Qwen3-Embedding-8B": { + "id": "Qwen3-Embedding-8B", + "name": "Qwen3-Embedding-8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 0 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "Qwen3.6-35B-A3B-FP8": { + "id": "Qwen3.6-35B-A3B-FP8", + "name": "Qwen3.6 35B A3B FP8", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "Ornith-1.0-35B-FP8": { + "id": "Ornith-1.0-35B-FP8", + "name": "Ornith-1.0-35B-FP8", + "description": "Large coding-reasoning model for agentic software tasks and RL search", + "family": "ornith", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-25", + "last_updated": "2026-06-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 100000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "Agents-A1": { + "id": "Agents-A1", + "name": "Agents-A1", + "description": "35B MoE agentic model built for long-horizon search, engineering, and scientific reasoning tasks", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "release_date": "2026-06-26", + "last_updated": "2026-06-26", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 100000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "deepseek-v4-flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 100000 + }, + "cost": { + "input": 0, + "output": 0 + } + } + } + }, + "opencode-go": { + "id": "opencode-go", + "env": [ + "OPENCODE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://opencode.ai/zen/go/v1", + "name": "OpenCode Go", + "doc": "https://opencode.ai/docs/go", + "models": { + "mimo-v2.6-pro": { + "id": "mimo-v2.6-pro", + "name": "MiMo-V2.6-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 + } + }, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "qwen3.7-max", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 + } + }, + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo V2.5", + "description": "MiMo omni model for text, image, video, audio, and agents", + "family": "mimo-v2.5", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 + } + }, + "grok-4.7": { + "id": "grok-4.7", + "name": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } + } + }, + "mimo-v2-omni": { + "id": "mimo-v2-omni", + "name": "MiMo V2 Omni", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo-v2-omni", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2, + "cache_read": 0.08 + } + }, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 + } + }, + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter multimodal flagship for coding, professional work, and long-horizon agentic workflows", + "family": "qwen3.8-max", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen3.5", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "status": "deprecated", + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 + } + }, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 32768 + }, + "status": "deprecated", + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.2 + } + }, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 + } + }, + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax-m2.5", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 65536 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "deepseek-v4-flash-vision-exp": { + "id": "deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 + } + }, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } + }, + "longcat-2.0": { + "id": "longcat-2.0", + "name": "LongCat-2.0", + "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", + "family": "longcat", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.006 + } + }, + "ox-alpha-free": { + "id": "ox-alpha-free", + "name": "Ox Alpha Free (Unlimited)", + "description": "Stealth reasoning model for coding, agentic tasks, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-08-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.3, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 0.6 + } + } + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "status": "deprecated", + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.1 + } + }, + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo V2 Pro", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo-v2-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 1, + "output": 3, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } + } + }, + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax-m2.7", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "space-bunny-free": { + "id": "space-bunny-free", + "name": "Space Bunny Free", + "description": "Anonymous preview reasoning model for coding, agentic tasks, tool use, and multimodal input", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "release_date": "2026-09-23", + "last_updated": "2026-09-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "input": 524288, + "output": 524288 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo V2.5 Pro", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "family": "mimo-v2.5-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 128000 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 + } + }, + "mimo-v2.6-flash": { + "id": "mimo-v2.6-flash", + "name": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 + } + }, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax-m3", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-31", + "last_updated": "2026-05-31", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "tiers": [ + { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12, + "tier": { + "type": "context", + "size": 512000 + } + } + ], + "context_over_200k": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 + } + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } + } + }, + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0.15, + "output": 0.47, + "cache_read": 0.016, + "cache_write": 0.2 + } + }, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "hy3": { + "id": "hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 192000, + "output": 128000 + }, + "cost": { + "input": 0.14, + "output": 0.58, + "cache_read": 0.035 + } + }, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 32768 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "omen-alpha": { + "id": "omen-alpha", + "name": "Omen Alpha", + "description": "oH man anothEr aLPha ModEl", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0.2, + "output": 0.66, + "cache_read": 0.04 + } + }, + "muse-spark-1.2-contributor": { + "id": "muse-spark-1.2-contributor", + "name": "Muse Spark 1.2 Contributor", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-05", + "last_updated": "2026-08-05", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 + } + }, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro (New)", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.66, + "output": 1.98, + "cache_read": 0.022 + } + }, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } + } + }, + "hy4-preview": { + "id": "hy4-preview", + "name": "Hy4 preview", + "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-08-28", + "last_updated": "2026-08-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1024000, + "output": 64000 + }, + "cost": { + "input": 0.834, + "output": 2.501, + "cache_read": 0.042 + } + }, + "muse-spark-1.3-contributor": { + "id": "muse-spark-1.3-contributor", + "name": "Muse Spark 1.3 Contributor", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for coding and agentic workflows.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 + } + }, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + }, + "grok-4.6": { + "id": "grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } + } + }, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.7-plus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.04, + "cache_write": 0.5, + "tiers": [ + { + "input": 1.2, + "output": 4.8, + "cache_read": 0.12, + "cache_write": 1.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.2, + "output": 4.8, + "cache_read": 0.12, + "cache_write": 1.5 + } + } + }, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 + } + } + } + }, + "openreason": { + "id": "openreason", + "env": [ + "OPENREASON_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.openreason.app/v1", + "name": "OpenReason", + "doc": "https://openreason.app/docs", + "models": { + "deepseek-ai/deepseek-v4-flash-0731": { + "id": "deepseek-ai/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.1371, + "output": 0.2743 + } + }, + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.0022, + "output": 4.22 + } + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.1055, + "output": 0.422 + } + } + } + }, + "lmstudio": { + "id": "lmstudio", + "env": [ + "LMSTUDIO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "http://127.0.0.1:1234/v1", + "name": "LMStudio", + "doc": "https://lmstudio.ai/models", + "models": { + "qwen/qwen3-30b-a3b-2507": { + "id": "qwen/qwen3-30b-a3b-2507", + "name": "Qwen3 30B A3B 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-30", + "last_updated": "2025-07-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen/qwen3-coder-30b": { + "id": "qwen/qwen3-coder-30b", + "name": "Qwen3 Coder 30B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + } + } + }, + "aki-io": { + "id": "aki-io", + "env": [ + "AKI_IO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://aki.io/v1", + "name": "AKI.IO", + "doc": "https://aki.io/docs/", + "models": { + "qwen3.8-27b": { + "id": "qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 2.2, + "cache_read": 0.1 + } + }, + "deepseek-v4-flash-0731-284b": { + "id": "deepseek-v4-flash-0731-284b", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 81920 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.1 + } + }, + "qwen3.6-35b": { + "id": "qwen3.6-35b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 0.5 + } + }, + "glm5.3-754b": { + "id": "glm5.3-754b", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 81920 + }, + "cost": { + "input": 1, + "output": 3.5, + "cache_read": 0.25 + } + }, + "mistral4-119b": { + "id": "mistral4-119b", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 81920 + }, + "cost": { + "input": 0.2, + "output": 0.6 + } + }, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 0.55 + } + }, + "gemma4-26b": { + "id": "gemma4-26b", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.5 + } + } + } + }, + "tensorx": { + "id": "tensorx", + "env": [ + "TENSORX_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.tensorx.ai/v1", + "name": "TensorX", + "doc": "https://docs.tensorx.ai/", + "models": { + "deepseek/deepseek-v4-flash-0731": { + "id": "deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.25, + "output": 0.3, + "cache_read": 0.06 + } + }, + "deepseek/deepseek-v4.1-flash": { + "id": "deepseek/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.13 + } + }, + "deepseek/deepseek-v4-pro-0813": { + "id": "deepseek/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 64000 + }, + "cost": { + "input": 2, + "output": 4, + "cache_read": 0.5 + } + }, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 1.75, + "output": 3.5, + "cache_read": 0.4375, + "cache_write": 2.185 + } + }, + "deepseek/deepseek-r1-0528": { + "id": "deepseek/deepseek-r1-0528", + "name": "DeepSeek R1-0528", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 164000, + "output": 8192 + }, + "cost": { + "input": 0.66, + "output": 2.6, + "cache_read": 0.165, + "cache_write": 0.825 + } + }, + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 163840 + }, + "cost": { + "input": 0.3, + "output": 0.5, + "cache_read": 0.075, + "cache_write": 0.375 + } + }, + "z-ai/glm-5v-turbo": { + "id": "z-ai/glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.3, + "cache_write": 1.5 + } + }, + "z-ai/glm-5.3-flash": { + "id": "z-ai/glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 64000 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } + }, + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 202752 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.25, + "cache_write": 1.25 + } + }, + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.5, + "output": 4.5, + "cache_read": 0.375 + } + }, + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 202752 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.35, + "cache_write": 1.75 + } + }, + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.3, + "cache_write": 1.5 + } + }, + "z-ai/glm-5.3": { + "id": "z-ai/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 64000 + }, + "cost": { + "input": 1.75, + "output": 4.5, + "cache_read": 0.44 + } + }, + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.75 + } + }, + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1, + "output": 4, + "cache_read": 0.25, + "cache_write": 1.25 + } + }, + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 2.8, + "cache_read": 0.125, + "cache_write": 0.625 + } + }, + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.25, + "output": 4.5, + "cache_read": 0.3125 + } + }, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.075, + "cache_write": 0.375 + } + }, + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.4, + "output": 2, + "cache_read": 0.1 + } + }, + "qwen/qwen3-235b-a22b-2507": { + "id": "qwen/qwen3-235b-a22b-2507", + "name": "Qwen3 235B-A22B-2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-06-30", + "release_date": "2025-07-21", + "last_updated": "2025-07-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131000, + "output": 262144 + }, + "cost": { + "input": 0.072, + "output": 0.464, + "cache_read": 0.018, + "cache_write": 0.09 + } + }, + "qwen/qwen3.8-27b": { + "id": "qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 2.4, + "cache_read": 0.1 + } + }, + "qwen/qwen3.8-2.4t-a95b": { + "id": "qwen/qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 64000 + }, + "cost": { + "input": 2.5, + "output": 6, + "cache_read": 0.63 + } + }, + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 3.5, + "cache_read": 0.125, + "cache_write": 0.625 + } + }, + "qwen/qwen3.5-9b": { + "id": "qwen/qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.15, + "output": 0.2, + "cache_read": 0.0375, + "cache_write": 0.1875 + } + }, + "qwen/qwen3.8-flash-next": { + "id": "qwen/qwen3.8-flash-next", + "name": "Qwen3.8 Flash Next", + "description": "Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 64000 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } + } + } + }, + "longcat": { + "id": "longcat", + "env": [ + "LONGCAT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.longcat.chat/openai", + "name": "LongCat", + "doc": "https://longcat.chat/platform/docs/", + "models": { + "LongCat-2.0": { + "id": "LongCat-2.0", + "name": "LongCat-2.0", + "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", + "family": "longcat", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.75, + "output": 2.95, + "cache_read": 0.015 + } + } + } + }, + "chutes": { + "id": "chutes", + "env": [ + "CHUTES_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://llm.chutes.ai/v1", + "name": "Chutes", + "doc": "https://llm.chutes.ai/v1/models", + "models": { + "Nemotron-3-Nano-Omni-30B-TEE": { + "id": "Nemotron-3-Nano-Omni-30B-TEE", + "name": "Nemotron 3 Nano Omni 30B TEE", + "description": "Omni-modal model for text, vision, audio, and multimodal agent tasks", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-07-23", + "last_updated": "2026-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 0 + }, + "cost": { + "input": 0.0245, + "output": 0.0978, + "cache_read": 0.0024499999999999995 + } + }, + "unsloth/Mistral-Nemo-Instruct-2407-TEE": { + "id": "unsloth/Mistral-Nemo-Instruct-2407-TEE", + "name": "Mistral Nemo Instruct 2407 TEE", + "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.0245, + "output": 0.0978, + "cache_read": 0.0024499999999999995 + } + }, + "google/gemma-4-31B-turbo-TEE": { + "id": "google/gemma-4-31B-turbo-TEE", + "name": "gemma 4 31B turbo TEE", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 65536 + }, + "cost": { + "input": 0.12, + "output": 0.37, + "cache_read": 0.011999999999999997 + } + }, + "Qwen/Qwen3-32B-TEE": { + "id": "Qwen/Qwen3-32B-TEE", + "name": "Qwen3 32B TEE", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 40960, + "output": 40960 + }, + "cost": { + "input": 0.104, + "output": 0.416, + "cache_read": 0.010399999999999998 + } + }, + "Qwen/Qwen3.6-27B-TEE": { + "id": "Qwen/Qwen3.6-27B-TEE", + "name": "Qwen3.6 27B TEE", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2, + "cache_read": 0.029999999999999992 + } + }, + "Qwen/Qwen3-235B-A22B-Thinking-2507-TEE": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507-TEE", + "name": "Qwen3 235B A22B Thinking 2507 TEE", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07", + "last_updated": "2026-06-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.2989, + "output": 1.1957, + "cache_read": 0.029889999999999993 + } + }, + "Qwen/Qwen3.8-27B-TEE": { + "id": "Qwen/Qwen3.8-27B-TEE", + "name": "Qwen3.8 27B TEE", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-16", + "last_updated": "2026-08-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.24, + "output": 2.2, + "cache_read": 0.023999999999999994 + } + }, + "Qwen/Qwen3.5-397B-A17B-TEE": { + "id": "Qwen/Qwen3.5-397B-A17B-TEE", + "name": "Qwen3.5 397B A17B TEE", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.45, + "output": 3, + "cache_read": 0.04499999999999999 + } + }, + "deepseek-ai/DeepSeek-V4-Flash-0731-TEE": { + "id": "deepseek-ai/DeepSeek-V4-Flash-0731-TEE", + "name": "DeepSeek V4 Flash 0731 TEE", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-08-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.44, + "output": 1.32, + "cache_read": 0.04399999999999999 + } + }, + "deepseek-ai/DeepSeek-V3.2-TEE": { + "id": "deepseek-ai/DeepSeek-V3.2-TEE", + "name": "DeepSeek V3.2 TEE", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12", + "last_updated": "2026-06-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 65536 + }, + "cost": { + "input": 1, + "output": 1, + "cache_read": 0.09999999999999998 + } + }, + "moonshotai/Kimi-K3-TEE": { + "id": "moonshotai/Kimi-K3-TEE", + "name": "Kimi K3 TEE", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-29", + "last_updated": "2026-07-29", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.29999999999999993 + } + }, + "moonshotai/Kimi-K2.6-TEE": { + "id": "moonshotai/Kimi-K2.6-TEE", + "name": "Kimi K2.6 TEE", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65535 + }, + "cost": { + "input": 0.5, + "output": 2.85, + "cache_read": 0.04999999999999999 + } + }, + "zai-org/GLM-5.2-TEE": { + "id": "zai-org/GLM-5.2-TEE", + "name": "GLM 5.2 TEE", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 1.25, + "output": 3.95, + "cache_read": 0.12499999999999997 + } + }, + "zai-org/GLM-5.1-TEE": { + "id": "zai-org/GLM-5.1-TEE", + "name": "GLM 5.1 TEE", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 65535 + }, + "cost": { + "input": 0.98, + "output": 3.08, + "cache_read": 0.09799999999999998 + } + } + } + }, + "edenai": { + "id": "edenai", + "env": [ + "EDENAI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.edenai.run/v3", + "name": "Eden AI", + "doc": "https://docs.edenai.co", + "models": { + "deepinfra/nemotron-3-ultra-550b-a55b": { + "id": "deepinfra/nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B (Deep Infra)", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 128000 + }, + "cost": { + "input": 0.5, + "output": 2.2, + "cache_read": 0.1 + } + }, + "deepinfra/meta-models/Muse-Glimmer-30B": { + "id": "deepinfra/meta-models/Muse-Glimmer-30B", + "name": "Muse Glimmer 30B (Deep Infra)", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.04 + } + }, + "deepinfra/tencent/Hy3": { + "id": "deepinfra/tencent/Hy3", + "name": "Hy3 (Deep Infra)", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 128000 + }, + "cost": { + "input": 0.13, + "output": 0.53, + "cache_read": 0.033 + } + }, + "deepinfra/meta-llama/Llama-Guard-3-8B": { + "id": "deepinfra/meta-llama/Llama-Guard-3-8B", + "name": "Llama-Guard-3-8B (Deep Infra)", + "description": "Llama 3.1-based safety classifier for moderating prompts and model responses", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 4096 + }, + "cost": { + "input": 0.055, + "output": 0.055 + } + }, + "deepinfra/meta-llama/Llama-3.2-11B-Vision-Instruct": { + "id": "deepinfra/meta-llama/Llama-3.2-11B-Vision-Instruct", + "name": "Llama-3.2-11B-Vision-Instruct (Deep Infra)", + "description": "Open multimodal Llama model for image understanding, captioning, and visual QA", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 4096 + }, + "cost": { + "input": 0.345, + "output": 0.345 + } + }, + "deepinfra/meta-llama/Llama-3.3-70B-Instruct": { + "id": "deepinfra/meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama-3.3-70B-Instruct (Deep Infra)", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 4096 + }, + "cost": { + "input": 0.1, + "output": 0.32 + } + }, + "deepinfra/thinkingmachines/Inkling-Small": { + "id": "deepinfra/thinkingmachines/Inkling-Small", + "name": "Inkling Small (Deep Infra)", + "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-30", + "last_updated": "2026-07-30", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 1048576 + }, + "cost": { + "input": 0.45, + "output": 1.2, + "cache_read": 0.1 + } + }, + "deepinfra/thinkingmachines/Inkling": { + "id": "deepinfra/thinkingmachines/Inkling", + "name": "Inkling (Deep Infra)", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 1048576 + }, + "cost": { + "input": 0.95, + "output": 4.05, + "cache_read": 0.16 + } + }, + "deepinfra/google/gemma-3-27b-it": { + "id": "deepinfra/google/gemma-3-27b-it", + "name": "Gemma 3 27B IT (Deep Infra)", + "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.08, + "output": 0.16 + } + }, + "deepinfra/google/gemma-3-12b-it": { + "id": "deepinfra/google/gemma-3-12b-it", + "name": "Gemma 3 12B IT (Deep Infra)", + "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.05, + "output": 0.15 + } + }, + "deepinfra/google/gemma-3-4b-it": { + "id": "deepinfra/google/gemma-3-4b-it", + "name": "Gemma 3 4B IT (Deep Infra)", + "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.05, + "output": 0.1 + } + }, + "deepinfra/deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "deepinfra/deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731 (Deep Infra)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.06, + "output": 0.18, + "cache_read": 0.015 + } + }, + "deepinfra/deepseek-ai/DeepSeek-R1": { + "id": "deepinfra/deepseek-ai/DeepSeek-R1", + "name": "DeepSeek-R1 (Deep Infra)", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 32768 + }, + "cost": { + "input": 0.7, + "output": 2.4 + } + }, + "deepinfra/deepseek-ai/DeepSeek-V4-Pro-0813": { + "id": "deepinfra/deepseek-ai/DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813 (Deep Infra)", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 1.3, + "output": 2.6, + "cache_read": 0.1 + } + }, + "deepinfra/deepseek-ai/DeepSeek-V4.1-Flash": { + "id": "deepinfra/deepseek-ai/DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash (Deep Infra)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.2, + "output": 0.6, + "cache_read": 0.006 + } + }, + "deepinfra/deepseek-ai/DeepSeek-V3": { + "id": "deepinfra/deepseek-ai/DeepSeek-V3", + "name": "DeepSeek-V3 (Deep Infra)", + "description": "Open DeepSeek MoE chat model for coding, math, and general reasoning", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2024-12-26", + "last_updated": "2024-12-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 8192 + }, + "cost": { + "input": 0.32, + "output": 0.89 + } + }, + "deepinfra/deepseek-ai/DeepSeek-V3-0324": { + "id": "deepinfra/deepseek-ai/DeepSeek-V3-0324", + "name": "DeepSeek V3 0324 (Deep Infra)", + "description": "March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-03-24", + "last_updated": "2025-03-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 163840 + }, + "cost": { + "input": 0.24, + "output": 0.9, + "cache_read": 0.135 + } + }, + "deepinfra/stepfun-ai/Step-3.7-Flash": { + "id": "deepinfra/stepfun-ai/Step-3.7-Flash", + "name": "Step 3.7 Flash (Deep Infra)", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 256000 + }, + "cost": { + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 + } + }, + "deepinfra/stepfun-ai/Step-3.5-Flash": { + "id": "deepinfra/stepfun-ai/Step-3.5-Flash", + "name": "Step 3.5 Flash (Deep Infra)", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 256000 + }, + "cost": { + "input": 0.09, + "output": 0.3, + "cache_read": 0.02 + } + }, + "deepinfra/moonshotai/Kimi-K2.5": { + "id": "deepinfra/moonshotai/Kimi-K2.5", + "name": "Kimi K2.5 (Deep Infra)", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.45, + "output": 2.25, + "cache_read": 0.07 + } + }, + "deepinfra/zai-org/GLM-4.7-Flash": { + "id": "deepinfra/zai-org/GLM-4.7-Flash", + "name": "GLM-4.7-Flash (Deep Infra)", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 0.06, + "output": 0.4, + "cache_read": 0.01 + } + }, + "deepinfra/nvidia/Nemotron-3-Nano-30B-A3B": { + "id": "deepinfra/nvidia/Nemotron-3-Nano-30B-A3B", + "name": "Nemotron 3 Nano 30B A3B (Deep Infra)", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-15", + "last_updated": "2025-12-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.05, + "output": 0.2, + "cache_read": 0.025 + } + }, + "deepinfra/nvidia/Llama-3.1-Nemotron-70B-Instruct": { + "id": "deepinfra/nvidia/Llama-3.1-Nemotron-70B-Instruct", + "name": "Llama 3.1 Nemotron 70B Instruct (Deep Infra)", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.6, + "output": 0.6 + } + }, + "deepinfra/ByteDance/Seed-2.0-code": { + "id": "deepinfra/ByteDance/Seed-2.0-code", + "name": "Seed 2.0 Code (Deep Infra)", + "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 131072 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.1 + } + }, + "deepinfra/ByteDance/Seed-2.0-mini": { + "id": "deepinfra/ByteDance/Seed-2.0-mini", + "name": "Seed 2.0 Mini (Deep Infra)", + "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.02 + } + }, + "deepinfra/openai/gpt-oss-20b": { + "id": "deepinfra/openai/gpt-oss-20b", + "name": "GPT OSS 20B (Deep Infra)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.03, + "output": 0.14 + } + }, + "deepinfra/openai/gpt-oss-120b": { + "id": "deepinfra/openai/gpt-oss-120b", + "name": "GPT OSS 120B (Deep Infra)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.037, + "output": 0.17 + } + }, + "cerebras/gpt-oss-120b": { + "id": "cerebras/gpt-oss-120b", + "name": "GPT OSS 120B (Cerebras)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.35, + "output": 0.75, + "cache_read": 0.35 + } + }, + "groq/openai/gpt-oss-20b": { + "id": "groq/openai/gpt-oss-20b", + "name": "GPT OSS 20B (Groq)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.075, + "output": 0.3, + "cache_read": 0.0375 + } + }, + "groq/openai/gpt-oss-safeguard-20b": { + "id": "groq/openai/gpt-oss-safeguard-20b", + "name": "GPT OSS Safeguard 20B (Groq)", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.075, + "output": 0.3, + "cache_read": 0.0375 + } + }, + "groq/openai/gpt-oss-120b": { + "id": "groq/openai/gpt-oss-120b", + "name": "GPT OSS 120B (Groq)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 + } + }, + "zai/glm-4.6v": { + "id": "zai/glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 0.9, + "cache_read": 0.05 + } + }, + "zai/glm-5v-turbo": { + "id": "zai/glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", + "modalities": { + "input": [ + "image", + "text", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.24 + } + }, + "zai/glm-5.3-flash": { + "id": "zai/glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 + } + }, + "zai/glm-4.6": { + "id": "zai/glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 + } + }, + "zai/glm-5": { + "id": "zai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.2 + } + }, + "zai/glm-4.7": { + "id": "zai/glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 + } + }, + "zai/glm-5.2": { + "id": "zai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "zai/glm-5.1": { + "id": "zai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "zai/glm-5-turbo": { + "id": "zai/glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.24 + } + }, + "zai/glm-5.3": { + "id": "zai/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "anthropic/claude-opus-latest": { + "id": "anthropic/claude-opus-latest", + "name": "Claude Opus Latest (Claude Opus 5)", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-opus-4-5": { + "id": "anthropic/claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-opus-5-5": { + "id": "anthropic/claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 + } + }, + "anthropic/claude-fable-5-1": { + "id": "anthropic/claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "anthropic/claude-sonnet-latest": { + "id": "anthropic/claude-sonnet-latest", + "name": "Claude Sonnet Latest (Claude Sonnet 5)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "anthropic/claude-opus-4-5-20251101": { + "id": "anthropic/claude-opus-4-5-20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-fable-latest": { + "id": "anthropic/claude-fable-latest", + "name": "Claude Fable Latest (Claude Fable 5.1)", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "anthropic/claude-opus-4-8": { + "id": "anthropic/claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "cohere/command-r-08-2024": { + "id": "cohere/command-r-08-2024", + "name": "Command R", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "cohere/command-a-03-2025": { + "id": "cohere/command-a-03-2025", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 288000, + "output": 8000 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "cohere/command-r7b-12-2024": { + "id": "cohere/command-r7b-12-2024", + "name": "Command R7B", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2024-12-02", + "last_updated": "2024-12-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 132000, + "output": 4000 + }, + "cost": { + "input": 0.0375, + "output": 0.15 + } + }, + "cohere/command-r-plus-08-2024": { + "id": "cohere/command-r-plus-08-2024", + "name": "Command R+", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "deepseek/deepseek-v4-flash-vision-exp": { + "id": "deepseek/deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 + } + }, + "deepseek/deepseek-chat": { + "id": "deepseek/deepseek-chat", + "name": "DeepSeek Chat", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 384000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 + } + }, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.66, + "output": 1.98, + "cache_read": 0.022 + } + }, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 + } + }, + "databricks/databricks-inkling": { + "id": "databricks/databricks-inkling", + "name": "Inkling (Databricks)", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 1048576 + }, + "cost": { + "input": 1, + "output": 4.05, + "cache_read": 0.1, + "cache_write": 1.00002 + } + }, + "databricks/databricks-gpt-oss-20b": { + "id": "databricks/databricks-gpt-oss-20b", + "name": "GPT OSS 20B (Databricks)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.07, + "output": 0.30002, + "cache_read": 0.007, + "cache_write": 0.07 + } + }, + "databricks/databricks-deepseek-v4-flash-0731": { + "id": "databricks/databricks-deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731 (Databricks)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.014, + "cache_write": 0.14 + } + }, + "databricks/databricks-gpt-oss-120b@eu": { + "id": "databricks/databricks-gpt-oss-120b@eu", + "name": "GPT OSS 120B (Databricks, EU)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15001, + "output": 0.59997, + "cache_read": 0.015001, + "cache_write": 0.15001 + } + }, + "databricks/databricks-gpt-oss-120b": { + "id": "databricks/databricks-gpt-oss-120b", + "name": "GPT OSS 120B (Databricks)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15001, + "output": 0.59997, + "cache_read": 0.015001, + "cache_write": 0.15001 + } + }, + "databricks/databricks-deepseek-v4-pro-0813": { + "id": "databricks/databricks-deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813 (Databricks)", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 1.32, + "output": 3.959999, + "cache_read": 0.132, + "cache_write": 1.31999 + } + }, + "databricks/databricks-gpt-oss-20b@eu": { + "id": "databricks/databricks-gpt-oss-20b@eu", + "name": "GPT OSS 20B (Databricks, EU)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.07, + "output": 0.30002, + "cache_read": 0.007, + "cache_write": 0.07 + } + }, + "together_ai/meta-models/Muse-Glimmer-30B": { + "id": "together_ai/meta-models/Muse-Glimmer-30B", + "name": "Muse Glimmer 30B (Together AI)", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.35, + "output": 1.5, + "cache_read": 0.04 + } + }, + "together_ai/thinkingmachines/Inkling": { + "id": "together_ai/thinkingmachines/Inkling", + "name": "Inkling (Together AI)", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 1048576 + }, + "cost": { + "input": 1, + "output": 4.05, + "cache_read": 0.17 + } + }, + "together_ai/deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "together_ai/deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731 (Together AI)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.03 + } + }, + "together_ai/deepseek-ai/DeepSeek-V4-Pro-0813": { + "id": "together_ai/deepseek-ai/DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813 (Together AI)", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 1.32, + "output": 3.96, + "cache_read": 0.13 + } + }, + "together_ai/deepseek-ai/DeepSeek-V4.1-Flash": { + "id": "together_ai/deepseek-ai/DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash (Together AI)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.006 + } + }, + "together_ai/openai/gpt-oss-120b": { + "id": "together_ai/openai/gpt-oss-120b", + "name": "GPT OSS 120B (Together AI)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "azure/gpt-5.2-codex": { + "id": "azure/gpt-5.2-codex", + "name": "GPT-5.2 Codex (Azure)", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "azure/gpt-5.1-codex": { + "id": "azure/gpt-5.1-codex", + "name": "GPT-5.1 Codex (Azure)", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "azure/gpt-5.1-codex-max": { + "id": "azure/gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max (Azure)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "azure/gpt-5.1-codex-mini": { + "id": "azure/gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini (Azure)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "flexai/Step-3.7-Flash": { + "id": "flexai/Step-3.7-Flash", + "name": "Step 3.7 Flash (FlexAI)", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 256000 + }, + "cost": { + "input": 0.2, + "output": 1.15, + "cache_read": 0.03 + } + }, + "flexai/DeepSeek-V4-Flash-0731": { + "id": "flexai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731 (FlexAI)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.065, + "output": 0.18, + "cache_read": 0.0097 + } + }, + "flexai/Muse-Glimmer-30B": { + "id": "flexai/Muse-Glimmer-30B", + "name": "Muse Glimmer 30B (FlexAI)", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.1, + "cache_read": 0.045 + } + }, + "flexai/gpt-oss-20b": { + "id": "flexai/gpt-oss-20b", + "name": "GPT OSS 20B (FlexAI)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.018, + "output": 0.09, + "cache_read": 0.0027 + } + }, + "flexai/gpt-oss-120b": { + "id": "flexai/gpt-oss-120b", + "name": "GPT OSS 120B (FlexAI)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.037, + "output": 0.17, + "cache_read": 0.0055 + } + }, + "google/gemini-flash-latest": { + "id": "google/gemini-flash-latest", + "name": "Gemini Flash Latest (Gemini 3.8 Flash)", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "google/gemini-3.1-flash-image-preview": { + "id": "google/gemini-3.1-flash-image-preview", + "name": "Nano Banana 2 Preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", + "modalities": { + "input": [ + "image", + "text" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3 + } + }, + "google/gemini-2.5-flash-image": { + "id": "google/gemini-2.5-flash-image", + "name": "Nano Banana", + "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", + "modalities": { + "input": [ + "audio", + "image", + "text", + "video" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333, + "input_audio": 1 + } + }, + "google/gemini-3.6-flash": { + "id": "google/gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "google/gemini-3-pro-image-preview": { + "id": "google/gemini-3-pro-image-preview", + "name": "Nano Banana Pro Preview", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "input_audio": 2 + } + }, + "google/gemini-3.5-flash-lite": { + "id": "google/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333, + "input_audio": 0.3 + } + }, + "google/gemini-3.1-flash-image": { + "id": "google/gemini-3.1-flash-image", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "image", + "text" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 3 + } + }, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "input_audio": 2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "reasoning": 9, + "cache_read": 0.15, + "cache_write": 0.083333, + "input_audio": 3 + } + }, + "google/gemini-3-pro-image": { + "id": "google/gemini-3-pro-image", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "input_audio": 2 + } + }, + "google/gemini-3.1-flash-lite-image": { + "id": "google/gemini-3.1-flash-lite-image", + "name": "Nano Banana 2 Lite", + "description": "Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 4096 + }, + "cost": { + "input": 0.25, + "output": 1.5 + } + }, + "google/gemini-3.7-flash": { + "id": "google/gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "input_audio": 2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "reasoning": 3, + "cache_read": 0.05, + "cache_write": 0.083333, + "input_audio": 1 + } + }, + "google/gemini-pro-latest": { + "id": "google/gemini-pro-latest", + "name": "Gemini Pro Latest (Gemini 3.1 Pro Preview)", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "input_audio": 2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "google/gemini-3.8-flash": { + "id": "google/gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.083333, + "input_audio": 0.5 + } + }, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.083333, + "input_audio": 0.5 + } + }, + "xai/grok-latest": { + "id": "xai/grok-latest", + "name": "Grok Latest (Grok 4.6)", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } + } + }, + "xai/grok-4.7": { + "id": "xai/grok-4.7", + "name": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 3.2, + "output": 9.6, + "cache_read": 0.8, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 3.2, + "output": 9.6, + "cache_read": 0.8 + } + } + }, + "xai/grok-4.3": { + "id": "xai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 30000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "xai/grok-4.20-0309-reasoning": { + "id": "xai/grok-4.20-0309-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 30000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "xai/grok-4.5": { + "id": "xai/grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.3, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 0.6 + } + } + }, + "xai/grok-4.20-0309-non-reasoning": { + "id": "xai/grok-4.20-0309-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 30000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "xai/grok-build-0.1": { + "id": "xai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 1, + "output": 2, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 4, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 4, + "cache_read": 0.4 + } + } + }, + "xai/grok-4.6": { + "id": "xai/grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } + } + }, + "nebius/google/gemma-3-27b-it": { + "id": "nebius/google/gemma-3-27b-it", + "name": "Gemma 3 27B IT (Nebius)", + "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 110000, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.3, + "cache_read": 0.1 + } + }, + "nebius/deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "nebius/deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731 (Nebius)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1024000, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.14 + } + }, + "nebius/deepseek-ai/DeepSeek-V4-Pro-0813": { + "id": "nebius/deepseek-ai/DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813 (Nebius)", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 979000, + "output": 384000 + }, + "cost": { + "input": 1.32, + "output": 3.96, + "cache_read": 1.32 + } + }, + "nebius/deepseek-ai/DeepSeek-V4.1-Flash": { + "id": "nebius/deepseek-ai/DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash (Nebius)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048000, + "output": 384000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.3 + } + }, + "nebius/nvidia/nemotron-3-super-120b-a12b": { + "id": "nebius/nvidia/nemotron-3-super-120b-a12b", + "name": "Nemotron 3 Super 120B A12B (Nebius)", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.3, + "output": 0.9, + "cache_read": 0.3 + } + }, + "nebius/nvidia/Nemotron-3-Ultra-550b-a55b": { + "id": "nebius/nvidia/Nemotron-3-Ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B (Nebius)", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 128000 + }, + "cost": { + "input": 1, + "output": 3, + "cache_read": 1 + } + }, + "nebius/openai/gpt-oss-120b": { + "id": "nebius/openai/gpt-oss-120b", + "name": "GPT OSS 120B (Nebius)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.15 + } + }, + "scaleway/gemma-3-27b-it": { + "id": "scaleway/gemma-3-27b-it", + "name": "Gemma 3 27B IT (Scaleway)", + "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 40000, + "output": 131072 + }, + "cost": { + "input": 0.287125, + "output": 0.57425 + } + }, + "scaleway/deepseek-v4-flash-0731": { + "id": "scaleway/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731 (Scaleway)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 384000 + }, + "cost": { + "input": 0.45644, + "output": 0.91288, + "cache_read": 0.091288 + } + }, + "scaleway/llama-3.3-70b-instruct": { + "id": "scaleway/llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct (Scaleway)", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 1.02699, + "output": 1.02699 + } + }, + "scaleway/gpt-oss-120b": { + "id": "scaleway/gpt-oss-120b", + "name": "GPT OSS 120B (Scaleway)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.171165, + "output": 0.68466 + } + }, + "ionos/meta-llama/Llama-3.3-70B-Instruct": { + "id": "ionos/meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama-3.3-70B-Instruct (IONOS)", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.741715, + "output": 0.741715 + } + }, + "ionos/openai/gpt-oss-120b": { + "id": "ionos/openai/gpt-oss-120b", + "name": "GPT OSS 120B (IONOS)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.171165, + "output": 0.741715 + } + }, + "vertex/gemini-flash-latest": { + "id": "vertex/gemini-flash-latest", + "name": "Gemini Flash Latest (Gemini 3.8 Flash, Vertex AI)", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "vertex/gemini-2.5-flash-image": { + "id": "vertex/gemini-2.5-flash-image", + "name": "Nano Banana (Vertex AI)", + "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333, + "input_audio": 1 + } + }, + "vertex/gemini-3.1-flash-lite@eu": { + "id": "vertex/gemini-3.1-flash-lite@eu", + "name": "Gemini 3.1 Flash Lite (Vertex AI, EU)", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.083333, + "input_audio": 0.5 + } + }, + "vertex/gemini-3.5-flash-lite@eu": { + "id": "vertex/gemini-3.5-flash-lite@eu", + "name": "Gemini 3.5 Flash Lite (Vertex AI, EU)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333, + "input_audio": 0.3 + } + }, + "vertex/gemini-3.6-flash@eu": { + "id": "vertex/gemini-3.6-flash@eu", + "name": "Gemini 3.6 Flash (Vertex AI, EU)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "vertex/gemini-3.6-flash": { + "id": "vertex/gemini-3.6-flash", + "name": "Gemini 3.6 Flash (Vertex AI)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "vertex/gemini-3.5-flash-lite": { + "id": "vertex/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite (Vertex AI)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333, + "input_audio": 0.3 + } + }, + "vertex/gemini-3.1-flash-image": { + "id": "vertex/gemini-3.1-flash-image", + "name": "Nano Banana 2 (Vertex AI)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "image", + "text" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 3 + } + }, + "vertex/gemini-3.1-pro-preview": { + "id": "vertex/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview (Vertex AI)", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "input_audio": 2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 0.25 + } + } + }, + "vertex/gemini-3.5-flash": { + "id": "vertex/gemini-3.5-flash", + "name": "Gemini 3.5 Flash (Vertex AI)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "reasoning": 9, + "cache_read": 0.15, + "cache_write": 0.083333, + "input_audio": 3 + } + }, + "vertex/gemini-3-pro-image": { + "id": "vertex/gemini-3-pro-image", + "name": "Nano Banana Pro (Vertex AI)", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "input_audio": 2 + } + }, + "vertex/gemini-3.7-flash@us": { + "id": "vertex/gemini-3.7-flash@us", + "name": "Gemini 3.7 Flash (Vertex AI, US)", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "vertex/gemini-3.7-flash@eu": { + "id": "vertex/gemini-3.7-flash@eu", + "name": "Gemini 3.7 Flash (Vertex AI, EU)", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "vertex/gemini-3.1-flash-lite-image": { + "id": "vertex/gemini-3.1-flash-lite-image", + "name": "Nano Banana 2 Lite (Vertex AI)", + "description": "Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 4096 + }, + "cost": { + "input": 0.25, + "output": 1.5 + } + }, + "vertex/gemini-3.8-flash@eu": { + "id": "vertex/gemini-3.8-flash@eu", + "name": "Gemini 3.8 Flash (Vertex AI, EU)", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "vertex/gemini-3.7-flash": { + "id": "vertex/gemini-3.7-flash", + "name": "Gemini 3.7 Flash (Vertex AI)", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "vertex/gemini-3.6-flash@us": { + "id": "vertex/gemini-3.6-flash@us", + "name": "Gemini 3.6 Flash (Vertex AI, US)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "vertex/gemini-3.5-flash-lite@us": { + "id": "vertex/gemini-3.5-flash-lite@us", + "name": "Gemini 3.5 Flash Lite (Vertex AI, US)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333, + "input_audio": 0.3 + } + }, + "vertex/gemini-3-flash-preview": { + "id": "vertex/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview (Vertex AI)", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "reasoning": 3, + "cache_read": 0.05, + "cache_write": 0.083333, + "input_audio": 1 + } + }, + "vertex/gemini-3.1-flash-lite@us": { + "id": "vertex/gemini-3.1-flash-lite@us", + "name": "Gemini 3.1 Flash Lite (Vertex AI, US)", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.083333, + "input_audio": 0.5 + } + }, + "vertex/gemini-pro-latest": { + "id": "vertex/gemini-pro-latest", + "name": "Gemini Pro Latest (Gemini 3.1 Pro Preview, Vertex AI)", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "input_audio": 2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 0.25 + } + } + }, + "vertex/gemini-3.8-flash": { + "id": "vertex/gemini-3.8-flash", + "name": "Gemini 3.8 Flash (Vertex AI)", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "vertex/gemini-3.5-flash@eu": { + "id": "vertex/gemini-3.5-flash@eu", + "name": "Gemini 3.5 Flash (Vertex AI, EU)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "reasoning": 9, + "cache_read": 0.15, + "cache_write": 0.083333, + "input_audio": 3 + } + }, + "vertex/gemini-3.8-flash@us": { + "id": "vertex/gemini-3.8-flash@us", + "name": "Gemini 3.8 Flash (Vertex AI, US)", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "reasoning": 3.75, + "cache_read": 0.075, + "cache_write": 0.041667, + "input_audio": 0.75 + } + }, + "vertex/gemini-3.5-flash@us": { + "id": "vertex/gemini-3.5-flash@us", + "name": "Gemini 3.5 Flash (Vertex AI, US)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "reasoning": 9, + "cache_read": 0.15, + "cache_write": 0.083333, + "input_audio": 3 + } + }, + "vertex/gemini-3.1-flash-lite": { + "id": "vertex/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite (Vertex AI)", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.083333, + "input_audio": 0.5 + } + }, + "perplexityai/sonar-pro": { + "id": "perplexityai/sonar-pro", + "name": "Sonar Pro", + "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", + "family": "sonar-pro", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "cost": { + "input": 3, + "output": 15 + } + }, + "perplexityai/sonar-deep-research": { + "id": "perplexityai/sonar-deep-research", + "name": "Sonar Deep Research", + "description": "Sonar search model for autonomous research and citation-backed long-form reports", + "family": "sonar", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-02-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "reasoning": 3 + } + }, + "perplexityai/sonar": { + "id": "perplexityai/sonar", + "name": "Sonar", + "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", + "family": "sonar", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 127072, + "output": 4096 + }, + "cost": { + "input": 1, + "output": 1 + } + }, + "perplexityai/sonar-reasoning-pro": { + "id": "perplexityai/sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", + "family": "sonar-reasoning", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 2, + "output": 8 + } + }, + "cloudflare/@cf/meta/llama-guard-3-8b": { + "id": "cloudflare/@cf/meta/llama-guard-3-8b", + "name": "Llama-Guard-3-8B (Cloudflare)", + "description": "Llama 3.1-based safety classifier for moderating prompts and model responses", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 4096 + }, + "cost": { + "input": 0.484, + "output": 0.03 + } + }, + "cloudflare/@cf/aisingapore/gemma-sea-lion-v4-27b-it": { + "id": "cloudflare/@cf/aisingapore/gemma-sea-lion-v4-27b-it", + "name": "Gemma-SEA-LION-v4-27B-IT (Cloudflare)", + "description": "Gemma 3 27B tuned by AI Singapore for Southeast Asian languages and instruction following", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.351, + "output": 0.555 + } + }, + "cloudflare/@cf/deepseek-ai/deepseek-v4-flash-0731": { + "id": "cloudflare/@cf/deepseek-ai/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731 (Cloudflare)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1310720, + "output": 384000 + }, + "cost": { + "input": 0.44, + "output": 1.32, + "cache_read": 0.014 + } + }, + "cloudflare/@cf/deepseek-ai/deepseek-v4-pro-0813": { + "id": "cloudflare/@cf/deepseek-ai/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813 (Cloudflare)", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 + } + }, + "cloudflare/@cf/zai-org/glm-4.7-flash": { + "id": "cloudflare/@cf/zai-org/glm-4.7-flash", + "name": "GLM-4.7-Flash (Cloudflare)", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.0605, + "output": 0.4 + } + }, + "cloudflare/@cf/qwen/qwen2.5-coder-32b-instruct": { + "id": "cloudflare/@cf/qwen/qwen2.5-coder-32b-instruct", + "name": "Qwen2.5-Coder-32B-Instruct (Cloudflare)", + "description": "Open coding-focused Qwen model for code generation, repair, and repository reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2024-11-12", + "last_updated": "2024-11-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 8192 + }, + "cost": { + "input": 0.66, + "output": 1 + } + }, + "cloudflare/@cf/openai/gpt-oss-20b": { + "id": "cloudflare/@cf/openai/gpt-oss-20b", + "name": "GPT OSS 20B (Cloudflare)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.2, + "output": 0.3 + } + }, + "cloudflare/@cf/openai/gpt-oss-120b": { + "id": "cloudflare/@cf/openai/gpt-oss-120b", + "name": "GPT OSS 120B (Cloudflare)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.35, + "output": 0.75 + } + }, + "minimax/MiniMax-M3": { + "id": "minimax/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 512000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "minimax/MiniMax-M2.1": { + "id": "minimax/MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 + } + }, + "minimax/MiniMax-M2.5": { + "id": "minimax/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 + } + }, + "minimax/MiniMax-M2.7": { + "id": "minimax/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "minimax/MiniMax-M2": { + "id": "minimax/MiniMax-M2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2 + } + }, + "tensorx/deepseek/deepseek-v4-flash-0731": { + "id": "tensorx/deepseek/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731 (TensorX)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.25, + "output": 0.3, + "cache_read": 0.0625 + } + }, + "tensorx/deepseek/deepseek-v4.1-flash": { + "id": "tensorx/deepseek/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (TensorX)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.125 + } + }, + "tensorx/deepseek/deepseek-v4-pro-0813": { + "id": "tensorx/deepseek/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813 (TensorX)", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 2, + "output": 4, + "cache_read": 0.5 + } + }, + "tensorx/moonshotai/kimi-k2.5": { + "id": "tensorx/moonshotai/kimi-k2.5", + "name": "Kimi K2.5 (TensorX)", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 2.8, + "cache_read": 0.125 + } + }, + "amazon/google.gemma-3-12b-it": { + "id": "amazon/google.gemma-3-12b-it", + "name": "Gemma 3 12B IT (Amazon Bedrock)", + "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 131072 + }, + "cost": { + "input": 0.09, + "output": 0.29 + } + }, + "amazon/google.gemma-3-4b-it": { + "id": "amazon/google.gemma-3-4b-it", + "name": "Gemma 3 4B IT (Amazon Bedrock)", + "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 131072 + }, + "cost": { + "input": 0.04, + "output": 0.08 + } + }, + "amazon/openai.gpt-oss-safeguard-20b@us": { + "id": "amazon/openai.gpt-oss-safeguard-20b@us", + "name": "GPT OSS Safeguard 20B (Amazon Bedrock, US)", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 131072 + }, + "cost": { + "input": 0.07, + "output": 0.2 + } + }, + "amazon/openai.gpt-oss-safeguard-20b": { + "id": "amazon/openai.gpt-oss-safeguard-20b", + "name": "GPT OSS Safeguard 20B (Amazon Bedrock)", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 131072 + }, + "cost": { + "input": 0.07, + "output": 0.2 + } + }, + "amazon/amazon.nova-lite-v1:0@us": { + "id": "amazon/amazon.nova-lite-v1:0@us", + "name": "Nova Lite (US)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.06, + "output": 0.24, + "cache_read": 0.015 + } + }, + "amazon/mistral.voxtral-mini-3b-2507@us": { + "id": "amazon/mistral.voxtral-mini-3b-2507@us", + "name": "Voxtral Mini 3B 2507 (Amazon Bedrock, US)", + "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", + "family": "voxtral", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.04, + "output": 0.04 + } + }, + "amazon/amazon.nova-micro-v1:0": { + "id": "amazon/amazon.nova-micro-v1:0", + "name": "Nova Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 10000 + }, + "cost": { + "input": 0.035, + "output": 0.14, + "cache_read": 0.00875 + } + }, + "amazon/amazon.nova-pro-v1:0": { + "id": "amazon/amazon.nova-pro-v1:0", + "name": "Nova Pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.8, + "output": 3.2, + "cache_read": 0.2 + } + }, + "amazon/mistral.voxtral-small-24b-2507": { + "id": "amazon/mistral.voxtral-small-24b-2507", + "name": "Voxtral Small 24B 2507 (Amazon Bedrock)", + "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", + "family": "voxtral", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "amazon/amazon.nova-lite-v1:0": { + "id": "amazon/amazon.nova-lite-v1:0", + "name": "Nova Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.06, + "output": 0.24, + "cache_read": 0.015 + } + }, + "amazon/mistral.voxtral-small-24b-2507@us": { + "id": "amazon/mistral.voxtral-small-24b-2507@us", + "name": "Voxtral Small 24B 2507 (Amazon Bedrock, US)", + "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", + "family": "voxtral", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "amazon/google.gemma-3-12b-it@us": { + "id": "amazon/google.gemma-3-12b-it@us", + "name": "Gemma 3 12B IT (Amazon Bedrock, US)", + "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 131072 + }, + "cost": { + "input": 0.09, + "output": 0.29 + } + }, + "amazon/google.gemma-3-27b-it": { + "id": "amazon/google.gemma-3-27b-it", + "name": "Gemma 3 27B IT (Amazon Bedrock)", + "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 131072 + }, + "cost": { + "input": 0.23, + "output": 0.38 + } + }, + "amazon/google.gemma-3-27b-it@us": { + "id": "amazon/google.gemma-3-27b-it@us", + "name": "Gemma 3 27B IT (Amazon Bedrock, US)", + "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 131072 + }, + "cost": { + "input": 0.23, + "output": 0.38 + } + }, + "amazon/mistral.pixtral-large-2502-v1:0": { + "id": "amazon/mistral.pixtral-large-2502-v1:0", + "name": "Pixtral Large (25.02) (Amazon Bedrock)", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-04-08", + "last_updated": "2025-04-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "amazon/amazon.nova-micro-v1:0@us": { + "id": "amazon/amazon.nova-micro-v1:0@us", + "name": "Nova Micro (US)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 10000 + }, + "cost": { + "input": 0.035, + "output": 0.14, + "cache_read": 0.00875 + } + }, + "amazon/zai.glm-4.7-flash@us": { + "id": "amazon/zai.glm-4.7-flash@us", + "name": "GLM-4.7-Flash (Amazon Bedrock, US)", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 203000, + "output": 131072 + }, + "cost": { + "input": 0.07, + "output": 0.4 + } + }, + "amazon/moonshotai.kimi-k2.5": { + "id": "amazon/moonshotai.kimi-k2.5", + "name": "Kimi K2.5 (Amazon Bedrock)", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 3 + } + }, + "amazon/amazon.nova-pro-v1:0@us": { + "id": "amazon/amazon.nova-pro-v1:0@us", + "name": "Nova Pro (US)", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.8, + "output": 3.2, + "cache_read": 0.2 + } + }, + "amazon/mistral.pixtral-large-2502-v1:0@us": { + "id": "amazon/mistral.pixtral-large-2502-v1:0@us", + "name": "Pixtral Large (25.02) (Amazon Bedrock, US)", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-04-08", + "last_updated": "2025-04-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "amazon/zai.glm-4.7-flash": { + "id": "amazon/zai.glm-4.7-flash", + "name": "GLM-4.7-Flash (Amazon Bedrock)", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 203000, + "output": 131072 + }, + "cost": { + "input": 0.07, + "output": 0.4 + } + }, + "amazon/moonshot.kimi-k2-thinking": { + "id": "amazon/moonshot.kimi-k2-thinking", + "name": "Kimi K2 Thinking (Amazon Bedrock)", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 2.5 + } + }, + "amazon/google.gemma-3-4b-it@us": { + "id": "amazon/google.gemma-3-4b-it@us", + "name": "Gemma 3 4B IT (Amazon Bedrock, US)", + "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 131072 + }, + "cost": { + "input": 0.04, + "output": 0.08 + } + }, + "amazon/mistral.voxtral-mini-3b-2507": { + "id": "amazon/mistral.voxtral-mini-3b-2507", + "name": "Voxtral Mini 3B 2507 (Amazon Bedrock)", + "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", + "family": "voxtral", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.04, + "output": 0.04 + } + }, + "ovhcloud/gpt-oss-20b": { + "id": "ovhcloud/gpt-oss-20b", + "name": "GPT OSS 20B (OVHcloud)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.05, + "output": 0.18 + } + }, + "ovhcloud/gpt-oss-120b": { + "id": "ovhcloud/gpt-oss-120b", + "name": "GPT OSS 120B (OVHcloud)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.09, + "output": 0.47 + } + }, + "fireworks_ai/gpt-oss-120b": { + "id": "fireworks_ai/gpt-oss-120b", + "name": "GPT OSS 120B (Fireworks AI)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.014 + } + }, + "fireworks_ai/accounts/fireworks/models/deepseek-v4-flash-0731": { + "id": "fireworks_ai/accounts/fireworks/models/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731 (Fireworks AI)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.22, + "output": 0.66, + "cache_read": 0.007 + } + }, + "fireworks_ai/accounts/fireworks/models/muse-glimmer-30b": { + "id": "fireworks_ai/accounts/fireworks/models/muse-glimmer-30b", + "name": "Muse Glimmer 30B (Fireworks AI)", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.35, + "output": 1.5, + "cache_read": 0.04 + } + }, + "fireworks_ai/accounts/fireworks/models/deepseek-v4-pro-0813": { + "id": "fireworks_ai/accounts/fireworks/models/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813 (Fireworks AI)", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 + } + }, + "fireworks_ai/accounts/fireworks/models/inkling": { + "id": "fireworks_ai/accounts/fireworks/models/inkling", + "name": "Inkling (Fireworks AI)", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 1, + "output": 4.05, + "cache_read": 0.17 + } + }, + "mistral/codestral-latest": { + "id": "mistral/codestral-latest", + "name": "Codestral (latest)", + "description": "Mistral code model for completions, refactors, and developer IDE workflows", + "family": "codestral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-05-29", + "last_updated": "2025-01-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 4096 + }, + "cost": { + "input": 0.3, + "output": 0.9, + "cache_read": 0.03 + } + }, + "mistral/mistral-large-latest": { + "id": "mistral/mistral-large-latest", + "name": "Mistral Large (latest)", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.2 + } + }, + "mistral/mistral-small-latest": { + "id": "mistral/mistral-small-latest", + "name": "Mistral Small (latest)", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 256000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 + } + }, + "mistral/devstral-medium-latest": { + "id": "mistral/devstral-medium-latest", + "name": "Devstral 2 (latest)", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.4, + "output": 2 + } + }, + "mistral/devstral-2512": { + "id": "mistral/devstral-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.4, + "output": 2, + "cache_read": 0.04 + } + }, + "mistral/mistral-medium-2505": { + "id": "mistral/mistral-medium-2505", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.4, + "output": 2 + } + }, + "mistral/magistral-medium-latest": { + "id": "mistral/magistral-medium-latest", + "name": "Magistral Medium (latest)", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-20", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 16384 + }, + "cost": { + "input": 1.5, + "output": 7.5, + "cache_read": 0.15 + } + }, + "mistral/mistral-small-2603": { + "id": "mistral/mistral-small-2603", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 256000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 + } + }, + "mistral/mistral-large-2512": { + "id": "mistral/mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.05 + } + }, + "mistral/voxtral-small-latest": { + "id": "mistral/voxtral-small-latest", + "name": "Voxtral Small (latest)", + "description": "Instruct model with native audio input for speech understanding and tool use", + "family": "voxtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32000 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 + } + }, + "mistral/mistral-medium-2604": { + "id": "mistral/mistral-medium-2604", + "name": "Mistral Medium 3.5", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.5, + "output": 7.5, + "cache_read": 0.15 + } + }, + "mistral/mistral-medium-latest": { + "id": "mistral/mistral-medium-latest", + "name": "Mistral Medium (latest)", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.5, + "output": 7.5, + "cache_read": 0.15 + } + }, + "qwen/qwen3-vl-235b-a22b-instruct": { + "id": "qwen/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 1.6 + } + }, + "qwen/qwq-plus": { + "id": "qwen/qwq-plus", + "name": "QwQ Plus", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-03-05", + "last_updated": "2025-03-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.8, + "output": 2.4 + } + }, + "qwen/qwen3.8-27b": { + "id": "qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.1, + "cache_write": 0.625 + } + }, + "qwen/qwen3-vl-235b-a22b-thinking": { + "id": "qwen/qwen3-vl-235b-a22b-thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 4 + } + }, + "qwen/qwen3.8-2.4t-a95b": { + "id": "qwen/qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 + } + }, + "qwen/deepseek-v4-flash-0731": { + "id": "qwen/deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731 (Alibaba)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.22, + "output": 0.66, + "cache_read": 0.022 + } + }, + "qwen/qwen3-235b-a22b-instruct-2507": { + "id": "qwen/qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B-A22B Instruct 2507", + "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-21", + "last_updated": "2025-07-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0.23, + "output": 0.92 + } + }, + "qwen/qwen-max": { + "id": "qwen/qwen-max", + "name": "Qwen Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 8192 + }, + "cost": { + "input": 1.6, + "output": 6.4, + "cache_read": 0.32 + } + }, + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 1.2 + } + }, + "qwen/qwen3.8-max": { + "id": "qwen/qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 + } + }, + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 1.5 + } + }, + "qwen/deepseek-v4.1-flash": { + "id": "qwen/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash (Alibaba)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 + } + }, + "qwen/qwen3-coder-next@eu": { + "id": "qwen/qwen3-coder-next@eu", + "name": "Qwen3 Coder Next (EU)", + "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 1.5 + } + }, + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 1.2, + "output": 6, + "cache_read": 0.24, + "cache_write": 1.5 + } + }, + "qwen/qwen3-coder-480b-a35b-instruct": { + "id": "qwen/qwen3-coder-480b-a35b-instruct", + "name": "Qwen3-Coder 480B-A35B Instruct", + "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 7.5 + } + }, + "qwen/qwen3-coder-30b-a3b-instruct": { + "id": "qwen/qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.45, + "output": 2.25 + } + }, + "qwen/qwen-vl-max": { + "id": "qwen/qwen-vl-max", + "name": "Qwen-VL Max", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-08", + "last_updated": "2025-08-13", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.8, + "output": 3.2, + "cache_read": 0.16 + } + }, + "qwen/qwen3-max@eu": { + "id": "qwen/qwen3-max@eu", + "name": "Qwen3 Max (EU)", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 1.2, + "output": 6, + "cache_read": 0.24, + "cache_write": 1.5 + } + }, + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 1.2 + } + }, + "qwen/qwen-vl-plus": { + "id": "qwen/qwen-vl-plus", + "name": "Qwen-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-08-15", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.21, + "output": 0.63, + "cache_read": 0.042 + } + }, + "qwen/qwen3-coder-flash": { + "id": "qwen/qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 1.5, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "qwen/deepseek-v4-pro-0813": { + "id": "qwen/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813 (Alibaba)", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.66, + "output": 1.98, + "cache_read": 0.066 + } + }, + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.2, + "cache_write": 1.25 + } + }, + "qwen/qwen3.8-flash": { + "id": "qwen/qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.47, + "cache_read": 0.016, + "cache_write": 0.2 + } + }, + "qwen/qwen3.8-max-0902": { + "id": "qwen/qwen3.8-max-0902", + "name": "Qwen3.8 Max 0902", + "description": "2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 + } + }, + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } + } + }, + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } + } + }, + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16385, + "output": 4096 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } + } + }, + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "pdf", + "image", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2 Pro", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "image", + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 21, + "output": 168 + } + }, + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", + "modalities": { + "input": [ + "image", + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 + } + }, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 + } + }, + "openai/gpt-4": { + "id": "openai/gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8191, + "output": 8192 + }, + "cost": { + "input": 30, + "output": 60 + } + }, + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "image", + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 + } + }, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 + } + }, + "openai/o1": { + "id": "openai/o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 15, + "output": 60, + "cache_read": 7.5 + } + }, + "openai/gpt-latest": { + "id": "openai/gpt-latest", + "name": "GPT Latest (GPT-6 Astra)", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } + } + }, + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", + "modalities": { + "input": [ + "image", + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 272000 + }, + "cost": { + "input": 15, + "output": 120 + } + }, + "openai/gpt-4o-2024-08-06": { + "id": "openai/gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-08-06", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "openai/gpt-6-astra": { + "id": "openai/gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } + } + }, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "image", + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 + } + }, + "openai/o3-pro": { + "id": "openai/o3-pro", + "name": "o3-pro", + "description": "High-effort o3 tier for difficult technical reasoning and careful answers", + "family": "o-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", + "modalities": { + "input": [ + "text", + "pdf", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 20, + "output": 80 + } + }, + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "pdf", + "image", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } + } + }, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "pdf", + "image", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "openai/gpt-4o-2024-11-20": { + "id": "openai/gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "openai/gpt-mini-latest": { + "id": "openai/gpt-mini-latest", + "name": "GPT Mini Latest (GPT-5.4 mini)", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "pdf", + "image", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 + } + }, + "openai/gpt-6-luna": { + "id": "openai/gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } + } + }, + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } + } + }, + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 10, + "output": 30 + } + }, + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10 + } + } + }, + "openai/gpt-6-sol": { + "id": "openai/gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } + } + }, + "openai/o1-pro": { + "id": "openai/o1-pro", + "name": "o1-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2025-03-19", + "last_updated": "2025-03-19", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 150, + "output": 600 + } + }, + "openai/gpt-pro-latest": { + "id": "openai/gpt-pro-latest", + "name": "GPT Pro Latest (GPT-5.5 Pro)", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } + } + }, + "infomaniak/mistralai/Ministral-3-14B-Instruct-2512": { + "id": "infomaniak/mistralai/Ministral-3-14B-Instruct-2512", + "name": "Ministral 3 14B (Infomaniak)", + "description": "Open vision-language model for efficient local deployment, instruction following, and tool use", + "family": "ministral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 100000, + "output": 262144 + }, + "cost": { + "input": 0.34233, + "output": 0.45644 + } + }, + "moonshot/kimi-k3": { + "id": "moonshot/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "moonshot/kimi-k2.7-code-highspeed": { + "id": "moonshot/kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.9, + "output": 8, + "cache_read": 0.38 + } + }, + "moonshot/kimi-k2.6": { + "id": "moonshot/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } + }, + "moonshot/kimi-k2.7-code": { + "id": "moonshot/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + } + } + }, + "stepfun": { + "id": "stepfun", + "env": [ + "STEPFUN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.stepfun.com/v1", + "name": "StepFun (China)", + "doc": "https://platform.stepfun.com/docs/zh/overview/concept", + "models": { + "step-1-32k": { + "id": "step-1-32k", + "name": "Step 1 (32K)", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-01-01", + "last_updated": "2026-02-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "input": 32768, + "output": 32768 + }, + "cost": { + "input": 2.05, + "output": 9.59, + "cache_read": 0.41 + } + }, + "step-tts-2": { + "id": "step-tts-2", + "name": "Step TTS 2", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "step", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-03-01", + "last_updated": "2026-07-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "step-5-preview": { + "id": "step-5-preview", + "name": "Step 5 Preview", + "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "release_date": "2026-09-16", + "last_updated": "2026-09-20", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 1000000, + "output": 1000000 + }, + "cost": { + "input": 0.959, + "output": 2.741, + "cache_read": 0.048 + } + }, + "step-3.5-flash-2603": { + "id": "step-3.5-flash-2603", + "name": "Step 3.5 Flash 2603", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 256000, + "output": 256000 + }, + "cost": { + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 + } + }, + "step-2-16k": { + "id": "step-2-16k", + "name": "Step 2 (16K)", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-01-01", + "last_updated": "2026-02-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16384, + "input": 16384, + "output": 8192 + }, + "cost": { + "input": 5.21, + "output": 16.44, + "cache_read": 1.04 + } + }, + "stepaudio-2.5-tts": { + "id": "stepaudio-2.5-tts", + "name": "StepAudio 2.5 TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "step", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-16", + "last_updated": "2026-07-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "stepaudio-2.5-asr": { + "id": "stepaudio-2.5-asr", + "name": "StepAudio 2.5 ASR", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "step", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-24", + "last_updated": "2026-07-02", + "modalities": { + "input": [ + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "step-3.5-flash": { + "id": "step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-06-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 256000, + "output": 256000 + }, + "cost": { + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 + } + }, + "step-3.7-flash": { + "id": "step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-06-29", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 256000, + "output": 256000 + }, + "cost": { + "input": 0.185, + "output": 1.11, + "cache_read": 0.037 + } + } + } + }, + "hpc-ai": { + "id": "hpc-ai", + "env": [ + "HPC_AI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.hpc-ai.com/inference/v1", + "name": "HPC-AI", + "doc": "https://www.hpc-ai.com/doc/docs/quickstart/", + "models": { + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5 + } + }, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1002000, + "output": 128000 + }, + "cost": { + "input": 1.74, + "output": 3.48, + "cache_read": 0.145 + } + }, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 128000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 + } + }, + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.1 + } + }, + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + }, + "zai-org/glm-5.2": { + "id": "zai-org/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "zai-org/glm-5.1": { + "id": "zai-org/glm-5.1", + "name": "GLM 5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-08", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202000, + "output": 202000 + }, + "cost": { + "input": 0.615, + "output": 2.46, + "cache_read": 0.133 + } + }, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196000, + "output": 195000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 + } + }, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + } + } + }, + "v0": { + "id": "v0", + "env": [ + "V0_API_KEY" + ], + "npm": "@ai-sdk/vercel", + "name": "v0", + "doc": "https://sdk.vercel.ai/providers/ai-sdk-providers/vercel", + "models": { + "v0-1.5-lg": { + "id": "v0-1.5-lg", + "name": "v0-1.5-lg", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "v0", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-06-09", + "last_updated": "2025-06-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 512000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75 + } + }, + "v0-1.5-md": { + "id": "v0-1.5-md", + "name": "v0-1.5-md", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "v0", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-06-09", + "last_updated": "2025-06-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 32000 + }, + "cost": { + "input": 3, + "output": 15 + } + }, + "v0-1.0-md": { + "id": "v0-1.0-md", + "name": "v0-1.0-md", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "v0", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 32000 + }, + "cost": { + "input": 3, + "output": 15 + } + } + } + }, + "tencent-coding-plan": { + "id": "tencent-coding-plan", + "env": [ + "TENCENT_CODING_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.lkeap.cloud.tencent.com/coding/v3", + "name": "Tencent Coding Plan (China)", + "doc": "https://cloud.tencent.com/document/product/1772/128947", + "models": { + "hunyuan-2.0-thinking": { + "id": "hunyuan-2.0-thinking", + "name": "Tencent HY 2.0 Think", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-03-08", + "last_updated": "2026-03-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 202752, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "hunyuan-2.0-instruct": { + "id": "hunyuan-2.0-instruct", + "name": "Tencent HY 2.0 Instruct", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-03-08", + "last_updated": "2026-03-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "tc-code-latest": { + "id": "tc-code-latest", + "name": "Auto", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-03-08", + "last_updated": "2026-03-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "hunyuan-turbos": { + "id": "hunyuan-turbos", + "name": "Hunyuan-TurboS", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-03-08", + "last_updated": "2026-03-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "hunyuan-t1": { + "id": "hunyuan-t1", + "name": "Hunyuan-T1", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-03-08", + "last_updated": "2026-03-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + } + } + }, + "tempr": { + "id": "tempr", + "env": [ + "TEMPR_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.temprhq.io/v1", + "name": "Tempr", + "doc": "https://temprhq.io/docs/gateway-reference.html", + "models": { + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "anthropic/claude-opus-4-5": { + "id": "anthropic/claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-sonnet-4-5": { + "id": "anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic/claude-fable-5-1": { + "id": "anthropic/claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "anthropic/claude-opus-4-5-20251101": { + "id": "anthropic/claude-opus-4-5-20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "anthropic/claude-opus-4-8": { + "id": "anthropic/claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-sonnet-4-5-20250929": { + "id": "anthropic/claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-haiku-4-5-20251001": { + "id": "anthropic/claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "google/gemini-flash-latest": { + "id": "google/gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "input_audio": 0.75 + } + }, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + } + }, + "google/gemini-flash-lite-latest": { + "id": "google/gemini-flash-lite-latest", + "name": "Gemini Flash-Lite Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 + } + }, + "google/gemini-embedding-2": { + "id": "google/gemini-embedding-2", + "name": "Gemini Embedding 2", + "description": "Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space", + "family": "gemini", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-11", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 1 + }, + "cost": { + "input": 0.2, + "output": 0 + } + }, + "google/gemini-3.6-flash": { + "id": "google/gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "input_audio": 0.75 + } + }, + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + } + }, + "google/gemini-3.5-flash-lite": { + "id": "google/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 + } + }, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "google/gemini-3.7-flash": { + "id": "google/gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "input_audio": 0.75 + } + }, + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 + } + }, + "google/gemini-3.8-flash": { + "id": "google/gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "input_audio": 0.75 + } + }, + "google/gemini-embedding-001": { + "id": "google/gemini-embedding-001", + "name": "Gemini Embedding 001", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "gemini", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-05", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2048, + "output": 1 + }, + "cost": { + "input": 0.15, + "output": 0 + } + }, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 + } + } + } + }, + "inception": { + "id": "inception", + "env": [ + "INCEPTION_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.inceptionlabs.ai/v1/", + "name": "Inception", + "doc": "https://docs.inceptionlabs.ai/get-started/models", + "models": { + "mercury-edit-2": { + "id": "mercury-edit-2", + "name": "Mercury Edit 2", + "description": "Code editing dLLM for autocomplete (FIM) and next-edit suggestions", + "family": "mercury", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-03-30", + "last_updated": "2026-03-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 8192 + }, + "cost": { + "input": 0.25, + "output": 0.75, + "cache_read": 0.025 + } + }, + "mercury-2.5": { + "id": "mercury-2.5", + "name": "Mercury 2.5", + "description": "Mercury 2.5 is the fastest reasoning LLM, and the latest diffusion LLM (dLLM) from Inception", + "family": "mercury", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-11-01", + "release_date": "2026-09-08", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 260000, + "output": 65536 + }, + "cost": { + "input": 0.04, + "output": 0.15, + "cache_read": 0.004 + } + }, + "mercury-2": { + "id": "mercury-2", + "name": "Mercury 2", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 50000 + }, + "cost": { + "input": 0.25, + "output": 0.75, + "cache_read": 0.025 + } + } + } + }, + "modelis": { + "id": "modelis", + "env": [ + "MODELIS_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://modelishub.com/v1", + "name": "Modelis", + "doc": "https://modelishub.com/pricing", + "models": { + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 10 + } + }, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5 + } + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50 + } + }, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + }, + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25 + } + }, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.435, + "output": 0.87 + } + }, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15 + } + }, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.0983, + "output": 0.1966 + } + }, + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 3, + "output": 9 + } + }, + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.768, + "output": 3.072 + } + } + } + }, + "opencode": { + "id": "opencode", + "env": [ + "OPENCODE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://opencode.ai/zen/v1", + "name": "OpenCode Zen", + "doc": "https://opencode.ai/docs/zen", + "models": { + "ling-3.0-flash-fin-free": { + "id": "ling-3.0-flash-fin-free", + "name": "Ling 3.0 Flash Fin Free", + "description": "Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution", + "family": "ling", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } + } + }, + "qwen3.6-plus-free": { + "id": "qwen3.6-plus-free", + "name": "Qwen3.6 Plus Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen-free", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 30, + "output": 180, + "cache_read": 30 + } + }, + "mimo-v2-pro-free": { + "id": "mimo-v2-pro-free", + "name": "MiMo V2 Pro Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo-pro-free", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 64000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "muse-spark-1.2-contributor-free": { + "id": "muse-spark-1.2-contributor-free", + "name": "Muse Spark 1.2 Free", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse-free", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-05", + "last_updated": "2026-08-05", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "muse-spark-1.3": { + "id": "muse-spark-1.3", + "name": "Muse Spark 1.3", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 + } + }, + "glm-5-free": { + "id": "glm-5-free", + "name": "GLM-5 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "glm-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "trinity-large-preview-free": { + "id": "trinity-large-preview-free", + "name": "Trinity Large Preview", + "description": "Legacy model retained for compatibility with older integrations", + "family": "trinity", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-01-27", + "last_updated": "2026-01-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-5.5-pro": { + "id": "gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 30, + "output": 180, + "cache_read": 30 + } + }, + "grok-4.7": { + "id": "grok-4.7", + "name": "Grok 4.7 (30% Off)", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.4, + "output": 4.2, + "cache_read": 0.35, + "tiers": [ + { + "input": 2.8, + "output": 8.4, + "cache_read": 0.7, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.8, + "output": 8.4, + "cache_read": 0.7 + } + } + }, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-01-14", + "last_updated": "2026-01-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "ling-3.0-tiny-free": { + "id": "ling-3.0-tiny-free", + "name": "Ling-3.0-tiny Free", + "description": "Compact MoE model for responsive agents, instruction following, and multi-turn conversations", + "family": "ling", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-06", + "last_updated": "2026-08-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.07, + "output": 8.5, + "cache_read": 0.107 + } + }, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 + } + }, + "kimi-k2.5-free": { + "id": "kimi-k2.5-free", + "name": "Kimi K2.5 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-free", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "glm-4.7-free": { + "id": "glm-4.7-free", + "name": "GLM-4.7 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "glm-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "glm-4.6": { + "id": "glm-4.6", + "name": "GLM-4.6", + "description": "Legacy model retained for compatibility with older integrations", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.1 + } + }, + "laguna-s-2.1-free": { + "id": "laguna-s-2.1-free", + "name": "Laguna S 2.1 Free", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 32000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "GPT-5 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.07, + "output": 8.5, + "cache_read": 0.107 + } + }, + "qwen3-coder": { + "id": "qwen3-coder", + "name": "Qwen3 Coder", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "status": "deprecated", + "cost": { + "input": 0.45, + "output": 1.8 + } + }, + "minimax-m3-free": { + "id": "minimax-m3-free", + "name": "MiniMax-M3 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax-m3-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-31", + "last_updated": "2026-05-31", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.5", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 + } + }, + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.2 + } + }, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.006 + } + }, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 + } + }, + "deepseek-v4-flash-vision-exp": { + "id": "deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 + } + }, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } + }, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "interleaved": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } + } + }, + "claude-opus-5-5": { + "id": "claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 + } + }, + "claude-fable-5-1": { + "id": "claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "gemini-3.6-flash": { + "id": "gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google" + }, + "cost": { + "input": 1.5, + "output": 7.5, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2.5, + "cache_read": 0.4 + } + }, + "minimax-m2.5-free": { + "id": "minimax-m2.5-free", + "name": "MiniMax-M2.5 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "gpt-5.3-codex-spark": { + "id": "gpt-5.3-codex-spark", + "name": "GPT-5.3 Codex Spark", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex-spark", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 128000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google" + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 + } + }, + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } + } + }, + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.3, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 0.6 + } + } + }, + "ring-2.6-1t-free": { + "id": "ring-2.6-1t-free", + "name": "Ring 2.6 1T Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "ring-1t-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-05-08", + "last_updated": "2026-05-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 66000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0 + } + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.08 + } + }, + "longcat-2.0-free": { + "id": "longcat-2.0-free", + "name": "LongCat-2.0 Free", + "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", + "family": "longcat", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.07, + "output": 8.5, + "cache_read": 0.107 + } + }, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "claude-3-5-haiku": { + "id": "claude-3-5-haiku", + "name": "Claude Haiku 3.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-haiku", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07-31", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 + } + }, + "hy3-preview-free": { + "id": "hy3-preview-free", + "name": "Hy3 preview Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "hy3-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 64000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "gemini-3-flash": { + "id": "gemini-3-flash", + "name": "Gemini 3 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google" + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05 + } + }, + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google" + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "space-bunny-free": { + "id": "space-bunny-free", + "name": "Space Bunny Free", + "description": "Anonymous preview reasoning model for coding, agentic tasks, tool use, and multimodal input", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "release_date": "2026-09-23", + "last_updated": "2026-09-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "input": 524288, + "output": 524288 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "nemotron-3-super-free": { + "id": "nemotron-3-super-free", + "name": "Nemotron 3 Super Free", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2026-02", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "minimax-m2.1-free": { + "id": "minimax-m2.1-free", + "name": "MiniMax-M2.1 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "mimo-v2.6-flash-free": { + "id": "mimo-v2.6-flash-free", + "name": "MiMo-V2.6-Flash Free", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "gemini-3-pro": { + "id": "gemini-3-pro", + "name": "Gemini 3 Pro", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/google" + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "nemotron-3-ultra-free": { + "id": "nemotron-3-ultra-free", + "name": "Nemotron 3 Ultra Free", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2026-02", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "claude-sonnet-4": { + "id": "claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } + } + }, + "muse-spark-1.2": { + "id": "muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-05", + "last_updated": "2026-08-05", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 + } + }, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Legacy model retained for compatibility with older integrations", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.1 + } + }, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 512000, + "output": 128000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } + } + }, + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0.15, + "output": 0.47, + "cache_read": 0.016, + "cache_write": 0.2 + } + }, + "gpt-5.1-codex-max": { + "id": "gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "mimo-v2.5-free": { + "id": "mimo-v2.5-free", + "name": "MiMo V2.5 Free", + "description": "MiMo omni model for text, image, video, audio, and agents", + "family": "mimo-v2.5-free", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "grok-code": { + "id": "grok-code", + "name": "Grok Code Fast 1", + "description": "Legacy model retained for compatibility with older integrations", + "family": "grok", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-20", + "last_updated": "2025-08-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "ling-2.6-flash-free": { + "id": "ling-2.6-flash-free", + "name": "Ling 2.6 Flash Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "ling-flash-free", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262100, + "output": 32800 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0 + } + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "deepseek-v4-flash-free": { + "id": "deepseek-v4-flash-free", + "name": "DeepSeek V4 Flash Free", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "hy3-free": { + "id": "hy3-free", + "name": "Hy3 Free", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hy3-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 190000, + "input": 192000, + "output": 64000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "nemotron-3.5-lightning-free": { + "id": "nemotron-3.5-lightning-free", + "name": "Nemotron 3.5 Lightning Free", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google" + }, + "cost": { + "input": 1.5, + "output": 7.5, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "kimi-k2": { + "id": "kimi-k2", + "name": "Kimi K2", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2.5, + "cache_read": 0.4 + } + }, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "mimo-v2-omni-free": { + "id": "mimo-v2-omni-free", + "name": "MiMo V2 Omni Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo-omni-free", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 64000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "grok-build-0.1": { + "id": "grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1, + "output": 2, + "cache_read": 0.2 + } + }, + "ling-3.0-flash-free": { + "id": "ling-3.0-flash-free", + "name": "Ling-3.0-flash Free", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-23", + "last_updated": "2026-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "gemini-3.8-flash": { + "id": "gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google" + }, + "cost": { + "input": 1.5, + "output": 7.5, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "north-mini-code-free": { + "id": "north-mini-code-free", + "name": "North Mini Code Free", + "description": "Cohere coding model for practical software engineering and agentic edits", + "family": "north-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09-23", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 64000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-6-luna": { + "id": "gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } + } + }, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 1.74, + "output": 3.84, + "cache_read": 0.145 + } + }, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "interleaved": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "big-pickle": { + "id": "big-pickle", + "name": "Big Pickle", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "big-pickle", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-10-17", + "last_updated": "2025-10-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "input": 160000, + "output": 32000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625 + } + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25 + } + } + }, + "mimo-v2-flash-free": { + "id": "mimo-v2-flash-free", + "name": "MiMo V2 Flash Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo-flash-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-12-16", + "last_updated": "2025-12-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + }, + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.07, + "output": 8.5, + "cache_read": 0.107 + } + }, + "grok-4.6": { + "id": "grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } + } + }, + "gemini-3.1-pro": { + "id": "gemini-3.1-pro", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google" + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "minimax-m2.1": { + "id": "minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.1 + } + }, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10 + } + } + }, + "x-preview-f-free": { + "id": "x-preview-f-free", + "name": "Ox Alpha Free (Unlimited)", + "description": "Stealth reasoning model for coding, agentic tasks, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-08-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "muse-spark-1.3-contributor-free": { + "id": "muse-spark-1.3-contributor-free", + "name": "Muse Spark 1.3 Free", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for coding and agentic workflows.", + "family": "muse-free", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 + } + }, + "gpt-6-sol": { + "id": "gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } + } + } + } + }, + "kenari": { + "id": "kenari", + "env": [ + "KENARI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://kenari.id/v1", + "name": "Kenari", + "doc": "https://kenari.id/docs", + "models": { + "gpt-5-4-mini": { + "id": "gpt-5-4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-5-6-luna": { + "id": "gpt-5-6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gemma-4-31b-it": { + "id": "gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "grok-4-5": { + "id": "grok-4-5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gemini-2-5-flash": { + "id": "gemini-2-5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nemotron-3-nano-30b-a3b": { + "id": "nemotron-3-nano-30b-a3b", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-12-15", + "last_updated": "2025-12-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "glm-5-3-flash": { + "id": "glm-5-3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "kimi-k2-7-code": { + "id": "kimi-k2-7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-5-6-terra": { + "id": "gpt-5-6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gemini-2-5-flash-lite": { + "id": "gemini-2-5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mimo-v2-5": { + "id": "mimo-v2-5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "grok-build-0-1": { + "id": "grok-build-0-1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "glm-4-7-flash:free": { + "id": "glm-4-7-flash:free", + "name": "GLM-4.7-Flash (Free)", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-image-2": { + "id": "gpt-image-2", + "name": "GPT-Image-2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 272000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen3-8-max": { + "id": "qwen3-8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "minimax-m2-7-highspeed": { + "id": "minimax-m2-7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mimo-v2-5-pro": { + "id": "mimo-v2-5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nemotron-3-super-120b-a12b": { + "id": "nemotron-3-super-120b-a12b", + "name": "Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gemini-3-7-flash": { + "id": "gemini-3-7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "step-3-7-flash:free": { + "id": "step-3-7-flash:free", + "name": "Step 3.7 Flash (Free)", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 256000, + "output": 256000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "kimi-k2-6:free": { + "id": "kimi-k2-6:free", + "name": "Kimi K2.6 (Free)", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-5-6-sol": { + "id": "gpt-5-6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nemotron-3-ultra-550b-a55b": { + "id": "nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gemini-3-1-pro": { + "id": "gemini-3-1-pro", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "glm-5-2": { + "id": "glm-5-2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistral-large:free": { + "id": "mistral-large:free", + "name": "Mistral Large (Free)", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mimo-v2-5:free": { + "id": "mimo-v2-5:free", + "name": "MiMo-V2.5 (Free)", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "minimax-m2-7": { + "id": "minimax-m2-7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "hy3:free": { + "id": "hy3:free", + "name": "Hy3 (Free)", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 192000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "kimi-k2-6": { + "id": "kimi-k2-6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistral-medium-3-5:free": { + "id": "mistral-medium-3-5:free", + "name": "Mistral Medium 3.5 (Free)", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 512000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "glm-5-3": { + "id": "glm-5-3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "deepseek-v4-1-flash": { + "id": "deepseek-v4-1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen3-7-plus": { + "id": "qwen3-7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gemini-3-1-flash-lite": { + "id": "gemini-3-1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "hy3": { + "id": "hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 192000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "glm-5-1": { + "id": "glm-5-1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nemotron-3-super-120b-a12b:free": { + "id": "nemotron-3-super-120b-a12b:free", + "name": "Nemotron 3 Super 120B A12B (Free)", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-5-5": { + "id": "gpt-5-5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "grok-imagine-image-2-0": { + "id": "grok-imagine-image-2-0", + "name": "Grok Imagine Image 2.0", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-08-07", + "last_updated": "2026-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 8000, + "output": 0 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "deepseek-v4-flash:free": { + "id": "deepseek-v4-flash:free", + "name": "DeepSeek V4 Flash (Free)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "grok-4-6": { + "id": "grok-4-6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "whisper-large-v3-turbo": { + "id": "whisper-large-v3-turbo", + "name": "Whisper Large v3 Turbo", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", + "modalities": { + "input": [ + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 448, + "output": 448 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gemini-3-6-flash": { + "id": "gemini-3-6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gemini-3-1-flash-tts": { + "id": "gemini-3-1-flash-tts", + "name": "Gemini 3.1 Flash TTS Preview", + "description": "Low-latency speech generation with steerable prompts and expressive audio tags", + "family": "gemini-flash", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-15", + "last_updated": "2026-04-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "kimi-k2-7-code:free": { + "id": "kimi-k2-7-code:free", + "name": "Kimi K2.7 Code (Free)", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gemini-3-5-flash": { + "id": "gemini-3-5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "step-3-7-flash": { + "id": "step-3-7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 256000, + "output": 256000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0, + "output": 0 + } + } + } + }, + "kimi-code-plan-global": { + "id": "kimi-code-plan-global", + "env": [ + "KIMI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.kimi.ai/coding/v1", + "name": "Kimi For Coding (kimi.ai)", + "doc": "https://www.kimi.ai/code/docs/en/kimi-code/models.html", + "models": { + "kimi-for-coding-highspeed": { + "id": "kimi-for-coding-highspeed", + "name": "Kimi For Coding HighSpeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "kimi-for-coding": { + "id": "kimi-for-coding", + "name": "kimi-for-coding", + "description": "Kimi coding model with more efficient thinking and up to 1M context, available through Kimi Code", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "k3-256k": { + "id": "k3-256k", + "name": "Kimi K3-256K", + "description": "256K-context version of Kimi K3, reducing token consumption for shorter coding sessions", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "k3": { + "id": "k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + } + } + }, + "trustedrouter": { + "id": "trustedrouter", + "env": [ + "TRUSTEDROUTER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.trustedrouter.com/v1", + "name": "TrustedRouter", + "doc": "https://trustedrouter.com/docs", + "models": { + "trustedrouter/cheap": { + "id": "trustedrouter/cheap", + "name": "Cheap", + "description": "TrustedRouter low-cost routing alias that prefers inexpensive healthy model endpoints.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-01", + "last_updated": "2026-06-27", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + } + }, + "trustedrouter/synth-code": { + "id": "trustedrouter/synth-code", + "name": "Synth Code", + "description": "TrustedRouter code synthesis orchestration alias that combines multiple model responses into one answer.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-20", + "last_updated": "2026-06-27", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + } + }, + "trustedrouter/e2e": { + "id": "trustedrouter/e2e", + "name": "End-to-End Encrypted", + "description": "TrustedRouter privacy routing alias for end-to-end encrypted provider routes where available.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-27", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + } + }, + "trustedrouter/zdr": { + "id": "trustedrouter/zdr", + "name": "Zero Data Retention", + "description": "TrustedRouter privacy routing alias that prefers zero data retention model endpoints.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-27", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + } + }, + "trustedrouter/fast": { + "id": "trustedrouter/fast", + "name": "Fast", + "description": "TrustedRouter speed routing alias that prefers low-latency healthy model endpoints.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-27", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + } + }, + "trustedrouter/synth": { + "id": "trustedrouter/synth", + "name": "Synth", + "description": "TrustedRouter synthesis orchestration alias that combines multiple model responses into one answer.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-20", + "last_updated": "2026-06-27", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + } + }, + "trustedrouter/auto": { + "id": "trustedrouter/auto", + "name": "Auto", + "description": "TrustedRouter automatic routing alias that chooses a healthy supported model endpoint for the request.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-01", + "last_updated": "2026-06-27", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + } + } + } + }, + "wafer.ai": { + "id": "wafer.ai", + "env": [ + "WAFER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://pass.wafer.ai/v1", + "name": "Wafer", + "doc": "https://docs.wafer.ai/wafer-pass", + "models": { + "Kimi-K2.6": { + "id": "Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Kimi K2.6 sparse MoE model with a 262K context window. Available serverless and not included in standard Wafer Pass. Non-ZDR only: requests with `Wafer-ZDR: required` are rejected.", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 1.14, + "output": 4.8, + "cache_read": 0.19, + "cache_write": 0 + } + }, + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 512000 + }, + "cost": { + "input": 0.33, + "output": 1.32, + "cache_read": 0.07, + "cache_write": 0, + "tiers": [ + { + "input": 0.66, + "output": 2.64, + "cache_read": 0.13, + "cache_write": 0, + "tier": { + "type": "context", + "size": 512000 + } + } + ], + "context_over_200k": { + "input": 0.66, + "output": 2.64, + "cache_read": 0.13, + "cache_write": 0 + } + } + }, + "GLM-5.1": { + "id": "GLM-5.1", + "name": "GLM-5.1", + "description": "General Language Model 5.1 — high-quality bilingual (EN/ZH) generation with strong coding and reasoning capabilities.", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-07", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.1, + "cache_write": 0 + } + }, + "GLM-5.2": { + "id": "GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4.1, + "cache_read": 0.2, + "cache_write": 0 + } + }, + "glm5.2-fast": { + "id": "glm5.2-fast", + "name": "GLM5.2-Fast", + "description": "The same model served for high TPS.", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 10.25, + "cache_read": 0.5, + "cache_write": 0 + } + } + } + }, + "zhipuai": { + "id": "zhipuai", + "env": [ + "ZHIPU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://open.bigmodel.cn/api/paas/v4", + "name": "Zhipu AI", + "doc": "https://docs.z.ai/guides/overview/pricing", + "models": { + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 5, + "output": 22, + "cache_read": 1.2, + "cache_write": 0 + } + }, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.03, + "cache_write": 0 + } + }, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0 + } + }, + "glm-5.3-flashx": { + "id": "glm-5.3-flashx", + "name": "GLM-5.3-FlashX", + "description": "High-speed GLM-5.3-Flash serving option for coding and agent workflows", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-18", + "last_updated": "2026-09-18", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.37, + "output": 1.25, + "cache_read": 0.075, + "cache_write": 0 + } + }, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 + } + }, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-27", + "last_updated": "2026-03-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 + } + }, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 + } + }, + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 98304 + }, + "cost": { + "input": 0.2, + "output": 1.1, + "cache_read": 0.03, + "cache_write": 0 + } + }, + "glm-4.7-flashx": { + "id": "glm-4.7-flashx", + "name": "GLM-4.7-FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0.07, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0 + } + }, + "glm-4.5-flash": { + "id": "glm-4.5-flash", + "name": "GLM-4.5-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 98304 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 + } + }, + "glm-4.7-flash": { + "id": "glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "glm-4.5v": { + "id": "glm-4.5v", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 64000, + "output": 16384 + }, + "cost": { + "input": 0.6, + "output": 1.8 + } + }, + "glm-4.6v-flash": { + "id": "glm-4.6v-flash", + "name": "GLM-4.6V-Flash", + "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "glm-4.6": { + "id": "glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 + } + }, + "glm-4.5": { + "id": "glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 98304 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 + } + }, + "glm-4.6v": { + "id": "glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 0.9 + } + } + } + }, + "lynkr": { + "id": "lynkr", + "env": [ + "LYNKR_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "http://127.0.0.1:8081/v1", + "name": "Lynkr", + "doc": "https://github.com/Fast-Editor/Lynkr", + "models": { + "lynkr-auto": { + "id": "lynkr-auto", + "name": "Lynkr Auto (complexity routing)", + "description": "Virtual model: Lynkr scores each request on complexity and routes it to the tier model the user configured (local Ollama/llama.cpp for simple requests, configured cloud providers for complex ones).", + "family": "auto", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-12-03", + "last_updated": "2026-07-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + } + } + }, + "meganova": { + "id": "meganova", + "env": [ + "MEGANOVA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.meganova.ai/v1", + "name": "Meganova", + "doc": "https://docs.meganova.ai", + "models": { + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "XiaomiMiMo/MiMo-V2-Flash": { + "id": "XiaomiMiMo/MiMo-V2-Flash", + "name": "MiMo V2 Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-12-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32000 + }, + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "Qwen/Qwen3.5-Plus": { + "id": "Qwen/Qwen3.5-Plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02", + "last_updated": "2026-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.4, + "output": 2.4, + "reasoning": 2.4 + } + }, + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 262000 + }, + "cost": { + "input": 0.09, + "output": 0.6 + } + }, + "Qwen/Qwen2.5-VL-32B-Instruct": { + "id": "Qwen/Qwen2.5-VL-32B-Instruct", + "name": "Qwen2.5 VL 32B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-03-24", + "last_updated": "2025-03-24", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 16384, + "output": 16384 + }, + "cost": { + "input": 0.2, + "output": 0.6 + } + }, + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-25", + "last_updated": "2025-08-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 164000, + "output": 164000 + }, + "cost": { + "input": 0.27, + "output": 1 + } + }, + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-03", + "last_updated": "2025-12-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 164000, + "output": 164000 + }, + "cost": { + "input": 0.26, + "output": 0.38 + } + }, + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 64000 + }, + "cost": { + "input": 0.5, + "output": 2.15 + } + }, + "deepseek-ai/DeepSeek-V3-0324": { + "id": "deepseek-ai/DeepSeek-V3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-03-24", + "last_updated": "2025-03-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 163840 + }, + "cost": { + "input": 0.25, + "output": 0.88 + } + }, + "deepseek-ai/DeepSeek-V3.2-Exp": { + "id": "deepseek-ai/DeepSeek-V3.2-Exp", + "name": "DeepSeek V3.2 Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-10", + "last_updated": "2025-10-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 164000, + "output": 164000 + }, + "cost": { + "input": 0.27, + "output": 0.4 + } + }, + "MiniMaxAI/MiniMax-M2.1": { + "id": "MiniMaxAI/MiniMax-M2.1", + "name": "MiniMax M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 131072 + }, + "cost": { + "input": 0.28, + "output": 1.2 + } + }, + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2 + } + }, + "mistralai/Mistral-Nemo-Instruct-2407": { + "id": "mistralai/Mistral-Nemo-Instruct-2407", + "name": "Mistral Nemo Instruct 2407", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 65536 + }, + "cost": { + "input": 0.02, + "output": 0.04 + } + }, + "mistralai/Mistral-Small-3.2-24B-Instruct-2506": { + "id": "mistralai/Mistral-Small-3.2-24B-Instruct-2506", + "name": "Mistral Small 3.2 24B Instruct", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2026-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.45, + "output": 2.8 + } + }, + "moonshotai/Kimi-K2-Thinking": { + "id": "moonshotai/Kimi-K2-Thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 2.6 + } + }, + "zai-org/GLM-4.7": { + "id": "zai-org/GLM-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 0.2, + "output": 0.8 + } + }, + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 0.8, + "output": 2.56 + } + }, + "zai-org/GLM-4.6": { + "id": "zai-org/GLM-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 0.45, + "output": 1.9 + } + } + } + }, + "ovhcloud": { + "id": "ovhcloud", + "env": [ + "OVHCLOUD_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://oai.endpoints.kepler.ai.cloud.ovh.net/v1", + "name": "OVHcloud AI Endpoints", + "doc": "https://www.ovhcloud.com/en/public-cloud/ai-endpoints/catalog//", + "models": { + "qwen3.8-27b": { + "id": "qwen3.8-27b", + "name": "Qwen3.8-27B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.47, + "output": 3.19 + } + }, + "qwen3guard-gen-8b": { + "id": "qwen3guard-gen-8b", + "name": "Qwen3Guard-Gen-8B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-01-22", + "last_updated": "2026-01-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistral-7b-instruct-v0.3": { + "id": "mistral-7b-instruct-v0.3", + "name": "Mistral-7B-Instruct-v0.3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-04-01", + "last_updated": "2025-04-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 65536, + "output": 65536 + }, + "cost": { + "input": 0.11, + "output": 0.11 + } + }, + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder-30B-A3B-Instruct", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-28", + "last_updated": "2025-10-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.07, + "output": 0.26 + } + }, + "qwen3guard-gen-0.6b": { + "id": "qwen3guard-gen-0.6b", + "name": "Qwen3Guard-Gen-0.6B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-01-22", + "last_updated": "2026-01-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "gpt-oss-20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2025-08-28", + "last_updated": "2025-08-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.05, + "output": 0.18 + } + }, + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5-9B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.12, + "output": 0.18 + } + }, + "meta-llama-3_3-70b-instruct": { + "id": "meta-llama-3_3-70b-instruct", + "name": "Meta-Llama-3_3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-04-01", + "last_updated": "2025-04-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.74, + "output": 0.74 + } + }, + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5-397B-A17B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-18", + "last_updated": "2026-05-18", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.71, + "output": 4.25 + } + }, + "qwen2.5-vl-72b-instruct": { + "id": "qwen2.5-vl-72b-instruct", + "name": "Qwen2.5-VL-72B-Instruct", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-03-31", + "last_updated": "2025-03-31", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 1.01, + "output": 1.01 + } + }, + "qwen3.6-27b": { + "id": "qwen3.6-27b", + "name": "Qwen3.6-27B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.47, + "output": 3.19 + } + }, + "mistral-small-3.2-24b-instruct-2506": { + "id": "mistral-small-3.2-24b-instruct-2506", + "name": "Mistral-Small-3.2-24B-Instruct-2506", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-16", + "last_updated": "2025-07-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.31 + } + }, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2025-08-28", + "last_updated": "2025-08-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.09, + "output": 0.47 + } + }, + "mistral-nemo-instruct-2407": { + "id": "mistral-nemo-instruct-2407", + "name": "Mistral-Nemo-Instruct-2407", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-11-20", + "last_updated": "2024-11-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 65536, + "output": 65536 + }, + "cost": { + "input": 0.14, + "output": 0.14 + } + } + } + }, + "requesty": { + "id": "requesty", + "env": [ + "REQUESTY_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://router.requesty.ai/v1", + "name": "Requesty", + "doc": "https://requesty.ai/solution/llm-routing/models", + "models": { + "ring-2.6-1t": { + "id": "ring-2.6-1t", + "name": "ring-2.6-1t", + "description": "Inclusion AI ring-2.6-1t", + "family": "ring", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "release_date": "2026-05-08", + "last_updated": "2026-05-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.3, + "output": 2.5 + } + }, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.75, + "output": 16.5, + "cache_read": 0.275 + } + }, + "nemotron-3-ultra-nvfp4": { + "id": "nemotron-3-ultra-nvfp4", + "name": "nemotron-3-ultra-nvfp4", + "description": "Nemotron-3-Ultra-550B-A55B-NVFP4 is a frontier-scale large language model (LLM) trained by NVIDIA, designed to deliver strong agentic, reasoning, and conversational capabilities. It is optimized for the most demanding workloads, including complex multi-step agents, long-context analysis, and high-accuracy reasoning over code, math, and science. The model employs a hybrid Latent Mixture-of-Experts (LatentMoE) architecture, utilizing interleaved Mamba-2 and MoE layers, along with select Attention layers. Like the Super model, the Ultra model incorporates Multi-Token Prediction (MTP) layers for faster text generation and improved quality, and it is trained using an NVFP4 pre-training recipe to maximize compute efficiency. The model has 55B active parameters and 550B parameters in total.", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 + } + }, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180, + "cache_read": 30 + } + }, + "claude-opus-5@eu": { + "id": "claude-opus-5@eu", + "name": "Claude Opus 5 (EU)", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "gemma-4-31b-it": { + "id": "gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "ling-2.6-1t": { + "id": "ling-2.6-1t", + "name": "ling-2.6-1t", + "description": "Inclusion AI ling-2.6-1t", + "family": "ling", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.3, + "output": 2.5 + } + }, + "gpt-4o-mini@eu": { + "id": "gpt-4o-mini@eu", + "name": "GPT-4o mini (EU)", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16000 + }, + "cost": { + "input": 0.165, + "output": 0.66, + "cache_read": 0.0825 + } + }, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 2.5, + "output": 7.5, + "cache_read": 0.25, + "cache_write": 3.125 + } + }, + "nemotron-3.5-lightning-30b-a3b": { + "id": "nemotron-3.5-lightning-30b-a3b", + "name": "nemotron-3.5-lightning-30b-a3b", + "description": "NVIDIA Nemotron 3.5 Lightning 30B-A3B is a hybrid Mamba-2 + MoE + Attention model with 30B total and 3B active parameters, pre-trained on over 20T tokens with an NVFP4 recipe and Multi-Token Prediction for fast generation. Up to 1M token context for long-running autonomous agents, sub-agent workhorse deployments, and agentic workflows. Supports reasoning and tool calling. English and coding languages plus Spanish, French, German, Italian, and Japanese. Open weights under the OpenMDW License Agreement v1.1. Part of the NVIDIA Nemotron family.", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 + } + }, + "gpt-5.5-pro": { + "id": "gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180 + } + }, + "mistral-medium-3-5@eu": { + "id": "mistral-medium-3-5@eu", + "name": "mistral-medium-3-5@eu", + "description": "Mistral Medium 3.5 is a dense 128B instruction following model from Mistral AI. It supports text and image inputs with text output, and is designed for agentic workflows, coding, and complex multi step reasoning. It is particularly strong at reliable multi tool calling and long horizon tasks, with a 256K context window, configurable reasoning effort per request, and a custom vision encoder that handles variable image sizes and aspect ratios. Self hostable on as few as four GPUs and available under open weights.", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.65, + "output": 8.25, + "cache_read": 1.65 + } + }, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "gpt-5@eu": { + "id": "gpt-5@eu", + "name": "GPT-5 (EU)", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.375, + "output": 11, + "cache_read": 0.1375 + } + }, + "deepseek-v4-flash-0731@eu": { + "id": "deepseek-v4-flash-0731@eu", + "name": "DeepSeek V4 Flash 0731 (EU)", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.28, + "output": 0.56, + "cache_read": 0.07 + } + }, + "seed-2.0-pro": { + "id": "seed-2.0-pro", + "name": "Seed 2.0 Pro", + "description": "Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.1 + } + }, + "kimi-k3@eu": { + "id": "kimi-k3@eu", + "name": "Kimi K3 (EU)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 262144 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.45 + } + }, + "seed-1.8": { + "id": "seed-1.8", + "name": "seed-1.8", + "description": "Optimized specifically for multimodal agent scenarios. It features enhanced agent capabilities, upgraded multimodal comprehension, and more flexible context management.", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "release_date": "2026-05-27", + "last_updated": "2026-05-27", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.05 + } + }, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 262144 + }, + "cost": { + "input": 0.2, + "output": 0.6, + "cache_read": 0.07 + } + }, + "qwen3.5-27b": { + "id": "qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.26, + "output": 2.6 + } + }, + "mistral-medium-latest@eu": { + "id": "mistral-medium-latest@eu", + "name": "Mistral Medium (latest) (EU)", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.44, + "output": 2.2, + "cache_read": 0.44 + } + }, + "glm-5.2-fast": { + "id": "glm-5.2-fast", + "name": "glm-5.2-fast", + "description": "GLM-5.2 introduces a robust 1M-token context and advanced, multi-effort coding capabilities to significantly enhance performance on long-horizon tasks. Its new IndexShare architecture and improved MTP layer simultaneously boost efficiency by reducing per-token FLOPs and increasing speculative decoding lengths. A 743B-parameter model in Zhipu AI's GLM series, designed to plan, execute, and iterate autonomously on extended, engineering-grade tasks.", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "release_date": "2026-07-13", + "last_updated": "2026-07-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 2.1, + "output": 6.6, + "cache_read": 0.21 + } + }, + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.28, + "output": 0.56, + "cache_read": 0.07 + } + }, + "deepseek-v4-pro-0813@eu": { + "id": "deepseek-v4-pro-0813@eu", + "name": "DeepSeek V4 Pro 0813 (EU)", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.75, + "output": 3.5, + "cache_read": 0.44 + } + }, + "inkling-256k": { + "id": "inkling-256k", + "name": "inkling-256k", + "description": "Inkling 256K is the extended context variant of Inkling, a large MoE hybrid reasoning model from Thinking Machines with audio and vision input support and a 256K context window.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 1.87, + "output": 4.68, + "cache_read": 0.374 + } + }, + "muse-glimmer-30b": { + "id": "muse-glimmer-30b", + "name": "Muse Glimmer 30B", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 20480 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "claude-opus-4-8@eu": { + "id": "claude-opus-4-8@eu", + "name": "Claude Opus 4.8 (EU)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 262144 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.45 + } + }, + "claude-fable-5@eu": { + "id": "claude-fable-5@eu", + "name": "Claude Fable 5 (EU)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 11, + "output": 55, + "cache_read": 1.1, + "cache_write": 13.75 + } + }, + "nemotron-3-nano-omni@eu": { + "id": "nemotron-3-nano-omni@eu", + "name": "nemotron-3-nano-omni@eu", + "description": "The most open, efficient, and accurate omni modal reasoning model for agentic AI.", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-05-20", + "last_updated": "2026-05-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 300000 + }, + "cost": { + "input": 0.06, + "output": 0.24, + "cache_read": 0.06 + } + }, + "gpt-4.1@eu": { + "id": "gpt-4.1@eu", + "name": "GPT-4.1 (EU)", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2.2, + "output": 8.8, + "cache_read": 0.55 + } + }, + "gemini-3.1-flash-lite@eu": { + "id": "gemini-3.1-flash-lite@eu", + "name": "Gemini 3.1 Flash Lite (EU)", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.275, + "output": 1.65, + "cache_read": 0.0275, + "cache_write": 0.091663 + } + }, + "nvidia-nemotron-3-super-120b-a12b": { + "id": "nvidia-nemotron-3-super-120b-a12b", + "name": "nvidia-nemotron-3-super-120b-a12b", + "description": "NVIDIA Nemotron 3 Super is a hybrid Mixture-of-Experts (MoE) model engineered for highest compute efficiency and accuracy in multi-agent applications and specialized agentic systems. It is optimized to run many collaborating agents per application on a single GPU, delivering high accuracy for reasoning, tool use, and instruction following.", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.1, + "output": 0.5 + } + }, + "gemini-3.5-flash-lite@eu": { + "id": "gemini-3.5-flash-lite@eu", + "name": "Gemini 3.5 Flash Lite (EU)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.33, + "output": 2.75, + "cache_read": 0.033 + } + }, + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "kimi-k2.7-code@eu": { + "id": "kimi-k2.7-code@eu", + "name": "Kimi K2.7 Code (EU)", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.25, + "output": 4.5, + "cache_read": 0.31 + } + }, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 393216 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.05 + } + }, + "glm-5.1@eu": { + "id": "glm-5.1@eu", + "name": "GLM-5.1 (EU)", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 200000 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 1.4 + } + }, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "claude-haiku-4-5@eu": { + "id": "claude-haiku-4-5@eu", + "name": "Claude Haiku 4.5 (latest) (EU)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1.1, + "output": 5.5, + "cache_read": 0.11, + "cache_write": 1.375 + } + }, + "claude-sonnet-4-5@eu": { + "id": "claude-sonnet-4-5@eu", + "name": "Claude Sonnet 4.5 (latest) (EU)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.3, + "cache_write": 4.125, + "tiers": [ + { + "input": 6.6, + "output": 24.75, + "cache_read": 0.6, + "cache_write": 8.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6.6, + "output": 24.75, + "cache_read": 0.6, + "cache_write": 8.25 + } + } + }, + "laguna-xs.2": { + "id": "laguna-xs.2", + "name": "Laguna XS.2", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-28", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "leanstral-1-5@eu": { + "id": "leanstral-1-5@eu", + "name": "leanstral-1-5@eu", + "description": "Leanstral 1.5 is an updated Lean 4 formal proof engineering model from Mistral AI, optimized for automated theorem proving and autoformalization. It has 119B total parameters with 6.5B active and supports a 256K token context window. It supports native function calling and structured output.", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "release_date": "2026-05-27", + "last_updated": "2026-05-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } + }, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } + } + }, + "nemotron-3-super-120b-a12b": { + "id": "nemotron-3-super-120b-a12b", + "name": "Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "claude-opus-5-5": { + "id": "claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "claude-sonnet-4-6@eu": { + "id": "claude-sonnet-4-6@eu", + "name": "Claude Sonnet 4.6 (EU)", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.3, + "cache_write": 4.125 + } + }, + "gemini-3.6-flash": { + "id": "gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 1.5, + "output": 7, + "cache_read": 0.15 + } + }, + "gpt-5.5@eu": { + "id": "gpt-5.5@eu", + "name": "GPT-5.5 (EU)", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "ling-3.0-tiny": { + "id": "ling-3.0-tiny", + "name": "ling-3.0-tiny", + "description": "Ling-3.0-tiny is an efficient 7.9B parameter MoE model from inclusionAI with only 1.3B active parameters per token. Built for responsive agents, reliable instruction following and multi turn conversation, with a 256K context window, native function calling, prompt caching and switchable Thinking and Instant modes.", + "family": "ling", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "release_date": "2026-08-05", + "last_updated": "2026-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "claude-opus-5-5@eu": { + "id": "claude-opus-5-5@eu", + "name": "Claude Opus 5.5 (EU)", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "gpt-5-mini@eu": { + "id": "gpt-5-mini@eu", + "name": "GPT-5 Mini (EU)", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 0.275, + "output": 2.2, + "cache_read": 0.0275 + } + }, + "mistral-medium-3-5": { + "id": "mistral-medium-3-5", + "name": "mistral-medium-3-5", + "description": "Mistral Medium 3.5 is a dense 128B instruction following model from Mistral AI. It supports text and image inputs with text output, and is designed for agentic workflows, coding, and complex multi step reasoning. It is particularly strong at reliable multi tool calling and long horizon tasks, with a 256K context window, configurable reasoning effort per request, and a custom vision encoder that handles variable image sizes and aspect ratios. Self hostable on as few as four GPUs and available under open weights.", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.65, + "output": 8.25, + "cache_read": 1.65 + } + }, + "grok-4.3": { + "id": "grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "cache_write": 1.25, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "cache_write": 2.5 + } + } + }, + "nemotron-3-ultra-550b-a55b": { + "id": "nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nemotron-3-nano-omni-30b-a3b-reasoning": { + "id": "nemotron-3-nano-omni-30b-a3b-reasoning", + "name": "Nemotron 3 Nano Omni 30B A3B Reasoning", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-28", + "last_updated": "2026-04-28", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 20480 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gemma-4-26b-a4b-it": { + "id": "gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.07, + "output": 0.34, + "cache_read": 0.07 + } + }, + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 + } + }, + "gpt-5.4@eu": { + "id": "gpt-5.4@eu", + "name": "GPT-5.4 (EU)", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } + } + }, + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "grok-4.2-beta": { + "id": "grok-4.2-beta", + "name": "grok-4.2-beta", + "description": "Grok 4.20 Beta is xAI's newest flagship model with industry-leading speed and agentic tool calling capabilities. It combines the lowest hallucination rate on the market with strict prompt adherance, delivering consistently precise and truthful responses.", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-03-19", + "last_updated": "2026-03-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 2000000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 0.4, + "cache_write": 4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 0.4, + "cache_write": 4 + } + } + }, + "gemini-3.1-flash-image": { + "id": "gemini-3.1-flash-image", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 2 + } + }, + "nemotron-lightning-3.5-30b-a3b": { + "id": "nemotron-lightning-3.5-30b-a3b", + "name": "nemotron-lightning-3.5-30b-a3b", + "description": "Nemotron-Lightning-3.5-30B-A3B is a 30B-parameter Mixture-of-Experts language model (3B active) from NVIDIA's Nemotron-H family, built on a hybrid Mamba-Transformer architecture for efficient long-context inference. Like other models in the family, it responds to queries by first generating a reasoning trace and then concluding with a final response, with reasoning behavior configurable through a flag in the chat template. It includes a multi-token prediction (MTP) speculative decoding head for low-latency serving.", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-08-15", + "last_updated": "2026-08-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.05, + "output": 0.2, + "cache_read": 0.01 + } + }, + "gpt-5-nano@eu": { + "id": "gpt-5-nano@eu", + "name": "GPT-5 Nano (EU)", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 0.055, + "output": 0.44, + "cache_read": 0.0055 + } + }, + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2 + } + } + }, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 4.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 9, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 9 + } + } + }, + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "cache_write": 2, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "cache_write": 4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1, + "cache_write": 4 + } + } + }, + "seed-2.0-mini": { + "id": "seed-2.0-mini", + "name": "Seed 2.0 Mini", + "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.02 + } + }, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "kimi-k2.6@eu": { + "id": "kimi-k2.6@eu", + "name": "Kimi K2.6 (EU)", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.95 + } + }, + "qwen3.5-2b": { + "id": "qwen3.5-2b", + "name": "qwen3.5-2b", + "description": "Qwen3.5-2B is a compact yet capable model from Alibaba's Qwen3.5 series. It features a 262K token context window, support for 201 languages, thinking/reasoning mode, and tool calling for agentic workflows. A strong choice for prototyping, fine-tuning, and efficient multilingual deployments.", + "family": "qwen3.5", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-03-10", + "last_updated": "2026-03-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.02, + "output": 0.1 + } + }, + "fugu-ultra": { + "id": "fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 1.2 + } + }, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 1.583 + } + }, + "ling-2.6-flash": { + "id": "ling-2.6-flash", + "name": "ling-2.6-flash", + "description": "Inclusion AI ling-2.6-flash", + "family": "ling", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "gemini-3-pro-image": { + "id": "gemini-3-pro-image", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 4.5 + } + }, + "gemini-3.7-flash@eu": { + "id": "gemini-3.7-flash@eu", + "name": "Gemini 3.7 Flash (EU)", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.825, + "output": 4.125, + "cache_read": 0.0825 + } + }, + "claude-fable-5.1": { + "id": "claude-fable-5.1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 + } + }, + "gpt-4.1-nano@eu": { + "id": "gpt-4.1-nano@eu", + "name": "GPT-4.1 nano (EU)", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.11, + "output": 0.44, + "cache_read": 0.0275 + } + }, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 + } + }, + "gpt-5.1@eu": { + "id": "gpt-5.1@eu", + "name": "GPT-5.1 (EU)", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.375, + "output": 11, + "cache_read": 0.1375 + } + }, + "gemini-2.5-flash-lite@eu": { + "id": "gemini-2.5-flash-lite@eu", + "name": "Gemini 2.5 Flash-Lite (EU)", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0.18333 + } + }, + "nemotron-3-nano-omni": { + "id": "nemotron-3-nano-omni", + "name": "nemotron-3-nano-omni", + "description": "The most open, efficient, and accurate omni modal reasoning model for agentic AI.", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-05-20", + "last_updated": "2026-05-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 300000 + }, + "cost": { + "input": 0.06, + "output": 0.24, + "cache_read": 0.06 + } + }, + "claude-sonnet-4@eu": { + "id": "claude-sonnet-4@eu", + "name": "Claude Sonnet 4 (latest) (EU)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } + } + }, + "inkling": { + "id": "inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 65536, + "output": 32768 + }, + "cost": { + "input": 1.87, + "output": 4.68, + "cache_read": 0.374 + } + }, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "deepseek-v4.1-flash@eu": { + "id": "deepseek-v4.1-flash@eu", + "name": "DeepSeek V4.1 Flash (EU)", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 393216 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.05 + } + }, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04 + } + } + }, + "qwen3.8-2.4T-A95B@eu": { + "id": "qwen3.8-2.4T-A95B@eu", + "name": "Qwen3.8 2.4T A95B (EU)", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 262144 + }, + "cost": { + "input": 2.5, + "output": 6, + "cache_read": 0.63 + } + }, + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.16, + "output": 0.47, + "cache_read": 0.016, + "cache_write": 0.2 + } + }, + "qwen3.8-flash-next": { + "id": "qwen3.8-flash-next", + "name": "Qwen3.8 Flash Next", + "description": "Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } + }, + "qwen3.8-flash-next@eu": { + "id": "qwen3.8-flash-next@eu", + "name": "Qwen3.8 Flash Next (EU)", + "description": "Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } + }, + "mistral-small-2603": { + "id": "mistral-small-2603", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.165, + "output": 0.66, + "cache_read": 0.165 + } + }, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 33, + "cache_read": 0.55 + } + }, + "claude-sonnet-5@eu": { + "id": "claude-sonnet-5@eu", + "name": "Claude Sonnet 5 (EU)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2.2, + "output": 11, + "cache_read": 0.22, + "cache_write": 2.75 + } + }, + "gemini-3.8-flash@eu": { + "id": "gemini-3.8-flash@eu", + "name": "Gemini 3.8 Flash (EU)", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.825, + "output": 4.125, + "cache_read": 0.0825 + } + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "qwen3.5-35b-a3b": { + "id": "qwen3.5-35b-a3b", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.14, + "output": 1, + "cache_read": 0.05 + } + }, + "glm-5.3@eu": { + "id": "glm-5.3@eu", + "name": "GLM-5.3 (EU)", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 1.2, + "output": 4.2, + "cache_read": 0.26 + } + }, + "gemini-2.5-flash@eu": { + "id": "gemini-2.5-flash@eu", + "name": "Gemini 2.5 Flash (EU)", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.075, + "cache_write": 0.55 + } + }, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 + } + }, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.8, + "output": 2.55, + "cache_read": 0.16 + } + }, + "hy3": { + "id": "hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.14, + "output": 0.58, + "cache_read": 0.035 + } + }, + "glm-5.2@eu": { + "id": "glm-5.2@eu", + "name": "GLM-5.2 (EU)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4.2, + "cache_read": 0.26 + } + }, + "gpt-5.6-sol@eu": { + "id": "gpt-5.6-sol@eu", + "name": "GPT-5.6 Sol (EU)", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 4.4, + "output": 22, + "cache_read": 0.44 + } + }, + "seed-2.0-code": { + "id": "seed-2.0-code", + "name": "Seed 2.0 Code", + "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.1 + } + }, + "devstral-latest@eu": { + "id": "devstral-latest@eu", + "name": "devstral-latest@eu", + "description": "An enterprise grade text model, that excels at using tools to explore codebases, editing multiple files and power software engineering agents.", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "release_date": "2026-05-27", + "last_updated": "2026-05-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.44, + "output": 2.2, + "cache_read": 0.44 + } + }, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "gemini-2.5-pro@eu": { + "id": "gemini-2.5-pro@eu", + "name": "Gemini 2.5 Pro (EU)", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.31, + "cache_write": 2.375, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.62, + "cache_write": 4.75, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.62, + "cache_write": 4.75 + } + } + }, + "gpt-5.6-terra@eu": { + "id": "gpt-5.6-terra@eu", + "name": "GPT-5.6 Terra (EU)", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.2, + "output": 13.2, + "cache_read": 0.22 + } + }, + "grok-build-0.1": { + "id": "grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 1, + "output": 2, + "cache_read": 0.1 + } + }, + "glm-5.3-flash@eu": { + "id": "glm-5.3-flash@eu", + "name": "GLM-5.3-Flash (EU)", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 262144 + }, + "cost": { + "input": 0.2, + "output": 0.6, + "cache_read": 0.07 + } + }, + "minimax-m2.7-highspeed": { + "id": "minimax-m2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 1.2 + } + }, + "deepseek-v4-pro@eu": { + "id": "deepseek-v4-pro@eu", + "name": "DeepSeek V4 Pro (EU)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 1.75, + "output": 3.5, + "cache_read": 0.44 + } + }, + "devstral-latest": { + "id": "devstral-latest", + "name": "devstral-latest", + "description": "An enterprise grade text model, that excels at using tools to explore codebases, editing multiple files and power software engineering agents.", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "release_date": "2026-05-27", + "last_updated": "2026-05-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.44, + "output": 2.2, + "cache_read": 0.44 + } + }, + "o4-mini@eu": { + "id": "o4-mini@eu", + "name": "o4-mini (EU)", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.21, + "output": 4.84, + "cache_read": 0.3025 + } + }, + "mistral-small-2603@eu": { + "id": "mistral-small-2603@eu", + "name": "Mistral Small 4 (EU)", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.165, + "output": 0.66, + "cache_read": 0.165 + } + }, + "qwen3.8-2.4T-A95B": { + "id": "qwen3.8-2.4T-A95B", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.2 + } + }, + "gemini-3.8-flash": { + "id": "gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 + } + }, + "gemini-3.5-flash@eu": { + "id": "gemini-3.5-flash@eu", + "name": "Gemini 3.5 Flash (EU)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 1.65, + "output": 9.9, + "cache_read": 0.165, + "cache_write": 1.7413 + } + }, + "gpt-6-luna": { + "id": "gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02 + } + } + }, + "nemotron-3.5-content-safety": { + "id": "nemotron-3.5-content-safety", + "name": "Nemotron 3.5 Content Safety", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "nemotron", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 + } + }, + "nvidia-nemotron-3-ultra": { + "id": "nvidia-nemotron-3-ultra", + "name": "nvidia-nemotron-3-ultra", + "description": "NVIDIA Nemotron 3 Ultra is NVIDIA's strongest open-weights reasoning model, positioned near GPT-5.4 Mini (xhigh) and ahead of DeepSeek V4-Flash and Qwen3.5-397B-A17B.", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": false, + "structured_output": false, + "release_date": "2026-06-23", + "last_updated": "2026-06-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.5, + "output": 2.5 + } + }, + "claude-fable-5.1@eu": { + "id": "claude-fable-5.1@eu", + "name": "Claude Fable 5.1 (EU)", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 11, + "output": 55, + "cache_read": 0.275, + "cache_write": 13.75 + } + }, + "step-3.7-flash": { + "id": "step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 256000 + }, + "cost": { + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 + } + }, + "gpt-6-luna@eu": { + "id": "gpt-6-luna@eu", + "name": "GPT-6 Luna (EU)", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.12, + "output": 0.6, + "cache_read": 0.012, + "tiers": [ + { + "input": 0.24, + "output": 0.9, + "cache_read": 0.024, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.24, + "output": 0.9, + "cache_read": 0.024 + } + } + }, + "claude-opus-4-6@eu": { + "id": "claude-opus-4-6@eu", + "name": "Claude Opus 4.6 (EU)", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "leanstral-1-5": { + "id": "leanstral-1-5", + "name": "leanstral-1-5", + "description": "Leanstral 1.5 is an updated Lean 4 formal proof engineering model from Mistral AI, optimized for automated theorem proving and autoformalization. It has 119B total parameters with 6.5B active and supports a 256K token context window. It supports native function calling and structured output.", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "release_date": "2026-05-27", + "last_updated": "2026-05-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625 + } + }, + "gpt-5.6-luna@eu": { + "id": "gpt-5.6-luna@eu", + "name": "GPT-5.6 Luna (EU)", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.22, + "output": 1.32, + "cache_read": 0.022 + } + }, + "claude-opus-4-7@eu": { + "id": "claude-opus-4-7@eu", + "name": "Claude Opus 4.7 (EU)", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "kat-coder-pro": { + "id": "kat-coder-pro", + "name": "kat-coder-pro", + "description": "KAT-Coder-Pro V2 by KwaiKAT is a non-reasoning model optimized for agentic coding. It delivers strong performance on reasoning-style tasks while requiring significantly fewer output tokens than peer models. With the 1210 release, it achieved a score of 64 on the Artificial Analysis Intelligence Index, placing it in the global Top 10 and ranking first among all non-reasoning models.", + "family": "kat-coder", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "release_date": "2026-03-27", + "last_updated": "2026-03-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.3, + "output": 1.2 + } + }, + "laguna-m.1": { + "id": "laguna-m.1", + "name": "Laguna M.1", + "description": "Poolside's open-weight model for agentic coding and long-horizon work", + "family": "laguna", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-04-28", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-6-sol@eu": { + "id": "gpt-6-sol@eu", + "name": "GPT-6 Sol (EU)", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.4, + "output": 12, + "cache_read": 0.24, + "tiers": [ + { + "input": 4.8, + "output": 18, + "cache_read": 0.48, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4.8, + "output": 18, + "cache_read": 0.48 + } + } + }, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 1.2, + "output": 4.2, + "cache_read": 0.26 + } + }, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + }, + "grok-4.6": { + "id": "grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "cache_write": 2, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "cache_write": 4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1, + "cache_write": 4 + } + } + }, + "minimax-m3@eu": { + "id": "minimax-m3@eu", + "name": "MiniMax-M3 (EU)", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 0.4, + "output": 2, + "cache_read": 0.1 + } + }, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.32, + "output": 1.28, + "cache_read": 0.032, + "cache_write": 0.4 + } + }, + "gpt-4.1-mini@eu": { + "id": "gpt-4.1-mini@eu", + "name": "GPT-4.1 mini (EU)", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.44, + "output": 1.76, + "cache_read": 0.11 + } + }, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10 + } + } + }, + "claude-opus-4-5@eu": { + "id": "claude-opus-4-5@eu", + "name": "Claude Opus 4.5 (latest) (EU)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.28, + "output": 0.56, + "cache_read": 0.07 + } + }, + "mistral-medium-latest": { + "id": "mistral-medium-latest", + "name": "Mistral Medium (latest)", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.44, + "output": 2.2, + "cache_read": 0.44 + } + }, + "gpt-6-sol": { + "id": "gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4 + } + } + }, + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 0.08333 + } + } + } + }, + "mistral": { + "id": "mistral", + "env": [ + "MISTRAL_API_KEY" + ], + "npm": "@ai-sdk/mistral", + "name": "Mistral", + "doc": "https://docs.mistral.ai/getting-started/models/", + "models": { + "open-mistral-nemo": { + "id": "open-mistral-nemo", + "name": "Open Mistral Nemo", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0.15, + "output": 0.15 + } + }, + "codestral-latest": { + "id": "codestral-latest", + "name": "Codestral (latest)", + "description": "Mistral code model for completions, refactors, and developer IDE workflows", + "family": "codestral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-05-29", + "last_updated": "2025-01-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 4096 + }, + "cost": { + "input": 0.3, + "output": 0.9 + } + }, + "mistral-large-2411": { + "id": "mistral-large-2411", + "name": "Mistral Large 2.1", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-18", + "last_updated": "2024-11-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "mistral-nemo": { + "id": "mistral-nemo", + "name": "Mistral Nemo", + "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.15, + "output": 0.15 + } + }, + "voxtral-mini-tts-latest": { + "id": "voxtral-mini-tts-latest", + "name": "Voxtral Mini TTS (latest)", + "description": "Multilingual text-to-speech model with zero-shot voice cloning", + "family": "voxtral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-03-01", + "last_updated": "2026-03-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "mistral-medium-2508": { + "id": "mistral-medium-2508", + "name": "Mistral Medium 3.1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-08-12", + "last_updated": "2025-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.4, + "output": 2 + } + }, + "mistral-large-latest": { + "id": "mistral-large-latest", + "name": "Mistral Large (latest)", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "mistral-small-latest": { + "id": "mistral-small-latest", + "name": "Mistral Small (latest)", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "zai-glm-5-2": { + "id": "zai-glm-5-2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "status": "beta", + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.14 + } + }, + "ministral-8b-latest": { + "id": "ministral-8b-latest", + "name": "Ministral 8B (latest)", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.1, + "output": 0.1 + } + }, + "devstral-medium-latest": { + "id": "devstral-medium-latest", + "name": "Devstral 2 (latest)", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2 + } + }, + "open-mixtral-8x22b": { + "id": "open-mixtral-8x22b", + "name": "Mixtral 8x22B", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mixtral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-17", + "last_updated": "2024-04-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 64000, + "output": 64000 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "devstral-2512": { + "id": "devstral-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2 + } + }, + "mistral-medium-2505": { + "id": "mistral-medium-2505", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.4, + "output": 2 + } + }, + "magistral-medium-latest": { + "id": "magistral-medium-latest", + "name": "Magistral Medium (latest)", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-medium", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2, + "output": 5 + } + }, + "pixtral-12b": { + "id": "pixtral-12b", + "name": "Pixtral 12B", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-09", + "release_date": "2024-09-01", + "last_updated": "2024-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.15, + "output": 0.15 + } + }, + "mistral-embed": { + "id": "mistral-embed", + "name": "Mistral Embed", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "mistral-embed", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2023-12-11", + "last_updated": "2023-12-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8000, + "output": 3072 + }, + "cost": { + "input": 0.1, + "output": 0 + } + }, + "devstral-small-2505": { + "id": "devstral-small-2505", + "name": "Devstral Small 2505", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "mistral-small-2603": { + "id": "mistral-small-2603", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "devstral-medium-2507": { + "id": "devstral-medium-2507", + "name": "Devstral Medium", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-07-10", + "last_updated": "2025-07-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2 + } + }, + "magistral-small": { + "id": "magistral-small", + "name": "Magistral Small", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-small", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "voxtral-mini-latest": { + "id": "voxtral-mini-latest", + "name": "Voxtral Mini (latest)", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "voxtral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-02-01", + "last_updated": "2026-02-01", + "modalities": { + "input": [ + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "labs-devstral-small-2512": { + "id": "labs-devstral-small-2512", + "name": "Devstral Small 2", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0 + } + }, + "mistral-large-2512": { + "id": "mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "devstral-latest": { + "id": "devstral-latest", + "name": "Devstral 2", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2 + } + }, + "devstral-small-2507": { + "id": "devstral-small-2507", + "name": "Devstral Small", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-07-10", + "last_updated": "2025-07-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "pixtral-large-latest": { + "id": "pixtral-large-latest", + "name": "Pixtral Large (latest)", + "description": "Mistral's larger vision model for document-heavy image understanding and chat", + "family": "pixtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2024-11-04", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "voxtral-small-latest": { + "id": "voxtral-small-latest", + "name": "Voxtral Small (latest)", + "description": "Instruct model with native audio input for speech understanding and tool use", + "family": "voxtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "open-mixtral-8x7b": { + "id": "open-mixtral-8x7b", + "name": "Mixtral 8x7B", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mixtral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-01", + "release_date": "2023-12-11", + "last_updated": "2023-12-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.7, + "output": 0.7 + } + }, + "mistral-small-2506": { + "id": "mistral-small-2506", + "name": "Mistral Small 3.2", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "zai-glm-5-3": { + "id": "zai-glm-5-3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.14 + } + }, + "mistral-medium-2604": { + "id": "mistral-medium-2604", + "name": "Mistral Medium 3.5", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.5, + "output": 7.5 + } + }, + "open-mistral-7b": { + "id": "open-mistral-7b", + "name": "Mistral 7B", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2023-09-27", + "last_updated": "2023-09-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8000, + "output": 8000 + }, + "cost": { + "input": 0.25, + "output": 0.25 + } + }, + "ministral-3b-latest": { + "id": "ministral-3b-latest", + "name": "Ministral 3B (latest)", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.04, + "output": 0.04 + } + }, + "mistral-medium-latest": { + "id": "mistral-medium-latest", + "name": "Mistral Medium (latest)", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.5, + "output": 7.5 + } + } + } + }, + "amazon-bedrock": { + "id": "amazon-bedrock", + "env": [ + "AWS_ACCESS_KEY_ID", + "AWS_SECRET_ACCESS_KEY", + "AWS_REGION", + "AWS_BEARER_TOKEN_BEDROCK" + ], + "npm": "@ai-sdk/amazon-bedrock", + "name": "Amazon Bedrock", + "doc": "https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html", + "models": { + "google.gemma-3-12b-it": { + "id": "google.gemma-3-12b-it", + "name": "Gemma 3 12B IT", + "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.09, + "output": 0.29 + } + }, + "google.gemma-3-4b-it": { + "id": "google.gemma-3-4b-it", + "name": "Gemma 3 4B IT", + "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 4096 + }, + "cost": { + "input": 0.04, + "output": 0.08 + } + }, + "eu.anthropic.claude-fable-5": { + "id": "eu.anthropic.claude-fable-5", + "name": "Claude Fable 5 (EU)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 11, + "output": 55, + "cache_read": 1.1, + "cache_write": 13.75 + } + }, + "qwen.qwen3-coder-480b-a35b-v1:0": { + "id": "qwen.qwen3-coder-480b-a35b-v1:0", + "name": "Qwen3-Coder 480B-A35B Instruct", + "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-09-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 65536 + }, + "cost": { + "input": 0.45, + "output": 1.8 + } + }, + "google.gemma-4-31b": { + "id": "google.gemma-4-31b", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 0.14, + "output": 0.4 + } + }, + "us.writer.palmyra-x5-v1:0": { + "id": "us.writer.palmyra-x5-v1:0", + "name": "Palmyra X5 (US)", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "palmyra", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-04-28", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1040000, + "input": 1040000, + "output": 8192 + }, + "cost": { + "input": 0.6, + "output": 6 + } + }, + "google.gemma-4-e2b": { + "id": "google.gemma-4-e2b", + "name": "Gemma 4 E2B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 0.04, + "output": 0.08 + } + }, + "us.anthropic.claude-opus-4-7": { + "id": "us.anthropic.claude-opus-4-7", + "name": "Claude Opus 4.7 (US)", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "global.openai.gpt-6-astra": { + "id": "global.openai.gpt-6-astra", + "name": "GPT-6 Astra (Global)", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } + } + }, + "eu.amazon.nova-lite-v1:0": { + "id": "eu.amazon.nova-lite-v1:0", + "name": "Nova Lite (EU)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.069, + "output": 0.276, + "cache_read": 0.01725, + "cache_write": 0.069 + } + }, + "global.anthropic.claude-opus-5-5": { + "id": "global.anthropic.claude-opus-5-5", + "name": "Claude Opus 5.5 (Global)", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 + } + }, + "deepseek.r1-v1:0": { + "id": "deepseek.r1-v1:0", + "name": "DeepSeek-R1", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 1.35, + "output": 5.4 + } + }, + "openai.gpt-oss-safeguard-20b": { + "id": "openai.gpt-oss-safeguard-20b", + "name": "GPT OSS Safeguard 20B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.07, + "output": 0.2 + } + }, + "anthropic.claude-opus-5-5": { + "id": "anthropic.claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 + } + }, + "global.anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "global.anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5 (Global)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "eu.amazon.nova-2-lite-v1:0": { + "id": "eu.amazon.nova-2-lite-v1:0", + "name": "Nova 2 Lite (EU)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-10", + "release_date": "2025-12-02", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65535 + }, + "cost": { + "input": 0.374, + "output": 3.157, + "cache_read": 0.0935, + "cache_write": 0.374 + } + }, + "us.amazon.nova-pro-v1:0": { + "id": "us.amazon.nova-pro-v1:0", + "name": "Nova Pro (US)", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.8, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0.8 + } + }, + "jp.anthropic.claude-opus-5-5": { + "id": "jp.anthropic.claude-opus-5-5", + "name": "Claude Opus 5.5 (JP)", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4.4, + "output": 22, + "cache_read": 0.22, + "cache_write": 5.5 + } + }, + "openai.gpt-5.6-luna": { + "id": "openai.gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 0.22, + "output": 1.32, + "cache_read": 0.022, + "cache_write": 0.275, + "tiers": [ + { + "input": 0.44, + "output": 1.98, + "cache_read": 0.044, + "cache_write": 0.55, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.44, + "output": 1.98, + "cache_read": 0.044, + "cache_write": 0.55 + } + } + }, + "us.meta.llama4-maverick-17b-instruct-v1:0": { + "id": "us.meta.llama4-maverick-17b-instruct-v1:0", + "name": "Llama 4 Maverick 17B Instruct (US)", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 8192 + }, + "cost": { + "input": 0.24, + "output": 0.97 + } + }, + "apac.amazon.nova-micro-v1:0": { + "id": "apac.amazon.nova-micro-v1:0", + "name": "Nova Micro (APAC)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 10000 + }, + "cost": { + "input": 0.037, + "output": 0.148, + "cache_read": 0.00925, + "cache_write": 0.037 + } + }, + "eu.anthropic.claude-sonnet-4-20250514-v1:0": { + "id": "eu.anthropic.claude-sonnet-4-20250514-v1:0", + "name": "Claude Sonnet 4 (EU)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "status": "deprecated", + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "eu.anthropic.claude-opus-4-5-20251101-v1:0": { + "id": "eu.anthropic.claude-opus-4-5-20251101-v1:0", + "name": "Claude Opus 4.5 (EU)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "global.openai.gpt-6-sol": { + "id": "global.openai.gpt-6-sol", + "name": "GPT-6 Sol (Global)", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } + } + }, + "qwen.qwen3-coder-30b-a3b-v1:0": { + "id": "qwen.qwen3-coder-30b-a3b-v1:0", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-31", + "last_updated": "2025-09-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "us.amazon.nova-premier-v1:0": { + "id": "us.amazon.nova-premier-v1:0", + "name": "Nova Premier (US)", + "description": "Multimodal model for complex analysis, long-context understanding, tool use, and model distillation", + "family": "nova", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-04-30", + "last_updated": "2025-04-30", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 10000 + }, + "status": "deprecated", + "cost": { + "input": 2.5, + "output": 12.5, + "cache_read": 0.625, + "cache_write": 2.5 + } + }, + "anthropic.claude-sonnet-4-6": { + "id": "anthropic.claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 + } + }, + "meta.llama3-1-70b-instruct-v1:0": { + "id": "meta.llama3-1-70b-instruct-v1:0", + "name": "Llama 3.1 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.72, + "output": 0.72 + } + }, + "us.writer.palmyra-x4-v1:0": { + "id": "us.writer.palmyra-x4-v1:0", + "name": "Palmyra X4 (US)", + "description": "Enterprise language model for workflow automation, coding, data analysis, and tool use", + "family": "palmyra", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2024-10-09", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 122880, + "output": 8192 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "apac.amazon.nova-pro-v1:0": { + "id": "apac.amazon.nova-pro-v1:0", + "name": "Nova Pro (APAC)", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.84, + "output": 3.36, + "cache_read": 0.21, + "cache_write": 0.84 + } + }, + "global.anthropic.claude-fable-5-1": { + "id": "global.anthropic.claude-fable-5-1", + "name": "Claude Fable 5.1 (Global)", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "us.anthropic.claude-sonnet-5": { + "id": "us.anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5 (US)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2.2, + "output": 11, + "cache_read": 0.22, + "cache_write": 2.75 + } + }, + "amazon.nova-micro-v1:0": { + "id": "amazon.nova-micro-v1:0", + "name": "Nova Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 10000 + }, + "cost": { + "input": 0.035, + "output": 0.14, + "cache_read": 0.00875, + "cache_write": 0.035 + } + }, + "minimax.minimax-m2.5": { + "id": "minimax.minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 98304 + }, + "cost": { + "input": 0.3, + "output": 1.2 + } + }, + "xai.grok-4.3": { + "id": "xai.grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-06-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 + } + }, + "us.amazon.nova-lite-v1:0": { + "id": "us.amazon.nova-lite-v1:0", + "name": "Nova Lite (US)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.06, + "output": 0.24, + "cache_read": 0.015, + "cache_write": 0.06 + } + }, + "amazon.nova-pro-v1:0": { + "id": "amazon.nova-pro-v1:0", + "name": "Nova Pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.8, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0.8 + } + }, + "qwen.qwen3-32b-v1:0": { + "id": "qwen.qwen3-32b-v1:0", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-09-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "openai.gpt-5.5": { + "id": "openai.gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 5.5, + "output": 33, + "cache_read": 0.55 + } + }, + "mistral.voxtral-small-24b-2507": { + "id": "mistral.voxtral-small-24b-2507", + "name": "Voxtral Small 24B 2507", + "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", + "family": "voxtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 8192 + }, + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "global.amazon.nova-2-lite-v1:0": { + "id": "global.amazon.nova-2-lite-v1:0", + "name": "Nova 2 Lite (Global)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-10", + "release_date": "2025-12-02", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65535 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.075, + "cache_write": 0.3 + } + }, + "us.openai.gpt-5.6-terra": { + "id": "us.openai.gpt-5.6-terra", + "name": "GPT-5.6 Terra (US)", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.2, + "output": 13.2, + "cache_read": 0.22, + "cache_write": 2.75, + "tiers": [ + { + "input": 4.4, + "output": 19.8, + "cache_read": 0.44, + "cache_write": 5.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4.4, + "output": 19.8, + "cache_read": 0.44, + "cache_write": 5.5 + } + } + }, + "mistral.devstral-2-123b": { + "id": "mistral.devstral-2-123b", + "name": "Devstral 2 123B", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.4, + "output": 2 + } + }, + "us.xai.grok-4.6": { + "id": "us.xai.grok-4.6", + "name": "Grok 4.6 (US)", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-18", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2.2, + "output": 6.6, + "cache_read": 0.55 + } + }, + "jp.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "jp.anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5 (JP)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 + } + }, + "anthropic.claude-fable-5": { + "id": "anthropic.claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "jp.anthropic.claude-opus-4-8": { + "id": "jp.anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8 (JP)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "us.openai.gpt-5.6-luna": { + "id": "us.openai.gpt-5.6-luna", + "name": "GPT-5.6 Luna (US)", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.22, + "output": 1.32, + "cache_read": 0.022, + "cache_write": 0.275, + "tiers": [ + { + "input": 0.44, + "output": 1.98, + "cache_read": 0.044, + "cache_write": 0.55, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.44, + "output": 1.98, + "cache_read": 0.044, + "cache_write": 0.55 + } + } + }, + "us.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "us.anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5 (US)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 + } + }, + "au.anthropic.claude-opus-5": { + "id": "au.anthropic.claude-opus-5", + "name": "Claude Opus 5 (AU)", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "nvidia.nemotron-nano-12b-v2": { + "id": "nvidia.nemotron-nano-12b-v2", + "name": "NVIDIA Nemotron Nano 12B v2 VL BF16", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-28", + "last_updated": "2025-10-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.2, + "output": 0.6 + } + }, + "us.amazon.nova-2-lite-v1:0": { + "id": "us.amazon.nova-2-lite-v1:0", + "name": "Nova 2 Lite (US)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-10", + "release_date": "2025-12-02", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65535 + }, + "cost": { + "input": 0.33, + "output": 2.75, + "cache_read": 0.0825, + "cache_write": 0.33 + } + }, + "us.openai.gpt-6-astra": { + "id": "us.openai.gpt-6-astra", + "name": "GPT-6 Astra (US)", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 11, + "output": 55, + "cache_read": 1.1, + "cache_write": 13.75, + "tiers": [ + { + "input": 22, + "output": 82.5, + "cache_read": 2.2, + "cache_write": 27.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 22, + "output": 82.5, + "cache_read": 2.2, + "cache_write": 27.5 + } + } + }, + "openai.gpt-5.6-terra": { + "id": "openai.gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 2.2, + "output": 13.2, + "cache_read": 0.22, + "cache_write": 2.75, + "tiers": [ + { + "input": 4.4, + "output": 19.8, + "cache_read": 0.44, + "cache_write": 5.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4.4, + "output": 19.8, + "cache_read": 0.44, + "cache_write": 5.5 + } + } + }, + "anthropic.claude-opus-4-6-v1": { + "id": "anthropic.claude-opus-4-6-v1", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "us.anthropic.claude-opus-4-1-20250805-v1:0": { + "id": "us.anthropic.claude-opus-4-1-20250805-v1:0", + "name": "Claude Opus 4.1 (US)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "us.meta.llama3-1-8b-instruct-v1:0": { + "id": "us.meta.llama3-1-8b-instruct-v1:0", + "name": "Llama 3.1 8B Instruct (US)", + "description": "Compact open Llama model for lightweight chat, drafting, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.22, + "output": 0.22 + } + }, + "eu.anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "eu.anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5 (EU)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1.1, + "output": 5.5, + "cache_read": 0.11, + "cache_write": 1.375 + } + }, + "global.anthropic.claude-opus-4-7": { + "id": "global.anthropic.claude-opus-4-7", + "name": "Claude Opus 4.7 (Global)", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "apac.anthropic.claude-sonnet-4-20250514-v1:0": { + "id": "apac.anthropic.claude-sonnet-4-20250514-v1:0", + "name": "Claude Sonnet 4 (APAC)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "status": "deprecated", + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "deepseek.v3.2": { + "id": "deepseek.v3.2", + "name": "DeepSeek V3.2", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2026-02-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 81920 + }, + "cost": { + "input": 0.62, + "output": 1.85 + } + }, + "qwen.qwen3-235b-a22b-2507-v1:0": { + "id": "qwen.qwen3-235b-a22b-2507-v1:0", + "name": "Qwen3 235B-A22B Instruct 2507", + "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-21", + "last_updated": "2025-09-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.22, + "output": 0.88 + } + }, + "amazon.nova-lite-v1:0": { + "id": "amazon.nova-lite-v1:0", + "name": "Nova Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.06, + "output": 0.24, + "cache_read": 0.015, + "cache_write": 0.06 + } + }, + "global.anthropic.claude-opus-5": { + "id": "global.anthropic.claude-opus-5", + "name": "Claude Opus 5 (Global)", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "openai.gpt-oss-20b-1:0": { + "id": "openai.gpt-oss-20b-1:0", + "name": "gpt-oss-20b", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.07, + "output": 0.3 + } + }, + "anthropic.claude-opus-5": { + "id": "anthropic.claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "meta.llama3-3-70b-instruct-v1:0": { + "id": "meta.llama3-3-70b-instruct-v1:0", + "name": "Llama 3.3 70B Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.72, + "output": 0.72 + } + }, + "minimax.minimax-m2": { + "id": "minimax.minimax-m2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204608, + "output": 128000 + }, + "cost": { + "input": 0.3, + "output": 1.2 + } + }, + "mistral.mistral-large-3-675b-instruct": { + "id": "mistral.mistral-large-3-675b-instruct", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "in.openai.gpt-5.6-luna": { + "id": "in.openai.gpt-5.6-luna", + "name": "GPT-5.6 Luna (India)", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.22, + "output": 1.32, + "cache_read": 0.022, + "cache_write": 0.275, + "tiers": [ + { + "input": 0.44, + "output": 1.98, + "cache_read": 0.044, + "cache_write": 0.55, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.44, + "output": 1.98, + "cache_read": 0.044, + "cache_write": 0.55 + } + } + }, + "eu.anthropic.claude-opus-4-8": { + "id": "eu.anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8 (EU)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "global.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "global.anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5 (Global)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "zai.glm-4.7": { + "id": "zai.glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.2 + } + }, + "au.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "au.anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5 (AU)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 + } + }, + "eu.anthropic.claude-opus-5-5": { + "id": "eu.anthropic.claude-opus-5-5", + "name": "Claude Opus 5.5 (EU)", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4.4, + "output": 22, + "cache_read": 0.22, + "cache_write": 5.5 + } + }, + "global.anthropic.claude-sonnet-4-6": { + "id": "global.anthropic.claude-sonnet-4-6", + "name": "Claude Sonnet 4.6 (Global)", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "jp.anthropic.claude-sonnet-4-6": { + "id": "jp.anthropic.claude-sonnet-4-6", + "name": "Claude Sonnet 4.6 (JP)", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 + } + }, + "google.gemma-3-27b-it": { + "id": "google.gemma-3-27b-it", + "name": "Gemma 3 27B IT", + "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 8192 + }, + "cost": { + "input": 0.23, + "output": 0.38 + } + }, + "eu.anthropic.claude-opus-4-6-v1": { + "id": "eu.anthropic.claude-opus-4-6-v1", + "name": "Claude Opus 4.6 (EU)", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "us.anthropic.claude-opus-5-5": { + "id": "us.anthropic.claude-opus-5-5", + "name": "Claude Opus 5.5 (US)", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4.4, + "output": 22, + "cache_read": 0.22, + "cache_write": 5.5 + } + }, + "us.anthropic.claude-opus-5": { + "id": "us.anthropic.claude-opus-5", + "name": "Claude Opus 5 (US)", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "us.openai.gpt-5.6-sol": { + "id": "us.openai.gpt-5.6-sol", + "name": "GPT-5.6 Sol (US)", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 4.4, + "output": 22, + "cache_read": 0.44, + "cache_write": 5.5, + "tiers": [ + { + "input": 8.8, + "output": 33, + "cache_read": 0.88, + "cache_write": 11, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8.8, + "output": 33, + "cache_read": 0.88, + "cache_write": 11 + } + } + }, + "us.moonshotai.kimi-k3": { + "id": "us.moonshotai.kimi-k3", + "name": "Kimi K3 (US)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 + } + }, + "anthropic.claude-sonnet-5": { + "id": "anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "mistral.magistral-small-2509": { + "id": "mistral.magistral-small-2509", + "name": "Magistral Small 1.2", + "description": "Open multimodal reasoning model for transparent analysis of text and images", + "family": "magistral", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-09-18", + "last_updated": "2025-09-18", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 40000 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "mistral.pixtral-large-2502-v1:0": { + "id": "mistral.pixtral-large-2502-v1:0", + "name": "Pixtral Large (25.02)", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-04-08", + "last_updated": "2025-04-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "anthropic.claude-opus-4-5-20251101-v1:0": { + "id": "anthropic.claude-opus-4-5-20251101-v1:0", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "nvidia.nemotron-nano-9b-v2": { + "id": "nvidia.nemotron-nano-9b-v2", + "name": "NVIDIA Nemotron Nano 9B v2", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-18", + "last_updated": "2025-08-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.06, + "output": 0.23 + } + }, + "eu.amazon.nova-micro-v1:0": { + "id": "eu.amazon.nova-micro-v1:0", + "name": "Nova Micro (EU)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 10000 + }, + "cost": { + "input": 0.04, + "output": 0.16, + "cache_read": 0.01, + "cache_write": 0.04 + } + }, + "zai.glm-5": { + "id": "zai.glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 3.2 + } + }, + "us.meta.llama3-1-70b-instruct-v1:0": { + "id": "us.meta.llama3-1-70b-instruct-v1:0", + "name": "Llama 3.1 70B Instruct (US)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.72, + "output": 0.72 + } + }, + "jp.anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "jp.anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5 (JP)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1.1, + "output": 5.5, + "cache_read": 0.11, + "cache_write": 1.375 + } + }, + "nvidia.nemotron-nano-3-30b": { + "id": "nvidia.nemotron-nano-3-30b", + "name": "NVIDIA Nemotron Nano 3 30B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-15", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 8192 + }, + "cost": { + "input": 0.06, + "output": 0.24 + } + }, + "mistral.ministral-3-3b-instruct": { + "id": "mistral.ministral-3-3b-instruct", + "name": "Ministral 3 3B", + "description": "Compact open vision-language model for edge deployment, instruction following, and tool use", + "family": "ministral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.1, + "output": 0.1 + } + }, + "us.openai.gpt-6-luna": { + "id": "us.openai.gpt-6-luna", + "name": "GPT-6 Luna (US)", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.11, + "output": 0.55, + "cache_read": 0.011, + "cache_write": 0.1375, + "tiers": [ + { + "input": 0.22, + "output": 0.825, + "cache_read": 0.022, + "cache_write": 0.275, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.22, + "output": 0.825, + "cache_read": 0.022, + "cache_write": 0.275 + } + } + }, + "global.openai.gpt-5.6-sol": { + "id": "global.openai.gpt-5.6-sol", + "name": "GPT-5.6 Sol (Global)", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10 + } + } + }, + "google.gemma-4-26b-a4b": { + "id": "google.gemma-4-26b-a4b", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 0.13, + "output": 0.4 + } + }, + "xai.grok-4.6": { + "id": "xai.grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-18", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 2.2, + "output": 6.6, + "cache_read": 0.55 + } + }, + "global.xai.grok-4.6": { + "id": "global.xai.grok-4.6", + "name": "Grok 4.6 (Global)", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-18", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "au.anthropic.claude-opus-4-6-v1": { + "id": "au.anthropic.claude-opus-4-6-v1", + "name": "AU Anthropic Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "qwen.qwen3-vl-235b-a22b": { + "id": "qwen.qwen3-vl-235b-a22b", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-11-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262000 + }, + "cost": { + "input": 0.53, + "output": 2.66 + } + }, + "anthropic.claude-fable-5-1": { + "id": "anthropic.claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "jp.anthropic.claude-sonnet-5": { + "id": "jp.anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5 (JP)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2.2, + "output": 11, + "cache_read": 0.22, + "cache_write": 2.75 + } + }, + "qwen.qwen3-coder-next": { + "id": "qwen.qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 1.2 + } + }, + "anthropic.claude-opus-4-8": { + "id": "anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "au.anthropic.claude-sonnet-5": { + "id": "au.anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5 (AU)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2.2, + "output": 11, + "cache_read": 0.22, + "cache_write": 2.75 + } + }, + "meta.llama4-scout-17b-instruct-v1:0": { + "id": "meta.llama4-scout-17b-instruct-v1:0", + "name": "Llama 4 Scout 17B Instruct", + "description": "Open Llama with long-context vision for efficient multimodal agents", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 10000000, + "output": 8192 + }, + "cost": { + "input": 0.17, + "output": 0.66 + } + }, + "openai.gpt-oss-safeguard-120b": { + "id": "openai.gpt-oss-safeguard-120b", + "name": "GPT OSS Safeguard 120B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "us.meta.llama3-3-70b-instruct-v1:0": { + "id": "us.meta.llama3-3-70b-instruct-v1:0", + "name": "Llama 3.3 70B Instruct (US)", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.72, + "output": 0.72 + } + }, + "global.openai.gpt-5.6-luna": { + "id": "global.openai.gpt-5.6-luna", + "name": "GPT-5.6 Luna (Global)", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } + } + }, + "au.anthropic.claude-sonnet-4-6": { + "id": "au.anthropic.claude-sonnet-4-6", + "name": "AU Anthropic Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 + } + }, + "openai.gpt-oss-120b": { + "id": "openai.gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/v1", + "shape": "responses" + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "deepseek.v3-v1:0": { + "id": "deepseek.v3-v1:0", + "name": "DeepSeek-V3.1", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-21", + "last_updated": "2025-09-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 81920 + }, + "cost": { + "input": 0.58, + "output": 1.68 + } + }, + "moonshotai.kimi-k2.5": { + "id": "moonshotai.kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-02-06", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262143, + "output": 16384 + }, + "cost": { + "input": 0.6, + "output": 3 + } + }, + "eu.anthropic.claude-sonnet-4-6": { + "id": "eu.anthropic.claude-sonnet-4-6", + "name": "Claude Sonnet 4.6 (EU)", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 + } + }, + "us-gov.openai.gpt-oss-120b-1:0": { + "id": "us-gov.openai.gpt-oss-120b-1:0", + "name": "gpt-oss-120b (GovCloud)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.18, + "output": 0.72 + } + }, + "us.anthropic.claude-fable-5": { + "id": "us.anthropic.claude-fable-5", + "name": "Claude Fable 5 (US)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 11, + "output": 55, + "cache_read": 1.1, + "cache_write": 13.75 + } + }, + "us.deepseek.r1-v1:0": { + "id": "us.deepseek.r1-v1:0", + "name": "DeepSeek-R1 (US)", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 1.35, + "output": 5.4 + } + }, + "eu.anthropic.claude-opus-5": { + "id": "eu.anthropic.claude-opus-5", + "name": "Claude Opus 5 (EU)", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "ca.amazon.nova-lite-v1:0": { + "id": "ca.amazon.nova-lite-v1:0", + "name": "Nova Lite (CA)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.064, + "output": 0.256, + "cache_read": 0.016, + "cache_write": 0.064 + } + }, + "us.openai.gpt-6-sol": { + "id": "us.openai.gpt-6-sol", + "name": "GPT-6 Sol (US)", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.2, + "output": 11, + "cache_read": 0.22, + "cache_write": 2.75, + "tiers": [ + { + "input": 4.4, + "output": 16.5, + "cache_read": 0.44, + "cache_write": 5.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4.4, + "output": 16.5, + "cache_read": 0.44, + "cache_write": 5.5 + } + } + }, + "au.anthropic.claude-opus-4-7": { + "id": "au.anthropic.claude-opus-4-7", + "name": "Claude Opus 4.7 (AU)", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "us.meta.llama4-scout-17b-instruct-v1:0": { + "id": "us.meta.llama4-scout-17b-instruct-v1:0", + "name": "Llama 4 Scout 17B Instruct (US)", + "description": "Open Llama with long-context vision for efficient multimodal agents", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 10000000, + "output": 8192 + }, + "cost": { + "input": 0.17, + "output": 0.66 + } + }, + "global.anthropic.claude-sonnet-5": { + "id": "global.anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5 (Global)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "global.anthropic.claude-fable-5": { + "id": "global.anthropic.claude-fable-5", + "name": "Claude Fable 5 (Global)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "apac.amazon.nova-lite-v1:0": { + "id": "apac.amazon.nova-lite-v1:0", + "name": "Nova Lite (APAC)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.063, + "output": 0.252, + "cache_read": 0.01575, + "cache_write": 0.063 + } + }, + "us.anthropic.claude-opus-4-8": { + "id": "us.anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8 (US)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "nvidia.nemotron-super-3-120b": { + "id": "nvidia.nemotron-super-3-120b", + "name": "NVIDIA Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.65 + } + }, + "global.openai.gpt-6-luna": { + "id": "global.openai.gpt-6-luna", + "name": "GPT-6 Luna (Global)", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } + } + }, + "us-gov.openai.gpt-oss-20b-1:0": { + "id": "us-gov.openai.gpt-oss-20b-1:0", + "name": "gpt-oss-20b (GovCloud)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.084, + "output": 0.36 + } + }, + "us.anthropic.claude-opus-4-6-v1": { + "id": "us.anthropic.claude-opus-4-6-v1", + "name": "Claude Opus 4.6 (US)", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "meta.llama3-1-8b-instruct-v1:0": { + "id": "meta.llama3-1-8b-instruct-v1:0", + "name": "Llama 3.1 8B Instruct", + "description": "Compact open Llama model for lightweight chat, drafting, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.22, + "output": 0.22 + } + }, + "writer.palmyra-x5-v1:0": { + "id": "writer.palmyra-x5-v1:0", + "name": "Palmyra X5", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "palmyra", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-04-28", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1040000, + "input": 1040000, + "output": 8192 + }, + "cost": { + "input": 0.6, + "output": 6 + } + }, + "anthropic.claude-opus-4-1-20250805-v1:0": { + "id": "anthropic.claude-opus-4-1-20250805-v1:0", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "us.anthropic.claude-sonnet-4-6": { + "id": "us.anthropic.claude-sonnet-4-6", + "name": "Claude Sonnet 4.6 (US)", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 + } + }, + "global.anthropic.claude-opus-4-6-v1": { + "id": "global.anthropic.claude-opus-4-6-v1", + "name": "Claude Opus 4.6 (Global)", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "eu.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "eu.anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5 (EU)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 + } + }, + "jp.anthropic.claude-opus-4-7": { + "id": "jp.anthropic.claude-opus-4-7", + "name": "Claude Opus 4.7 (JP)", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "au.anthropic.claude-opus-5-5": { + "id": "au.anthropic.claude-opus-5-5", + "name": "Claude Opus 5.5 (AU)", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4.4, + "output": 22, + "cache_read": 0.22, + "cache_write": 5.5 + } + }, + "writer.palmyra-x4-v1:0": { + "id": "writer.palmyra-x4-v1:0", + "name": "Palmyra X4", + "description": "Enterprise language model for workflow automation, coding, data analysis, and tool use", + "family": "palmyra", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2024-10-09", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 122880, + "output": 8192 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "global.anthropic.claude-opus-4-5-20251101-v1:0": { + "id": "global.anthropic.claude-opus-4-5-20251101-v1:0", + "name": "Claude Opus 4.5 (Global)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic.claude-opus-4-7": { + "id": "anthropic.claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "us.anthropic.claude-sonnet-4-20250514-v1:0": { + "id": "us.anthropic.claude-sonnet-4-20250514-v1:0", + "name": "Claude Sonnet 4 (US)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "status": "deprecated", + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "eu.anthropic.claude-sonnet-5": { + "id": "eu.anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5 (EU)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2.2, + "output": 11, + "cache_read": 0.22, + "cache_write": 2.75 + } + }, + "zai.glm-4.7-flash": { + "id": "zai.glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0.07, + "output": 0.4 + } + }, + "moonshot.kimi-k2-thinking": { + "id": "moonshot.kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262143, + "output": 16000 + }, + "cost": { + "input": 0.6, + "output": 2.5 + } + }, + "eu.amazon.nova-pro-v1:0": { + "id": "eu.amazon.nova-pro-v1:0", + "name": "Nova Pro (EU)", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.92, + "output": 3.68, + "cache_read": 0.23, + "cache_write": 0.92 + } + }, + "global.openai.gpt-5.6-terra": { + "id": "global.openai.gpt-5.6-terra", + "name": "GPT-5.6 Terra (Global)", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } + } + }, + "openai.gpt-6-astra": { + "id": "openai.gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 11, + "output": 55, + "cache_read": 1.1, + "cache_write": 13.75, + "tiers": [ + { + "input": 22, + "output": 82.5, + "cache_read": 2.2, + "cache_write": 27.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 22, + "output": 82.5, + "cache_read": 2.2, + "cache_write": 27.5 + } + } + }, + "us.anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "us.anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5 (US)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1.1, + "output": 5.5, + "cache_read": 0.11, + "cache_write": 1.375 + } + }, + "meta.llama4-maverick-17b-instruct-v1:0": { + "id": "meta.llama4-maverick-17b-instruct-v1:0", + "name": "Llama 4 Maverick 17B Instruct", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 8192 + }, + "cost": { + "input": 0.24, + "output": 0.97 + } + }, + "qwen.qwen3-next-80b-a3b": { + "id": "qwen.qwen3-next-80b-a3b", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-11", + "last_updated": "2025-11-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262000 + }, + "cost": { + "input": 0.15, + "output": 1.2 + } + }, + "eu.mistral.pixtral-large-2502-v1:0": { + "id": "eu.mistral.pixtral-large-2502-v1:0", + "name": "Pixtral Large (25.02) (EU)", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-04-08", + "last_updated": "2025-04-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "au.anthropic.claude-opus-4-8": { + "id": "au.anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8 (AU)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "us.anthropic.claude-fable-5-1": { + "id": "us.anthropic.claude-fable-5-1", + "name": "Claude Fable 5.1 (US)", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 11, + "output": 55, + "cache_read": 0.275, + "cache_write": 13.75 + } + }, + "global.moonshotai.kimi-k3": { + "id": "global.moonshotai.kimi-k3", + "name": "Kimi K3 (Global)", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "in.openai.gpt-5.6-terra": { + "id": "in.openai.gpt-5.6-terra", + "name": "GPT-5.6 Terra (India)", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.2, + "output": 13.2, + "cache_read": 0.22, + "cache_write": 2.75, + "tiers": [ + { + "input": 4.4, + "output": 19.8, + "cache_read": 0.44, + "cache_write": 5.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4.4, + "output": 19.8, + "cache_read": 0.44, + "cache_write": 5.5 + } + } + }, + "openai.gpt-oss-20b": { + "id": "openai.gpt-oss-20b", + "name": "gpt-oss-20b", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/v1", + "shape": "responses" + }, + "cost": { + "input": 0.07, + "output": 0.3 + } + }, + "mistral.ministral-3-14b-instruct": { + "id": "mistral.ministral-3-14b-instruct", + "name": "Ministral 14B 3.0", + "description": "Open vision-language model for efficient local deployment, instruction following, and tool use", + "family": "ministral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.2, + "output": 0.2 + } + }, + "au.anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "au.anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5 (AU)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1.1, + "output": 5.5, + "cache_read": 0.11, + "cache_write": 1.375 + } + }, + "openai.gpt-5.6-sol": { + "id": "openai.gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 4.4, + "output": 22, + "cache_read": 0.44, + "cache_write": 5.5, + "tiers": [ + { + "input": 8.8, + "output": 33, + "cache_read": 0.88, + "cache_write": 11, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8.8, + "output": 33, + "cache_read": 0.88, + "cache_write": 11 + } + } + }, + "minimax.minimax-m2.1": { + "id": "minimax.minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2 + } + }, + "us.anthropic.claude-opus-4-5-20251101-v1:0": { + "id": "us.anthropic.claude-opus-4-5-20251101-v1:0", + "name": "Claude Opus 4.5 (US)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "eu.anthropic.claude-opus-4-7": { + "id": "eu.anthropic.claude-opus-4-7", + "name": "Claude Opus 4.7 (EU)", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "jp.amazon.nova-2-lite-v1:0": { + "id": "jp.amazon.nova-2-lite-v1:0", + "name": "Nova 2 Lite (JP)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-10", + "release_date": "2025-12-02", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65535 + }, + "cost": { + "input": 0.396, + "output": 3.311, + "cache_read": 0.099, + "cache_write": 0.396 + } + }, + "us.mistral.pixtral-large-2502-v1:0": { + "id": "us.mistral.pixtral-large-2502-v1:0", + "name": "Pixtral Large (25.02) (US)", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-04-08", + "last_updated": "2025-04-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "jp.anthropic.claude-opus-5": { + "id": "jp.anthropic.claude-opus-5", + "name": "Claude Opus 5 (JP)", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 + } + }, + "global.anthropic.claude-sonnet-4-20250514-v1:0": { + "id": "global.anthropic.claude-sonnet-4-20250514-v1:0", + "name": "Claude Sonnet 4 (Global)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "status": "deprecated", + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "global.anthropic.claude-opus-4-8": { + "id": "global.anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8 (Global)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "openai.gpt-5.4": { + "id": "openai.gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 2.75, + "output": 16.5, + "cache_read": 0.275 + } + }, + "mistral.voxtral-mini-3b-2507": { + "id": "mistral.voxtral-mini-3b-2507", + "name": "Voxtral Mini 3B 2507", + "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", + "family": "voxtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 4096 + }, + "cost": { + "input": 0.04, + "output": 0.04 + } + }, + "openai.gpt-oss-120b-1:0": { + "id": "openai.gpt-oss-120b-1:0", + "name": "gpt-oss-120b", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "amazon.nova-2-lite-v1:0": { + "id": "amazon.nova-2-lite-v1:0", + "name": "Nova 2 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-10", + "release_date": "2025-12-02", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65535 + }, + "cost": { + "input": 0.33, + "output": 2.75, + "cache_read": 0.0825, + "cache_write": 0.33 + } + }, + "us.amazon.nova-micro-v1:0": { + "id": "us.amazon.nova-micro-v1:0", + "name": "Nova Micro (US)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 10000 + }, + "cost": { + "input": 0.035, + "output": 0.14, + "cache_read": 0.00875, + "cache_write": 0.035 + } + }, + "mistral.ministral-3-8b-instruct": { + "id": "mistral.ministral-3-8b-instruct", + "name": "Ministral 3 8B", + "description": "Compact open vision-language model for edge deployment, instruction following, and tool use", + "family": "ministral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.15, + "output": 0.15 + } + } + } + }, + "synthetic": { + "id": "synthetic", + "env": [ + "SYNTHETIC_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.synthetic.new/openai/v1", + "name": "Synthetic", + "doc": "https://synthetic.new/pricing", + "models": { + "hf:moonshotai/Kimi-K3": { + "id": "hf:moonshotai/Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-16", + "last_updated": "2026-07-27", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 65536 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.45 + } + }, + "hf:moonshotai/Kimi-K2.7-Code": { + "id": "hf:moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.95 + } + }, + "hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4": { + "id": "hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4", + "name": "Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 1, + "cache_read": 0.3 + } + }, + "hf:zai-org/GLM-5.2": { + "id": "hf:zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 65536 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 1.4 + } + }, + "hf:zai-org/GLM-5.3-Flash": { + "id": "hf:zai-org/GLM-5.3-Flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 65536 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.04 + } + }, + "hf:zai-org/GLM-4.7-Flash": { + "id": "hf:zai-org/GLM-4.7-Flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.5, + "cache_read": 0.1 + } + }, + "hf:deepseek-ai/DeepSeek-V4.1-Flash": { + "id": "hf:deepseek-ai/DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 65536 + }, + "cost": { + "input": 0.6, + "output": 1.2, + "cache_read": 0.03 + } + }, + "hf:Qwen/Qwen3.6-27B": { + "id": "hf:Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.45, + "output": 3.6, + "cache_read": 0.45 + } + }, + "hf:openai/gpt-oss-120b": { + "id": "hf:openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.1, + "cache_read": 0.1 + } + }, + "hf:MiniMaxAI/MiniMax-M3": { + "id": "hf:MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 65536 + }, + "cost": { + "input": 0.6, + "output": 1.2, + "cache_read": 0.6 + } + } + } + }, + "llmgateway": { + "id": "llmgateway", + "env": [ + "LLMGATEWAY_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.llmgateway.io/v1", + "name": "DevPass (LLM Gateway)", + "doc": "https://llmgateway.io/docs", + "models": { + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "Grok 4.1 Fast Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 30000 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } + }, + "glm-4.6v": { + "id": "glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 0.9, + "cache_read": 0.05 + } + }, + "qwen-flash": { + "id": "qwen-flash", + "name": "Qwen Flash", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 32768 + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0.0625 + } + }, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } + }, + "glm-4.5": { + "id": "glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131000, + "output": 98304 + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 + } + }, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180 + } + }, + "mimo-v2.6-pro": { + "id": "mimo-v2.6-pro", + "name": "MiMo-V2.6-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 + } + }, + "gemma-4-31b-it": { + "id": "gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.25, + "cache_read": 0.01 + } + }, + "grok-4-5": { + "id": "grok-4-5", + "name": "Grok 4.5", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.3 + } + }, + "gpt-3.5-turbo": { + "id": "gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16385, + "output": 4096 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0 + } + }, + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "Grok 4.1 Fast Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 2000000 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } + }, + "qwen3-vl-235b-a22b-instruct": { + "id": "qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-09-15", + "last_updated": "2025-09-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 8192 + }, + "cost": { + "input": 0.2, + "output": 0.88, + "cache_read": 0.11 + } + }, + "seed-1-6-flash-250715": { + "id": "seed-1-6-flash-250715", + "name": "Seed 1.6 Flash (250715)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.07, + "output": 0.3, + "cache_read": 0.015 + } + }, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 3.75, + "cache_read": 0.25, + "cache_write": 3.125 + } + }, + "muse-spark-1.3": { + "id": "muse-spark-1.3", + "name": "Muse Spark 1.3", + "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 + } + }, + "Qwen3.8-27B": { + "id": "Qwen3.8-27B", + "name": "Qwen3.8 27B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.08, + "output": 0.35, + "cache_read": 0.05 + } + }, + "gpt-4o-transcribe": { + "id": "gpt-4o-transcribe", + "name": "GPT-4o Transcribe", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-03-20", + "last_updated": "2025-03-20", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16000, + "output": 16000 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028, + "tiers": [ + { + "input": 0.8, + "output": 4, + "cache_read": 0.16, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.8, + "output": 4, + "cache_read": 0.16 + } + } + }, + "gpt-5.5-pro": { + "id": "gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } + } + }, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "mistral-large-latest": { + "id": "mistral-large-latest", + "name": "Mistral Large (latest)", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 262144 + }, + "cost": { + "input": 4, + "output": 12 + } + }, + "nemotron-3.5-lightning": { + "id": "nemotron-3.5-lightning", + "name": "Nemotron 3.5 Lightning 30B A3B", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.08, + "output": 0.2 + } + }, + "qwen3-vl-235b-a22b-thinking": { + "id": "qwen3-vl-235b-a22b-thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-09-15", + "last_updated": "2025-09-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.98, + "output": 3.95 + } + }, + "qwen3.8-2.4t-a95b": { + "id": "qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1010000, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25 + } + }, + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "qwen-coder-plus": { + "id": "qwen-coder-plus", + "name": "Qwen Coder Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-09-18", + "last_updated": "2024-09-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.502, + "output": 1.004 + } + }, + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.24 + } + }, + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "qwen3-vl-plus": { + "id": "qwen3-vl-plus", + "name": "Qwen3-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.2, + "output": 1.6, + "reasoning": 4.8, + "cache_read": 0.04, + "cache_write": 0.25 + } + }, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.088, + "output": 0.25, + "cache_read": 0.025 + } + }, + "fugu-max": { + "id": "fugu-max", + "name": "Fugu Max", + "description": "Multi-agent model for routing expert agents across complex analytical tasks", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25 + } + }, + "ministral-8b-2512": { + "id": "ministral-8b-2512", + "name": "Ministral 8B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "mistral", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.15 + } + }, + "glm-4.6": { + "id": "glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.55, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 + } + }, + "glm-5.2-fast": { + "id": "glm-5.2-fast", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 2.2, + "output": 6.5, + "cache_read": 0.45 + } + }, + "muse-glimmer-30b": { + "id": "muse-glimmer-30b", + "name": "Muse Glimmer 30B", + "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-01-04", + "release_date": "2026-08-10", + "last_updated": "2026-08-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.04 + } + }, + "grok-build-0-1": { + "id": "grok-build-0-1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 1, + "output": 2, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 4, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 4, + "cache_read": 0.4 + } + } + }, + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct (2507)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-08", + "last_updated": "2025-07-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 8192 + }, + "cost": { + "input": 0.09, + "output": 0.58 + } + }, + "qwen-max": { + "id": "qwen-max", + "name": "Qwen Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 8192 + }, + "cost": { + "input": 1.6, + "output": 6.4 + } + }, + "qwen3-next-80b-a3b-thinking": { + "id": "qwen3-next-80b-a3b-thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 1.2 + } + }, + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max Preview", + "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-07-19", + "last_updated": "2026-07-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.25, + "cache_write": 2.5 + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.108, + "output": 0.675, + "cache_read": 0.06 + } + }, + "glm-4.5-x": { + "id": "glm-4.5-x", + "name": "GLM-4.5 X", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "status": "beta", + "cost": { + "input": 2.2, + "output": 8.9, + "cache_read": 0.45 + } + }, + "grok-4": { + "id": "grok-4", + "name": "Grok 4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-07-09", + "last_updated": "2025-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.75 + } + }, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "llama-4-maverick-17b-instruct": { + "id": "llama-4-maverick-17b-instruct", + "name": "Llama 4 Maverick 17B Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-04-05", + "last_updated": "2025-04-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 2048 + }, + "cost": { + "input": 0.27, + "output": 0.85 + } + }, + "gpt-5.2-pro": { + "id": "gpt-5.2-pro", + "name": "GPT-5.2 Pro", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 21, + "output": 168 + } + }, + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 + } + }, + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 + } + }, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 203000, + "output": 131072 + }, + "cost": { + "input": 0.72, + "output": 2.3, + "cache_read": 0.144, + "cache_write": 0 + } + }, + "fugu-ultra-v2.0": { + "id": "fugu-ultra-v2.0", + "name": "Fugu Ultra v2.0", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-11", + "last_updated": "2026-09-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + } + }, + "gpt-4": { + "id": "gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 30, + "output": 60 + } + }, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1050000, + "output": 384000 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.003 + } + }, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 + } + }, + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 228700, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "kimi-k3-fast": { + "id": "kimi-k3-fast", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1040384, + "output": 131072 + }, + "cost": { + "input": 4.5, + "output": 22.5, + "cache_read": 0.45 + } + }, + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 40960, + "output": 16384 + }, + "cost": { + "input": 0.36, + "output": 0.87, + "reasoning": 8.4 + } + }, + "muse-spark-1.1": { + "id": "muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-08", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 + } + }, + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 + } + }, + "qwen-plus": { + "id": "qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 1.2, + "reasoning": 4, + "cache_read": 0.08, + "cache_write": 0.5 + } + }, + "kimi-k2.7-code-highspeed": { + "id": "kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.9, + "output": 8, + "cache_read": 0.38 + } + }, + "qwen3.7-flash": { + "id": "qwen3.7-flash", + "name": "Qwen3.7 Flash", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-07-27", + "last_updated": "2026-07-27", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 0.03, + "output": 0.13, + "cache_read": 0.006, + "cache_write": 0.0375 + } + }, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 3.05, + "cache_read": 0.13 + } + }, + "minimax-m2.1-lightning": { + "id": "minimax-m2.1-lightning", + "name": "MiniMax M2.1 Lightning", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 131072 + }, + "cost": { + "input": 0.12, + "output": 0.48 + } + }, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.845, + "output": 3.38, + "cache_read": 0.6, + "cache_write": 3.75 + } + }, + "claude-opus-5-5": { + "id": "claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 + } + }, + "glm-4.5-airx": { + "id": "glm-4.5-airx", + "name": "GLM-4.5 AirX", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 1.1, + "output": 4.5, + "cache_read": 0.22 + } + }, + "glm-4.5v": { + "id": "glm-4.5v", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.6, + "output": 1.8, + "cache_read": 0.11 + } + }, + "claude-fable-5-1": { + "id": "claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "devstral-2512": { + "id": "devstral-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2 + } + }, + "llama-4-scout-17b-instruct": { + "id": "llama-4-scout-17b-instruct", + "name": "Llama 4 Scout 17B Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-04-05", + "last_updated": "2025-04-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 2048 + }, + "cost": { + "input": 0.18, + "output": 0.59 + } + }, + "gemini-3.6-flash": { + "id": "gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.08333 + } + }, + "qwen3-235b-a22b-fp8": { + "id": "qwen3-235b-a22b-fp8", + "name": "Qwen3 235B A22B FP8", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-04-28", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 40960, + "output": 8192 + }, + "cost": { + "input": 0.2, + "output": 0.8 + } + }, + "o1": { + "id": "o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 15, + "output": 60, + "cache_read": 7.5 + } + }, + "glm-4-32b-0414-128k": { + "id": "glm-4-32b-0414-128k", + "name": "GLM-4 32B (0414-128k)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.1, + "output": 0.1 + } + }, + "qwen3-coder-480b-a35b-instruct": { + "id": "qwen3-coder-480b-a35b-instruct", + "name": "Qwen3-Coder 480B-A35B Instruct", + "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.38, + "output": 1.55 + } + }, + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 65536 + }, + "cost": { + "input": 0.07, + "output": 0.27 + } + }, + "sonar-pro": { + "id": "sonar-pro", + "name": "Sonar Pro", + "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", + "family": "sonar-pro", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "cost": { + "input": 3, + "output": 15 + } + }, + "ling-3.0-flash": { + "id": "ling-3.0-flash", + "name": "InclusionAI Ling 3.0 Flash", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-02", + "last_updated": "2026-08-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.06, + "output": 0.18, + "cache_read": 0.012 + } + }, + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 2.5, + "cache_read": 0.06 + } + }, + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 272000 + }, + "cost": { + "input": 15, + "output": 120 + } + }, + "qwen3-vl-30b-a3b-instruct": { + "id": "qwen3-vl-30b-a3b-instruct", + "name": "Qwen3 VL 30B A3B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-02", + "last_updated": "2025-10-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "seed-1-6-250915": { + "id": "seed-1-6-250915", + "name": "Seed 1.6 (250915)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-09-15", + "last_updated": "2025-09-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.05 + } + }, + "gemma-4-26b-a4b-it": { + "id": "gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.07, + "output": 0.34 + } + }, + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "cache_write": 0.08333 + } + }, + "llama-3-70b-instruct": { + "id": "llama-3-70b-instruct", + "name": "Llama 3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2024-04-18", + "last_updated": "2024-04-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8000 + }, + "cost": { + "input": 0.51, + "output": 0.74 + } + }, + "qwen3-vl-flash": { + "id": "qwen3-vl-flash", + "name": "Qwen3 VL Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-09", + "last_updated": "2025-10-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32000 + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.01 + } + }, + "qwen3-235b-a22b-thinking-2507": { + "id": "qwen3-235b-a22b-thinking-2507", + "name": "Qwen3 235B A22B Thinking (2507)", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-08", + "last_updated": "2025-07-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 3 + } + }, + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-09-03", + "last_updated": "2026-09-03", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "output": 1050000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "minimax-text-01": { + "id": "minimax-text-01", + "name": "MiniMax Text 01", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-01-15", + "last_updated": "2025-01-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.2, + "output": 1.1 + } + }, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "glm-4.6v-flashx": { + "id": "glm-4.6v-flashx", + "name": "GLM-4.6V FlashX", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16000 + }, + "cost": { + "input": 0.04, + "output": 0.4, + "cache_read": 0.004 + } + }, + "grok-4-20-beta-0309-non-reasoning": { + "id": "grok-4-20-beta-0309-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 30000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.405, + "output": 1.98, + "cache_read": 0.225 + } + }, + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 4096 + }, + "cost": { + "input": 0.135, + "output": 0.4 + } + }, + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 + } + }, + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.248, + "output": 1.485 + } + }, + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 1.2 + } + }, + "fugu-ultra": { + "id": "fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-06-22", + "last_updated": "2026-06-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + } + }, + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32766 + }, + "cost": { + "input": 0.04, + "output": 0.19, + "cache_read": 0.01 + } + }, + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.08, + "output": 0.32, + "cache_read": 0.017, + "cache_write": 0.375 + } + }, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 0.08333 + } + }, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } + } + }, + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.05, + "cache_write": 0.3125 + } + }, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1, + "max": 24576 + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 + } + }, + "qwen3-coder-flash": { + "id": "qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 1.5, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "nemotron-3-ultra-550b": { + "id": "nemotron-3-ultra-550b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 128000 + }, + "cost": { + "input": 0.5, + "output": 2.2, + "cache_read": 0.1 + } + }, + "llama-3.2-11b-instruct": { + "id": "llama-3.2-11b-instruct", + "name": "Llama 3.2 11B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2024-09-25", + "last_updated": "2024-09-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.07, + "output": 0.33 + } + }, + "inkling-small": { + "id": "inkling-small", + "name": "Inkling Small", + "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-07-30", + "last_updated": "2026-07-30", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 1048576 + }, + "cost": { + "input": 0.45, + "output": 1.2, + "cache_read": 0.1 + } + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } + } + }, + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.2, + "cache_write": 1.25 + } + }, + "glm-4.7-flash": { + "id": "glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0.06, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0 + } + }, + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.15 + } + }, + "inkling": { + "id": "inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 1048576 + }, + "cost": { + "input": 0.95, + "output": 4.05, + "cache_read": 0.16 + } + }, + "muse-spark-1.2": { + "id": "muse-spark-1.2", + "name": "Muse Spark 1.2", + "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-05", + "last_updated": "2026-08-05", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 + } + }, + "mimo-v2.6-flash": { + "id": "mimo-v2.6-flash", + "name": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 + } + }, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "seed-1-8-251228": { + "id": "seed-1-8-251228", + "name": "Seed 1.8 (251228)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-18", + "last_updated": "2025-12-18", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.05 + } + }, + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.38, + "output": 1.98, + "cache_read": 0.19, + "cache_write": 0 + } + }, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 512000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 + } + }, + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.47, + "cache_read": 0.016, + "cache_write": 0.2 + } + }, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "seed-1-6-250615": { + "id": "seed-1-6-250615", + "name": "Seed 1.6 (250615)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-06-25", + "last_updated": "2025-06-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.05 + } + }, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "qwen35-397b-a17b": { + "id": "qwen35-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.6, + "output": 3.6 + } + }, + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": false, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "qwen3.6-max-preview": { + "id": "qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 1.3, + "output": 7.8, + "cache_read": 0.13, + "cache_write": 1.625 + } + }, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.08333 + } + }, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.8, + "output": 2.55, + "cache_read": 0.16, + "cache_write": 0 + } + }, + "ernie-4.5-vl-424b-a47b": { + "id": "ernie-4.5-vl-424b-a47b", + "name": "ERNIE 4.5 VL 424B A47B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ernie", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-06-30", + "last_updated": "2025-06-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 123000, + "output": 123000 + }, + "cost": { + "input": 0.42, + "output": 1.25 + } + }, + "hy3": { + "id": "hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "input": 192000, + "output": 128000 + }, + "cost": { + "input": 0.132, + "output": 0.528, + "cache_read": 0.033 + } + }, + "kimi-k2": { + "id": "kimi-k2", + "name": "Kimi K2", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-11", + "last_updated": "2025-07-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 16384 + }, + "cost": { + "input": 0.57, + "output": 2.3, + "cache_read": 0.5 + } + }, + "qwen-omni-turbo": { + "id": "qwen-omni-turbo", + "name": "Qwen-Omni Turbo", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-01-19", + "last_updated": "2025-03-26", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text", + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 2048 + }, + "cost": { + "input": 0.2, + "output": 0.8 + } + }, + "hy-mt2-plus": { + "id": "hy-mt2-plus", + "name": "Hy-MT2 Plus", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "Hy", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0.074, + "output": 0.295 + } + }, + "grok-4-3": { + "id": "grok-4-3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 30000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.931, + "output": 2.93, + "cache_read": 0.173, + "cache_write": 0 + } + }, + "grok-4-20-non-reasoning": { + "id": "grok-4-20-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 30000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05 + } + }, + "grok-4-20-beta-0309-reasoning": { + "id": "grok-4-20-beta-0309-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 30000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "minimax-m2.7-highspeed": { + "id": "minimax-m2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 + } + }, + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "glm-4.7-flashx": { + "id": "glm-4.7-flashx", + "name": "GLM-4.7-FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0.07, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0 + } + }, + "muse-spark-1.2-contributor": { + "id": "muse-spark-1.2-contributor", + "name": "Muse Spark 1.2 Contributor", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-06", + "last_updated": "2026-08-06", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 + } + }, + "sonar": { + "id": "sonar", + "name": "Sonar", + "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", + "family": "sonar", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 130000, + "output": 4096 + }, + "cost": { + "input": 1, + "output": 1 + } + }, + "mistral-large-2512": { + "id": "mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "grok-4-6": { + "id": "grok-4-6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "gemini-pro-latest": { + "id": "gemini-pro-latest", + "name": "Gemini Pro Latest", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-27", + "last_updated": "2026-02-27", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2 + } + }, + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 + } + }, + "minimax-m2.5-highspeed": { + "id": "minimax-m2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-13", + "last_updated": "2026-02-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.03, + "cache_write": 0.375 + } + }, + "gemini-3.8-flash": { + "id": "gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.08333 + } + }, + "ministral-14b-2512": { + "id": "ministral-14b-2512", + "name": "Ministral 14B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "mistral", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 8192 + }, + "cost": { + "input": 0.2, + "output": 0.2 + } + }, + "glm-5-turbo": { + "id": "glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.24 + } + }, + "gpt-6-luna": { + "id": "gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125 + } + }, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1050000, + "output": 384000 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 + } + }, + "step-3.7-flash": { + "id": "step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2026-03-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 256000 + }, + "cost": { + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 + } + }, + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 + } + }, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "gpt-4o-mini-transcribe": { + "id": "gpt-4o-mini-transcribe", + "name": "GPT-4o Mini Transcribe", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-03-20", + "last_updated": "2025-03-20", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16000, + "output": 16000 + }, + "cost": { + "input": 1.25, + "output": 5 + } + }, + "sonar-reasoning-pro": { + "id": "sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", + "family": "sonar-reasoning", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 2, + "output": 8 + } + }, + "llama-3.2-3b-instruct": { + "id": "llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2024-09-18", + "last_updated": "2024-09-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32000 + }, + "cost": { + "input": 0.03, + "output": 0.05 + } + }, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } + } + }, + "mistral-small-2506": { + "id": "mistral-small-2506", + "name": "Mistral Small 3.2", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "grok-4-20-reasoning": { + "id": "grok-4-20-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 30000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "grok-4-7": { + "id": "grok-4-7", + "name": "Grok 4.7", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-09-21", + "last_updated": "2026-09-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 16384 + }, + "cost": { + "input": 0.26, + "output": 0.38, + "cache_read": 0.13 + } + }, + "gpt-4-turbo": { + "id": "gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 10, + "output": 30 + } + }, + "atria-dawn-preview": { + "id": "atria-dawn-preview", + "name": "Atria Dawn Preview", + "description": "Preview model for early access evaluation, prototyping, and compatibility testing", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-09-12", + "last_updated": "2026-09-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "hy4-preview": { + "id": "hy4-preview", + "name": "Hy4 preview", + "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-08-28", + "last_updated": "2026-08-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.834, + "output": 2.501, + "cache_read": 0.042 + } + }, + "muse-spark-1.3-contributor": { + "id": "muse-spark-1.3-contributor", + "name": "Muse Spark 1.3 Contributor", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "muse", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 0.1, + "output": 0.2, + "cache_read": 0.002 + } + }, + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131000, + "output": 98304 + }, + "cost": { + "input": 0.13, + "output": 0.85, + "cache_read": 0.025, + "cache_write": 0 + } + }, + "qwen-plus-latest": { + "id": "qwen-plus-latest", + "name": "Qwen Plus Latest", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-01-25", + "last_updated": "2025-01-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 8192 + }, + "cost": { + "input": 0.4, + "output": 1.2, + "cache_read": 0.08, + "cache_write": 0.5 + } + }, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32766 + }, + "cost": { + "input": 0.032, + "output": 0.14, + "cache_read": 0.032 + } + }, + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "ministral-3b-2512": { + "id": "ministral-3b-2512", + "name": "Ministral 3B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "mistral", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.1, + "output": 0.1 + } + }, + "o3": { + "id": "o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.2 + } + }, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + }, + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "cache_write": 0.5 + } + }, + "minimax-m2.1": { + "id": "minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.27, + "output": 1.1 + } + }, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5 + } + }, + "auto": { + "id": "auto", + "name": "Auto Route", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "llama-3.1-70b-instruct": { + "id": "llama-3.1-70b-instruct", + "name": "Llama 3.1 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 2048 + }, + "status": "beta", + "cost": { + "input": 0.72, + "output": 0.72 + } + }, + "minimax-m2": { + "id": "minimax-m2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 131072 + }, + "cost": { + "input": 0.2, + "output": 1, + "cache_read": 0.03 + } + }, + "custom": { + "id": "custom", + "name": "Custom Model", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "codestral-2508": { + "id": "codestral-2508", + "name": "Codestral", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "mistral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-30", + "last_updated": "2025-07-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 16384 + }, + "cost": { + "input": 0.3, + "output": 0.9 + } + }, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1050000, + "output": 384000 + }, + "cost": { + "input": 0.05, + "output": 0.1, + "cache_read": 0.01 + } + }, + "gpt-6-sol": { + "id": "gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 0.08333 + } + } + } + }, + "sap-ai-core": { + "id": "sap-ai-core", + "env": [ + "AICORE_SERVICE_KEY" + ], + "npm": "@jerome-benoit/sap-ai-provider-v2", + "name": "SAP AI Core", + "doc": "https://help.sap.com/docs/sap-ai-core", + "models": { + "anthropic--claude-4-opus": { + "id": "anthropic--claude-4-opus", + "name": "anthropic--claude-4-opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "gpt-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } + }, + "nvidia--llama-3.2-nv-embedqa-1b": { + "id": "nvidia--llama-3.2-nv-embedqa-1b", + "name": "nvidia--llama-3.2-nv-embedqa-1b", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-09-25", + "last_updated": "2024-09-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 4096 + }, + "cost": { + "input": 0.07, + "output": 0 + } + }, + "text-embedding-3-small": { + "id": "text-embedding-3-small", + "name": "text-embedding-3-small", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-01-25", + "last_updated": "2024-01-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8191, + "output": 1536 + }, + "cost": { + "input": 0.02, + "output": 0 + } + }, + "sap-abap-1": { + "id": "sap-abap-1", + "name": "sap-abap-1", + "description": "SAP-hosted model for ABAP code generation and enterprise development tasks", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-11-26", + "last_updated": "2025-11-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 4096 + }, + "cost": { + "input": 0.48, + "output": 1.7 + } + }, + "anthropic--claude-4.8-opus": { + "id": "anthropic--claude-4.8-opus", + "name": "anthropic--claude-4.8-opus", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "gpt-5-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "anthropic--claude-3-haiku": { + "id": "anthropic--claude-3-haiku", + "name": "anthropic--claude-3-haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-08-31", + "release_date": "2024-03-13", + "last_updated": "2024-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 4096 + }, + "cost": { + "input": 0.25, + "output": 1.25, + "cache_read": 0.03, + "cache_write": 0.3 + } + }, + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "gpt-4.1-nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.08, + "output": 0.26 + } + }, + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "gpt-5-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 + } + }, + "anthropic--claude-4.5-haiku": { + "id": "anthropic--claude-4.5-haiku", + "name": "anthropic--claude-4.5-haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "anthropic--claude-4.7-opus": { + "id": "anthropic--claude-4.7-opus", + "name": "anthropic--claude-4.7-opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gemini-embedding": { + "id": "gemini-embedding", + "name": "Gemini Embedding 001", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "gemini", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-05", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2048, + "output": 1 + } + }, + "gemini-embedding-2": { + "id": "gemini-embedding-2", + "name": "Gemini Embedding 2", + "description": "Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space", + "family": "gemini", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-11", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 3072 + } + }, + "amazon--nova-pro": { + "id": "amazon--nova-pro", + "name": "amazon--nova-pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 8192 + }, + "cost": { + "input": 0.56, + "output": 2.13 + } + }, + "amazon--nova-lite": { + "id": "amazon--nova-lite", + "name": "amazon--nova-lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.3, + "output": 2.37 + } + }, + "sonar-pro": { + "id": "sonar-pro", + "name": "sonar-pro", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "cost": { + "input": 3, + "output": 15 + } + }, + "anthropic--claude-3.5-sonnet": { + "id": "anthropic--claude-3.5-sonnet", + "name": "anthropic--claude-3.5-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04-30", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "status": "deprecated", + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "text-embedding-3-large": { + "id": "text-embedding-3-large", + "name": "text-embedding-3-large", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-01-25", + "last_updated": "2024-01-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8191, + "output": 3072 + }, + "cost": { + "input": 0.09, + "output": 0 + } + }, + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + } + }, + "amazon--nova-micro": { + "id": "amazon--nova-micro", + "name": "amazon--nova-micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.03, + "output": 0.1 + } + }, + "anthropic--claude-4.5-opus": { + "id": "anthropic--claude-4.5-opus", + "name": "anthropic--claude-4.5-opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic--claude-4-sonnet": { + "id": "anthropic--claude-4-sonnet", + "name": "anthropic--claude-4-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "gemini-3.5-flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "gemini-2.5-pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-03-25", + "last_updated": "2025-06-05", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "anthropic--claude-3-sonnet": { + "id": "anthropic--claude-3-sonnet", + "name": "anthropic--claude-3-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-08-31", + "release_date": "2024-03-04", + "last_updated": "2024-03-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 4096 + }, + "status": "deprecated", + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "sonar-deep-research": { + "id": "sonar-deep-research", + "name": "sonar-deep-research", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar-deep-research", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-02-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "reasoning": 3 + } + }, + "mistralai--mistral-small": { + "id": "mistralai--mistral-small", + "name": "mistralai--mistral-small", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.07, + "output": 0.28 + } + }, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "gemini-2.5-flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-04-17", + "last_updated": "2025-06-05", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 + } + }, + "anthropic--claude-3.7-sonnet": { + "id": "anthropic--claude-3.7-sonnet", + "name": "anthropic--claude-3.7-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-10-31", + "release_date": "2025-02-24", + "last_updated": "2025-02-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "status": "deprecated", + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "gpt-5.6-luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 1, + "output": 6, + "cache_read": 0.1, + "tiers": [ + { + "input": 2, + "output": 9, + "cache_read": 0.2, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 9, + "cache_read": 0.2 + } + } + }, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "gpt-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 9.44, + "cache_read": 0.12 + } + }, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "gpt-5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + } + }, + "anthropic--claude-4.6-opus": { + "id": "anthropic--claude-4.6-opus", + "name": "anthropic--claude-4.6-opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "mistralai--mistral-medium-instruct": { + "id": "mistralai--mistral-medium-instruct", + "name": "mistralai--mistral-medium-instruct", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.36, + "output": 1.22 + } + }, + "gpt-4.1": { + "id": "gpt-4.1", + "name": "gpt-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.32 + } + }, + "anthropic--claude-3-opus": { + "id": "anthropic--claude-3-opus", + "name": "anthropic--claude-3-opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-08-31", + "release_date": "2024-02-29", + "last_updated": "2024-02-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 4096 + }, + "status": "deprecated", + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "anthropic--claude-4.6-sonnet": { + "id": "anthropic--claude-4.6-sonnet", + "name": "anthropic--claude-4.6-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "sonar": { + "id": "sonar", + "name": "sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 1, + "output": 1 + } + }, + "amazon--titan-embed-text": { + "id": "amazon--titan-embed-text", + "name": "amazon--titan-embed-text", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-04-30", + "last_updated": "2024-04-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 1536 + }, + "cost": { + "input": 0.14, + "output": 0 + } + }, + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "gpt-4.1-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 + } + }, + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "gemini-2.5-flash-lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 512, + "max": 24576 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "input_audio": 0.3 + } + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "gpt-5.6-terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } + } + }, + "mistralai--mistral-medium": { + "id": "mistralai--mistral-medium", + "name": "Mistral Medium 3.5", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + } + }, + "gpt-5": { + "id": "gpt-5", + "name": "gpt-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "anthropic--claude-4.5-sonnet": { + "id": "anthropic--claude-4.5-sonnet", + "name": "anthropic--claude-4.5-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "gpt-5.6-sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "cohere--command-a-reasoning": { + "id": "cohere--command-a-reasoning", + "name": "cohere--command-a-reasoning", + "description": "Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows", + "family": "command-a", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 32000 + }, + "cost": { + "input": 0.63, + "output": 5.05 + } + }, + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "gemini-3.1-flash-lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 + } + } + } + }, + "vivgrid": { + "id": "vivgrid", + "env": [ + "VIVGRID_API_KEY" + ], + "npm": "@ai-sdk/openai", + "api": "https://api.vivgrid.com/v1", + "name": "Vivgrid", + "doc": "https://docs.vivgrid.com/models", + "models": { + "viv-fast": { + "id": "viv-fast", + "name": "Viv Fast", + "description": "Fast coding model", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-09", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 256000 + }, + "cost": { + "input": 0.13, + "output": 0.4, + "cache_read": 0.05 + } + }, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } + }, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-01-14", + "last_updated": "2026-01-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.04 + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.03 + } + }, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.31, + "output": 1.23, + "cache_read": 0.01 + } + }, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "claude-fable-5-1": { + "id": "claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.5, + "cache_write": 12.5 + } + }, + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } + } + }, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1.25, + "cache_write": 12.5 + } + }, + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 1.35, + "output": 3, + "reasoning": 3, + "cache_read": 0.05 + } + }, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT 5.6 Luna", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 1, + "output": 6, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "gpt-5.1-codex-max": { + "id": "gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075 + } + }, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 1.2, + "output": 4.2, + "cache_read": 0.3 + } + }, + "gemini-3.8-flash": { + "id": "gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.15 + } + }, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 + } + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT 5.6 Terra", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125 + } + }, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 0.28, + "output": 0.42 + } + }, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4.2, + "cache_read": 0.26 + } + }, + "gemini-3.1-flash-lite-preview": { + "id": "gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 1 + } + }, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT 5.6 Sol", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.15, + "output": 0.3, + "reasoning": 0.3, + "cache_read": 0.03 + } + } + } + }, + "klokintegration": { + "id": "klokintegration", + "env": [ + "KLOKINTEGRATION_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api-gw.klok.ipaas.se/proxy/kloker-key/v1", + "name": "klokintegration.se", + "doc": "https://klokintegration.se/docs/ai-api", + "models": { + "Kloker-Integration-Developer": { + "id": "Kloker-Integration-Developer", + "name": "Kloker Integration Developer", + "description": "Knows the customer integration environment and Klok best practices. Opinionated about implementation. The gateway runs ecosystem lookup tools server-side and appends a system-prompt injection. Client system prompts and OpenAI tool calls are preserved.", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-29", + "last_updated": "2026-08-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 50000 + }, + "status": "beta", + "cost": { + "input": 0.23, + "output": 1.16 + } + }, + "Kloker": { + "id": "Kloker", + "name": "Kloker", + "description": "Cheap general model with a clean context. Nothing from the customer environment is packed in. It tracks the current best open source model. The Klok team verifies it and upgrades it periodically.", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-29", + "last_updated": "2026-08-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 50000 + }, + "status": "beta", + "cost": { + "input": 0.23, + "output": 1.16 + } + }, + "Kloker-Integration-Architect": { + "id": "Kloker-Integration-Architect", + "name": "Kloker Integration Architect", + "description": "Knows the customer integration environment and Klok best practices. Opinionated about structure. The gateway runs ecosystem lookup tools server-side and appends a system-prompt injection (data contracts, CloudEvents, event-driven flows). Client system prompts and OpenAI tool calls are preserved.", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2026-08-29", + "last_updated": "2026-08-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 50000 + }, + "status": "beta", + "cost": { + "input": 0.23, + "output": 1.16 + } + } + } + }, + "google-vertex": { + "id": "google-vertex", + "env": [ + "GOOGLE_VERTEX_PROJECT", + "GOOGLE_VERTEX_LOCATION", + "GOOGLE_APPLICATION_CREDENTIALS" + ], + "npm": "@ai-sdk/google-vertex", + "name": "Vertex", + "doc": "https://cloud.google.com/vertex-ai/generative-ai/docs/models", + "models": { + "claude-opus-4-8@default": { + "id": "claude-opus-4-8@default", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "gemini-2.5-flash-tts": { + "id": "gemini-2.5-flash-tts", + "name": "Gemini 2.5 Flash TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gemini-flash", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-09-30", + "last_updated": "2025-12-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 0.5, + "output": 10 + } + }, + "gemini-flash-latest": { + "id": "gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "claude-sonnet-4-5@20250929": { + "id": "claude-sonnet-4-5@20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "gemini-2.5-flash-image": { + "id": "gemini-2.5-flash-image", + "name": "Nano Banana", + "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", + "family": "gemini-flash", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 30 + } + }, + "claude-opus-4-6@default": { + "id": "claude-opus-4-6@default", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "claude-opus-4@20250514": { + "id": "claude-opus-4@20250514", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "gemini-flash-lite-latest": { + "id": "gemini-flash-lite-latest", + "name": "Gemini Flash-Lite Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 + } + }, + "claude-opus-5-5@default": { + "id": "claude-opus-5-5@default", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 + } + }, + "gemini-3.6-flash": { + "id": "gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "input_audio": 0.75 + } + }, + "claude-opus-5@default": { + "id": "claude-opus-5@default", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "claude-sonnet-4-6@default": { + "id": "claude-sonnet-4-6@default", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } + } + }, + "claude-haiku-4-5@20251001": { + "id": "claude-haiku-4-5@20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 + } + }, + "gemini-3.1-flash-image": { + "id": "gemini-3.1-flash-image", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 60, + "cache_read": 0.05 + } + }, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "claude-opus-4-1@20250805": { + "id": "claude-opus-4-1@20250805", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "claude-opus-4-5@20251101": { + "id": "claude-opus-4-5@20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "claude-opus-4-7@default": { + "id": "claude-opus-4-7@default", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } + } + }, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 + } + }, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } + } + }, + "gemini-3-pro-image": { + "id": "gemini-3-pro-image", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 120, + "cache_read": 0.2 + } + }, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 + } + }, + "claude-fable-5-1@default": { + "id": "claude-fable-5-1@default", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "input_audio": 0.75 + } + }, + "gemini-3.1-pro-preview-customtools": { + "id": "gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } + } + }, + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 + } + }, + "claude-fable-5@default": { + "id": "claude-fable-5@default", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "gemini-3.8-flash": { + "id": "gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "input_audio": 0.75 + } + }, + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 512, + "max": 24576 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "input_audio": 0.3 + } + }, + "gemini-2.5-pro-tts": { + "id": "gemini-2.5-pro-tts", + "name": "Gemini 2.5 Pro TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gemini-pro", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-09-30", + "last_updated": "2025-12-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 1, + "output": 20 + } + }, + "gemini-embedding-001": { + "id": "gemini-embedding-001", + "name": "Gemini Embedding 001", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "gemini", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-05", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2048, + "output": 1 + }, + "cost": { + "input": 0.15, + "output": 0 + } + }, + "claude-sonnet-4@20250514": { + "id": "claude-sonnet-4@20250514", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "gemini-3.1-flash-lite-preview": { + "id": "gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "status": "deprecated", + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 + } + }, + "claude-sonnet-5@default": { + "id": "claude-sonnet-5@default", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 + } + }, + "meta/llama-4-maverick-17b-128e-instruct-maas": { + "id": "meta/llama-4-maverick-17b-128e-instruct-maas", + "name": "Llama 4 Maverick 17B 128E Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 8192 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.35, + "output": 1.15 + } + }, + "meta/llama-3.3-70b-instruct-maas": { + "id": "meta/llama-3.3-70b-instruct-maas", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.72, + "output": 0.72 + } + }, + "xai/grok-4.20-reasoning": { + "id": "xai/grok-4.20-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 30000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "xai/grok-4.20-non-reasoning": { + "id": "xai/grok-4.20-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 30000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "xai/grok-4.1-fast-reasoning": { + "id": "xai/grok-4.1-fast-reasoning", + "name": "Grok 4.1 Fast (Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 30000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } + }, + "xai/grok-4.3": { + "id": "xai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 30000 + }, + "status": "beta", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } + } + }, + "xai/grok-4.1-fast-non-reasoning": { + "id": "xai/grok-4.1-fast-non-reasoning", + "name": "Grok 4.1 Fast", + "description": "Fast Grok model for responsive chat, tool-assisted work, and low-latency responses", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 30000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } + }, + "xai/grok-4.6": { + "id": "xai/grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 524288, + "output": 500000 + }, + "status": "beta", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } + } + }, + "deepseek-ai/deepseek-v3.1-maas": { + "id": "deepseek-ai/deepseek-v3.1-maas", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-28", + "last_updated": "2025-08-28", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 32768 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.6, + "output": 1.7, + "cache_read": 0.06 + } + }, + "deepseek-ai/deepseek-v3.2-maas": { + "id": "deepseek-ai/deepseek-v3.2-maas", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-17", + "last_updated": "2026-04-04", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 65536 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.56, + "output": 1.68, + "cache_read": 0.056 + } + }, + "moonshotai/kimi-k2-thinking-maas": { + "id": "moonshotai/kimi-k2-thinking-maas", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.6, + "output": 2.5, + "cache_read": 0.06 + } + }, + "zai-org/glm-5-maas": { + "id": "zai-org/glm-5-maas", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.1 + } + }, + "zai-org/glm-4.7-maas": { + "id": "zai-org/glm-4.7-maas", + "name": "GLM-4.7", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-06", + "last_updated": "2026-01-06", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.6, + "output": 2.2, + "cache_read": 0.06 + } + }, + "zai-org/glm-5.2-maas": { + "id": "zai-org/glm-5.2-maas", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 64000 + }, + "status": "beta", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.14 + } + }, + "qwen/qwen3-235b-a22b-instruct-2507-maas": { + "id": "qwen/qwen3-235b-a22b-instruct-2507-maas", + "name": "Qwen3 235B A22B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-13", + "last_updated": "2025-08-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 16384 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.22, + "output": 0.88 + } + }, + "openai/gpt-oss-20b-maas": { + "id": "openai/gpt-oss-20b-maas", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.07, + "output": 0.25, + "cache_read": 0.007 + } + }, + "openai/gpt-oss-120b-maas": { + "id": "openai/gpt-oss-120b-maas", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.09, + "output": 0.36 + } + } + } + }, + "evroc": { + "id": "evroc", + "env": [ + "EVROC_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://models.think.evroc.com/v1", + "name": "evroc", + "doc": "https://docs.evroc.com/products/think/overview.html", + "models": { + "google/gemma-4-26B-A4B-it": { + "id": "google/gemma-4-26B-A4B-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.144, + "output": 0.575 + } + }, + "intfloat/multilingual-e5-large-instruct": { + "id": "intfloat/multilingual-e5-large-instruct", + "name": "E5 Multi-Lingual Large Embeddings 0.6B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2024-06-01", + "last_updated": "2024-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 512, + "output": 512 + }, + "cost": { + "input": 0.114, + "output": 0.114 + } + }, + "Qwen/Qwen3.8-27B": { + "id": "Qwen/Qwen3.8-27B", + "name": "Qwen3.8-27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.87, + "output": 3.5 + } + }, + "Qwen/Qwen3-Reranker-4B": { + "id": "Qwen/Qwen3-Reranker-4B", + "name": "Qwen3 Reranker 4B", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2025-07-30", + "last_updated": "2025-07-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32000, + "output": 4096 + }, + "cost": { + "input": 0.0575, + "output": 0 + } + }, + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.345, + "output": 1.38 + } + }, + "Qwen/Qwen3-Embedding-8B": { + "id": "Qwen/Qwen3-Embedding-8B", + "name": "Qwen3 Embedding 8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2025-07-30", + "last_updated": "2025-07-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 40960, + "output": 4096 + }, + "cost": { + "input": 0.115, + "output": 0.115 + } + }, + "mistralai/Mistral-Medium-3.5-128B": { + "id": "mistralai/Mistral-Medium-3.5-128B", + "name": "Mistral Medium 3.5", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.725, + "output": 6.9 + } + }, + "mistralai/Voxtral-Small-24B-2507": { + "id": "mistralai/Voxtral-Small-24B-2507", + "name": "Voxtral Small 24B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "voxtral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2025-03-01", + "last_updated": "2025-03-01", + "modalities": { + "input": [ + "audio", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.0023, + "output": 0.0023, + "output_audio": 2.3 + } + }, + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.4375, + "output": 5.75 + } + }, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 131072 + }, + "cost": { + "input": 1.4375, + "output": 5.75 + } + }, + "nvidia/Llama-3.3-70B-Instruct-FP8": { + "id": "nvidia/Llama-3.3-70B-Instruct-FP8", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 1.15, + "output": 1.15 + } + }, + "KBLab/kb-whisper-large": { + "id": "KBLab/kb-whisper-large", + "name": "KB Whisper", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", + "modalities": { + "input": [ + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 448, + "output": 448 + }, + "cost": { + "input": 0.0023, + "output": 0.0023, + "output_audio": 2.3 + } + }, + "evroc/roc": { + "id": "evroc/roc", + "name": "roc", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2026-01", + "release_date": "2026-06-06", + "last_updated": "2026-06-06", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 2.875, + "output": 11.516 + } + }, + "openai/whisper-large-v3": { + "id": "openai/whisper-large-v3", + "name": "Whisper 3 Large", + "description": "Open Whisper checkpoint for robust multilingual transcription and captioning", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", + "modalities": { + "input": [ + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 448, + "output": 4096 + }, + "cost": { + "input": 0.0023, + "output": 0.0023, + "output_audio": 2.3 + } + }, + "openai/whisper-large-v3-turbo": { + "id": "openai/whisper-large-v3-turbo", + "name": "Whisper Large v3 Turbo", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", + "modalities": { + "input": [ + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 448, + "output": 448 + }, + "cost": { + "input": 0.0023, + "output": 0.0023, + "output_audio": 2.3 + } + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 65536, + "output": 65536 + }, + "cost": { + "input": 0.23, + "output": 0.92 + } + } + } + }, + "tokengo": { + "id": "tokengo", + "env": [ + "TOKENGO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.tokengo.com/v1", + "name": "TokenGo", + "doc": "https://www.tokengo.com/docs", + "models": { + "deepseek/deepseek-v3.1": { + "id": "deepseek/deepseek-v3.1", + "name": "DeepSeek-V3.1", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-21", + "last_updated": "2025-08-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.19, + "output": 0.71, + "cache_read": 0.06 + } + }, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.435, + "output": 0.87 + } + }, + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 64000 + }, + "cost": { + "input": 0.2174, + "output": 0.326, + "cache_read": 0.06 + } + }, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.098, + "output": 0.196, + "cache_read": 0.028 + } + }, + "z-ai/glm-5.3-flash": { + "id": "z-ai/glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0.075, + "output": 0.025, + "cache_read": 0.015 + } + }, + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.89, + "output": 3.2647, + "cache_read": 0.2226 + } + }, + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "z-ai/glm-5.3": { + "id": "z-ai/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } + }, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 + } + }, + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.4, + "output": 2.65, + "cache_read": 0.2 + } + } + } + }, + "submodel": { + "id": "submodel", + "env": [ + "SUBMODEL_INSTAGEN_ACCESS_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://llm.submodel.ai/v1", + "name": "submodel", + "doc": "https://submodel.gitbook.io", + "models": { + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen3 235B A22B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.2, + "output": 0.6 + } + }, + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.2, + "output": 0.3 + } + }, + "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.2, + "output": 0.8 + } + }, + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 75000, + "output": 163840 + }, + "cost": { + "input": 0.2, + "output": 0.8 + } + }, + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 75000, + "output": 163840 + }, + "cost": { + "input": 0.5, + "output": 2.15 + } + }, + "deepseek-ai/DeepSeek-V3-0324": { + "id": "deepseek-ai/DeepSeek-V3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 75000, + "output": 163840 + }, + "cost": { + "input": 0.2, + "output": 0.8 + } + }, + "zai-org/GLM-4.5-FP8": { + "id": "zai-org/GLM-4.5-FP8", + "name": "GLM 4.5 FP8", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.2, + "output": 0.8 + } + }, + "zai-org/GLM-4.5-Air": { + "id": "zai-org/GLM-4.5-Air", + "name": "GLM 4.5 Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.5 + } + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.5 + } + } + } + }, + "kosmik": { + "id": "kosmik", + "env": [ + "KOSMIK_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.koscompute.com/v1", + "name": "Kosmik Compute", + "doc": "https://api.koscompute.com/docs/", + "models": { + "qwen/qwen3.8-27b": { + "id": "qwen/qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.35, + "output": 2.2, + "cache_read": 0.09 + } + } + } + }, + "tencent-token-plan": { + "id": "tencent-token-plan", + "env": [ + "TENCENT_TOKEN_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.lkeap.cloud.tencent.com/plan/v3", + "name": "Tencent Token Plan", + "doc": "https://cloud.tencent.com/document/product/1823/130060", + "models": { + "hy3": { + "id": "hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 192000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "hy4-preview": { + "id": "hy4-preview", + "name": "Hy4 preview", + "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-08-28", + "last_updated": "2026-08-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1024000, + "output": 64000 + }, + "cost": { + "input": 0.834, + "output": 2.501, + "cache_read": 0.042 + } + } + } + }, + "togetherai": { + "id": "togetherai", + "env": [ + "TOGETHER_API_KEY" + ], + "npm": "@ai-sdk/togetherai", + "name": "Together AI", + "doc": "https://docs.together.ai/docs/serverless-models", + "models": { + "meta-llama/Llama-3.3-70B-Instruct-Turbo": { + "id": "meta-llama/Llama-3.3-70B-Instruct-Turbo", + "name": "Llama 3.3 70B", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2026-07-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 1.04, + "output": 1.04 + } + }, + "meta-llama/Meta-Llama-3-8B-Instruct-Lite": { + "id": "meta-llama/Meta-Llama-3-8B-Instruct-Lite", + "name": "Meta Llama 3 8B Instruct Lite", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-04-18", + "last_updated": "2024-04-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0.14, + "output": 0.14 + } + }, + "pearl-ai/gemma-4-31b-it": { + "id": "pearl-ai/gemma-4-31b-it", + "name": "Pearl AI Gemma 4 31B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.28, + "output": 0.86 + } + }, + "deepcogito/cogito-v2-1-671b": { + "id": "deepcogito/cogito-v2-1-671b", + "name": "Cogito v2.1 671B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "cogito", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "temperature": true, + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 163840, + "output": 163840 + }, + "cost": { + "input": 1.25, + "output": 1.25 + } + }, + "thinkingmachines/Inkling": { + "id": "thinkingmachines/Inkling", + "name": "Inkling", + "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", + "family": "ling", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "max", + "xhigh", + "high", + "medium", + "low", + "none" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 131072 + }, + "cost": { + "input": 1, + "output": 4.05, + "cache_read": 0.17 + } + }, + "essentialai/Rnj-1-Instruct": { + "id": "essentialai/Rnj-1-Instruct", + "name": "Rnj-1 Instruct", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "rnj", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-05", + "last_updated": "2025-12-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "status": "deprecated", + "cost": { + "input": 0.15, + "output": 0.15 + } + }, + "google/gemma-3n-E4B-it": { + "id": "google/gemma-3n-E4B-it", + "name": "Gemma 3N E4B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-05-20", + "last_updated": "2025-05-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.06, + "output": 0.12 + } + }, + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.39, + "output": 0.97 + } + }, + "Qwen/Qwen3.7-Max": { + "id": "Qwen/Qwen3.7-Max", + "name": "Qwen3.7 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-07-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 500000 + }, + "cost": { + "input": 1.25, + "output": 3.75, + "cache_read": 0.125 + } + }, + "Qwen/Qwen2.5-7B-Instruct-Turbo": { + "id": "Qwen/Qwen2.5-7B-Instruct-Turbo", + "name": "Qwen 2.5 7B Instruct Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-09-19", + "last_updated": "2024-09-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 0.3 + } + }, + "Qwen/Qwen3-Coder-Next-FP8": { + "id": "Qwen/Qwen3-Coder-Next-FP8", + "name": "Qwen3 Coder Next FP8", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2026-02-03", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.5, + "output": 1.2 + } + }, + "Qwen/Qwen3-235B-A22B-Instruct-2507-tput": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507-tput", + "name": "Qwen3 235B A22B Instruct 2507 FP8", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.2, + "output": 0.6 + } + }, + "Qwen/Qwen3.5-9B": { + "id": "Qwen/Qwen3.5-9B", + "name": "Qwen3.5 9B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-03", + "last_updated": "2026-03-03", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.17, + "output": 0.25 + } + }, + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen3.5 397B A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-16", + "last_updated": "2026-06-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 130000 + }, + "status": "deprecated", + "cost": { + "input": 0.6, + "output": 3.6, + "cache_read": 0.35 + } + }, + "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 2, + "output": 2 + } + }, + "Qwen/Qwen3.6-Plus": { + "id": "Qwen/Qwen3.6-Plus", + "name": "Qwen3.6 Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 500000 + }, + "cost": { + "input": 0.5, + "output": 3 + } + }, + "LiquidAI/LFM2-24B-A2B": { + "id": "LiquidAI/LFM2-24B-A2B", + "name": "LFM2-24B-A2B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "liquid", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-02-25", + "last_updated": "2026-02-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.03, + "output": 0.12 + } + }, + "deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.03 + } + }, + "deepseek-ai/DeepSeek-R1": { + "id": "deepseek-ai/DeepSeek-R1", + "name": "DeepSeek-R1", + "description": "Legacy model retained for compatibility with older integrations", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-03-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163839, + "output": 163839 + }, + "status": "deprecated", + "cost": { + "input": 3, + "output": 7 + } + }, + "deepseek-ai/DeepSeek-V4-Pro-0813": { + "id": "deepseek-ai/DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 1.32, + "output": 3.96, + "cache_read": 0.13 + } + }, + "deepseek-ai/DeepSeek-V4.1-Flash": { + "id": "deepseek-ai/DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.006 + } + }, + "deepseek-ai/DeepSeek-V3-1": { + "id": "deepseek-ai/DeepSeek-V3-1", + "name": "DeepSeek V3.1", + "description": "Legacy model retained for compatibility with older integrations", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0.6, + "output": 1.7 + } + }, + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 512000, + "output": 384000 + }, + "cost": { + "input": 1.74, + "output": 3.48, + "cache_read": 0.2 + } + }, + "deepseek-ai/DeepSeek-V3": { + "id": "deepseek-ai/DeepSeek-V3", + "name": "DeepSeek-V3", + "description": "Legacy model retained for compatibility with older integrations", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-12-26", + "last_updated": "2025-05-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 1.25, + "output": 1.25 + } + }, + "MiniMaxAI/MiniMax-M3": { + "id": "MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 524288, + "output": 250000 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "MiniMaxAI/MiniMax-M2.7": { + "id": "MiniMaxAI/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131000 + }, + "cost": { + "input": 1.2, + "output": 4.5, + "cache_read": 0.2 + } + }, + "moonshotai/Kimi-K3": { + "id": "moonshotai/Kimi-K3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3 + } + }, + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Kimi coding model for software agents, refactors, and repository reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-14", + "last_updated": "2026-06-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.19 + } + }, + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi K2.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-k2", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "temperature": true, + "knowledge": "2026-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.5, + "output": 2.8 + } + }, + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-11", + "release_date": "2026-04-07", + "last_updated": "2026-07-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-16", + "last_updated": "2026-06-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 512000, + "output": 164000 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "status": "deprecated", + "cost": { + "input": 1, + "output": 3.2 + } + }, + "zai-org/GLM-5.3-Flash": { + "id": "zai-org/GLM-5.3-Flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048575, + "output": 400000 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 + } + }, + "zai-org/GLM-5.3": { + "id": "zai-org/GLM-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 262144 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 + } + }, + "nvidia/nemotron-3-ultra-550b-a55b": { + "id": "nvidia/nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 512300, + "output": 512300 + }, + "cost": { + "input": 0.6, + "output": 3.6, + "cache_read": 0.2 + } + }, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.05, + "output": 0.2 + } + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + } + } + }, + "helicone": { + "id": "helicone", + "env": [ + "HELICONE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://ai-gateway.helicone.ai/v1", + "name": "Helicone", + "doc": "https://helicone.ai/models", + "models": { + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "xAI Grok 4.1 Fast Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-11", + "release_date": "2025-11-17", + "last_updated": "2025-11-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 2000000 + }, + "cost": { + "input": 0.19999999999999998, + "output": 0.5, + "cache_read": 0.049999999999999996 + } + }, + "grok-3": { + "id": "grok-3", + "name": "xAI Grok 3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.75 + } + }, + "llama-4-scout": { + "id": "llama-4-scout", + "name": "Meta Llama 4 Scout 17B 16E", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.08, + "output": 0.3 + } + }, + "llama-guard-4": { + "id": "llama-guard-4", + "name": "Meta Llama Guard 4 12B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 1024 + }, + "cost": { + "input": 0.21, + "output": 0.21 + } + }, + "mistral-large-2411": { + "id": "mistral-large-2411", + "name": "Mistral-Large", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-24", + "last_updated": "2024-07-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "xAI Grok 4.1 Fast Non-Reasoning", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-11", + "release_date": "2025-11-17", + "last_updated": "2025-11-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 30000 + }, + "cost": { + "input": 0.19999999999999998, + "output": 0.5, + "cache_read": 0.049999999999999996 + } + }, + "qwen3-vl-235b-a22b-instruct": { + "id": "qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 16384 + }, + "cost": { + "input": 0.3, + "output": 1.5 + } + }, + "mistral-nemo": { + "id": "mistral-nemo", + "name": "Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16400 + }, + "cost": { + "input": 20, + "output": 40 + } + }, + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "OpenAI: GPT-5.1 Codex", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-codex", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 + } + }, + "gemini-3-pro-preview": { + "id": "gemini-3-pro-preview", + "name": "Google Gemini 3 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-11", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.19999999999999998 + } + }, + "gpt-4o": { + "id": "gpt-4o", + "name": "OpenAI GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-05", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "gpt-4.1-mini-2025-04-14": { + "id": "gpt-4.1-mini-2025-04-14", + "name": "OpenAI GPT-4.1 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.39999999999999997, + "output": 1.5999999999999999, + "cache_read": 0.09999999999999999 + } + }, + "glm-4.6": { + "id": "glm-4.6", + "name": "Zai GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.44999999999999996, + "output": 1.5 + } + }, + "claude-3.5-haiku": { + "id": "claude-3.5-haiku", + "name": "Anthropic: Claude 3.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "cost": { + "input": 0.7999999999999999, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 + } + }, + "o1-mini": { + "id": "o1-mini", + "name": "OpenAI: o1-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 65536 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 + } + }, + "grok-4": { + "id": "grok-4", + "name": "xAI Grok 4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-09", + "last_updated": "2024-07-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.75 + } + }, + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "OpenAI: GPT-5 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 + } + }, + "llama-4-maverick": { + "id": "llama-4-maverick", + "name": "Meta Llama 4 Maverick 17B 128E", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "qwen3-coder": { + "id": "qwen3-coder", + "name": "Qwen3 Coder 480B A35B Instruct Turbo", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 16384 + }, + "cost": { + "input": 0.22, + "output": 0.95 + } + }, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "OpenAI GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.024999999999999998 + } + }, + "o4-mini": { + "id": "o4-mini", + "name": "OpenAI o4 Mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 + } + }, + "o3-mini": { + "id": "o3-mini", + "name": "OpenAI o3 Mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2023-10", + "release_date": "2023-10-01", + "last_updated": "2023-10-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 + } + }, + "kimi-k2-0711": { + "id": "kimi-k2-0711", + "name": "Kimi K2 (07/11)", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0.5700000000000001, + "output": 2.3 + } + }, + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "OpenAI GPT-4.1 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.09999999999999999, + "output": 0.39999999999999997, + "cache_read": 0.024999999999999998 + } + }, + "grok-4-fast-reasoning": { + "id": "grok-4-fast-reasoning", + "name": "xAI: Grok 4 Fast Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-09-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 2000000 + }, + "cost": { + "input": 0.19999999999999998, + "output": 0.5, + "cache_read": 0.049999999999999996 + } + }, + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 40960 + }, + "cost": { + "input": 0.29, + "output": 0.59 + } + }, + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "OpenAI GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 0.049999999999999996, + "output": 0.39999999999999997, + "cache_read": 0.005 + } + }, + "claude-3.5-sonnet-v2": { + "id": "claude-3.5-sonnet-v2", + "name": "Anthropic: Claude 3.5 Sonnet v2", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.30000000000000004, + "cache_write": 3.75 + } + }, + "gpt-5.1-chat-latest": { + "id": "gpt-5.1-chat-latest", + "name": "OpenAI GPT-5.1 Chat", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-codex", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 + } + }, + "claude-4.5-sonnet": { + "id": "claude-4.5-sonnet", + "name": "Anthropic: Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.30000000000000004, + "cache_write": 3.75 + } + }, + "deepseek-tng-r1t2-chimera": { + "id": "deepseek-tng-r1t2-chimera", + "name": "DeepSeek TNG R1T2 Chimera", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-02", + "last_updated": "2025-07-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 130000, + "output": 163840 + }, + "cost": { + "input": 0.3, + "output": 1.2 + } + }, + "deepseek-v3": { + "id": "deepseek-v3", + "name": "DeepSeek V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2024-12-26", + "last_updated": "2024-12-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.56, + "output": 1.68, + "cache_read": 0.07 + } + }, + "deepseek-r1-distill-llama-70b": { + "id": "deepseek-r1-distill-llama-70b", + "name": "DeepSeek R1 Distill Llama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.03, + "output": 0.13 + } + }, + "o1": { + "id": "o1", + "name": "OpenAI: o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 15, + "output": 60, + "cache_read": 7.5 + } + }, + "mistral-small": { + "id": "mistral-small", + "name": "Mistral Small 3.2", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.075, + "output": 0.2 + } + }, + "chatgpt-4o-latest": { + "id": "chatgpt-4o-latest", + "name": "OpenAI ChatGPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2024-08-14", + "last_updated": "2024-08-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 5, + "output": 20, + "cache_read": 2.5 + } + }, + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3 Coder 30B A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-31", + "last_updated": "2025-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.09999999999999999, + "output": 0.3 + } + }, + "sonar-pro": { + "id": "sonar-pro", + "name": "Perplexity Sonar Pro", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 4096 + }, + "cost": { + "input": 3, + "output": 15 + } + }, + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-11", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 262144 + }, + "cost": { + "input": 0.48, + "output": 2 + } + }, + "grok-4-fast-non-reasoning": { + "id": "grok-4-fast-non-reasoning", + "name": "xAI Grok 4 Fast Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-09-19", + "last_updated": "2025-09-19", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 2000000 + }, + "cost": { + "input": 0.19999999999999998, + "output": 0.5, + "cache_read": 0.049999999999999996 + } + }, + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "Anthropic: Claude Opus 4.1 (20250805)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "OpenAI: GPT-5 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 15, + "output": 120 + } + }, + "gpt-5-chat-latest": { + "id": "gpt-5-chat-latest", + "name": "OpenAI GPT-5 Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2024-09", + "release_date": "2024-09-30", + "last_updated": "2024-09-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 + } + }, + "deepseek-v3.1-terminus": { + "id": "deepseek-v3.1-terminus", + "name": "DeepSeek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.27, + "output": 1, + "cache_read": 0.21600000000000003 + } + }, + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Anthropic: Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "gpt-5.1": { + "id": "gpt-5.1", + "name": "OpenAI GPT-5.1", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 + } + }, + "llama-3.3-70b-versatile": { + "id": "llama-3.3-70b-versatile", + "name": "Meta Llama 3.3 70B Versatile", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32678 + }, + "cost": { + "input": 0.59, + "output": 0.7899999999999999 + } + }, + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Meta Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16400 + }, + "cost": { + "input": 0.13, + "output": 0.39 + } + }, + "grok-3-mini": { + "id": "grok-3-mini", + "name": "xAI Grok 3 Mini", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 0.5, + "cache_read": 0.075 + } + }, + "gemma-3-12b-it": { + "id": "gemma-3-12b-it", + "name": "Google Gemma 3 12B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.049999999999999996, + "output": 0.09999999999999999 + } + }, + "qwen2.5-coder-7b-fast": { + "id": "qwen2.5-coder-7b-fast", + "name": "Qwen2.5 Coder 7B fast", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-09", + "release_date": "2024-09-15", + "last_updated": "2024-09-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 8192 + }, + "cost": { + "input": 0.03, + "output": 0.09 + } + }, + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "OpenAI GPT-4o-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 + } + }, + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262000, + "output": 16384 + }, + "cost": { + "input": 0.14, + "output": 1.4 + } + }, + "o3-pro": { + "id": "o3-pro", + "name": "OpenAI o3 Pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 20, + "output": 80 + } + }, + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "OpenAI GPT-OSS 20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.049999999999999996, + "output": 0.19999999999999998 + } + }, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Google Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.3125, + "cache_write": 1.25 + } + }, + "sonar-deep-research": { + "id": "sonar-deep-research", + "name": "Perplexity Sonar Deep Research", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar-deep-research", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 127000, + "output": 4096 + }, + "cost": { + "input": 2, + "output": 8 + } + }, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Google Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.075, + "cache_write": 0.3 + } + }, + "claude-4.5-haiku": { + "id": "claude-4.5-haiku", + "name": "Anthropic: Claude 4.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-10", + "release_date": "2025-10-01", + "last_updated": "2025-10-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.09999999999999999, + "cache_write": 1.25 + } + }, + "gemma2-9b-it": { + "id": "gemma2-9b-it", + "name": "Google Gemma 2", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2024-06-25", + "last_updated": "2024-06-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0.01, + "output": 0.03 + } + }, + "claude-3.7-sonnet": { + "id": "claude-3.7-sonnet", + "name": "Anthropic: Claude 3.7 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-02", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.30000000000000004, + "cache_write": 3.75 + } + }, + "hermes-2-pro-llama-3-8b": { + "id": "hermes-2-pro-llama-3-8b", + "name": "Hermes 2 Pro Llama 3 8B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-05", + "release_date": "2024-05-27", + "last_updated": "2024-05-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.14 + } + }, + "claude-sonnet-4": { + "id": "claude-sonnet-4", + "name": "Anthropic: Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-14", + "last_updated": "2025-05-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.30000000000000004, + "cache_write": 3.75 + } + }, + "qwen3-235b-a22b-thinking": { + "id": "qwen3-235b-a22b-thinking", + "name": "Qwen3 235B A22B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 81920 + }, + "cost": { + "input": 0.3, + "output": 2.9000000000000004 + } + }, + "grok-code-fast-1": { + "id": "grok-code-fast-1", + "name": "xAI Grok Code Fast 1", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2024-08-25", + "last_updated": "2024-08-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 10000 + }, + "cost": { + "input": 0.19999999999999998, + "output": 1.5, + "cache_read": 0.02 + } + }, + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "Anthropic: Claude Sonnet 4.5 (20250929)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.30000000000000004, + "cache_write": 3.75 + } + }, + "gpt-4.1": { + "id": "gpt-4.1", + "name": "OpenAI GPT-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "qwen3-30b-a3b": { + "id": "qwen3-30b-a3b", + "name": "Qwen3 30B A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-06-01", + "last_updated": "2025-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 41000, + "output": 41000 + }, + "cost": { + "input": 0.08, + "output": 0.29 + } + }, + "kimi-k2-0905": { + "id": "kimi-k2-0905", + "name": "Kimi K2 (09/05)", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 16384 + }, + "cost": { + "input": 0.5, + "output": 2, + "cache_read": 0.39999999999999997 + } + }, + "llama-3.1-8b-instruct": { + "id": "llama-3.1-8b-instruct", + "name": "Meta Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16384, + "output": 16384 + }, + "cost": { + "input": 0.02, + "output": 0.049999999999999996 + } + }, + "llama-3.1-8b-instruct-turbo": { + "id": "llama-3.1-8b-instruct-turbo", + "name": "Meta Llama 3.1 8B Instruct Turbo", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.02, + "output": 0.03 + } + }, + "llama-prompt-guard-2-86m": { + "id": "llama-prompt-guard-2-86m", + "name": "Meta Llama Prompt Guard 2 86M", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 512, + "output": 2 + }, + "cost": { + "input": 0.01, + "output": 0.01 + } + }, + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Anthropic: Claude 4.5 Haiku (20251001)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-10", + "release_date": "2025-10-01", + "last_updated": "2025-10-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.09999999999999999, + "cache_write": 1.25 + } + }, + "sonar": { + "id": "sonar", + "name": "Perplexity Sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 127000, + "output": 4096 + }, + "cost": { + "input": 1, + "output": 1 + } + }, + "deepseek-reasoner": { + "id": "deepseek-reasoner", + "name": "DeepSeek Reasoner", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 64000 + }, + "cost": { + "input": 0.56, + "output": 1.68, + "cache_read": 0.07 + } + }, + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "OpenAI GPT-4.1 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.39999999999999997, + "output": 1.5999999999999999, + "cache_read": 0.09999999999999999 + } + }, + "ernie-4.5-21b-a3b-thinking": { + "id": "ernie-4.5-21b-a3b-thinking", + "name": "Baidu Ernie 4.5 21B A3B Thinking", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "ernie", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-03-16", + "last_updated": "2025-03-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 8000 + }, + "cost": { + "input": 0.07, + "output": 0.28 + } + }, + "claude-4.5-opus": { + "id": "claude-4.5-opus", + "name": "Anthropic: Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-11", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Google Gemini 2.5 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 512, + "max": 24576 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.09999999999999999, + "output": 0.39999999999999997, + "cache_read": 0.024999999999999998, + "cache_write": 0.09999999999999999 + } + }, + "claude-3-haiku-20240307": { + "id": "claude-3-haiku-20240307", + "name": "Anthropic: Claude 3 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-03", + "release_date": "2024-03-07", + "last_updated": "2024-03-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 4096 + }, + "cost": { + "input": 0.25, + "output": 1.25, + "cache_read": 0.03, + "cache_write": 0.3 + } + }, + "sonar-reasoning-pro": { + "id": "sonar-reasoning-pro", + "name": "Perplexity Sonar Reasoning Pro", + "description": "Web-grounded reasoning model for multi-step research and cited answers", + "family": "sonar-reasoning", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 127000, + "output": 4096 + }, + "cost": { + "input": 2, + "output": 8 + } + }, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 163840, + "output": 65536 + }, + "cost": { + "input": 0.27, + "output": 0.41 + } + }, + "llama-prompt-guard-2-22m": { + "id": "llama-prompt-guard-2-22m", + "name": "Meta Llama Prompt Guard 2 22M", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 512, + "output": 2 + }, + "cost": { + "input": 0.01, + "output": 0.01 + } + }, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "OpenAI GPT-OSS 120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.04, + "output": 0.16 + } + }, + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "OpenAI: GPT-5.1 Codex Mini", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-codex", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.024999999999999998 + } + }, + "o3": { + "id": "o3", + "name": "OpenAI o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "gpt-5": { + "id": "gpt-5", + "name": "OpenAI GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 + } + }, + "llama-3.1-8b-instant": { + "id": "llama-3.1-8b-instant", + "name": "Meta Llama 3.1 8B Instant", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 32678 + }, + "cost": { + "input": 0.049999999999999996, + "output": 0.08 + } + }, + "sonar-reasoning": { + "id": "sonar-reasoning", + "name": "Perplexity Sonar Reasoning", + "description": "Web-grounded reasoning model for multi-step research and cited answers", + "family": "sonar-reasoning", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 127000, + "output": 4096 + }, + "cost": { + "input": 1, + "output": 5 + } + }, + "claude-opus-4": { + "id": "claude-opus-4", + "name": "Anthropic: Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-14", + "last_updated": "2025-05-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + } + } + }, + "cortecs": { + "id": "cortecs", + "env": [ + "CORTECS_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.cortecs.ai/v1", + "name": "Cortecs", + "doc": "https://api.cortecs.ai/v1/models", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "output": 128000 + }, + "cost": { + "input": 2.898, + "output": 15.453, + "cache_read": 0.242 + } + }, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 0.996, + "output": 4.982, + "cache_read": 0.099, + "cache_write": 1.186 + } + }, + "gpt-oss-safeguard-120b": { + "id": "gpt-oss-safeguard-120b", + "name": "GPT OSS Safeguard 120B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.179, + "output": 0.697 + } + }, + "gemma-4-31b-it": { + "id": "gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 262000 + }, + "cost": { + "input": 0.223, + "output": 0.39 + } + }, + "qwen3.8-27b": { + "id": "qwen3.8-27b", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.04 + } + }, + "gemma-3-27b-it": { + "id": "gemma-3-27b-it", + "name": "Gemma 3 27B IT", + "description": "Gemma 3 is a family of lightweight, multimodal models from Google, supporting text and image inputs, multilingual capabilities, and a 131K context window.", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, + "temperature": false, + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131000, + "output": 110000 + }, + "cost": { + "input": 0.099, + "output": 0.299 + } + }, + "qwen3.8-2.4t-a95b": { + "id": "qwen3.8-2.4t-a95b", + "name": "Qwen3.8 2.4T A95B", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 2.5, + "output": 6, + "cache_read": 0.625 + } + }, + "mistral-large-2402": { + "id": "mistral-large-2402", + "name": "mistral-large-2402", + "description": "Mistral Large (24.02) is Mistral AI’s most advanced language model, built for complex multilingual reasoning, code generation, and deep text understanding.", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-05-26", + "last_updated": "2025-05-26", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 8192 + }, + "cost": { + "input": 4.284, + "output": 12.952 + } + }, + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 202752, + "output": 202752 + }, + "cost": { + "input": 1.186, + "output": 3.955, + "cache_read": 0.296, + "cache_write": 1.544 + } + }, + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16000 + }, + "cost": { + "input": 2.659, + "output": 10.635, + "cache_read": 1.33 + } + }, + "qwen3guard-gen-8b": { + "id": "qwen3guard-gen-8b", + "name": "qwen3guard-gen-8b", + "description": "Qwen3Guard-Gen-8B is a large-scale multilingual safety moderation model designed for high-accuracy prompt and response classification.", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": false, + "release_date": "2026-02-04", + "last_updated": "2026-02-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 0.1, + "output": 0.35, + "cache_read": 0.018 + } + }, + "ministral-8b-2512": { + "id": "ministral-8b-2512", + "name": "ministral-8b-2512", + "description": "Ministral 3 8B is a balanced, efficient multimodal model offering strong text and vision capabilities, optimized for edge and local deployment.", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-12-03", + "last_updated": "2025-12-03", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.179, + "output": 0.179, + "cache_read": 0.017 + } + }, + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 0.055, + "output": 0.174, + "cache_read": 0.009 + } + }, + "mistral-7b-instruct-v0.2": { + "id": "mistral-7b-instruct-v0.2", + "name": "mistral-7b-instruct-v0.2", + "description": "Mistral 7B Instruct is a compact, 7B parameter model optimized for fast and efficient text and code generation with a 32K token context window.", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": false, + "release_date": "2025-05-26", + "last_updated": "2025-05-26", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 8192 + }, + "cost": { + "input": 0.159, + "output": 0.219 + } + }, + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B-A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-07-21", + "last_updated": "2025-07-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 262000 + }, + "cost": { + "input": 0.069, + "output": 0.455, + "cache_read": 0.018 + } + }, + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 3, + "output": 14.999 + } + }, + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.167, + "output": 0.891 + } + }, + "nova-lite-v1": { + "id": "nova-lite-v1", + "name": "nova-lite-v1", + "description": "Nova Lite is a fast, low-cost multimodal foundation model capable of reasoning over text, images, and video in 200+ languages.", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.069, + "output": 0.275 + } + }, + "apertus-70b": { + "id": "apertus-70b", + "name": "Apertus 70B", + "description": "Apertus 70B is an open, multilingual language model designed for research, long-context reasoning, and sovereignty-focused AI systems.", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-09", + "release_date": "2025-09-02", + "last_updated": "2025-09-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 65536, + "output": 16384 + }, + "cost": { + "input": 1.393, + "output": 2.228 + } + }, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.279, + "output": 2.192, + "cache_read": 0.056 + } + }, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 202752 + }, + "cost": { + "input": 0.988, + "output": 3.164, + "cache_read": 0.247 + } + }, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 64000 + }, + "cost": { + "input": 0.5, + "output": 1.499, + "cache_read": 0.13 + } + }, + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.111, + "output": 0.434, + "cache_read": 0.056 + } + }, + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196000, + "output": 196000 + }, + "cost": { + "input": 0.296, + "output": 1.186, + "cache_read": 0.075 + } + }, + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 16384, + "output": 16384 + }, + "cost": { + "input": 0.179, + "output": 0.697 + } + }, + "mistral-medium-3.5": { + "id": "mistral-medium-3.5", + "name": "mistral-medium-3.5", + "description": "Mistral Medium 3.5 is a frontier multimodal 128B model combining reasoning, coding, and instruction-following with strong agentic performance and efficient deployment.", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 1.532, + "output": 7.843, + "cache_read": 0.154 + } + }, + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.06, + "output": 0.439, + "cache_read": 0.019 + } + }, + "claude-4-6-sonnet": { + "id": "claude-4-6-sonnet", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3.196, + "output": 15.94, + "cache_read": 0.32, + "cache_write": 3.999 + } + }, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.516, + "output": 2.869, + "cache_read": 0.115 + } + }, + "claude-4-5-sonnet": { + "id": "claude-4-5-sonnet", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 2.989, + "output": 14.945, + "cache_read": 0.326, + "cache_write": 4.078 + } + }, + "devstral-2512": { + "id": "devstral-2512", + "name": "Devstral 2", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.478, + "output": 2.392, + "cache_read": 0.045 + } + }, + "gemini-3.6-flash": { + "id": "gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.038 + } + }, + "mistral-7b-instruct-v0.3": { + "id": "mistral-7b-instruct-v0.3", + "name": "mistral-7b-instruct-v0.3", + "description": "Mistral-7B-Instruct-v0.3 model is a fine-tuned version of the Mistral 7B base model, optimized for instruction-following tasks. Released in 2023, it is intended for demonstration purposes and does not include built-in guardrails or moderation features.", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-05-26", + "last_updated": "2025-05-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 127000, + "output": 127000 + }, + "cost": { + "input": 0.111, + "output": 0.111 + } + }, + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.067, + "output": 0.245, + "cache_read": 0.014 + } + }, + "mixtral-8x7B-instruct-v0.1": { + "id": "mixtral-8x7B-instruct-v0.1", + "name": "Mixtral 8x7B Instruct v0.1", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2023-12-11", + "last_updated": "2023-12-11", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32000, + "output": 4096 + }, + "cost": { + "input": 0.488, + "output": 0.758 + } + }, + "gemma-4-26b-a4b-it": { + "id": "gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 81920 + }, + "cost": { + "input": 0.111, + "output": 0.557 + } + }, + "gemini-3.5-flash-lite": { + "id": "gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.33, + "output": 2.749, + "cache_read": 0.033 + } + }, + "voxtral-small-2507": { + "id": "voxtral-small-2507", + "name": "voxtral-small-2507", + "description": "Voxtral Small is a multimodal model with audio input, combining advanced speech capabilities with strong text performance for transcription, translation, and audio understanding.", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-02-02", + "last_updated": "2026-02-02", + "modalities": { + "input": [ + "text", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.123, + "output": 0.368, + "cache_read": 0.012 + } + }, + "qwen3.5-122b-a10b": { + "id": "qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.495, + "output": 3.46, + "cache_read": 0.124 + } + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.495, + "output": 2.768, + "cache_read": 0.124 + } + }, + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.375, + "output": 10.96, + "cache_read": 0.156 + } + }, + "claude-opus-5": { + "id": "claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 27.498, + "cache_read": 0.55, + "cache_write": 6.874 + } + }, + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131000, + "output": 131000 + }, + "cost": { + "input": 0.724, + "output": 0.724 + } + }, + "qwen3guard-gen-0.6b": { + "id": "qwen3guard-gen-0.6b", + "name": "qwen3guard-gen-0.6b", + "description": "Qwen3Guard-Gen-0.6B is a lightweight multilingual safety moderation model that classifies prompts and responses into safe, controversial, or unsafe categories.", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": false, + "release_date": "2026-02-04", + "last_updated": "2026-02-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16000 + }, + "cost": { + "input": 0.159, + "output": 0.638, + "cache_read": 0.081 + } + }, + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 32768 + }, + "cost": { + "input": 0.167, + "output": 0.557 + } + }, + "pixtral-12b-2409": { + "id": "pixtral-12b-2409", + "name": "pixtral-12b-2409", + "description": "Pixtral 2409 12B is a state-of-the-art multimodal model with 12B parameters and a 400M vision encoder, natively trained on interleaved text and image data. It excels in tasks spanning vision-language reasoning, instruction following, and pure text understanding, making it highly effective for real-world multimodal applications.", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2024-11-09", + "last_updated": "2024-11-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.223, + "output": 0.223 + } + }, + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131000, + "output": 131000 + }, + "cost": { + "input": 0.045, + "output": 0.167 + } + }, + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 196608 + }, + "cost": { + "input": 0.668, + "output": 2.674 + } + }, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 1.649, + "output": 9.899, + "cache_read": 0.165, + "cache_write": 1 + } + }, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 1.495, + "output": 9.964, + "cache_read": 0.242, + "cache_write": 0.434 + } + }, + "hermes-4-405b": { + "id": "hermes-4-405b", + "name": "hermes-4-405b", + "description": "Hermes 4 405B is a frontier hybrid-mode reasoning model built on Llama 3.1, optimized for advanced logic, math, coding, and structured output generation.", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2024-08-13", + "last_updated": "2024-08-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.996, + "output": 2.989 + } + }, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.299, + "output": 2.491, + "cache_read": 0.029, + "cache_write": 0.097 + } + }, + "qwen3-vl-235b-a22b": { + "id": "qwen3-vl-235b-a22b", + "name": "qwen3-vl-235b-a22b", + "description": "Qwen3 VL 235B A22B is a 235B-parameter MoE vision-language flagship model (≈22B active) designed for frontier-level multimodal understanding across text, images, documents, and long videos.", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-01-13", + "last_updated": "2026-01-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.617, + "output": 3.119, + "cache_read": 0.052 + } + }, + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 64000 + }, + "cost": { + "input": 2, + "output": 3.999, + "cache_read": 0.5 + } + }, + "glm-4.7-flash": { + "id": "glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 203000, + "output": 203000 + }, + "cost": { + "input": 0.08, + "output": 0.478 + } + }, + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.111, + "output": 0.167 + } + }, + "claude-opus4-6": { + "id": "claude-opus4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.313, + "output": 26.561, + "cache_read": 0.531, + "cache_write": 6.645 + } + }, + "claude-sonnet-4": { + "id": "claude-sonnet-4", + "name": "Claude Sonnet 4 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 65000 + }, + "cost": { + "input": 2.898, + "output": 14.493, + "cache_read": 0.29, + "cache_write": 3.624 + } + }, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 0.395, + "output": 1.977, + "cache_read": 0.099 + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 0.219, + "output": 1.32, + "cache_read": 0.022, + "cache_write": 0.275 + } + }, + "qwen3.8-flash-next": { + "id": "qwen3.8-flash-next", + "name": "Qwen3.8 Flash Next", + "description": "Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 64000 + }, + "cost": { + "input": 0.201, + "output": 0.5, + "cache_read": 0.05 + } + }, + "mistral-small-2603": { + "id": "mistral-small-2603", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 256000 + }, + "cost": { + "input": 0.156, + "output": 0.625, + "cache_read": 0.016 + } + }, + "nemotron-nano-v2-12b": { + "id": "nemotron-nano-v2-12b", + "name": "nemotron-nano-v2-12b", + "description": "NVIDIA Nemotron Nano v2 12B is a 12-billion-parameter multimodal reasoning model designed for advanced video understanding, document intelligence, and visual reasoning, built with a hybrid Transformer-Mamba architecture for high efficiency and low latency.", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-10-31", + "last_updated": "2025-10-31", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.24, + "output": 0.707 + } + }, + "mistral-small-2503": { + "id": "mistral-small-2503", + "name": "mistral-small-2503", + "description": "Combines advanced text and vision capabilities with 24 billion parameters, supporting multilingual tasks and long contexts up to 131k tokens, making it versatile for various applications without sacrificing performance.", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-03-20", + "last_updated": "2025-03-20", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.111, + "output": 0.334 + } + }, + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2.192, + "output": 8.769, + "cache_read": 0.546 + } + }, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2.2, + "output": 11, + "cache_read": 0.219, + "cache_write": 2.749 + } + }, + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 262000 + }, + "cost": { + "input": 0.668, + "output": 4.01 + } + }, + "gemini-3.7-flash": { + "id": "gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.038 + } + }, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1000000 + }, + "cost": { + "input": 1.1, + "output": 2.99, + "cache_read": 0.18 + } + }, + "pixtral-large-2502": { + "id": "pixtral-large-2502", + "name": "Pixtral Large (25.02)", + "description": "Pixtral Large (25.02) is a 124B open-weight multimodal model built on Mistral Large 2, offering advanced image understanding and strong performance across text and code tasks.", + "family": "pixtral", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-04-08", + "last_updated": "2025-04-08", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 1.993, + "output": 5.978 + } + }, + "claude-opus4-7": { + "id": "claude-opus4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.437, + "output": 27.186, + "cache_read": 0.544, + "cache_write": 6.797 + } + }, + "minicpm-v-4.5": { + "id": "minicpm-v-4.5", + "name": "minicpm-v-4.5", + "description": "MiniCPM-V 4.5 is a compact, high-performance vision-language model excelling in video understanding, OCR, and multimodal reasoning with efficient deployment.", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.651, + "output": 1.097 + } + }, + "llama-3.1-8b-instruct": { + "id": "llama-3.1-8b-instruct", + "name": "Llama-3.1-8B-Instruct", + "description": "Optimized for dialogue, this LLM by Meta outperforms other open-source chat models in benchmarks while prioritizing helpfulness and safety.", + "family": "llama", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.167, + "output": 0.167 + } + }, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 202752 + }, + "cost": { + "input": 1.384, + "output": 4.348, + "cache_read": 0.346 + } + }, + "qwen3-30b-a3b-instruct-2507": { + "id": "qwen3-30b-a3b-instruct-2507", + "name": "qwen3-30b-a3b-instruct-2507", + "description": "Qwen3-30B-A3B-Instruct-2507 is an advanced Mixture-of-Experts model optimized for reasoning, coding, and multilingual instruction following.", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-07-28", + "last_updated": "2025-07-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262000, + "output": 262000 + }, + "cost": { + "input": 0.099, + "output": 0.299 + } + }, + "qwen2.5-vl-72b-instruct": { + "id": "qwen2.5-vl-72b-instruct", + "name": "qwen2.5-vl-72b-instruct", + "description": "Qwen2.5-VL is a powerful vision-language model with advanced capabilities in visual understanding, long video reasoning, and structured output generation.", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, + "temperature": false, + "release_date": "2025-01-27", + "last_updated": "2025-01-27", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 1.014, + "output": 1.014 + } + }, + "mistral-large-2512": { + "id": "mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.613, + "output": 1.838, + "cache_read": 0.061 + } + }, + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.434, + "output": 1.704, + "cache_read": 0.134 + } + }, + "gemini-3.8-flash": { + "id": "gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.825, + "output": 4.125, + "cache_read": 0.082, + "cache_write": 0.084 + } + }, + "ministral-14b-2512": { + "id": "ministral-14b-2512", + "name": "ministral-14b-2512", + "description": "Ministral 3 14B is a frontier-level 14B multimodal model optimized for local deployment, delivering state-of-the-art text and vision reasoning with a 256K context window and strong agentic capabilities.", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-12-03", + "last_updated": "2025-12-03", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.24, + "output": 0.24, + "cache_read": 0.022 + } + }, + "glm-5-turbo": { + "id": "glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 202752, + "output": 202752 + }, + "cost": { + "input": 1.186, + "output": 3.955, + "cache_read": 0.296, + "cache_write": 1.544 + } + }, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 1.73, + "output": 3.46, + "cache_read": 0.432 + } + }, + "qwen3.6-27b": { + "id": "qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262000, + "output": 262000 + }, + "cost": { + "input": 0.446, + "output": 3.008 + } + }, + "claude-opus4-8": { + "id": "claude-opus4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.437, + "output": 27.186, + "cache_read": 0.544, + "cache_write": 6.797 + } + }, + "mistral-small-3.2-24b-instruct-2506": { + "id": "mistral-small-3.2-24b-instruct-2506", + "name": "mistral-small-3.2-24b-instruct-2506", + "description": "Mistral-Small-3.2-24B-Instruct-2506 is a 24B parameter instruction-tuned model with enhanced long-context support (128k) and state-of-the-art vision understanding.", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-05-26", + "last_updated": "2025-05-26", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131000, + "output": 131000 + }, + "cost": { + "input": 0.1, + "output": 0.312 + } + }, + "deepseek-r1-0528": { + "id": "deepseek-r1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 164000, + "output": 164000 + }, + "cost": { + "input": 0.652, + "output": 2.57, + "cache_read": 0.163 + } + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.2, + "output": 13.199, + "cache_read": 0.219, + "cache_write": 2.749 + } + }, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 163840 + }, + "cost": { + "input": 0.296, + "output": 0.495, + "cache_read": 0.075 + } + }, + "nova-2-lite": { + "id": "nova-2-lite", + "name": "Nova 2 Lite", + "description": "Nova 2 Lite is an advanced multimodal reasoning model that combines efficiency and performance, delivering reliable AI for agentic workflows and enterprise applications.", + "family": "nova", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-10", + "release_date": "2025-12-02", + "last_updated": "2025-12-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65535 + }, + "cost": { + "input": 0.373, + "output": 3.144 + } + }, + "nova-pro-v1": { + "id": "nova-pro-v1", + "name": "Nova Pro 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 10000 + }, + "cost": { + "input": 0.918, + "output": 3.671 + } + }, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131000, + "output": 131000 + }, + "cost": { + "input": 0.089, + "output": 0.446, + "cache_read": 0.01 + } + }, + "ministral-3b-2512": { + "id": "ministral-3b-2512", + "name": "ministral-3b-2512", + "description": "Ministral 3 3B is a compact, efficient multimodal model with strong language, vision capabilities, and ideal for custom fine-tuning.", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-12-03", + "last_updated": "2025-12-03", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.123, + "output": 0.123, + "cache_read": 0.012 + } + }, + "claude-opus4-5": { + "id": "claude-opus4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5.313, + "output": 26.568, + "cache_read": 0.531, + "cache_write": 6.645 + } + }, + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 1048576 + }, + "cost": { + "input": 1.114, + "output": 3.899, + "cache_read": 0.279 + } + }, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.706, + "output": 3.208, + "cache_read": 0.18 + } + }, + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.375, + "output": 10.96, + "cache_read": 0.156 + } + }, + "minimax-m2.1": { + "id": "minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196000, + "output": 196000 + }, + "cost": { + "input": 0.359, + "output": 1.435 + } + }, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5.5, + "output": 32.998, + "cache_read": 0.55, + "cache_write": 6.879 + } + }, + "mistral-nemo-instruct-2407": { + "id": "mistral-nemo-instruct-2407", + "name": "mistral-nemo-instruct-2407", + "description": "A 12B parameter, instruct-tuned language model by Mistral AI and NVIDIA, designed for advanced instruction following, multi-turn conversations, and generating text and code across multiple languages.", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2024-08-07", + "last_updated": "2024-08-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.145, + "output": 0.145, + "cache_read": 0.014 + } + }, + "minimax-m2": { + "id": "minimax-m2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 400000, + "output": 196000 + }, + "cost": { + "input": 0.349, + "output": 1.405 + } + }, + "nova-micro-v1": { + "id": "nova-micro-v1", + "name": "nova-micro-v1", + "description": "Nova Micro is a multilingual text-to-text foundation model with strong reasoning capabilities and broad language coverage across 200+ languages.", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 10000 + }, + "cost": { + "input": 0.04, + "output": 0.159 + } + }, + "codestral-2508": { + "id": "codestral-2508", + "name": "Codestral 2508", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "mistral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-07-30", + "last_updated": "2025-07-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.368, + "output": 1.103, + "cache_read": 0.037 + } + }, + "nvidia-nemotron-3-nano-30b-a3b": { + "id": "nvidia-nemotron-3-nano-30b-a3b", + "name": "nvidia-nemotron-3-nano-30b-a3b", + "description": "Nemotron-Nano-3-30B-A3B is a compact Mixture-of-Experts model optimized for efficient reasoning, chat, and coding, with strong multilingual support and long-context RAG and agent workflows.", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-01-12", + "last_updated": "2026-01-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.06, + "output": 0.24 + } + }, + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text", + "image", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65535 + }, + "cost": { + "input": 0.272, + "output": 1.631, + "cache_read": 0.025, + "cache_write": 0.082 + } + } + } + }, + "agnes": { + "id": "agnes", + "env": [ + "AGNES_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://apihub.agnes-ai.com/v1", + "name": "Agnes AI", + "doc": "https://agnes-ai.com/doc", + "models": { + "agnes-2.5-pro-alpha": { + "id": "agnes-2.5-pro-alpha", + "name": "Agnes 2.5 Pro Alpha", + "description": "Paid reasoning model for advanced coding, scientific reasoning, long-context analysis, agentic workflows, and multimodal understanding.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.45, + "output": 0.9, + "cache_read": 0.0038 + } + }, + "agnes-2.0-flash": { + "id": "agnes-2.0-flash", + "name": "Agnes 2.0 Flash", + "description": "Fast and efficient model for agent workflows, tool calling, coding, and image understanding.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-25", + "last_updated": "2026-05-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 512000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "agnes-2.5-flash": { + "id": "agnes-2.5-flash", + "name": "Agnes 2.5 Flash", + "description": "Upgraded model with improved coding, agent workflows, tool calling, and multimodal understanding.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-07", + "last_updated": "2026-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 512000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + } + } + }, + "daoxe": { + "id": "daoxe", + "env": [ + "DAOXE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://daoxe.com/v1", + "name": "DaoXE", + "doc": "https://daoxe.com/pricing", + "models": { + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } + }, + "grok-4.3": { + "id": "grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 30000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 + } + }, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2 + } + }, + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 + } + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.1 + } + }, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + } + }, + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 5 + } + }, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + } + } + }, + "morph": { + "id": "morph", + "env": [ + "MORPH_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.morphllm.com/v1", + "name": "Morph", + "doc": "https://docs.morphllm.com/api-reference/introduction", + "models": { + "morph-v3-large": { + "id": "morph-v3-large", + "name": "Morph v3 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "morph", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-08-15", + "last_updated": "2024-08-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.9, + "output": 1.9 + } + }, + "morph-v3-fast": { + "id": "morph-v3-fast", + "name": "Morph v3 Fast", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "morph", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-08-15", + "last_updated": "2024-08-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16000, + "output": 16000 + }, + "cost": { + "input": 0.8, + "output": 1.2 + } + }, + "auto": { + "id": "auto", + "name": "Auto", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-06-01", + "last_updated": "2024-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.85, + "output": 1.55 + } + } + } + }, + "openai": { + "id": "openai", + "env": [ + "OPENAI_API_KEY" + ], + "npm": "@ai-sdk/openai", + "name": "OpenAI", + "doc": "https://platform.openai.com/docs/models", + "models": { + "chatgpt-image-latest": { + "id": "chatgpt-image-latest", + "name": "chatgpt-image-latest", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "input": 0, + "output": 0 + } + }, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } + } + }, + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } + } + }, + "gpt-3.5-turbo": { + "id": "gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16385, + "output": 4096 + }, + "status": "deprecated", + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0 + } + }, + "gpt-5.5-pro": { + "id": "gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } + } + }, + "text-embedding-3-small": { + "id": "text-embedding-3-small", + "name": "text-embedding-3-small", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2024-01", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8191, + "output": 1536 + }, + "cost": { + "input": 0.02, + "output": 0 + } + }, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 + } + }, + "gpt-realtime-2.1": { + "id": "gpt-realtime-2.1", + "name": "GPT-Realtime-2.1", + "description": "Realtime speech-to-speech model with configurable reasoning, tool use, and robust voice-agent behavior", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2026-07-06", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text", + "audio", + "image" + ], + "output": [ + "text", + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 96000, + "output": 32000 + }, + "cost": { + "input": 4, + "output": 24, + "cache_read": 0.4, + "input_audio": 32, + "output_audio": 64 + } + }, + "gpt-4o-2024-05-13": { + "id": "gpt-4o-2024-05-13", + "name": "GPT-4o (2024-05-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "status": "deprecated", + "cost": { + "input": 5, + "output": 15 + } + }, + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.025 + } + }, + "gpt-image-2": { + "id": "gpt-image-2", + "name": "gpt-image-2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "input": 0, + "output": 0 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 1.25 + } + }, + "gpt-5.2-pro": { + "id": "gpt-5.2-pro", + "name": "GPT-5.2 Pro", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 21, + "output": 168 + } + }, + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "status": "deprecated", + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 + } + }, + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "status": "deprecated", + "cost": { + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 + } + }, + "text-embedding-ada-002": { + "id": "text-embedding-ada-002", + "name": "text-embedding-ada-002", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2022-12", + "release_date": "2022-12-15", + "last_updated": "2022-12-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 1536 + }, + "cost": { + "input": 0.1, + "output": 0 + } + }, + "gpt-4": { + "id": "gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 8192 + }, + "status": "deprecated", + "cost": { + "input": 30, + "output": 60 + } + }, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "status": "deprecated", + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 + } + }, + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 + } + }, + "gpt-5.6": { + "id": "gpt-5.6", + "name": "GPT-5.6", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 8, + "output": 40, + "cache_read": 0.8, + "cache_write": 10 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10 + } + } + }, + "gpt-5.2-chat-latest": { + "id": "gpt-5.2-chat-latest", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "status": "deprecated", + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "o1": { + "id": "o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "status": "deprecated", + "cost": { + "input": 15, + "output": 60, + "cache_read": 7.5 + } + }, + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 272000 + }, + "cost": { + "input": 15, + "output": 120 + } + }, + "gpt-5.3-codex-spark": { + "id": "gpt-5.3-codex-spark", + "name": "GPT-5.3 Codex Spark", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex-spark", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 100000, + "output": 32000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "text-embedding-3-large": { + "id": "text-embedding-3-large", + "name": "text-embedding-3-large", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2024-01", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8191, + "output": 3072 + }, + "cost": { + "input": 0.13, + "output": 0 + } + }, + "gpt-4o-2024-08-06": { + "id": "gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-08-06", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 20, + "output": 100, + "cache_read": 2, + "cache_write": 25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } + } + }, + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 + } + }, + "o3-pro": { + "id": "o3-pro", + "name": "o3-pro", + "description": "High-effort o3 tier for difficult technical reasoning and careful answers", + "family": "o-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 20, + "output": 80 + } + }, + "gpt-image-1": { + "id": "gpt-image-1", + "name": "gpt-image-1", + "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", + "family": "gpt-image", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-04-24", + "last_updated": "2025-04-24", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "input": 0, + "output": 0 + }, + "status": "deprecated" + }, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 0.4, + "output": 2.4, + "cache_read": 0.04, + "cache_write": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } + } + }, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + }, + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "gpt-4o-2024-11-20": { + "id": "gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10, + "cache_read": 1.25 + } + }, + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 + } + }, + "gpt-6-luna": { + "id": "gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 0.2, + "output": 1, + "cache_read": 0.02, + "cache_write": 0.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } + }, + "cost": { + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } + } + }, + "gpt-5.3-chat-latest": { + "id": "gpt-5.3-chat-latest", + "name": "GPT-5.3 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "status": "deprecated", + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 + } + }, + "gpt-image-1-mini": { + "id": "gpt-image-1-mini", + "name": "gpt-image-1-mini", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-09-26", + "last_updated": "2025-09-26", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "input": 0, + "output": 0 + } + }, + "gpt-image-1.5": { + "id": "gpt-image-1.5", + "name": "gpt-image-1.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-11-25", + "last_updated": "2025-11-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "input": 0, + "output": 0 + } + }, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 4, + "output": 24, + "cache_read": 0.4, + "cache_write": 5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } + } + }, + "gpt-4-turbo": { + "id": "gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "status": "deprecated", + "cost": { + "input": 10, + "output": 30 + } + }, + "o3": { + "id": "o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 8, + "output": 40, + "cache_read": 0.8, + "cache_write": 10 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10 + } + } + }, + "gpt-6-sol": { + "id": "gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } + } + }, + "o1-pro": { + "id": "o1-pro", + "name": "o1-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2025-03-19", + "last_updated": "2025-03-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "status": "deprecated", + "cost": { + "input": 150, + "output": 600 + } + } + } + }, + "alibaba-coding-plan-cn": { + "id": "alibaba-coding-plan-cn", + "env": [ + "ALIBABA_CODING_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://coding.dashscope.aliyuncs.com/v1", + "name": "Alibaba Coding Plan (China)", + "doc": "https://help.aliyun.com/zh/model-studio/coding-plan", + "models": { + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 + } + }, + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-03", + "last_updated": "2026-02-03", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 202752, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "qwen3-max-2026-01-23": { + "id": "qwen3-max-2026-01-23", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-23", + "last_updated": "2026-01-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.1875, + "output": 1.125, + "cache_write": 0.234375 + } + }, + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 24576 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + } + } + }, + "io-net": { + "id": "io-net", + "env": [ + "IOINTELLIGENCE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.intelligence.io.solutions/api/v1", + "name": "IO.NET", + "doc": "https://io.net/docs/guides/intelligence/io-intelligence", + "models": { + "meta-llama/Llama-3.2-90B-Vision-Instruct": { + "id": "meta-llama/Llama-3.2-90B-Vision-Instruct", + "name": "Llama 3.2 90B Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 16000, + "output": 4096 + }, + "cost": { + "input": 0.35, + "output": 0.4, + "cache_read": 0.175, + "cache_write": 0.7 + } + }, + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.13, + "output": 0.38, + "cache_read": 0.065, + "cache_write": 0.26 + } + }, + "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { + "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", + "name": "Llama 4 Maverick 17B 128E Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-15", + "last_updated": "2025-01-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 430000, + "output": 4096 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075, + "cache_write": 0.3 + } + }, + "Qwen/Qwen3-Next-80B-A3B-Instruct": { + "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", + "name": "Qwen 3 Next 80B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-10", + "last_updated": "2025-01-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 4096 + }, + "cost": { + "input": 0.1, + "output": 0.8, + "cache_read": 0.05, + "cache_write": 0.2 + } + }, + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen 3 235B Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 4096 + }, + "cost": { + "input": 0.11, + "output": 0.6, + "cache_read": 0.055, + "cache_write": 0.22 + } + }, + "Qwen/Qwen2.5-VL-32B-Instruct": { + "id": "Qwen/Qwen2.5-VL-32B-Instruct", + "name": "Qwen 2.5 VL 32B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-09", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32000, + "output": 4096 + }, + "cost": { + "input": 0.05, + "output": 0.22, + "cache_read": 0.025, + "cache_write": 0.1 + } + }, + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", + "name": "DeepSeek R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 2, + "output": 8.75, + "cache_read": 1, + "cache_write": 4 + } + }, + "mistralai/Mistral-Nemo-Instruct-2407": { + "id": "mistralai/Mistral-Nemo-Instruct-2407", + "name": "Mistral Nemo Instruct 2407", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-05", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.02, + "output": 0.04, + "cache_read": 0.01, + "cache_write": 0.04 + } + }, + "mistralai/Magistral-Small-2506": { + "id": "mistralai/Magistral-Small-2506", + "name": "Magistral Small 2506", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-small", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-01", + "last_updated": "2025-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.25, + "cache_write": 1 + } + }, + "mistralai/Devstral-Small-2505": { + "id": "mistralai/Devstral-Small-2505", + "name": "Devstral Small 2505", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-05-01", + "last_updated": "2025-05-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.05, + "output": 0.22, + "cache_read": 0.025, + "cache_write": 0.1 + } + }, + "mistralai/Mistral-Large-Instruct-2411": { + "id": "mistralai/Mistral-Large-Instruct-2411", + "name": "Mistral Large Instruct 2411", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 1, + "cache_write": 4 + } + }, + "Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar": { + "id": "Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar", + "name": "Qwen 3 Coder 480B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-15", + "last_updated": "2025-01-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 106000, + "output": 4096 + }, + "cost": { + "input": 0.22, + "output": 0.95, + "cache_read": 0.11, + "cache_write": 0.44 + } + }, + "moonshotai/Kimi-K2-Thinking": { + "id": "moonshotai/Kimi-K2-Thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 4096 + }, + "cost": { + "input": 0.55, + "output": 2.25, + "cache_read": 0.275, + "cache_write": 1.1 + } + }, + "moonshotai/Kimi-K2-Instruct-0905": { + "id": "moonshotai/Kimi-K2-Instruct-0905", + "name": "Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2024-09-05", + "last_updated": "2024-09-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 4096 + }, + "cost": { + "input": 0.39, + "output": 1.9, + "cache_read": 0.195, + "cache_write": 0.78 + } + }, + "zai-org/GLM-4.6": { + "id": "zai-org/GLM-4.6", + "name": "GLM 4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-11-15", + "last_updated": "2024-11-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 4096 + }, + "cost": { + "input": 0.4, + "output": 1.75, + "cache_read": 0.2, + "cache_write": 0.8 + } + }, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT-OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 64000, + "output": 4096 + }, + "cost": { + "input": 0.03, + "output": 0.14, + "cache_read": 0.015, + "cache_write": 0.06 + } + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT-OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 4096 + }, + "cost": { + "input": 0.04, + "output": 0.4, + "cache_read": 0.02, + "cache_write": 0.08 + } + } + } + }, + "infomaniak": { + "id": "infomaniak", + "env": [ + "INFOMANIAK_API_KEY", + "INFOMANIAK_PRODUCT_ID" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.infomaniak.com/2/ai/${INFOMANIAK_PRODUCT_ID}/openai/v1", + "name": "Infomaniak", + "doc": "https://www.infomaniak.com/en/hosting/ai-services/open-source-models", + "models": { + "mini_lm_l12_v2": { + "id": "mini_lm_l12_v2", + "name": "All-MiniLM-L12-v2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2021-08-30", + "last_updated": "2026-08-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128, + "input": 128, + "output": 384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "bge_multilingual_gemma2": { + "id": "bge_multilingual_gemma2", + "name": "BGE Multilingual Gemma2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-07-25", + "last_updated": "2026-08-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8000, + "input": 8000, + "output": 3584 + }, + "cost": { + "input": 0.08, + "output": 0 + } + }, + "swiss-ai/Apertus-v1.5-70B": { + "id": "swiss-ai/Apertus-v1.5-70B", + "name": "Apertus v1.5 70B", + "description": "Open, ethically-sourced Swiss AI model for multilingual, multimodal chat and instruction following", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-07-24", + "last_updated": "2026-08-01", + "modalities": { + "input": [ + "text", + "image", + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 100000, + "input": 100000, + "output": 8192 + }, + "status": "beta", + "cost": { + "input": 0.87, + "output": 3.1 + } + }, + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-08-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 100000, + "input": 100000, + "output": 32768 + }, + "cost": { + "input": 0.25, + "output": 0.5 + } + }, + "Qwen/Qwen3.5-122B-A10B-FP8": { + "id": "Qwen/Qwen3.5-122B-A10B-FP8", + "name": "Qwen3.5 122B-A10B FP8", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-08-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "input": 200000, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3.97 + } + }, + "Qwen/Qwen3.5-397B-A17B-FP8": { + "id": "Qwen/Qwen3.5-397B-A17B-FP8", + "name": "Qwen3.5 397B-A17B FP8", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-08-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "input": 200000, + "output": 65536 + }, + "status": "beta", + "cost": { + "input": 0.99, + "output": 4.46 + } + }, + "mistralai/Mistral-Small-4-119B-2603": { + "id": "mistralai/Mistral-Small-4-119B-2603", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-08-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 256000, + "output": 256000 + }, + "cost": { + "input": 0.25, + "output": 0.93 + } + }, + "mistralai/Ministral-3-14B-Instruct-2512": { + "id": "mistralai/Ministral-3-14B-Instruct-2512", + "name": "Ministral 3 14B Instruct", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2026-08-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 100000, + "input": 100000, + "output": 25600 + }, + "status": "beta", + "cost": { + "input": 0.37, + "output": 0.5 + } + }, + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-08-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 256000, + "output": 256000 + }, + "status": "beta", + "cost": { + "input": 0.74, + "output": 3.72 + } + }, + "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8": { + "id": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8", + "name": "Nemotron 3 Nano 30B A3B FP8", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "temperature": true, + "release_date": "2025-12-15", + "last_updated": "2026-08-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "input": 1000000, + "output": 262144 + }, + "status": "beta", + "cost": { + "input": 0.06, + "output": 0.25 + } + } + } + }, + "llmtech": { + "id": "llmtech", + "env": [ + "LLMTECH_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.llmtech.eu/v1", + "name": "LLM Tech", + "doc": "https://llmtech.eu/models/qwen3.8-27b", + "models": { + "nvidia/Qwen3.8-27B-NVFP4": { + "id": "nvidia/Qwen3.8-27B-NVFP4", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.25, + "output": 2.09, + "cache_read": 0.04 + } + } + } + }, + "crossmodel": { + "id": "crossmodel", + "env": [ + "CROSSMODEL_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.crossmodel.ai/v1", + "name": "CrossModel", + "doc": "https://www.crossmodel.ai/docs", + "models": { + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, + "anthropic/claude-opus-5-5": { + "id": "anthropic/claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 + } + }, + "anthropic/claude-fable-5-1": { + "id": "anthropic/claude-fable-5-1", + "name": "Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 + } + }, + "anthropic/claude-opus-5": { + "id": "anthropic/claude-opus-5", + "name": "Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + } + }, + "anthropic/claude-opus-4-8": { + "id": "anthropic/claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 + } + }, + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "deepseek/deepseek-v4.1-flash": { + "id": "deepseek/deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.27, + "output": 1.08, + "cache_read": 0.0054, + "cache_write": 0.27 + } + }, + "deepseek/deepseek-v4-flash-vision-exp": { + "id": "deepseek/deepseek-v4-flash-vision-exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.27, + "output": 1.08, + "cache_read": 0.0054, + "cache_write": 0.27 + } + }, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 1.215, + "output": 3.645, + "cache_read": 0.0405, + "cache_write": 1.215 + } + }, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.27, + "output": 1.08, + "cache_read": 0.0054, + "cache_write": 0.27 + } + }, + "tencent/hy3": { + "id": "tencent/hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "input": 192000, + "output": 131072 + }, + "cost": { + "input": 0.16, + "output": 0.64, + "cache_read": 0.04, + "cache_write": 0.16 + } + }, + "tencent/hy4-preview": { + "id": "tencent/hy4-preview", + "name": "Hy4 preview", + "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", + "family": "Hy", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-08-28", + "last_updated": "2026-08-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.96, + "output": 2.88, + "cache_read": 0.048, + "cache_write": 0.96 + } + }, + "z-ai/glm-5.3-flash": { + "id": "z-ai/glm-5.3-flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 0.15, + "output": 0.5, + "cache_read": 0.03, + "cache_write": 0.15 + } + }, + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.16, + "cache_write": 0.6, + "tiers": [ + { + "input": 0.8, + "output": 3.4, + "cache_read": 0.2, + "cache_write": 0.8, + "tier": { + "type": "context", + "size": 32000 + } + } + ] + } + }, + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.47, + "output": 2.16, + "cache_read": 0.1, + "cache_write": 0.47, + "tiers": [ + { + "input": 0.62, + "output": 2.47, + "cache_read": 0.13, + "cache_write": 0.62, + "tier": { + "type": "context", + "size": 32000 + } + } + ] + } + }, + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 1.2, + "output": 4.4, + "cache_read": 0.3, + "cache_write": 1.2 + } + }, + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 1, + "output": 3.8, + "cache_read": 0.2, + "cache_write": 1, + "tiers": [ + { + "input": 1.2, + "output": 4.4, + "cache_read": 0.3, + "cache_write": 1.2, + "tier": { + "type": "context", + "size": 32000 + } + } + ] + } + }, + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 0.9, + "output": 3.7, + "cache_read": 0.18, + "cache_write": 0.9, + "tiers": [ + { + "input": 1.1, + "output": 4.3, + "cache_read": 0.27, + "cache_write": 1.1, + "tier": { + "type": "context", + "size": 32000 + } + } + ] + } + }, + "z-ai/glm-5.3": { + "id": "z-ai/glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 1.2, + "output": 4.4, + "cache_read": 0.3, + "cache_write": 1.2 + } + }, + "x-ai/grok-4.7": { + "id": "x-ai/grok-4.7", + "name": "Grok 4.7", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-05", + "release_date": "2026-09-21", + "last_updated": "2026-09-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "cache_write": 2, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "cache_write": 4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1, + "cache_write": 4 + } + } + }, + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 1000000 + }, + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "cache_write": 1.25, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "cache_write": 2.5 + } + } + }, + "x-ai/grok-4.5": { + "id": "x-ai/grok-4.5", + "name": "Grok 4.5", + "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.3, + "cache_write": 2, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 0.6, + "cache_write": 4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 0.6, + "cache_write": 4 + } + } + }, + "x-ai/grok-build-0.1": { + "id": "x-ai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 1, + "output": 2, + "cache_read": 0.2, + "cache_write": 1, + "tiers": [ + { + "input": 2, + "output": 4, + "cache_read": 0.4, + "cache_write": 2, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 4, + "cache_read": 0.4, + "cache_write": 2 + } + } + }, + "x-ai/grok-4.6": { + "id": "x-ai/grok-4.6", + "name": "Grok 4.6", + "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02-01", + "release_date": "2026-08-12", + "last_updated": "2026-08-12", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 500000, + "output": 500000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5, + "cache_write": 2, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "cache_write": 4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1, + "cache_write": 4 + } + } + }, + "xiaomi/mimo-v2.6-pro": { + "id": "xiaomi/mimo-v2.6-pro", + "name": "MiMo-V2.6-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 262144 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.13, - "output": 0.5 + "input": 0.47, + "output": 0.94, + "cache_read": 0.005, + "cache_write": 0.47 } }, - "qwen/qwq-32b-preview": { - "id": "qwen/qwq-32b-preview", - "name": "Qwen QwQ 32B Preview", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 0.16, + "output": 0.32, + "cache_read": 0.004, + "cache_write": 0.16 } }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B (Thinking)", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -188437,56 +316789,76 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.65 + "input": 0.47, + "output": 0.94, + "cache_read": 0.005, + "cache_write": 0.47 } }, - "qwen/qwen3-coder": { - "id": "qwen/qwen3-coder", - "name": "Qwen 3 Coder 480B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "attachment": false, - "reasoning": false, + "xiaomi/mimo-v2.6-flash": { + "id": "xiaomi/mimo-v2.6-flash", + "name": "MiMo-V2.6-Flash", + "description": "MiMo Flash model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "input": 262000, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.13, - "output": 0.5 + "input": 0.16, + "output": 0.32, + "cache_read": 0.004, + "cache_write": 0.16 } }, - "qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen 3 235b A22B 2507 Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-11", - "last_updated": "2025-09-11", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -188495,27 +316867,35 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 262144 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 0.5 + "input": 0.33, + "output": 1.32, + "cache_read": 0.066, + "cache_write": 0.42 } }, - "qwen/qwen3.5-plus": { - "id": "qwen/qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", @@ -188526,39 +316906,68 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 983616, - "input": 983616, - "output": 65536 + "context": 1024000, + "output": 512000 }, "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.04 + "input": 0.33, + "output": 1.32, + "cache_read": 0.066, + "cache_write": 0.33, + "tiers": [ + { + "input": 0.66, + "output": 2.63, + "cache_read": 0.132, + "cache_write": 0.66, + "tier": { + "type": "context", + "size": 512000 + } + } + ], + "context_over_200k": { + "input": 0.66, + "output": 2.63, + "cache_read": 0.132, + "cache_write": 0.66 + } } }, - "qwen/qwen3.5-plus-thinking": { - "id": "qwen/qwen3.5-plus-thinking", - "name": "Qwen3.5 Plus Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "gemini/gemini-3.6-flash": { + "id": "gemini/gemini-3.6-flash", + "name": "Gemini 3.6 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ @@ -188567,30 +316976,49 @@ }, "open_weights": false, "limit": { - "context": 983616, - "input": 983616, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.04 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.75 } }, - "qwen/qwen3-235b-a22b": { - "id": "qwen/qwen3-235b-a22b", - "name": "Qwen 3 235b A22B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "gemini/gemini-3.5-flash-lite": { + "id": "gemini/gemini-3.5-flash-lite", + "name": "Gemini 3.5 Flash Lite", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash-lite", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-07-21", + "last_updated": "2026-07-21", "modalities": { "input": [ "text", - "pdf" + "image", + "audio", + "video" ], "output": [ "text" @@ -188598,57 +317026,45 @@ }, "open_weights": false, "limit": { - "context": 41000, - "input": 41000, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { "input": 0.3, - "output": 0.5 - } - }, - "qwen/qwen-2.5-72b-instruct": { - "id": "qwen/qwen-2.5-72b-instruct", - "name": "Qwen2.5 72B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-03", - "last_updated": "2025-07-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "input": 131072, - "output": 8192 - }, - "cost": { - "input": 0.357, - "output": 0.408 + "output": 2.5, + "cache_read": 0.03, + "cache_write": 0.3 } }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "attachment": false, - "reasoning": false, + "gemini/gemini-3.1-pro-preview": { + "id": "gemini/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -188656,19 +317072,39 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 4 + } } }, - "qwen/qwen3.5-9b": { - "id": "qwen/qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "gemini/gemini-3.5-flash": { + "id": "gemini/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -188676,19 +317112,27 @@ "type": "toggle" }, { - "type": "budget_tokens", - "min": 1024, - "max": 81920 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-03-10", - "last_updated": "2026-03-10", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -188696,60 +317140,113 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.05, - "output": 0.15 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 1.5 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "gemini/gemini-2.5-pro": { + "id": "gemini/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 258048, - "input": 258048, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "cache_write": 1.25, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 2.5 + } } }, - "qwen/qwen3-coder-flash": { - "id": "qwen/qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-17", - "last_updated": "2025-09-17", + "gemini/gemini-2.5-flash": { + "id": "gemini/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -188757,57 +317254,45 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, + "context": 1048576, "output": 65536 }, "cost": { "input": 0.3, - "output": 1.5 - } - }, - "qwen/Qwen2.5-Coder-32B-Instruct": { - "id": "qwen/Qwen2.5-Coder-32B-Instruct", - "name": "Qwen 2.5 Coder 32b", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-03", - "last_updated": "2025-07-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "input": 32000, - "output": 8192 - }, - "cost": { - "input": 0.2006, - "output": 0.2006 + "output": 2.5, + "cache_read": 0.03, + "cache_write": 0.3 } }, - "qwen/Qwen3-8B": { - "id": "qwen/Qwen3-8B", - "name": "Qwen 3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "gemini/gemini-3.7-flash": { + "id": "gemini/gemini-3.7-flash", + "name": "Gemini 3.7 Flash", + "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-03", + "release_date": "2026-08-13", + "last_updated": "2026-08-13", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -188815,28 +317300,49 @@ }, "open_weights": false, "limit": { - "context": 41000, - "input": 41000, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.47, - "output": 0.47 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.75 } }, - "qwen/qwen3-14b": { - "id": "qwen/qwen3-14b", - "name": "Qwen 3 14b", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "gemini/gemini-3-flash-preview": { + "id": "gemini/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -188844,28 +317350,44 @@ }, "open_weights": false, "limit": { - "context": 41000, - "input": 41000, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.08, - "output": 0.24 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.5 } }, - "qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen 3 235b A22B 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, + "gemini/gemini-3.8-flash": { + "id": "gemini/gemini-3.8-flash", + "name": "Gemini 3.8 Flash", + "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "temperature": true, + "release_date": "2026-09-02", + "last_updated": "2026-09-02", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -188873,28 +317395,47 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 262144 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.13, - "output": 0.5 + "input": 0.75, + "output": 3.75, + "cache_read": 0.075, + "cache_write": 0.75 } }, - "qwen/qwen3-30b-a3b": { - "id": "qwen/qwen3-30b-a3b", - "name": "Qwen3 30B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "gemini/gemini-2.5-flash-lite": { + "id": "gemini/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -188902,56 +317443,33 @@ }, "open_weights": false, "limit": { - "context": 41000, - "input": 41000, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { "input": 0.1, - "output": 0.3 - } - }, - "amazon/nova-lite-v1": { - "id": "amazon/nova-lite-v1", - "name": "Amazon Nova Lite 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "input": 300000, - "output": 5120 - }, - "cost": { - "input": 0.0595, - "output": 0.238 + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0.1 } }, - "amazon/nova-pro-v1": { - "id": "amazon/nova-pro-v1", - "name": "Amazon Nova Pro 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -188962,29 +317480,49 @@ }, "open_weights": false, "limit": { - "context": 300000, - "input": 300000, - "output": 32000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.7989999999999999, - "output": 3.1959999999999997 + "input": 1.88, + "output": 5.63, + "cache_read": 0.375, + "cache_write": 2.35 } }, - "amazon/nova-micro-v1": { - "id": "amazon/nova-micro-v1", - "name": "Amazon Nova Micro 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "qwen/qwen3.8-max": { + "id": "qwen/qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -188992,29 +317530,38 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 5120 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.0357, - "output": 0.1394 + "input": 1.88, + "output": 5.63, + "cache_read": 0.23, + "cache_write": 2.35 } }, - "amazon/nova-2-lite-v1": { - "id": "amazon/nova-2-lite-v1", - "name": "Amazon Nova 2 Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "qwen/qwen3.7-flash": { + "id": "qwen/qwen3.7-flash", + "name": "Qwen3.7 Flash", + "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -189023,29 +317570,69 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 1000000, - "output": 65535 + "input": 991000, + "output": 65536 }, "cost": { - "input": 0.5099999999999999, - "output": 4.25 + "input": 0.04, + "output": 0.13, + "cache_read": 0.01, + "cache_write": 0.04, + "tiers": [ + { + "input": 0.1, + "output": 0.37, + "cache_read": 0.02, + "cache_write": 0.12, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.19, + "output": 0.74, + "cache_read": 0.04, + "cache_write": 0.24, + "tier": { + "type": "context", + "size": 256000 + } + } + ] } }, - "alibaba/qwen3.6-flash": { - "id": "alibaba/qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "qwen/qwen3.8-omni-flash": { + "id": "qwen/qwen3.8-omni-flash", + "name": "Qwen3.8 Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "release_date": "2026-09-17", + "last_updated": "2026-09-17", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ @@ -189054,24 +317641,33 @@ }, "open_weights": false, "limit": { - "context": 991800, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.19, - "output": 1.16 + "input": 0.13, + "output": 0.43, + "cache_read": 0.016, + "cache_write": 0.13 } }, - "alibaba/qwen3.6-27b": { - "id": "alibaba/qwen3.6-27b", - "name": "Qwen3.6 27B", + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen3.6 Flash", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", @@ -189084,32 +317680,58 @@ }, "open_weights": false, "limit": { - "context": 260096, - "input": 260096, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.203, - "output": 2.24 + "input": 0.19, + "output": 1.13, + "cache_read": 0.019, + "cache_write": 0.24, + "tiers": [ + { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075, + "cache_write": 0.94, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075, + "cache_write": 0.94 + } } }, - "alibaba/qwen3.6-27b:thinking": { - "id": "alibaba/qwen3.6-27b:thinking", - "name": "Qwen3.6 27B Thinking", + "qwen/qwen3.8-flash": { + "id": "qwen/qwen3.8-flash", + "name": "Qwen3.8 Flash", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 131072 + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", @@ -189122,29 +317744,39 @@ }, "open_weights": false, "limit": { - "context": 260096, - "input": 260096, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.203, - "output": 2.24 + "input": 0.13, + "output": 0.43, + "cache_read": 0.016, + "cache_write": 0.2 } }, - "aion-labs/aion-rp-llama-3.1-8b": { - "id": "aion-labs/aion-rp-llama-3.1-8b", - "name": "Llama 3.1 8b (uncensored)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -189152,28 +317784,57 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.2006, - "output": 0.2006 + "input": 0.32, + "output": 1.88, + "cache_read": 0.032, + "cache_write": 0.4, + "tiers": [ + { + "input": 1.25, + "output": 7.5, + "cache_read": 0.124, + "cache_write": 1.57, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.25, + "output": 7.5, + "cache_read": 0.124, + "cache_write": 1.57 + } } }, - "aion-labs/aion-2.5": { - "id": "aion-labs/aion-2.5", - "name": "AionLabs: Aion-2.5", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -189181,30 +317842,63 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.35 + "input": 0.32, + "output": 1.25, + "cache_read": 0.032, + "cache_write": 0.4, + "tiers": [ + { + "input": 0.96, + "output": 3.75, + "cache_read": 0.096, + "cache_write": 1.2, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.96, + "output": 3.75, + "cache_read": 0.096, + "cache_write": 1.2 + } } }, - "aion-labs/aion-1.0-mini": { - "id": "aion-labs/aion-1.0-mini", - "name": "Aion 1.0 mini (DeepSeek)", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-20", - "last_updated": "2025-02-20", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189212,28 +317906,62 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 8192 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.7989999999999999, - "output": 1.394 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 2.5, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 5 + } } }, - "aion-labs/aion-2.0": { - "id": "aion-labs/aion-2.0", - "name": "AionLabs: Aion-2.0", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189241,29 +317969,58 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.8, - "output": 1.6 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "aion-labs/aion-1.0": { - "id": "aion-labs/aion-1.0", - "name": "Aion 1.0", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-01", - "last_updated": "2025-02-01", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189271,30 +318028,46 @@ }, "open_weights": false, "limit": { - "context": 65536, - "input": 65536, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 3.995, - "output": 7.99 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02, + "cache_write": 0.2 } }, - "pamanseau/OpenReasoning-Nemotron-32B": { - "id": "pamanseau/OpenReasoning-Nemotron-32B", - "name": "OpenReasoning Nemotron 32B", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, + "openai/gpt-6-astra": { + "id": "openai/gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189302,29 +318075,52 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } } }, - "LatitudeGames/Wayfarer-Large-70B-Llama-3.3": { - "id": "LatitudeGames/Wayfarer-Large-70B-Llama-3.3", - "name": "Llama 3.3 70B Wayfarer", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-20", - "last_updated": "2025-02-20", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189332,29 +318128,45 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, + "context": 128000, "output": 16384 }, "cost": { - "input": 0.700000007, - "output": 0.700000007 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075, + "cache_write": 0.15 } }, - "baseten/Kimi-K2-Instruct-FP4": { - "id": "baseten/Kimi-K2-Instruct-FP4", - "name": "Kimi K2 0711 Instruct FP4", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-11", - "last_updated": "2025-07-11", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189362,29 +318174,47 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 2 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075, + "cache_write": 0.75 } }, - "inflection/inflection-3-pi": { - "id": "inflection/inflection-3-pi", - "name": "Inflection 3 Pi", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-10-11", - "last_updated": "2024-10-11", + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189392,29 +318222,64 @@ }, "open_weights": false, "limit": { - "context": 8000, - "input": 8000, - "output": 4096 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 2.499, - "output": 9.996 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "cache_write": 0.5 + } } }, - "inflection/inflection-3-productivity": { - "id": "inflection/inflection-3-productivity", - "name": "Inflection 3 Productivity", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-10-11", - "last_updated": "2024-10-11", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189422,59 +318287,65 @@ }, "open_weights": false, "limit": { - "context": 8000, - "input": 8000, - "output": 4096 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 2.499, - "output": 9.996 - } - }, - "ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0": { - "id": "ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0", - "name": "Omega Directive 24B Unslop v2.0", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", - "modalities": { - "input": [ - "text" + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16384, - "input": 16384, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 0.5 + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 10 + } } }, - "MiniMaxAI/MiniMax-M1-80k": { - "id": "MiniMaxAI/MiniMax-M1-80k", - "name": "MiniMax M1 80K", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-16", - "last_updated": "2025-06-16", + "openai/gpt-6-luna": { + "id": "openai/gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189482,28 +318353,65 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.6052, - "output": 2.4225000000000003 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } } }, - "upstage/solar-pro-3": { - "id": "upstage/solar-pro-3", - "name": "Solar Pro 3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189511,31 +318419,65 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "allenai/olmo-3-32b-think": { - "id": "allenai/olmo-3-32b-think", - "name": "Olmo 3 32B Think", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "allenai", - "attachment": false, + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189543,29 +318485,65 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 0.44999999999999996 + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.8, + "cache_write": 10 + } } }, - "essentialai/rnj-1-instruct": { - "id": "essentialai/rnj-1-instruct", - "name": "RNJ-1 Instruct 8B", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "rnj", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-13", - "last_updated": "2025-12-13", + "openai/gpt-6-sol": { + "id": "openai/gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189573,146 +318551,182 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "attachment": false, + "moonshot/kimi-k3": { + "id": "moonshot/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", "high", - "xhigh" + "max" ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "input": 1048576, - "output": 384000 + "output": 1048576 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3 } }, - "deepseek/deepseek-v4-flash:thinking": { - "id": "deepseek/deepseek-v4-flash:thinking", - "name": "DeepSeek V4 Flash (Thinking)", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "attachment": false, + "moonshot/kimi-k2.6": { + "id": "moonshot/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 384000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 1, + "output": 4.16, + "cache_read": 0.18, + "cache_write": 1 } }, - "deepseek/deepseek-v4-pro-cheaper:thinking": { - "id": "deepseek/deepseek-v4-pro-cheaper:thinking", - "name": "DeepSeek V4 Pro Cheaper (Thinking)", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "attachment": false, + "moonshot/kimi-k2.7-code": { + "id": "moonshot/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2026-04-25", - "last_updated": "2026-04-25", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 384000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 1, + "output": 4.16, + "cache_read": 0.18, + "cache_write": 1 } - }, - "deepseek/deepseek-prover-v2-671b": { - "id": "deepseek/deepseek-prover-v2-671b", - "name": "DeepSeek Prover v2 671B", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", + } + } + }, + "arcee": { + "id": "arcee", + "env": [ + "ARCEE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.arcee.ai/api/v1", + "name": "Arcee", + "doc": "https://docs.arcee.ai", + "models": { + "trinity-large-thinking": { + "id": "trinity-large-thinking", + "name": "Trinity Large Thinking", + "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", + "family": "trinity", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-30", - "last_updated": "2025-04-30", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-05-28", "modalities": { "input": [ "text" @@ -189721,21 +318735,23 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 160000, - "input": 160000, - "output": 16384 + "context": 262144, + "output": 262144 }, + "status": "beta", "cost": { - "input": 1, - "output": 2.5 + "input": 0.25, + "output": 0.8, + "cache_read": 0.06 } }, - "deepseek/deepseek-latest": { - "id": "deepseek/deepseek-latest", - "name": "DeepSeek Latest", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "deepseek/deepseek-v4-pro-0813": { + "id": "deepseek/deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -189743,14 +318759,18 @@ "type": "effort", "values": [ "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ "text" @@ -189759,38 +318779,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "input": 1048576, "output": 384000 }, + "status": "beta", "cost": { - "input": 1.1, - "output": 2.2, - "cache_read": 0.11 + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 } }, "deepseek/deepseek-v4-pro": { "id": "deepseek/deepseek-v4-pro", "name": "DeepSeek V4 Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, + "temperature": true, + "knowledge": "2025-05", "release_date": "2026-04-24", "last_updated": "2026-04-24", "modalities": { @@ -189801,22 +318824,23 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, + "context": 512000, "output": 384000 }, + "status": "beta", "cost": { - "input": 1.1, - "output": 2.2, - "cache_read": 0.11 + "input": 1.74, + "output": 3.48, + "cache_read": 0.2 } }, - "deepseek/deepseek-v4-pro:thinking": { - "id": "deepseek/deepseek-v4-pro:thinking", - "name": "DeepSeek V4 Pro (Thinking)", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "deepseek/deepseek-v4-flash-latest": { + "id": "deepseek/deepseek-v4-flash-latest", + "name": "DeepSeek V4 Flash Latest", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -189824,14 +318848,19 @@ "type": "effort", "values": [ "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -189840,72 +318869,78 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "input": 1048576, "output": 384000 }, + "status": "beta", "cost": { - "input": 1.1, - "output": 2.2, - "cache_read": 0.11 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "deepseek/deepseek-v3.2-speciale": { - "id": "deepseek/deepseek-v3.2-speciale", - "name": "DeepSeek V3.2 Speciale", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": true, + "thinkingmachines/inkling-small": { + "id": "thinkingmachines/inkling-small", + "name": "Inkling Small", + "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", + "family": "ling", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-07-30", + "last_updated": "2026-07-30", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163000, - "input": 163000, - "output": 65536 + "context": 262144, + "output": 262144 }, + "status": "beta", "cost": { - "input": 0.27999999999999997, - "output": 0.42000000000000004 + "input": 0.5, + "output": 1.2, + "cache_read": 0.1 } }, - "deepseek/deepseek-v4-pro-cheaper": { - "id": "deepseek/deepseek-v4-pro-cheaper", - "name": "DeepSeek V4 Pro Cheaper", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "moonshotai/kimi-k3": { + "id": "moonshotai/kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-04-25", - "last_updated": "2026-04-25", + "temperature": true, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text" @@ -189914,115 +318949,106 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "input": 1048576, - "output": 384000 + "output": 131072 }, + "status": "beta", "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "deepseek/deepseek-v3.2:thinking": { - "id": "deepseek/deepseek-v3.2:thinking", - "name": "DeepSeek V3.2 Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": true, + "zai-org/glm-5.2": { + "id": "zai-org/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163000, - "input": 163000, - "output": 65536 + "context": 262144, + "output": 131072 }, + "status": "beta", "cost": { - "input": 0.27999999999999997, - "output": 0.42000000000000004 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } - }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", + } + } + }, + "drun": { + "id": "drun", + "env": [ + "DRUN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://chat.d.run/v1", + "name": "D.Run (China)", + "doc": "https://www.d.run", + "models": { + "public/deepseek-v3": { + "id": "public/deepseek-v3", + "name": "DeepSeek V3", "description": "DeepSeek chat model for instruction following, coding, and analysis", "family": "deepseek", - "attachment": true, + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-12-26", + "last_updated": "2024-12-26", "modalities": { "input": [ - "text", - "pdf" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 163000, - "input": 163000, - "output": 65536 - }, - "cost": { - "input": 0.27999999999999997, - "output": 0.42000000000000004 - } - }, - "minimax/minimax-m3:thinking": { - "id": "minimax/minimax-m3:thinking", - "name": "MiniMax M3 Thinking", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 512000, - "input": 512000, - "output": 80000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.28, + "output": 1.1 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", + "public/minimax-m25": { + "id": "public/minimax-m25", "name": "MiniMax M2.5", "description": "MiniMax model for chat, coding, office work, and agentic tasks", "family": "minimax", @@ -190030,9 +319056,12 @@ "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "interleaved": { + "field": "reasoning_details" + }, + "temperature": true, + "release_date": "2025-03-01", + "last_updated": "2025-03-01", "modalities": { "input": [ "text" @@ -190044,26 +319073,29 @@ "open_weights": false, "limit": { "context": 204800, - "input": 204800, "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.29, + "output": 1.16 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "public/deepseek-r1": { + "id": "public/deepseek-r1", + "name": "DeepSeek R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2025-12-19", - "last_updated": "2025-12-19", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -190072,28 +319104,50 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 131072 + "context": 131072, + "output": 32000 }, "cost": { - "input": 0.33, - "output": 1.32 + "input": 0.55, + "output": 2.2 } - }, - "minimax/minimax-m2.7-turbo": { - "id": "minimax/minimax-m2.7-turbo", - "name": "MiniMax M2.7 Turbo", - "description": "Efficient MiniMax model for quick assistance, coding, and routine automation", + } + } + }, + "amd": { + "id": "amd", + "env": [ + "AMD_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://developer.amd.com.cn/radeon/api/v1", + "name": "AMD", + "doc": "https://developer.amd.com.cn/radeon/tokenfactory", + "models": { + "Qwen3.8-27B": { + "id": "Qwen3.8-27B", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium" + ] + } + ], "tool_call": true, + "interleaved": true, "structured_output": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -190102,58 +319156,43 @@ "text" ] }, - "open_weights": false, - "limit": { - "context": 204800, - "input": 204800, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.4 - } - }, - "minimax/minimax-01": { - "id": "minimax/minimax-01", - "name": "MiniMax 01", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-15", - "last_updated": "2025-01-15", - "modalities": { - "input": [ - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000192, - "input": 1000192, - "output": 16384 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.1394, - "output": 1.1219999999999999 + "input": 0, + "output": 0 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "DeepSeek-V4.1-Flash": { + "id": "DeepSeek-V4.1-Flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -190163,29 +319202,44 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 512000, - "input": 512000, - "output": 80000 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "minimax/minimax-m2-her": { - "id": "minimax/minimax-m2-her", - "name": "MiniMax M2-her", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "DeepSeek-V4-Flash": { + "id": "DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-01-24", - "last_updated": "2026-01-24", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -190194,32 +319248,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65532, - "input": 65532, - "output": 2048 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0.30200000000000005, - "output": 1.2069999999999999 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "minimax/minimax-latest": { - "id": "minimax/minimax-latest", - "name": "MiniMax Latest", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "Qwen3.8-Flash-Next": { + "id": "Qwen3.8-Flash-Next", + "name": "Qwen3.8 Flash Next", + "description": "Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "temperature": true, + "release_date": "2026-08-27", + "last_updated": "2026-08-27", "modalities": { "input": [ "text", @@ -190229,30 +319290,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 512000, - "input": 512000, - "output": 80000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.15, + "output": 0.47, + "cache_read": 0.016 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "MiniCPM5-2B": { + "id": "MiniCPM5-2B", + "name": "MiniCPM5-2B", + "description": "Dense 2B-class open-source model for on-device and resource-constrained use, with native long-context support, tool calling, and agentic tasks", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": true, + "release_date": "2026-09-06", + "last_updated": "2026-09-12", "modalities": { "input": [ "text" @@ -190261,57 +319320,85 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "input": 204800, + "context": 131072, "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.124, + "output": 0.7425, + "cache_read": 0.124 } }, - "kwaipilot/kat-coder-pro-v2": { - "id": "kwaipilot/kat-coder-pro-v2", - "name": "KAT Coder Pro V2", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-03-28", - "last_updated": "2026-03-28", + "DeepSeek-V4-Flash-Vision-Exp": { + "id": "DeepSeek-V4-Flash-Vision-Exp", + "name": "DeepSeek V4 Flash Vision Exp", + "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", + "family": "deepseek-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-21", + "last_updated": "2026-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 80000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } - }, - "mlabonne/NeuralDaredevil-8B-abliterated": { - "id": "mlabonne/NeuralDaredevil-8B-abliterated", - "name": "Neural Daredevil 8B abliterated", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + } + } + }, + "qvac": { + "id": "qvac", + "env": [ + "QVAC_API_KEY" + ], + "npm": "@qvac/ai-sdk-provider", + "name": "QVAC", + "doc": "https://www.npmjs.com/package/@qvac/ai-sdk-provider", + "models": { + "qwen3.5-0.8b": { + "id": "qwen3.5-0.8b", + "name": "Qwen3.5 0.8B", + "description": "Qwen instruction model for multilingual chat and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "tool_call": true, + "temperature": true, + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text" @@ -190320,88 +319407,93 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "input": 8192, + "context": 32768, "output": 8192 }, "cost": { - "input": 0.44, - "output": 0.44 + "input": 0, + "output": 0 } }, - "VongolaChouko/Starcannon-Unleashed-12B-v1.0": { - "id": "VongolaChouko/Starcannon-Unleashed-12B-v1.0", - "name": "Mistral Nemo Starcannon 12b v1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "gemma4-31b": { + "id": "gemma4-31b", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 0, + "output": 0 } }, - "TEE/minimax-m2.5": { - "id": "TEE/minimax-m2.5", - "name": "MiniMax M2.5 TEE", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": false, + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 196608, - "input": 196608, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 1.38 + "input": 0, + "output": 0 } }, - "TEE/glm-4.7": { - "id": "TEE/glm-4.7", - "name": "GLM 4.7 TEE", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "qwen3.5-2b": { + "id": "qwen3.5-2b", + "name": "Qwen3.5 2B", + "description": "Qwen instruction model for multilingual chat and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-01-29", - "last_updated": "2026-01-29", + "tool_call": true, + "temperature": true, + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text" @@ -190410,28 +319502,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "input": 131000, - "output": 65535 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.85, - "output": 3.3 + "input": 0, + "output": 0 } }, - "TEE/llama3-3-70b": { - "id": "TEE/llama3-3-70b", - "name": "Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-03", - "last_updated": "2025-07-03", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -190440,62 +319533,68 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 131072, + "output": 32768 }, "cost": { - "input": 2, - "output": 2 + "input": 0, + "output": 0 } }, - "TEE/gemma-4-31b-it": { - "id": "TEE/gemma-4-31b-it", - "name": "Gemma 4 31B IT TEE", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": false, + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "max" + ] } ], "tool_call": true, - "structured_output": false, - "release_date": "2026-05-26", - "last_updated": "2026-05-26", + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 262144 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 0.46 + "input": 0, + "output": 0 } }, - "TEE/qwen3-30b-a3b-instruct-2507": { - "id": "TEE/qwen3-30b-a3b-instruct-2507", - "name": "Qwen3 30B A3B Instruct 2507 TEE", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen3.5-4b": { + "id": "qwen3.5-4b", + "name": "Qwen3.5 4B", + "description": "Qwen instruction model for multilingual chat and tool use", "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "tool_call": true, + "temperature": true, + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text" @@ -190504,62 +319603,63 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "input": 262000, - "output": 32768 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 0.44999999999999996 + "input": 0, + "output": 0 } }, - "TEE/glm-5.1": { - "id": "TEE/glm-5.1", - "name": "GLM 5.1 TEE", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "qwen3.6-27b": { + "id": "qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202752, - "input": 202752, - "output": 65535 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.5, - "output": 5.25, - "cache_read": 0.3 + "input": 0, + "output": 0 } }, - "TEE/deepseek-v4-pro": { - "id": "TEE/deepseek-v4-pro", - "name": "DeepSeek V4 Pro TEE", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2026-04-25", - "last_updated": "2026-04-25", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -190568,32 +319668,45 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 800000, - "input": 800000, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 5.25, - "cache_read": 0.15 + "input": 0, + "output": 0 } - }, - "TEE/deepseek-v4-pro:thinking": { - "id": "TEE/deepseek-v4-pro:thinking", - "name": "DeepSeek V4 Pro Thinking TEE", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "attachment": false, + } + } + }, + "claudinio": { + "id": "claudinio", + "env": [ + "CLAUDINIO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.claudin.io/v1", + "name": "Claudinio", + "doc": "https://claudin.io", + "models": { + "claudinio": { + "id": "claudinio", + "name": "Claudinio", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "knowledge": "2026-05", + "release_date": "2026-05-12", + "last_updated": "2026-06-02", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -190601,30 +319714,32 @@ }, "open_weights": false, "limit": { - "context": 800000, - "input": 800000, - "output": 65536 + "context": 256000, + "output": 64000 }, "cost": { - "input": 1.5, - "output": 5.25, + "input": 0.5, + "output": 2, "cache_read": 0.15 } }, - "TEE/gemma4-31b:thinking": { - "id": "TEE/gemma4-31b:thinking", - "name": "Gemma 4 31B Thinking TEE", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": false, + "claudius": { + "id": "claudius", + "name": "Claudius", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "release_date": "2026-05-02", - "last_updated": "2026-05-02", + "tool_call": true, + "knowledge": "2026-05", + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -190632,26 +319747,49 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 131072 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.45, - "output": 1 + "input": 3, + "output": 8, + "cache_read": 0.9 } - }, - "TEE/kimi-k2.5": { - "id": "TEE/kimi-k2.5", - "name": "Kimi K2.5 TEE", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + } + } + }, + "runinfra": { + "id": "runinfra", + "env": [ + "RUNINFRA_GATEWAY_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.runinfra.ai/v1", + "name": "RunInfra", + "doc": "https://runinfra.ai/docs", + "models": { + "Inferact/Qwen3.8-2.4T-A95B-NVFP4": { + "id": "Inferact/Qwen3.8-2.4T-A95B-NVFP4", + "name": "Qwen3.8 2.4T A95B (NVFP4)", + "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-01-29", - "last_updated": "2026-01-29", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-12", "modalities": { "input": [ "text" @@ -190660,28 +319798,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 65535 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.9 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "TEE/qwen2.5-vl-72b-instruct": { - "id": "TEE/qwen2.5-vl-72b-instruct", - "name": "Qwen2.5 VL 72B TEE", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "Qwen/Qwen3.8-27B": { + "id": "Qwen/Qwen3.8-27B", + "name": "Qwen3.8 27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-01", - "last_updated": "2025-02-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", @@ -190691,28 +319841,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "input": 65536, - "output": 8192 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.7, - "output": 0.7 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 } }, - "TEE/gpt-oss-120b": { - "id": "TEE/gpt-oss-120b", - "name": "GPT-OSS 120B TEE", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "deepseek-ai/DeepSeek-V4-Flash-0731": { + "id": "deepseek-ai/DeepSeek-V4-Flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -190721,58 +319883,76 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 16384 + "context": 1048576, + "output": 32768 }, "cost": { - "input": 2, - "output": 2 + "input": 0.13, + "output": 0.27, + "cache_read": 0.01 } }, - "TEE/qwen3.5-27b": { - "id": "TEE/qwen3.5-27b", - "name": "Qwen3.5 27B TEE", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-03-13", - "last_updated": "2026-03-13", + "deepseek-ai/DeepSeek-V4-Pro-0813": { + "id": "deepseek-ai/DeepSeek-V4-Pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 65536 + "context": 1048576, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 2.4 + "input": 0.6, + "output": 1.9, + "cache_read": 0.03 } }, - "TEE/gemma-4-26b-a4b-uncensored": { - "id": "TEE/gemma-4-26b-a4b-uncensored", - "name": "Gemma 4 26B A4B Uncensored TEE", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "zai-org/GLM-5.3-Flash": { + "id": "zai-org/GLM-5.3-Flash", + "name": "GLM-5.3-Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": true, "structured_output": true, - "release_date": "2026-05-23", - "last_updated": "2026-05-23", + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", @@ -190782,57 +319962,118 @@ "text" ] }, - "open_weights": false, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 + } + }, + "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16": { + "id": "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16", + "name": "Nemotron 3.5 Lightning 30B A3B", + "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-11", + "last_updated": "2026-08-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, "limit": { - "context": 65536, - "input": 65536, - "output": 65536 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.7 + "input": 0.05, + "output": 0.15, + "cache_read": 0.01 } }, - "TEE/deepseek-v3.1": { - "id": "TEE/deepseek-v3.1", - "name": "DeepSeek V3.1 TEE", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "ornith-ai/Ornith-1.5-35B-A3B": { + "id": "ornith-ai/Ornith-1.5-35B-A3B", + "name": "Ornith 1.5 35B A3B", + "description": "Mixture-of-experts coding-reasoning model for agentic software tasks, tool use, and image understanding", + "family": "ornith", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-18", + "last_updated": "2026-08-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 164000, - "input": 164000, - "output": 8192 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1, - "output": 2.5 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 } - }, - "TEE/kimi-k2.6": { - "id": "TEE/kimi-k2.6", - "name": "Kimi K2.6 TEE", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + } + } + }, + "hetzner": { + "id": "hetzner", + "env": [ + "HETZNER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.hetzner.com/api/v1", + "name": "Hetzner", + "doc": "https://experiments.hetzner.com/docs/inference", + "models": { + "Qwen3.8-27B": { + "id": "Qwen3.8-27B", + "name": "Qwen3.8-27B", + "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "structured_output": true, + "temperature": true, + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text", @@ -190842,60 +320083,90 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "input": 262144, - "output": 65536 + "output": 262144 }, + "status": "beta", "cost": { - "input": 1.5, - "output": 5.25, - "cache_read": 0.375 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "TEE/qwen3.5-397b-a17b": { - "id": "TEE/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B TEE", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "Qwen/Qwen3.6-35B-A3B-FP8": { + "id": "Qwen/Qwen3.6-35B-A3B-FP8", + "name": "Qwen3.6 35B A3B FP8", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-28", - "last_updated": "2026-02-28", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 258048, - "input": 258048, - "output": 65536 + "context": 262144, + "output": 262144 }, + "status": "beta", "cost": { - "input": 0.6, - "output": 3.6 + "input": 0, + "output": 0, + "cache_read": 0 } - }, - "TEE/kimi-k2.5-thinking": { - "id": "TEE/kimi-k2.5-thinking", - "name": "Kimi K2.5 Thinking TEE", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + } + } + }, + "digitalocean": { + "id": "digitalocean", + "env": [ + "DIGITALOCEAN_ACCESS_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.do-ai.run/v1", + "name": "DigitalOcean", + "doc": "https://docs.digitalocean.com/products/gradient-ai-platform/details/models/", + "models": { + "deepseek-3.2": { + "id": "deepseek-3.2", + "name": "Deepseek 3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2026-01-29", - "last_updated": "2026-01-29", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-12-02", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" @@ -190904,37 +320175,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 65535 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0.3, - "output": 1.9 + "input": 0.5, + "output": 1.6, + "cache_read": 0.15 } }, - "TEE/qwen3.6-35b-a3b-uncensored": { - "id": "TEE/qwen3.6-35b-a3b-uncensored", - "name": "Qwen3.6 35B A3B Uncensored TEE", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "openai-gpt-5.5": { + "id": "openai-gpt-5.5", + "name": "OpenAI GPT-5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 131072 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-05-23", - "last_updated": "2026-05-23", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", @@ -190946,29 +320222,58 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "TEE/glm-4.7-flash": { - "id": "TEE/glm-4.7-flash", - "name": "GLM 4.7 Flash TEE", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "openai-gpt-5.2-pro": { + "id": "openai-gpt-5.2-pro", + "name": "OpenAI GPT-5.2 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -190976,25 +320281,25 @@ }, "open_weights": false, "limit": { - "context": 203000, - "input": 203000, - "output": 65535 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.5 + "input": 21, + "output": 168 } }, - "TEE/gemma4-31b": { - "id": "TEE/gemma4-31b", - "name": "Gemma 4 31B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "all-mini-lm-l6-v2": { + "id": "all-mini-lm-l6-v2", + "name": "All-MiniLM-L6-v2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, - "release_date": "2026-04-04", - "last_updated": "2026-04-04", + "temperature": false, + "release_date": "2021-08-30", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -191003,28 +320308,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 131072 + "context": 256, + "output": 384 }, "cost": { - "input": 0.45, - "output": 1 + "input": 0.009, + "output": 0 } }, - "TEE/gemma-3-27b-it": { - "id": "TEE/gemma-3-27b-it", - "name": "Gemma 3 27B TEE", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "multi-qa-mpnet-base-dot-v1": { + "id": "multi-qa-mpnet-base-dot-v1", + "name": "Multi-QA-mpnet-base-dot-v1", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-03-10", - "last_updated": "2025-03-10", + "temperature": false, + "release_date": "2021-08-30", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -191033,61 +320337,63 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 8192 + "context": 512, + "output": 768 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0.009, + "output": 0 } }, - "TEE/gpt-oss-20b": { - "id": "TEE/gpt-oss-20b", - "name": "GPT-OSS 20B TEE", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "openai-gpt-image-1": { + "id": "openai-gpt-image-1", + "name": "GPT Image 1", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "release_date": "2025-04-24", + "last_updated": "2025-04-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 8192 + "context": 0, + "output": 0 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 5, + "output": 40, + "cache_read": 1.25 } }, - "TEE/glm-5": { - "id": "TEE/glm-5", - "name": "GLM 5 TEE", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "anthropic-claude-3-opus": { + "id": "anthropic-claude-3-opus", + "name": "Claude 3 Opus", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-opus", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "tool_call": true, + "temperature": true, + "knowledge": "2023-08", + "release_date": "2024-02-29", + "last_updated": "2024-02-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -191095,29 +320401,44 @@ }, "open_weights": false, "limit": { - "context": 203000, - "input": 203000, - "output": 65535 + "context": 200000, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 1.2, - "output": 3.5 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "TEE/glm-5.1-thinking": { - "id": "TEE/glm-5.1-thinking", - "name": "GLM 5.1 Thinking TEE", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": false, + "anthropic-claude-4.6-sonnet": { + "id": "anthropic-claude-4.6-sonnet", + "name": "Anthropic Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -191125,64 +320446,60 @@ }, "open_weights": false, "limit": { - "context": 202752, - "input": 202752, - "output": 65535 + "context": 200000, + "output": 8192 }, "cost": { - "input": 1.5, - "output": 5.25, - "cache_read": 0.3 - } - }, - "TEE/deepseek-v3.2": { - "id": "TEE/deepseek-v3.2", - "name": "DeepSeek V3.2 TEE", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text" + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 164000, - "input": 164000, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 1 + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "TEE/qwen3.5-122b-a10b": { - "id": "TEE/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B A10B TEE", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "anthropic-claude-haiku-4.5": { + "id": "anthropic-claude-haiku-4.5", + "name": "Anthropic Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": false, - "release_date": "2026-05-26", - "last_updated": "2026-05-26", + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -191190,59 +320507,88 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 262144 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.46, - "output": 3.68 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "Sao10K/L3.1-70B-Hanami-x1": { - "id": "Sao10K/L3.1-70B-Hanami-x1", - "name": "Llama 3.1 70B Hanami", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "glm-5.3-flash": { + "id": "glm-5.3-flash", + "name": "GLM5.3 Flash", + "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", + "family": "glm-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 0.15, + "output": 0.5, + "cache_read": 0.03 } }, - "Sao10K/L3-8B-Stheno-v3.2": { - "id": "Sao10K/L3-8B-Stheno-v3.2", - "name": "Sao10K Stheno 8b", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-11-29", - "last_updated": "2024-11-29", + "openai-o1": { + "id": "openai-o1", + "name": "OpenAI o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -191250,26 +320596,38 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.2006, - "output": 0.2006 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "Sao10K/L3.3-70B-Euryale-v2.3": { - "id": "Sao10K/L3.3-70B-Euryale-v2.3", - "name": "Llama 3.3 70B Euryale", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "deepseek-v4-flash-0731": { + "id": "deepseek-v4-flash-0731", + "name": "DeepSeek V4 Flash 0731", + "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-07-31", + "last_updated": "2026-07-31", "modalities": { "input": [ "text" @@ -191278,31 +320636,44 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 20480, - "input": 20480, - "output": 16384 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "Sao10K/L3.1-70B-Euryale-v2.2": { - "id": "Sao10K/L3.1-70B-Euryale-v2.2", - "name": "Llama 3.1 70B Euryale", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "anthropic-claude-opus-4.7": { + "id": "anthropic-claude-opus-4.7", + "name": "Anthropic Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -191310,59 +320681,74 @@ }, "open_weights": false, "limit": { - "context": 20480, - "input": 20480, - "output": 16384 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.306, - "output": 0.357 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "Steelskull/L3.3-MS-Evalebis-70b": { - "id": "Steelskull/L3.3-MS-Evalebis-70b", - "name": "MS Evalebis 70b", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "stable-diffusion-3.5-large": { + "id": "stable-diffusion-3.5-large", + "name": "Stable Diffusion 3.5 Large", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "stable-diffusion", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "temperature": false, + "release_date": "2024-10-22", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 256, + "output": 1 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 0.08, + "output": 0 } }, - "Steelskull/L3.3-MS-Nevoria-70b": { - "id": "Steelskull/L3.3-MS-Nevoria-70b", - "name": "Steelskull Nevoria 70b", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "openai-gpt-6-luna": { + "id": "openai-gpt-6-luna", + "name": "OpenAI GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -191370,29 +320756,60 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02 + } } }, - "Steelskull/L3.3-Cu-Mai-R1-70b": { - "id": "Steelskull/L3.3-Cu-Mai-R1-70b", - "name": "Llama 3.3 70B Cu Mai", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8-Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -191400,59 +320817,88 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 1000000, + "output": 262144 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "Steelskull/L3.3-MS-Evayale-70B": { - "id": "Steelskull/L3.3-MS-Evayale-70B", - "name": "Evayale 70b ", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "Steelskull/L3.3-Nevoria-R1-70b": { - "id": "Steelskull/L3.3-Nevoria-R1-70b", - "name": "Steelskull Nevoria R1 70b", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "anthropic-claude-opus-5.5": { + "id": "anthropic-claude-opus-5.5", + "name": "Anthropic Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -191460,103 +320906,135 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.49299999999999994, - "output": 0.49299999999999994 + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.4, + "cache_write": 10, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.4, + "cache_write": 10 + } } }, - "Steelskull/L3.3-Electra-R1-70b": { - "id": "Steelskull/L3.3-Electra-R1-70b", - "name": "Steelskull Electra R1 70b", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "llama-4-maverick": { + "id": "llama-4-maverick", + "name": "Llama 4 Maverick", + "description": "Open multimodal Llama model for strong reasoning and fast responses", "family": "llama", - "attachment": false, + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, + "context": 128000, "output": 16384 }, "cost": { - "input": 0.69989, - "output": 0.69989 + "input": 0.25, + "output": 0.87 } - } - } - }, - "moark": { - "id": "moark", - "env": [ - "MOARK_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://moark.com/v1", - "name": "Moark", - "doc": "https://moark.com/docs/openapi/v1#tag/%E6%96%87%E6%9C%AC%E7%94%9F%E6%88%90", - "models": { - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + }, + "openai-o3-mini": { + "id": "openai-o3-mini", + "name": "OpenAI o3 mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "output": 100000 }, "cost": { - "input": 2.1, - "output": 8.4 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "GLM-4.7": { - "id": "GLM-4.7", - "name": "GLM-4.7", + "glm-5": { + "id": "glm-5", + "name": "GLM 5", "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-02-11", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -191567,47 +321045,39 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 64000, + "output": 64000 }, "cost": { - "input": 3.5, - "output": 14 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } - } - } - }, - "lilac": { - "id": "lilac", - "env": [ - "LILAC_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.getlilac.com/v1", - "name": "Lilac", - "doc": "https://docs.getlilac.com/inference/models", - "models": { - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + }, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ "text", @@ -191619,37 +321089,41 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.7, - "output": 3.5, - "cache_read": 0.2 + "input": 0.3, + "output": 1.2, + "cache_read": 0.006 } }, - "minimaxai/minimax-m3": { - "id": "minimaxai/minimax-m3", - "name": "MiniMax M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax-m3", + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax M2.5 (Public Preview)", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax-m2.5", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "knowledge": "2025-08", + "release_date": "2026-02-12", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -191657,164 +321131,231 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0.28, - "output": 1.1, - "cache_read": 0.05 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "anthropic-claude-4.1-opus": { + "id": "anthropic-claude-4.1-opus", + "name": "Anthropic Claude 4.1 Opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "wan2-2-t2v-a14b": { + "id": "wan2-2-t2v-a14b", + "name": "Wan2.2-T2V-A14B", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-07-28", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "video" + ] + }, + "open_weights": true, + "limit": { + "context": 100, + "output": 1 }, + "cost": { + "input": 0.6, + "output": 0 + } + }, + "openai-gpt-4o": { + "id": "openai-gpt-4o", + "name": "OpenAI GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262100, - "output": 262100 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.11, - "output": 0.35 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "zai-org/glm-5.2": { - "id": "zai-org/glm-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "anthropic-claude-4.5-haiku": { + "id": "anthropic-claude-4.5-haiku", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "max" + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 524288 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.9, - "output": 3, - "cache_read": 0.27 + "input": 1, + "output": 5, + "cache_read": 1, + "cache_write": 1.25 } - } - } - }, - "ambient": { - "id": "ambient", - "env": [ - "AMBIENT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.ambient.xyz/v1", - "name": "Ambient", - "doc": "https://ambient.xyz", - "models": { - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + }, + "anthropic-claude-fable-5.1": { + "id": "anthropic-claude-fable-5.1", + "name": "Anthropic Claude Fable 5.1", + "description": "Claude model for demanding reasoning and long-horizon agentic work", + "family": "claude-fable", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-01", + "last_updated": "2026-09-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.75, - "output": 3.5, - "cache_read": 0.16, - "cache_write": 0 + "input": 10, + "output": 50, + "cache_read": 0.25, + "cache_write": 12.5 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", + "kimi-k2.6": { + "id": "kimi-k2.6", "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-01", "release_date": "2026-04-21", "last_updated": "2026-04-21", @@ -191835,26 +321376,21 @@ "cost": { "input": 0.95, "output": 4, - "cache_read": 0.2, - "cache_write": 0 + "cache_read": 0.19 } }, - "zai-org/GLM-5.1-FP8": { - "id": "zai-org/GLM-5.1-FP8", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "deepseek-v3": { + "id": "deepseek-v3", + "name": "DeepSeek V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2024-07", + "release_date": "2024-12-26", + "last_updated": "2025-03-24", "modalities": { "input": [ "text" @@ -191865,103 +321401,80 @@ }, "open_weights": true, "limit": { - "context": 202752, + "context": 163840, "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0, - "cache_write": 0 } - } - } - }, - "neon": { - "id": "neon", - "env": [ - "NEON_AI_GATEWAY_BASE_URL", - "NEON_AI_GATEWAY_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/mlflow/v1", - "name": "Neon", - "doc": "https://neon.com/docs", - "models": { - "gemini-3-flash": { - "id": "gemini-3-flash", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, + }, + "deepseek-r1-distill-llama-70b": { + "id": "deepseek-r1-distill-llama-70b", + "name": "DeepSeek R1 Distill Llama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2025-01-30", + "last_updated": "2025-01-30", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 32678, + "output": 8192 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 + "input": 0.99, + "output": 0.99 } }, - "claude-sonnet-4": { - "id": "claude-sonnet-4", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "openai-gpt-5-nano": { + "id": "openai-gpt-5-nano", + "name": "OpenAI GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -191969,28 +321482,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "qwen3-embedding-0.6b": { + "id": "qwen3-embedding-0.6b", + "name": "Qwen3 Embedding 0.6B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "tool_call": false, + "temperature": false, + "release_date": "2025-06-03", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -192001,30 +321512,29 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 8000, + "output": 1024 }, + "status": "beta", "cost": { - "input": 0.15, - "output": 1.2 + "input": 0.04, + "output": 0 } }, - "llama-4-maverick": { - "id": "llama-4-maverick", - "name": "Llama 4 Maverick 17B Instruct", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", - "attachment": true, + "mistral-7b-instruct-v0.3": { + "id": "mistral-7b-instruct-v0.3", + "name": "Mistral 7B Instruct v0.3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2024-05-22", + "last_updated": "2024-05-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -192032,41 +321542,63 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 16384 + "context": 32768, + "output": 32768 + } + }, + "qwen-2.5-14b-instruct": { + "id": "qwen-2.5-14b-instruct", + "name": "Qwen 2.5 14B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-09", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "cost": { - "input": 0.5, - "output": 1.5 + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 } }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai-o3": { + "id": "openai-o3", + "name": "OpenAI o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -192075,19 +321607,18 @@ "open_weights": false, "limit": { "context": 200000, - "output": 64000 + "output": 100000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "openai-gpt-5": { + "id": "openai-gpt-5", + "name": "OpenAI GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "family": "gpt", "attachment": true, "reasoning": true, @@ -192114,40 +321645,33 @@ "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" - }, "cost": { "input": 1.25, "output": 10, "cache_read": 0.125 } }, - "gemma-3-12b": { - "id": "gemma-3-12b", - "name": "Gemma 3 12B", - "description": "Google's open-weight Gemma 3 vision-language model for text and image understanding", - "family": "gemma", + "openai-gpt-4.1": { + "id": "openai-gpt-4.1", + "name": "OpenAI GPT-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -192157,19 +321681,20 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.5 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "gpt-5-1-codex-mini": { - "id": "gpt-5-1-codex-mini", - "name": "GPT-5.1 Codex mini", + "openai-gpt-5.1-codex-max": { + "id": "openai-gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "family": "gpt-codex", "attachment": true, @@ -192180,7 +321705,8 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh" ] } ], @@ -192196,32 +321722,92 @@ "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "anthropic-claude-3.5-haiku": { + "id": "anthropic-claude-3.5-haiku", + "name": "Claude 3.5 Haiku", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-haiku", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-11-05", + "last_updated": "2024-11-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 } }, - "gpt-5-1": { - "id": "gpt-5-1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", + "anthropic-claude-3.5-sonnet": { + "id": "anthropic-claude-3.5-sonnet", + "name": "Claude 3.5 Sonnet", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-sonnet", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-06-20", + "last_updated": "2024-10-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "status": "deprecated", + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -192236,73 +321822,83 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.5, + "output": 2.7, + "cache_read": 0.203 } }, - "meta-llama-3-3-70b-instruct": { - "id": "meta-llama-3-3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, + "anthropic-claude-opus-4": { + "id": "anthropic-claude-opus-4", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "gemini-3-pro": { - "id": "gemini-3-pro", - "name": "Gemini 3 Pro Preview", - "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", - "family": "gemini-pro", + "openai-gpt-6-astra": { + "id": "openai-gpt-6-astra", + "name": "OpenAI GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -192311,23 +321907,22 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "temperature": false, + "knowledge": "2026-04-30", + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -192335,88 +321930,100 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, + "input": 10, + "output": 50, + "cache_read": 1, "tiers": [ { - "input": 4, - "output": 18, - "cache_read": 0.4, + "input": 20, + "output": 75, + "cache_read": 2, "tier": { "type": "context", - "size": 200000 + "size": 272000 } } ], "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 + "input": 20, + "output": 75, + "cache_read": 2 } } }, - "gpt-5-2": { - "id": "gpt-5-2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "openai-gpt-image-2": { + "id": "openai-gpt-image-2", + "name": "OpenAI GPT Image 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2025-04-24", + "last_updated": "2025-04-24", "modalities": { "input": [ "text", "image" ], "output": [ + "image", + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 16384 + }, + "cost": { + "input": 8, + "output": 30 + } + }, + "openai-gpt-image-1.5": { + "id": "openai-gpt-image-1.5", + "name": "OpenAI GPT Image 1.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-11-25", + "last_updated": "2025-11-25", + "modalities": { + "input": [ "text", "image" + ], + "output": [ + "image", + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" + "context": 0, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 5, + "output": 10, + "cache_read": 1 } }, - "gemini-3-1-flash-lite": { - "id": "gemini-3-1-flash-lite", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "openai-gpt-4o-mini": { + "id": "openai-gpt-4o-mini", + "name": "OpenAI GPT-4o mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -192432,16 +322039,13 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -192449,92 +322053,60 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "qwen3-coder-flash": { + "id": "qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.45, + "output": 1.7, + "cache_read": 0.09 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "deepseek-4-flash": { + "id": "deepseek-4-flash", + "name": "Deepseek V4 Flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "release_date": "2026-05-27", + "last_updated": "2026-05-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -192542,43 +322114,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "gemini-2-5-pro": { - "id": "gemini-2-5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "nemotron-3-ultra-550b": { + "id": "nemotron-3-ultra-550b", + "name": "Nemotron 3 Ultra", + "description": "Flagship Nemotron model for high-throughput reasoning and complex agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "release_date": "2026-06-04", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -192586,63 +322144,42 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.9, + "output": 1.7 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai-gpt-5.2": { + "id": "openai-gpt-5.2", + "name": "OpenAI GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -192650,22 +322187,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "gpt-5-3-codex": { - "id": "gpt-5-3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "nvidia-nemotron-3-super-120b": { + "id": "nvidia-nemotron-3-super-120b", + "name": "NVIDIA Nemotron 3 Super 120B (Public Preview)", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -192675,50 +322211,40 @@ "low", "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" + "context": 1000000, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.3, + "output": 0.65, + "cache_read": 0.06 } }, - "gpt-5-4-nano": { - "id": "gpt-5-4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, + "deepseek-v4-pro-0813": { + "id": "deepseek-v4-pro-0813", + "name": "DeepSeek V4 Pro 0813", + "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -192726,245 +322252,239 @@ "values": [ "none", "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-08-12", + "last_updated": "2026-08-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 1.32, + "output": 3.96, + "cache_read": 0.044 } }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo V2.5 Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.8, + "output": 3, + "cache_read": 0.16 } }, - "gpt-5-4-mini": { - "id": "gpt-5-4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "mistral-3-14B": { + "id": "mistral-3-14B", + "name": "Ministral 3 14B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-15", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 128000 + }, + "cost": { + "input": 0.2, + "output": 0.2 + } + }, + "anthropic-claude-opus-5": { + "id": "anthropic-claude-opus-5", + "name": "Anthropic Claude Opus 5", + "description": "Strongest Claude Opus model for coding, agents, and professional work", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "knowledge": "2026-05", + "release_date": "2026-07-24", + "last_updated": "2026-07-24", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" - }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-2-5-flash": { - "id": "gemini-2-5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", + "nemotron-3-nano-omni": { + "id": "nemotron-3-nano-omni", + "name": "Nemotron 3 Nano Omni", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ "text", "image", - "audio", - "video", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 65536, "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 + "input": 0.5, + "output": 0.9 } }, - "gpt-5-2-codex": { - "id": "gpt-5-2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", - "attachment": true, + "nemotron-3-nano-30b": { + "id": "nemotron-3-nano-30b", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "context": 262144, + "output": 262144 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "arcee-trinity-large-thinking": { + "id": "arcee-trinity-large-thinking", + "name": "Arcee Trinity Large Thinking (Public Preview)", + "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", + "family": "trinity", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -192978,10 +322498,9 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-01", + "last_updated": "2026-05-28", "modalities": { "input": [ "text" @@ -192992,45 +322511,41 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 32000 }, "cost": { - "input": 0.072, - "output": 0.28 + "input": 0.25, + "output": 0.9, + "cache_read": 0.06 } }, - "gemini-3-5-flash": { - "id": "gemini-3-5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "anthropic-claude-5-sonnet": { + "id": "anthropic-claude-5-sonnet", + "name": "Anthropic Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -193038,46 +322553,42 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "gemini-3-1-pro": { - "id": "gemini-3-1-pro", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "openai-gpt-5.4-pro": { + "id": "openai-gpt-5.4-pro", + "name": "OpenAI GPT-5.4 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -193085,131 +322596,99 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, + "input": 30, + "output": 180, "tiers": [ { - "input": 4, - "output": 18, - "cache_read": 0.4, + "input": 60, + "output": 270, "tier": { "type": "context", - "size": 200000 + "size": 272000 } } ], "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 + "input": 60, + "output": 270 } } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "llama3-8b-instruct": { + "id": "llama3-8b-instruct", + "name": "Llama 3.1 Instruct (8B)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.198, + "output": 0.198 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 - } - ], + "llama3.3-70b-instruct": { + "id": "llama3.3-70b-instruct", + "name": "Llama 3.3 Instruct (70B)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.65, + "output": 0.65 } }, - "qwen35-122b-a10b": { - "id": "qwen35-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai-gpt-oss-120b": { + "id": "openai-gpt-oss-120b", + "name": "OpenAI GPT-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -193217,10 +322696,14 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2024-06", + "release_date": "2025-08-05", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -193231,27 +322714,39 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 8000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.22, - "output": 2.2 + "input": 0.1, + "output": 0.7, + "cache_read": 0.02 } }, - "meta-llama-3-1-8b-instruct": { - "id": "meta-llama-3-1-8b-instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Meta's compact open-weight Llama 3.1 model for fast, low-cost text generation", - "family": "llama", + "alibaba-qwen3-32b": { + "id": "alibaba-qwen3-32b", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2025-04-30", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -193262,77 +322757,71 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.45 + "input": 0.25, + "output": 0.55 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", + "openai-gpt-5.3-codex": { + "id": "openai-gpt-5.3-codex", + "name": "OpenAI GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" - }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen 3.5 397B A17B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen3.5", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", "high" @@ -193341,64 +322830,56 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 0.55, + "output": 3.5, + "cache_read": 0.111 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "anthropic-claude-3.7-sonnet": { + "id": "anthropic-claude-3.7-sonnet", + "name": "Claude 3.7 Sonnet", + "description": "Legacy model retained for compatibility with older integrations", "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2024-11", + "release_date": "2025-02-24", + "last_updated": "2025-02-24", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -193406,9 +322887,10 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 200000, "output": 64000 }, + "status": "deprecated", "cost": { "input": 3, "output": 15, @@ -193416,319 +322898,162 @@ "cache_write": 3.75 } }, - "gpt-5-4": { - "id": "gpt-5-4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "anthropic-claude-opus-4.8": { + "id": "anthropic-claude-opus-4.8", + "name": "Anthropic Claude Opus 4.8", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2026-05-28", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" - }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gpt-5-1-codex-max": { - "id": "gpt-5-1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai-gpt-5.6-luna": { + "id": "openai-gpt-5.6-luna", + "name": "OpenAI GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-luna", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, + "input": 922000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", - "shape": "responses" - }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "tiers": [ + { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.8, + "cache_read": 0.04 + } } }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "qwen3-tts-voicedesign": { + "id": "qwen3-tts-voicedesign", + "name": "Qwen3 TTS VoiceDesign", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.05, - "output": 0.2 - } - } - } - }, - "upstage": { - "id": "upstage", - "env": [ - "UPSTAGE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.upstage.ai/v1/solar", - "name": "Upstage", - "doc": "https://developers.upstage.ai/docs/apis/chat", - "models": { - "solar-pro2": { - "id": "solar-pro2", - "name": "solar-pro2", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "solar-pro", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 8192 - }, - "cost": { - "input": 0.25, - "output": 0.25 + "context": 32768, + "output": 1 } }, - "solar-pro3": { - "id": "solar-pro3", - "name": "solar-pro3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "solar-pro", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.25, - "output": 0.25 - } - }, - "solar-mini": { - "id": "solar-mini", - "name": "solar-mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "solar-mini", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-06-12", - "last_updated": "2025-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 4096 - }, - "cost": { - "input": 0.15, - "output": 0.15 - } - } - } - }, - "zhipuai-coding-plan": { - "id": "zhipuai-coding-plan", - "env": [ - "ZHIPU_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://open.bigmodel.cn/api/coding/paas/v4", - "name": "Zhipu AI Coding Plan", - "doc": "https://docs.bigmodel.cn/cn/coding-plan/overview", - "models": { - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -193737,43 +323062,46 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.21 } }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "openai-gpt-6-sol": { + "id": "openai-gpt-6-sol", + "name": "OpenAI GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" @@ -193781,39 +323109,58 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2, + "output": 10, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4 + } } }, - "glm-5-turbo": { - "id": "glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, + "anthropic-claude-4.5-sonnet": { + "id": "anthropic-claude-4.5-sonnet", + "name": "Anthropic Claude 4.5 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -193822,32 +323169,44 @@ "open_weights": false, "limit": { "context": 200000, - "output": 131072 + "output": 64000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", + "bge-m3": { + "id": "bge-m3", + "name": "BGE M3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "bge", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-01-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" @@ -193858,38 +323217,29 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 8192, + "output": 1024 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.02, + "output": 0 } }, - "glm-4.6v": { - "id": "glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "ministral-3-8b-instruct-2512": { + "id": "ministral-3-8b-instruct-2512", + "name": "Ministral 3 8B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -193897,78 +323247,98 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 0.9 + "context": 262144, + "output": 262144 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "openai-gpt-5.6-sol": { + "id": "openai-gpt-5.6-sol", + "name": "OpenAI GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt-sol", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 4, + "output": 20, + "cache_read": 0.4, + "tiers": [ + { + "input": 8, + "output": 30, + "cache_read": 0.8, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 8, + "output": 30, + "cache_read": 0.8 + } } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -193979,88 +323349,74 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.3, + "output": 4.3, + "cache_read": 0.26 } - } - } - }, - "chutes": { - "id": "chutes", - "env": [ - "CHUTES_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://llm.chutes.ai/v1", - "name": "Chutes", - "doc": "https://llm.chutes.ai/v1/models", - "models": { - "moonshotai/Kimi-K2.6-TEE": { - "id": "moonshotai/Kimi-K2.6-TEE", - "name": "Kimi K2.6 TEE", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + }, + "openai-gpt-5.4-nano": { + "id": "openai-gpt-5.4-nano", + "name": "OpenAI GPT-5.4 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65535 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.66, - "output": 3.5, - "cache_read": 0.33 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "moonshotai/Kimi-K2.5-TEE": { - "id": "moonshotai/Kimi-K2.5-TEE", - "name": "Kimi K2.5 TEE", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-01", - "last_updated": "2026-01", + "bge-reranker-v2-m3": { + "id": "bge-reranker-v2-m3", + "name": "BGE Reranker v2 M3", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "bge", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-03-12", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -194068,32 +323424,42 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65535 + "context": 8192, + "output": 1 }, "cost": { - "input": 0.44, - "output": 2, - "cache_read": 0.22 + "input": 0.01, + "output": 0 } }, - "google/gemma-4-31B-turbo-TEE": { - "id": "google/gemma-4-31B-turbo-TEE", - "name": "gemma 4 31B turbo TEE", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "nemotron-nano-12b-v2-vl": { + "id": "nemotron-nano-12b-v2-vl", + "name": "Nemotron-nano 12b v2-vl", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-10-28", + "last_updated": "2025-10-28", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -194101,61 +323467,83 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.12, - "output": 0.37, - "cache_read": 0.06 + "input": 0.2, + "output": 0.6 } }, - "Qwen/Qwen3-32B-TEE": { - "id": "Qwen/Qwen3-32B-TEE", - "name": "Qwen3 32B TEE", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, + "openai-gpt-5.4-mini": { + "id": "openai-gpt-5.4-mini", + "name": "OpenAI GPT-5.4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 40960 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.104, - "output": 0.416, - "cache_read": 0.052 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "Qwen/Qwen3.6-27B-TEE": { - "id": "Qwen/Qwen3.6-27B-TEE", - "name": "Qwen3.6 27B TEE", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai-gpt-5-mini": { + "id": "openai-gpt-5-mini", + "name": "OpenAI GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -194165,69 +323553,40 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.3, + "input": 0.25, "output": 2, - "cache_read": 0.15 - } - }, - "Qwen/Qwen3-235B-A22B-Thinking-2507-TEE": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507-TEE", - "name": "Qwen3 235B A22B Thinking 2507 TEE", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07", - "last_updated": "2026-06-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.2989, - "output": 1.1957, - "cache_read": 0.14945 + "cache_read": 0.025 } }, - "Qwen/Qwen3.5-397B-A17B-TEE": { - "id": "Qwen/Qwen3.5-397B-A17B-TEE", - "name": "Qwen3.5 397B A17B TEE", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "anthropic-claude-opus-4.6": { + "id": "anthropic-claude-opus-4.6", + "name": "Anthropic Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -194237,29 +323596,47 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.45, - "output": 3, - "cache_read": 0.225 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "unsloth/Mistral-Nemo-Instruct-2407-TEE": { - "id": "unsloth/Mistral-Nemo-Instruct-2407-TEE", - "name": "Mistral Nemo Instruct 2407 TEE", - "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", - "family": "mistral-nemo", + "e5-large-v2": { + "id": "e5-large-v2", + "name": "E5 Large v2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "temperature": false, + "release_date": "2023-05-19", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" @@ -194270,66 +323647,85 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 512, + "output": 1024 }, "cost": { - "input": 0.0245, - "output": 0.0978, - "cache_read": 0.01225 + "input": 0.02, + "output": 0 } }, - "zai-org/GLM-5-TEE": { - "id": "zai-org/GLM-5-TEE", - "name": "GLM 5 TEE", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "anthropic-claude-opus-4.5": { + "id": "anthropic-claude-opus-4.5", + "name": "Anthropic Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 65535 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 2.55, - "cache_read": 0.475 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "zai-org/GLM-5.1-TEE": { - "id": "zai-org/GLM-5.1-TEE", - "name": "GLM 5.1 TEE", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "Deepseek V4 Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -194340,34 +323736,32 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 65535 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0.98, - "output": 3.08, - "cache_read": 0.49 + "input": 1.74, + "output": 3.48, + "cache_read": 0.348 } }, - "zai-org/GLM-5.2-TEE": { - "id": "zai-org/GLM-5.2-TEE", - "name": "GLM 5.2 TEE", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "gemma-4-31B-it": { + "id": "gemma-4-31B-it", + "name": "Gemma 4", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-04-22", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -194375,66 +323769,100 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 65535 + "context": 256000, + "output": 8192 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.7 + "input": 0.18, + "output": 0.5, + "cache_read": 0.036 } }, - "deepseek-ai/DeepSeek-V3.2-TEE": { - "id": "deepseek-ai/DeepSeek-V3.2-TEE", - "name": "DeepSeek V3.2 TEE", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "openai-gpt-5.6-terra": { + "id": "openai-gpt-5.6-terra", + "name": "OpenAI GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-terra", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2025-12", - "last_updated": "2026-06-21", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1, - "output": 1, - "cache_read": 0.5 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "MiniMaxAI/MiniMax-M2.5-TEE": { - "id": "MiniMaxAI/MiniMax-M2.5-TEE", - "name": "MiniMax M2.5 TEE", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -194445,39 +323873,26 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 65536 + "context": 1048576, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 1.2, - "cache_read": 0.075 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } - } - } - }, - "minimax-cn-coding-plan": { - "id": "minimax-cn-coding-plan", - "env": [ - "MINIMAX_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://api.minimaxi.com/anthropic/v1", - "name": "MiniMax Token Plan (minimaxi.com)", - "doc": "https://platform.minimaxi.com/docs/token-plan/intro", - "models": { - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + }, + "gte-large-en-v1.5": { + "id": "gte-large-en-v1.5", + "name": "GTE Large (v1.5)", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-03-27", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -194488,58 +323903,81 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 8192, + "output": 1024 }, "cost": { - "input": 0, + "input": 0.09, "output": 0 } }, - "MiniMax-M2.5-highspeed": { - "id": "MiniMax-M2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, + "anthropic-claude-fable-5": { + "id": "anthropic-claude-fable-5", + "name": "Anthropic Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "temperature": false, + "release_date": "2026-06-09", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "MiniMax-M2.7-highspeed": { - "id": "MiniMax-M2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "openai-gpt-oss-20b": { + "id": "openai-gpt-oss-20b", + "name": "OpenAI GPT-oss-20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-06", + "release_date": "2025-08-05", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -194550,58 +323988,86 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.05, + "output": 0.45 } }, - "MiniMax-M2": { - "id": "MiniMax-M2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "openai-gpt-5.4": { + "id": "openai-gpt-5.4", + "name": "OpenAI GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, + "context": 400000, "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "mistral-nemo-instruct-2407": { + "id": "mistral-nemo-instruct-2407", + "name": "Mistral Nemo Instruct", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mistral", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text" @@ -194612,275 +324078,239 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.3, + "output": 0.3 } }, - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + "anthropic-claude-sonnet-4": { + "id": "anthropic-claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-25", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 64000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.3, + "cache_write": 3.75, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.3, + "cache_write": 3.75 + } } }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "fal-ai/fast-sdxl": { + "id": "fal-ai/fast-sdxl", + "name": "Fast SDXL", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "stable-diffusion", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2023-07-26", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "context": 0, + "output": 0 } - } - } - }, - "deepseek": { - "id": "deepseek", - "env": [ - "DEEPSEEK_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.deepseek.com", - "name": "DeepSeek", - "doc": "https://api-docs.deepseek.com/quick_start/pricing", - "models": { - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + }, + "fal-ai/elevenlabs/tts/multilingual-v2": { + "id": "fal-ai/elevenlabs/tts/multilingual-v2", + "name": "ElevenLabs Multilingual TTS v2", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "elevenlabs", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2023-08-22", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "context": 0, + "output": 0 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "fal-ai/flux/schnell": { + "id": "fal-ai/flux/schnell", + "name": "FLUX.1 [schnell]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-08-01", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "context": 0, + "output": 0 } }, - "deepseek-reasoner": { - "id": "deepseek-reasoner", - "name": "DeepSeek Reasoner", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "fal-ai/stable-audio-25/text-to-audio": { + "id": "fal-ai/stable-audio-25/text-to-audio", + "name": "Stable Audio 2.5 (Text-to-Audio)", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-10-08", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "context": 0, + "output": 0 } - }, - "deepseek-chat": { - "id": "deepseek-chat", - "name": "DeepSeek Chat", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + } + } + }, + "aixy": { + "id": "aixy", + "env": [ + "AIXY_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.aixy-gateway.com/v1", + "name": "Aixy", + "doc": "https://docs.aixy-gateway.com/integrations/overview", + "models": { + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } } } }, - "wafer.ai": { - "id": "wafer.ai", + "alibaba-cn": { + "id": "alibaba-cn", "env": [ - "WAFER_API_KEY" + "DASHSCOPE_API_KEY" ], "npm": "@ai-sdk/openai-compatible", - "api": "https://pass.wafer.ai/v1", - "name": "Wafer", - "doc": "https://docs.wafer.ai/wafer-pass", + "api": "https://dashscope.aliyuncs.com/compatible-mode/v1", + "name": "Alibaba (China)", + "doc": "https://www.alibabacloud.com/help/en/model-studio/models", "models": { - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", + "qwen-flash": { + "id": "qwen-flash", + "name": "Qwen Flash", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -194888,22 +324318,15 @@ "type": "toggle" }, { - "type": "effort", - "values": [ - "high", - "max" - ] + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-05-30", + "knowledge": "2024-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -194912,21 +324335,19 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 32768 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.01, - "cache_write": 0 + "input": 0.022, + "output": 0.216 } }, - "Qwen3.6-35B-A3B": { - "id": "Qwen3.6-35B-A3B", - "name": "Qwen3.6-35B-A3B", + "qwen3.5-flash": { + "id": "qwen3.5-flash", + "name": "Qwen3.5 Flash", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, @@ -194934,14 +324355,21 @@ "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2026-05-11", - "last_updated": "2026-05-30", + "release_date": "2026-02-23", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", @@ -194952,36 +324380,49 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 229376, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.15, - "output": 1, - "cache_read": 0.02, - "cache_write": 0 + "input": 0.029, + "output": 0.287, + "reasoning": 0.287, + "tiers": [ + { + "input": 0.115, + "output": 1.147, + "reasoning": 1.147, + "tier": { + "type": "context", + "size": 128000 + } + }, + { + "input": 0.172, + "output": 1.72, + "reasoning": 1.72, + "tier": { + "type": "context", + "size": 256000 + } + } + ] } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7-Max", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "qwen3.7-max", + "qwen2-5-coder-32b-instruct": { + "id": "qwen2-5-coder-32b-instruct", + "name": "Qwen2.5-Coder 32B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-30", + "knowledge": "2024-04", + "release_date": "2024-11", + "last_updated": "2024-11", "modalities": { "input": [ "text" @@ -194990,46 +324431,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 131072, + "output": 8192 }, "cost": { - "input": 5, - "output": 15, - "cache_read": 0.5, - "cache_write": 0 + "input": 0.287, + "output": 0.861 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", + "deepseek-v3-2-exp": { + "id": "deepseek-v3-2-exp", + "name": "DeepSeek V3.2 Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-05-30", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -195038,129 +324460,88 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.02, - "cache_write": 0 + "input": 0.287, + "output": 0.431 } }, - "Qwen3.5-397B-A17B": { - "id": "Qwen3.5-397B-A17B", - "name": "Qwen3.5-397B-A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "deepseek-r1-distill-qwen-14b": { + "id": "deepseek-r1-distill-qwen-14b", + "name": "DeepSeek R1 Distill Qwen 14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-06-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 32768, + "output": 16384 }, "cost": { - "input": 0.43, - "output": 2.6, - "cache_read": 0.04, - "cache_write": 0 + "input": 0.144, + "output": 0.431 } }, - "Kimi-K2.6": { - "id": "Kimi-K2.6", - "name": "Kimi-K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen-math-plus": { + "id": "qwen-math-plus", + "name": "Qwen Math Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-13", - "last_updated": "2026-06-01", + "knowledge": "2024-04", + "release_date": "2024-08-16", + "last_updated": "2024-09-19", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 4096, + "output": 3072 }, "cost": { - "input": 0.68, - "output": 3.15, - "cache_read": 0.07, - "cache_write": 0 + "input": 0.574, + "output": 1.721 } }, - "GLM-5.2": { - "id": "GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "qwen-deep-research": { + "id": "qwen-deep-research", + "name": "Qwen Deep Research", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-22", + "knowledge": "2024-04", + "release_date": "2024-01", + "last_updated": "2024-01", "modalities": { "input": [ "text" @@ -195169,36 +324550,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 1.2, - "output": 4.1, - "cache_read": 0.2, - "cache_write": 0 + "input": 7.742, + "output": 23.367 } }, - "GLM-5.1": { - "id": "GLM-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-07", - "last_updated": "2026-06-01", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -195207,42 +324588,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.1, - "cache_write": 0 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } - } - } - }, - "minimax": { - "id": "minimax", - "env": [ - "MINIMAX_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://api.minimax.io/anthropic/v1", - "name": "MiniMax (minimax.io)", - "doc": "https://platform.minimax.io/docs/guides/quickstart", - "models": { - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", + }, + "qwen2-5-32b-instruct": { + "id": "qwen2-5-32b-instruct", + "name": "Qwen2.5 32B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -195253,26 +324622,27 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.287, + "output": 0.861 } }, - "MiniMax-M2.5-highspeed": { - "id": "MiniMax-M2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "qwq-plus": { + "id": "qwq-plus", + "name": "QwQ Plus", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "knowledge": "2024-04", + "release_date": "2025-03-05", + "last_updated": "2025-03-05", "modalities": { "input": [ "text" @@ -195281,30 +324651,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.23, + "output": 0.574 } }, - "MiniMax-M2.7-highspeed": { - "id": "MiniMax-M2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", + "qwq-32b": { + "id": "qwq-32b", + "name": "QwQ 32B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-04", + "release_date": "2024-12", + "last_updated": "2024-12", "modalities": { "input": [ "text" @@ -195315,31 +324684,30 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.287, + "output": 0.861 } }, - "MiniMax-M2": { - "id": "MiniMax-M2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", + "qwen2-5-vl-72b-instruct": { + "id": "qwen2-5-vl-72b-instruct", + "name": "Qwen2.5-VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -195347,26 +324715,25 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 2.294, + "output": 6.881 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "deepseek-v3-1": { + "id": "deepseek-v3-1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -195375,83 +324742,69 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.574, + "output": 1.721 } }, - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, + "qwen3-vl-plus": { + "id": "qwen3-vl-plus", + "name": "Qwen3-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-25", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "tiers": [ - { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12, - "tier": { - "type": "context", - "size": 512000 - } - } - ], - "context_over_200k": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 - } + "input": 0.143353, + "output": 1.433525, + "reasoning": 4.300576 } }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "qwen-plus-character": { + "id": "qwen-plus-character", + "name": "Qwen Plus Character", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-04", + "release_date": "2024-01", + "last_updated": "2024-01", "modalities": { "input": [ "text" @@ -195460,90 +324813,31 @@ "text" ] }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 - } - } - } - }, - "github-copilot": { - "id": "github-copilot", - "env": [ - "GITHUB_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.githubcopilot.com", - "name": "GitHub Copilot", - "doc": "https://docs.github.com/en/copilot", - "models": { - "claude-sonnet-4.5": { - "id": "claude-sonnet-4.5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 32000 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, "open_weights": false, "limit": { - "context": 200000, - "input": 168000, - "output": 32000 + "context": 32768, + "output": 4096 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.115, + "output": 0.287 } }, - "claude-sonnet-4": { - "id": "claude-sonnet-4", - "name": "Claude Sonnet 4 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen-max": { + "id": "qwen-max", + "name": "Qwen Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "knowledge": "2024-04", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -195551,100 +324845,87 @@ }, "open_weights": false, "limit": { - "context": 216000, - "input": 128000, - "output": 16000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.345, + "output": 1.377 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "qwen3-next-80b-a3b-thinking": { + "id": "qwen3-next-80b-a3b-thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 64000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.144, + "output": 1.434 } }, - "claude-haiku-4.5": { - "id": "claude-haiku-4.5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "qwen3.8-max": { + "id": "qwen3.8-max", + "name": "Qwen3.8 Max", + "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "xhigh" + ] + }, { "type": "budget_tokens", - "min": 1024, - "max": 32000 + "min": 0, + "max": 262144 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-08-03", + "last_updated": "2026-08-03", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ @@ -195653,89 +324934,76 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 136000, - "output": 64000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 1.77744, + "output": 5.33231, + "cache_read": 0.22218, + "cache_write": 2.22179 } }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "kimi-k3": { + "id": "kimi-k3", + "name": "Kimi K3", + "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", + "family": "kimi-k3", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", - "medium", - "high" + "high", + "max" ] - }, - { - "type": "budget_tokens", - "min": 256, - "max": 24000 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "temperature": false, + "release_date": "2026-07-16", + "last_updated": "2026-07-16", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 128000, - "output": 64000 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 + "input": 2.827, + "output": 14.133, + "cache_read": 0.283 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen2-5-math-7b-instruct": { + "id": "qwen2-5-math-7b-instruct", + "name": "Qwen2.5-Math 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -195743,45 +325011,44 @@ }, "open_weights": true, "limit": { - "context": 256000, - "input": 224000, - "output": 32000 + "context": 4096, + "output": 3072 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.144, + "output": 0.287 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -195790,79 +325057,90 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 65536 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.115, + "output": 0.688, + "reasoning": 0.688, + "tiers": [ + { + "input": 0.287, + "output": 1.72, + "reasoning": 1.72, + "tier": { + "type": "context", + "size": 128000 + } + }, + { + "input": 0.573, + "output": 3.44, + "reasoning": 3.44, + "tier": { + "type": "context", + "size": 256000 + } + } + ] } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen2-5-72b-instruct": { + "id": "qwen2-5-72b-instruct", + "name": "Qwen2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.574, + "output": 1.721 } }, - "claude-opus-4.7": { - "id": "claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 32768 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -195870,98 +325148,88 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 168000, - "output": 32000 + "context": 202752, + "output": 16384 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } + "cost": { + "input": 0.573, + "output": 2.58, + "tiers": [ + { + "input": 0.86, + "output": 3.154, + "tier": { + "type": "context", + "size": 32000 } } - } - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + ] } }, - "mai-code-1-flash-picker": { - "id": "mai-code-1-flash-picker", - "name": "MAI-Code-1-Flash", - "description": "Microsoft coding model built for fast, efficient assistance in everyday developer workflows", - "family": "mai", - "attachment": false, + "deepseek-v4.1-flash": { + "id": "deepseek-v4.1-flash", + "name": "DeepSeek V4.1 Flash", + "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", + "family": "deepseek-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-06-02", - "last_updated": "2026-06-08", + "knowledge": "2025-05", + "release_date": "2026-09-10", + "last_updated": "2026-09-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 128000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.29754, + "output": 1.19015, + "cache_read": 0.01488 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, + "deepseek-r1-distill-llama-8b": { + "id": "deepseek-r1-distill-llama-8b", + "name": "DeepSeek R1 Distill Llama 8B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -195969,91 +325237,108 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 32768, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0, + "output": 0 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 38912 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.287, + "output": 1.147, + "reasoning": 2.868 } }, - "claude-opus-4.8": { - "id": "claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "qwen2-5-coder-7b-instruct": { + "id": "qwen2-5-coder-7b-instruct", + "name": "Qwen2.5-Coder 7B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-11", + "last_updated": "2024-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.144, + "output": 0.287 + } + }, + "qwen-plus": { + "id": "qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -196061,66 +325346,43 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 168000, - "output": 64000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "context": 1000000, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.115, + "output": 0.287, + "reasoning": 1.147, + "cache_read": 0.012, + "cache_write": 0.144 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "qwen3.7-flash": { + "id": "qwen3.7-flash", + "name": "Qwen3.7 Flash", + "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "structured_output": true, + "temperature": true, + "release_date": "2026-07-15", + "last_updated": "2026-07-15", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -196129,87 +325391,111 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "input": 991000, + "output": 65536 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.02962, + "output": 0.1185, + "cache_read": 0.002962, + "cache_write": 0.03703, + "tiers": [ + { + "input": 0.08887, + "output": 0.35549, + "cache_read": 0.008887, + "cache_write": 0.11109, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.17774, + "output": 0.71098, + "cache_read": 0.017774, + "cache_write": 0.22218, + "tier": { + "type": "context", + "size": 256000 + } + } + ] } }, - "claude-opus-4.5": { - "id": "claude-opus-4.5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Moonshot Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "budget_tokens", - "min": 1024, - "max": 32000 + "max": 81920 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 168000, - "output": 32000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.929, + "output": 3.858 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-23", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -196217,96 +325503,84 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, + "input": 0.359, + "output": 1.434, + "reasoning": 1.434, "tiers": [ { - "input": 5, - "output": 22.5, - "cache_read": 0.5, + "input": 0.574, + "output": 2.294, + "reasoning": 2.294, "tier": { "type": "context", - "size": 272000 + "size": 32000 + } + }, + { + "input": 1.004, + "output": 4.014, + "reasoning": 4.014, + "tier": { + "type": "context", + "size": 128000 } } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + ] } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen-omni-turbo-realtime": { + "id": "qwen-omni-turbo-realtime", + "name": "Qwen-Omni Turbo Realtime", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-05-08", + "last_updated": "2025-05-08", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 32768, + "output": 2048 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.23, + "output": 0.918, + "input_audio": 3.584, + "output_audio": 7.168 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, + "deepseek-v3": { + "id": "deepseek-v3", + "name": "DeepSeek V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -196314,51 +325588,29 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 65536, + "output": 8192 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.287, + "output": 1.147 } }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, + "deepseek-r1-distill-llama-70b": { + "id": "deepseek-r1-distill-llama-70b", + "name": "DeepSeek R1 Distill Llama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 256, - "max": 32000 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -196366,65 +325618,86 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 136000, - "output": 64000 + "context": 32768, + "output": 16384 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, + "input": 0.287, + "output": 0.861 + } + }, + "qwen3-coder-480b-a35b-instruct": { + "id": "qwen3-coder-480b-a35b-instruct", + "name": "Qwen3-Coder 480B-A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.861, + "output": 3.441, "tiers": [ { - "input": 4, - "output": 18, - "cache_read": 0.4, + "input": 1.291, + "output": 5.161, "tier": { "type": "context", - "size": 200000 + "size": 32000 + } + }, + { + "input": 2.151, + "output": 8.602, + "tier": { + "type": "context", + "size": 128000 } } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + ] } }, - "claude-sonnet-4.6": { - "id": "claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, + "qwen-turbo": { + "id": "qwen-turbo", + "name": "Qwen Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] + "type": "toggle" }, { "type": "budget_tokens", - "min": 1024, - "max": 32000 + "max": 38912 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2024-04", + "release_date": "2024-11-01", + "last_updated": "2025-07-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -196432,44 +325705,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 168000, - "output": 32000 + "context": 1000000, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.044, + "output": 0.087, + "reasoning": 0.431 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, + "deepseek-r1-distill-qwen-7b": { + "id": "deepseek-r1-distill-qwen-7b", + "name": "DeepSeek R1 Distill Qwen 7B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -196477,154 +325736,107 @@ }, "open_weights": false, "limit": { - "context": 264000, - "input": 128000, - "output": 64000 + "context": 32768, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.072, + "output": 0.144 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 256, - "max": 32000 - } - ], + "qwen2-5-math-72b-instruct": { + "id": "qwen2-5-math-72b-instruct", + "name": "Qwen2.5-Math 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 64000 + "context": 4096, + "output": 3072 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 + "input": 0.574, + "output": 1.721 } }, - "claude-opus-4.6": { - "id": "claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 168000, - "output": 32000 + "context": 262144, + "output": 65536 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } + "cost": { + "input": 0.216, + "output": 0.861, + "tiers": [ + { + "input": 0.323, + "output": 1.291, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.538, + "output": 2.151, + "tier": { + "type": "context", + "size": 128000 } } - } - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + ] } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen-doc-turbo": { + "id": "qwen-doc-turbo", + "name": "Qwen Doc Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01", + "last_updated": "2024-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -196632,108 +325844,65 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.087, + "output": 0.144 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Moonshot Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.574, + "output": 2.294 } - } - } - }, - "clarifai": { - "id": "clarifai", - "env": [ - "CLARIFAI_PAT" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.clarifai.com/v2/ext/openai/v1", - "name": "Clarifai", - "doc": "https://docs.clarifai.com/compute/inference/", - "models": { - "moonshotai/chat-completion/models/Kimi-K2_6": { - "id": "moonshotai/chat-completion/models/Kimi-K2_6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + }, + "qvq-max": { + "id": "qvq-max", + "name": "QVQ Max", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qvq", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2024-04", + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ "text", @@ -196743,28 +325912,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 4 + "input": 1.147, + "output": 4.588 } }, - "minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput": { - "id": "minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput", - "name": "MiniMax-M2.5 High Throughput", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "qwen2-5-7b-instruct": { + "id": "qwen2-5-7b-instruct", + "name": "Qwen2.5 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-25", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -196775,74 +325944,107 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.072, + "output": 0.144 } }, - "openai/chat-completion/models/gpt-oss-120b-high-throughput": { - "id": "openai/chat-completion/models/gpt-oss-120b-high-throughput", - "name": "GPT OSS 120B High Throughput", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "qwen3-omni-flash": { + "id": "qwen3-omni-flash", + "name": "Qwen3-Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2026-02-25", + "knowledge": "2024-04", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" + ], + "output": [ + "text", + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 65536, + "output": 16384 + }, + "cost": { + "input": 0.058, + "output": 0.23, + "input_audio": 3.584, + "output_audio": 7.168 + } + }, + "qwen-vl-max": { + "id": "qwen-vl-max", + "name": "Qwen-VL Max", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-08", + "last_updated": "2025-08-13", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 131072, - "output": 16384 + "output": 8192 }, "cost": { - "input": 0.09, - "output": 0.36 + "input": 0.23, + "output": 0.574 } }, - "openai/chat-completion/models/gpt-oss-20b": { - "id": "openai/chat-completion/models/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "qwen3-235b-a22b": { + "id": "qwen3-235b-a22b", + "name": "Qwen3 235B-A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 38912 } ], "tool_call": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-12-12", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -196857,23 +326059,24 @@ "output": 16384 }, "cost": { - "input": 0.045, - "output": 0.18 + "input": 0.287, + "output": 1.147, + "reasoning": 2.868 } }, - "mistralai/completion/models/Ministral-3-14B-Reasoning-2512": { - "id": "mistralai/completion/models/Ministral-3-14B-Reasoning-2512", - "name": "Ministral 3 14B Reasoning 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, + "qwen3-vl-30b-a3b": { + "id": "qwen3-vl-30b-a3b", + "name": "Qwen3-VL 30B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-01", - "last_updated": "2025-12-12", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text", @@ -196885,30 +326088,45 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 1.7 + "input": 0.108, + "output": 0.431, + "reasoning": 1.076 } }, - "mistralai/completion/models/Ministral-3-3B-Reasoning-2512": { - "id": "mistralai/completion/models/Ministral-3-3B-Reasoning-2512", - "name": "Ministral 3 3B Reasoning 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Moonshot Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "release_date": "2025-12", - "last_updated": "2026-02-25", + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -196917,28 +326135,28 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "output": 32768 }, "cost": { - "input": 1.039, - "output": 0.54825 + "input": 0.574, + "output": 2.411 } }, - "deepseek-ai/deepseek-ocr/models/DeepSeek-OCR": { - "id": "deepseek-ai/deepseek-ocr/models/DeepSeek-OCR", - "name": "DeepSeek OCR", - "description": "OCR model for extracting structured text from documents and screenshots", - "family": "deepseek", - "attachment": true, + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-10-20", - "last_updated": "2026-02-25", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -196946,57 +326164,57 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 0.7 + "input": 0.144, + "output": 0.574 } }, - "qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507": { - "id": "qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507", - "name": "Qwen3 30B A3B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "qwen-vl-plus": { + "id": "qwen-vl-plus", + "name": "Qwen-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-07-31", - "last_updated": "2026-02-25", + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-08-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.36, - "output": 1.3 + "input": 0.115, + "output": 0.287 } }, - "qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507": { - "id": "qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen3 30B A3B Instruct 2507", + "deepseek-r1-distill-qwen-1-5b": { + "id": "deepseek-r1-distill-qwen-1-5b", + "name": "DeepSeek R1 Distill Qwen 1.5B", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-07-30", - "last_updated": "2026-02-25", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -197005,59 +326223,70 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 32768, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 0.5 + "input": 0, + "output": 0 } }, - "qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct": { - "id": "qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct", - "name": "Qwen3 Coder 30B A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 131072 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-31", - "last_updated": "2026-02-12", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.11458, - "output": 0.74812 + "input": 0.1875, + "output": 1.125, + "cache_write": 0.234375 } }, - "arcee_ai/AFM/models/trinity-mini": { - "id": "arcee_ai/AFM/models/trinity-mini", - "name": "Trinity Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "trinity-mini", + "qwen3-coder-flash": { + "id": "qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12", - "last_updated": "2026-02-25", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -197066,83 +326295,63 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.045, - "output": 0.15 + "input": 0.144, + "output": 0.574 } }, - "clarifai/main/models/mm-poly-8b": { - "id": "clarifai/main/models/mm-poly-8b", - "name": "MM Poly 8B", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "mm-poly", - "attachment": true, + "qwen2-5-omni-7b": { + "id": "qwen2-5-omni-7b", + "name": "Qwen2.5-Omni 7B", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-06", - "last_updated": "2026-02-25", + "knowledge": "2024-04", + "release_date": "2024-12", + "last_updated": "2024-12", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 32768, - "output": 4096 + "output": 2048 }, "cost": { - "input": 0.658, - "output": 1.11 + "input": 0.087, + "output": 0.345, + "input_audio": 5.448 } - } - } - }, - "the-grid-ai": { - "id": "the-grid-ai", - "env": [ - "THEGRIDAI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.thegrid.ai/v1", - "name": "The Grid AI", - "doc": "https://thegrid.ai/docs", - "models": { - "agent-prime": { - "id": "agent-prime", - "name": "Agent Prime", - "description": "Reliable models for dependable agentic applications, multi-step tool use, and reasoning workflows. Any model that meets the contract spec can serve your request.", + }, + "qwen-mt-plus": { + "id": "qwen-mt-plus", + "name": "Qwen-MT Plus", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-05-04", - "last_updated": "2026-07-06", + "knowledge": "2024-04", + "release_date": "2025-01", + "last_updated": "2025-01", "modalities": { "input": [ "text" @@ -197153,45 +326362,29 @@ }, "open_weights": false, "limit": { - "context": 196608, - "input": 120000, - "output": 30000 + "context": 16384, + "output": 8192 }, - "status": "beta" + "cost": { + "input": 0.259, + "output": 0.775 + } }, - "agent-max": { - "id": "agent-max", - "name": "Agent Max", - "description": "Frontier models for autonomous research, deep multi-step tool chains, and complex long-horizon tasks. Any model that meets the contract spec can serve your request.", + "tongyi-intent-detect-v3": { + "id": "tongyi-intent-detect-v3", + "name": "Tongyi Intent Detect V3", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "yi", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-05-04", - "last_updated": "2026-07-06", + "knowledge": "2024-04", + "release_date": "2024-01", + "last_updated": "2024-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -197199,75 +326392,59 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 + "context": 8192, + "output": 1024 }, - "status": "beta" + "cost": { + "input": 0.058, + "output": 0.144 + } }, - "text-standard": { - "id": "text-standard", - "name": "Text Standard", - "description": "Price-optimized models with low-latency, high-throughput and shorter maximum outputs. Any model that meets the contract spec can serve your request.", + "qwen3-vl-235b-a22b": { + "id": "qwen3-vl-235b-a22b", + "name": "Qwen3-VL 235B-A22B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-07-06", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 120000, - "output": 16000 + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.286705, + "output": 1.14682, + "reasoning": 2.867051 } }, - "code-prime": { - "id": "code-prime", - "name": "Code Prime", - "description": "Reliable models for everyday software tasks, code completion, review, and standard debugging. Any model that meets the contract spec can serve your request.", + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-05-04", - "last_updated": "2026-07-06", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2026-09-11", "modalities": { "input": [ "text" @@ -197276,38 +326453,54 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 196608, - "input": 120000, - "output": 30000 + "context": 1048576, + "output": 65536 }, - "status": "beta" + "cost": { + "input": 0.574, + "output": 2.296, + "tiers": [ + { + "input": 0.861, + "output": 3.444, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 1.435, + "output": 5.74, + "tier": { + "type": "context", + "size": 128000 + } + }, + { + "input": 2.87, + "output": 28.7, + "tier": { + "type": "context", + "size": 256000 + } + } + ] + } }, - "text-prime": { - "id": "text-prime", - "name": "Text Prime", - "description": "Reliable models for everyday text generation, editing, and analysis across diverse workflows. Any model that meets the contract spec can serve your request.", + "moonshot-kimi-k2-instruct": { + "id": "moonshot-kimi-k2-instruct", + "name": "Moonshot Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-07-06", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -197316,83 +326509,61 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 196608, - "input": 120000, - "output": 30000 + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.574, + "output": 2.294 } }, - "code-max": { - "id": "code-max", - "name": "Code Max", - "description": "Frontier models for complex research, architectural decisions, debugging, and multi-file development. Any model that meets the contract spec can serve your request.", + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": true, "temperature": true, - "release_date": "2026-05-04", - "last_updated": "2026-07-06", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 + "context": 204800, + "output": 131072 }, - "status": "beta" + "cost": { + "input": 0.3, + "output": 1.2 + } }, - "agent-standard": { - "id": "agent-standard", - "name": "Agent Standard", - "description": "Price-optimized models for fast tool calls, simple agent loops, high-throughput automation, and orchestration. Any model that meets the contract spec can serve your request.", + "qwen2-5-14b-instruct": { + "id": "qwen2-5-14b-instruct", + "name": "Qwen2.5 14B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-05-04", - "last_updated": "2026-07-06", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -197401,19 +326572,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 120000, - "output": 16000 + "context": 131072, + "output": 8192 }, - "status": "beta" + "cost": { + "input": 0.144, + "output": 0.431 + } }, - "text-max": { - "id": "text-max", - "name": "Text Max", - "description": "Frontier models for deep reasoning, long context, and complex workflows. Any model that meets the contract spec can serve your request.", - "attachment": false, + "qwen3.8-flash": { + "id": "qwen3.8-flash", + "name": "Qwen3.8 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -197424,10 +326598,12 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "xhigh" ] + }, + { + "type": "budget_tokens", + "max": 262144 } ], "tool_call": true, @@ -197435,13 +326611,13 @@ "field": "reasoning_content" }, "structured_output": true, - "temperature": true, - "release_date": "2026-03-24", - "last_updated": "2026-07-06", + "release_date": "2026-08-26", + "last_updated": "2026-08-26", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -197450,88 +326626,27 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 922000, - "output": 128000 + "output": 131072 + }, + "cost": { + "input": 0.11875, + "output": 0.40073, + "cache_read": 0.01187, + "cache_write": 0.14844 } }, - "code-standard": { - "id": "code-standard", - "name": "Code Standard", - "description": "Price-optimized models for rapid autocomplete, linting, high-frequency suggestions, and batch edits. Any model that meets the contract spec can serve your request.", + "qwen2-5-vl-7b-instruct": { + "id": "qwen2-5-vl-7b-instruct", + "name": "Qwen2.5-VL 7B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-05-04", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "input": 120000, - "output": 16000 - }, - "status": "beta" - } - } - }, - "synthetic": { - "id": "synthetic", - "env": [ - "SYNTHETIC_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.synthetic.new/openai/v1", - "name": "Synthetic", - "doc": "https://synthetic.new/pricing", - "models": { - "hf:moonshotai/Kimi-K2.7-Code": { - "id": "hf:moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text", @@ -197543,35 +326658,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.95 + "input": 0.287, + "output": 0.717 } }, - "hf:zai-org/GLM-4.7-Flash": { - "id": "hf:zai-org/GLM-4.7-Flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", + "qwen-math-turbo": { + "id": "qwen-math-turbo", + "name": "Qwen Math Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2024-04", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", "modalities": { "input": [ "text" @@ -197580,42 +326686,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 65536 + "context": 4096, + "output": 3072 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.1 + "input": 0.287, + "output": 0.861 } }, - "hf:zai-org/GLM-5.2": { - "id": "hf:zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "xhigh" - ] - } - ], + "qwen-long": { + "id": "qwen-long", + "name": "Qwen Long", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-04", + "release_date": "2025-01-25", + "last_updated": "2025-01-25", "modalities": { "input": [ "text" @@ -197624,83 +326716,68 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 65536 + "context": 10000000, + "output": 8192 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 1.4 + "input": 0.072, + "output": 0.287 } }, - "hf:MiniMaxAI/MiniMax-M3": { - "id": "hf:MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, + "qwen3.6-max-preview": { + "id": "qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 131072 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2026-04-20", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, + "context": 245800, "output": 65536 }, "cost": { - "input": 0.6, - "output": 1.2, - "cache_read": 0.6 + "input": 1.32, + "output": 7.9, + "cache_read": 0.132 } }, - "hf:openai/gpt-oss-120b": { - "id": "hf:openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "qwen-mt-turbo": { + "id": "qwen-mt-turbo", + "name": "Qwen-MT Turbo", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-04", + "release_date": "2025-01", + "last_updated": "2025-01", "modalities": { "input": [ "text" @@ -197709,47 +326786,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 16384, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.1 + "input": 0.101, + "output": 0.28 } }, - "hf:Qwen/Qwen3.6-27B": { - "id": "hf:Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -197761,78 +326833,43 @@ "output": 65536 }, "cost": { - "input": 0.45, - "output": 3.6, - "cache_read": 0.45 + "input": 0.172, + "output": 1.032, + "reasoning": 1.032, + "tiers": [ + { + "input": 0.43, + "output": 2.58, + "reasoning": 2.58, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4": { - "id": "hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4", - "name": "Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "qwen3-8b": { + "id": "qwen3-8b", + "name": "Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 38912 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 1, - "cache_read": 0.3 - } - } - } - }, - "iflowcn": { - "id": "iflowcn", - "env": [ - "IFLOW_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://apis.iflow.cn/v1", - "name": "iFlow", - "doc": "https://platform.iflow.cn/en/docs", - "models": { - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3-Coder-Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -197843,26 +326880,31 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.072, + "output": 0.287, + "reasoning": 0.717 } }, - "deepseek-v3": { - "id": "deepseek-v3", - "name": "DeepSeek-V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-26", - "last_updated": "2024-12-26", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -197873,56 +326915,76 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.1, + "output": 3.851, + "cache_read": 0.275, + "cache_write": 0 } }, - "kimi-k2": { - "id": "kimi-k2", - "name": "Kimi-K2", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "qwen-omni-turbo": { + "id": "qwen-omni-turbo", + "name": "Qwen-Omni Turbo", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "knowledge": "2024-04", + "release_date": "2025-01-19", + "last_updated": "2025-03-26", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 64000 + "context": 32768, + "output": 2048 }, "cost": { - "input": 0, - "output": 0 + "input": 0.058, + "output": 0.23, + "input_audio": 3.584, + "output_audio": 7.168 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3-32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 131072 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "release_date": "2026-04-14", + "last_updated": "2026-04-14", "modalities": { "input": [ "text" @@ -197933,59 +326995,75 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 202752, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.825, + "output": 3.301, + "cache_read": 0.17, + "tiers": [ + { + "input": 1.1, + "output": 3.851, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "qwen3-max-preview": { - "id": "qwen3-max-preview", - "name": "Qwen3-Max-Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "qwen3-omni-flash-realtime": { + "id": "qwen3-omni-flash-realtime", + "name": "Qwen3-Omni Flash Realtime", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2024-04", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 65536, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.23, + "output": 0.918, + "input_audio": 3.584, + "output_audio": 7.168 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3-Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "qwen3-asr-flash": { + "id": "qwen3-asr-flash", + "name": "Qwen3-ASR Flash", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "tool_call": false, + "temperature": false, + "knowledge": "2024-04", + "release_date": "2025-09-08", + "last_updated": "2025-09-08", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" @@ -197993,17 +327071,17 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 53248, + "output": 4096 }, "cost": { - "input": 0, - "output": 0 + "input": 0.032, + "output": 0.032 } }, - "qwen3-235b": { - "id": "qwen3-235b", - "name": "Qwen3-235B-A22B", + "deepseek-r1-distill-qwen-32b": { + "id": "deepseek-r1-distill-qwen-32b", + "name": "DeepSeek R1 Distill Qwen 32B", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, @@ -198011,9 +327089,8 @@ "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -198022,29 +327099,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 32768, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.287, + "output": 0.861 } }, - "glm-4.6": { - "id": "glm-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2025-11-13", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -198055,27 +327131,42 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.574, + "output": 2.294 } }, - "qwen3-235b-a22b-thinking-2507": { - "id": "qwen3-235b-a22b-thinking-2507", - "name": "Qwen3-235B-A22B-Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -198086,17 +327177,18 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek-R1", + "deepseek-r1-0528": { + "id": "deepseek-r1-0528", + "name": "DeepSeek R1 0528", "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", "family": "deepseek-thinking", "attachment": false, @@ -198104,9 +327196,8 @@ "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -198115,32 +327206,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.574, + "output": 2.294 } }, - "qwen3-vl-plus": { - "id": "qwen3-vl-plus", - "name": "Qwen3-VL-Plus", + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": true, - "reasoning": false, + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -198148,26 +327249,55 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "qwen3-235b-a22b-instruct": { - "id": "qwen3-235b-a22b-instruct", - "name": "Qwen3-235B-A22B-Instruct", + "qwen3-14b": { + "id": "qwen3-14b", + "name": "Qwen3 14B", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 38912 + } + ], "tool_call": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -198178,26 +327308,40 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.144, + "output": 0.574, + "reasoning": 1.434 } }, - "kimi-k2-0905": { - "id": "kimi-k2-0905", - "name": "Kimi-K2-0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "glm-5.3": { + "id": "glm-5.3", + "name": "GLM-5.3", + "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "release_date": "2026-08-14", + "last_updated": "2026-08-14", "modalities": { "input": [ "text" @@ -198206,106 +327350,101 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.1, + "output": 3.851, + "cache_read": 0.275, + "cache_write": 0 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek-V3.2-Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 1000000, "output": 64000 }, "cost": { - "input": 0, - "output": 0 - } - } - } - }, - "xiaomi-token-plan-sgp": { - "id": "xiaomi-token-plan-sgp", - "env": [ - "XIAOMI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://token-plan-sgp.xiaomimimo.com/v1", - "name": "Xiaomi Token Plan (Singapore)", - "doc": "https://platform.xiaomimimo.com/#/docs", - "models": { - "mimo-v2.5-tts": { - "id": "mimo-v2.5-tts", - "name": "MiMo-V2.5-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 128000 + } + } ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 } }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -198316,38 +327455,31 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", + "qwen-vl-ocr": { + "id": "qwen-vl-ocr", + "name": "Qwen-VL OCR", + "description": "OCR model for extracting structured text from documents and screenshots", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-04", + "release_date": "2024-10-28", + "last_updated": "2026-09-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -198355,50 +327487,50 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 34096, + "output": 4096 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.043, + "output": 0.072 } }, - "mimo-v2-tts": { - "id": "mimo-v2-tts", - "name": "MiMo-V2-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", + "siliconflow/deepseek-v3-0324": { + "id": "siliconflow/deepseek-v3-0324", + "name": "siliconflow/deepseek-v3-0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": false, - "tool_call": false, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-12-26", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0, - "output": 0 + "input": 0.25, + "output": 1 } }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, + "siliconflow/deepseek-v3.1-terminus": { + "id": "siliconflow/deepseek-v3.1-terminus", + "name": "siliconflow/deepseek-v3.1-terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -198406,167 +327538,166 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2025-09-29", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 163840, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.27, + "output": 1 } }, - "mimo-v2.5-tts-voicedesign": { - "id": "mimo-v2.5-tts-voicedesign", - "name": "MiMo-V2.5-TTS-VoiceDesign", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", + "siliconflow/deepseek-r1-0528": { + "id": "siliconflow/deepseek-r1-0528", + "name": "siliconflow/deepseek-r1-0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-05-28", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 163840, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.5, + "output": 2.18 } }, - "mimo-v2.5-tts-voiceclone": { - "id": "mimo-v2.5-tts-voiceclone", - "name": "MiMo-V2.5-TTS-VoiceClone", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", + "siliconflow/deepseek-v3.2": { + "id": "siliconflow/deepseek-v3.2", + "name": "siliconflow/deepseek-v3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-03", + "last_updated": "2025-12-03", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 163840, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.27, + "output": 0.42 } - } - } - }, - "claudinio": { - "id": "claudinio", - "env": [ - "CLAUDINIO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.claudin.io/v1", - "name": "Claudinio", - "doc": "https://claudin.io", - "models": { - "claudinio": { - "id": "claudinio", - "name": "Claudinio", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + }, + "kimi/kimi-k2.5": { + "id": "kimi/kimi-k2.5", + "name": "kimi/kimi-k2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "knowledge": "2026-05", - "release_date": "2026-05-12", - "last_updated": "2026-06-02", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.5, - "output": 2, - "cache_read": 0.15 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "claudius": { - "id": "claudius", - "name": "Claudius", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "MiniMax/MiniMax-M2.7": { + "id": "MiniMax/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "knowledge": "2026-05", - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 3, - "output": 8, - "cache_read": 0.9 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 } } } diff --git a/packages/agentlayer-core/src/models.ts b/packages/agentlayer-core/src/models.ts index 8a2c82b..8cb40b9 100644 --- a/packages/agentlayer-core/src/models.ts +++ b/packages/agentlayer-core/src/models.ts @@ -37,10 +37,12 @@ export type ModelKey = `${string}/${string}` * Effective Codex context window, per model. * * Codex reserves 5% of the raw backend window for system prompt, tool overhead, and output: - * gpt-5.4/5.5 and GPT-6 Astra use 272,000 raw tokens; gpt-5.6 uses 372,000. + * gpt-5.4/5.5 and GPT-6 Astra/Sol/Luna use 272,000 raw tokens; gpt-5.6 uses 372,000. */ export const CODEX_CONTEXT_WINDOWS = { 'gpt-6-astra': 258_400, + 'gpt-6-sol': 258_400, + 'gpt-6-luna': 258_400, 'gpt-5.6-sol': 353_400, 'gpt-5.6-terra': 353_400, 'gpt-5.6-luna': 353_400, diff --git a/packages/agentlayer-core/test/compaction-engine.test.ts b/packages/agentlayer-core/test/compaction-engine.test.ts index 333672d..338cfc2 100644 --- a/packages/agentlayer-core/test/compaction-engine.test.ts +++ b/packages/agentlayer-core/test/compaction-engine.test.ts @@ -13,6 +13,7 @@ import { serializeConversation, } from '../src/compaction' import { assistantMessage, toolCall, toolResult, userMessage } from '../src/messages' +import { getCodexContextWindow } from '../src/models' describe('compaction policy', () => { test('uses Fold-compatible budget arithmetic', () => { @@ -21,6 +22,13 @@ describe('compaction policy', () => { expect(compactionUsableTokens({ contextWindow: 1 })).toBe(1) }) + test.each(['gpt-6-sol', 'gpt-6-luna'])('%s reserves output and summary headroom', (modelId) => { + const contextWindow = getCodexContextWindow(modelId) + + expect(contextWindow).toBe(258_400) + expect(compactionUsableTokens({ contextWindow })).toBe(210_016) + }) + test('caps history and turn-prefix summary output against reserved and model budgets', () => { expect(resolveCompactionMaxOutputTokens({ reserveTokens: 20_000 })).toBe(16_000) expect(resolveCompactionMaxOutputTokens({ reserveTokens: 20_000, turnPrefix: true })).toBe(10_000) diff --git a/packages/agentlayer-core/test/models.test.ts b/packages/agentlayer-core/test/models.test.ts index a45756e..811d530 100644 --- a/packages/agentlayer-core/test/models.test.ts +++ b/packages/agentlayer-core/test/models.test.ts @@ -35,6 +35,16 @@ describe('ModelProvider.getModelLimits', () => { expect(limits?.output).toBe(128_000) }) + test.each(['gpt-6-sol', 'gpt-6-luna'] as const)( + 'codex/%s uses the effective Codex window while retaining public output metadata', + (modelId) => { + const limits = provider.getModelLimits(`codex/${modelId}`) + + expect(limits?.context).toBe(258_400) + expect(limits?.output).toBe(128_000) + }, + ) + test('openai/gpt-5.5 keeps the public OpenAI API context window', () => { const limits = provider.getModelLimits('openai/gpt-5.5') @@ -47,8 +57,8 @@ describe('ModelProvider.getModelLimits', () => { // reported as `undefined`, so Azure AI Foundry sessions recorded tokens and no // dollars at all — silently, and unrecoverably, since cost is frozen at ingest. expect(provider.getModelPricing(`${CUSTOM_RESPONSES_PROVIDER}/gpt-5.6-sol`)).toMatchObject({ - input: 5, - output: 30, + input: 4, + output: 20, }) // The public Responses API, not the private Codex one, so it keeps the public window. @@ -60,9 +70,9 @@ describe('ModelProvider.getModelLimits', () => { expect(limits?.context).toBe(1_050_000) expect(limits?.output).toBe(128_000) - expect(provider.getModelPricing('openai/gpt-5.6-sol')).toMatchObject({ input: 5, output: 30 }) - expect(provider.getModelPricing('openai/gpt-5.6-terra')).toMatchObject({ input: 2.5, output: 15 }) - expect(provider.getModelPricing('openai/gpt-5.6-luna')).toMatchObject({ input: 1, output: 6 }) + expect(provider.getModelPricing('openai/gpt-5.6-sol')).toMatchObject({ input: 4, output: 20 }) + expect(provider.getModelPricing('openai/gpt-5.6-terra')).toMatchObject({ input: 2, output: 12 }) + expect(provider.getModelPricing('openai/gpt-5.6-luna')).toMatchObject({ input: 0.2, output: 1.2 }) }) test('openai/gpt-6-astra keeps the public OpenAI API context window and pricing', () => { @@ -72,4 +82,24 @@ describe('ModelProvider.getModelLimits', () => { expect(limits?.output).toBe(128_000) expect(provider.getModelPricing('openai/gpt-6-astra')).toMatchObject({ input: 10, output: 50 }) }) + + test.each([ + ['gpt-6-sol', 2, 10], + ['gpt-6-luna', 0.1, 0.5], + ] as const)('%s keeps its public OpenAI limits and pricing', (modelId, input, output) => { + expect(provider.getModelLimits(`openai/${modelId}`)).toMatchObject({ + context: 1_050_000, + input: 922_000, + output: 128_000, + }) + expect(provider.getModelPricing(`openai/${modelId}`)).toMatchObject({ input, output }) + }) + + test('claude-opus-5-5 exposes public Anthropic limits and pricing', () => { + expect(provider.getModelLimits('anthropic/claude-opus-5-5')).toMatchObject({ + context: 1_000_000, + output: 128_000, + }) + expect(provider.getModelPricing('anthropic/claude-opus-5-5')).toMatchObject({ input: 4, output: 20 }) + }) }) diff --git a/packages/agentlayer-provider-openai-codex/test/codex-ws-adapter.test.ts b/packages/agentlayer-provider-openai-codex/test/codex-ws-adapter.test.ts index 1bee7b7..1fee2ff 100644 --- a/packages/agentlayer-provider-openai-codex/test/codex-ws-adapter.test.ts +++ b/packages/agentlayer-provider-openai-codex/test/codex-ws-adapter.test.ts @@ -33,8 +33,10 @@ function makeOptions(overrides?: Partial): LanguageM } } -describe('GPT-5.6 max reasoning', () => { - test.each(['gpt-5.6-sol', 'gpt-5.6-terra', 'gpt-5.6-luna'])('%s accepts max effort', (modelId) => { +describe('model-specific max reasoning', () => { + const maxReasoningModels = ['gpt-5.6-sol', 'gpt-5.6-terra', 'gpt-5.6-luna', 'gpt-6-sol', 'gpt-6-luna'] + + test.each(maxReasoningModels)('%s accepts max effort', (modelId) => { expect(isReasoningEffortForModel(modelId, 'max')).toBe(true) }) @@ -42,45 +44,42 @@ describe('GPT-5.6 max reasoning', () => { expect(isReasoningEffortForModel('gpt-5.4', 'max')).toBe(false) }) - test.each(['gpt-5.6-sol', 'gpt-5.6-terra', 'gpt-5.6-luna'])( - '%s uses the regular Responses request shape', - async (modelId) => { - const request = convertCallOptionsToLLMRequest( - modelId, - makeOptions({ - prompt: [ - { role: 'system', content: 'Use the repository tools.' }, - { role: 'user', content: [{ type: 'text', text: 'Locate the implementation.' }] }, - ], - tools: [ - { - type: 'function', - name: 'search', - description: 'Search the repository', - inputSchema: { type: 'object', properties: { query: { type: 'string' } } }, - }, - ], - toolChoice: { type: 'auto' }, - providerOptions: { - openai: { reasoningEffort: 'max', promptCacheKey: 'cache-key' }, + test.each(maxReasoningModels)('%s uses the regular Responses request shape', async (modelId) => { + const request = convertCallOptionsToLLMRequest( + modelId, + makeOptions({ + prompt: [ + { role: 'system', content: 'Use the repository tools.' }, + { role: 'user', content: [{ type: 'text', text: 'Locate the implementation.' }] }, + ], + tools: [ + { + type: 'function', + name: 'search', + description: 'Search the repository', + inputSchema: { type: 'object', properties: { query: { type: 'string' } } }, }, - }), - makeConfig(), - ) - - const body = await Effect.runPromise(webSocketRoute.body.from(request)) - - expect(body.instructions).toBe('Use the repository tools.') - expect(body.tools).toHaveLength(1) - expect(body.tool_choice).toBe('auto') - expect(body.prompt_cache_key).toBe('cache-key') - expect(body.reasoning).toEqual({ effort: 'max', summary: 'detailed' }) - expect(body.parallel_tool_calls).toBeUndefined() - expect(body.input).toEqual([ - { role: 'user', content: [{ type: 'input_text', text: 'Locate the implementation.' }] }, - ]) - }, - ) + ], + toolChoice: { type: 'auto' }, + providerOptions: { + openai: { reasoningEffort: 'max', promptCacheKey: 'cache-key' }, + }, + }), + makeConfig(), + ) + + const body = await Effect.runPromise(webSocketRoute.body.from(request)) + + expect(body.instructions).toBe('Use the repository tools.') + expect(body.tools).toHaveLength(1) + expect(body.tool_choice).toBe('auto') + expect(body.prompt_cache_key).toBe('cache-key') + expect(body.reasoning).toEqual({ effort: 'max', summary: 'detailed' }) + expect(body.parallel_tool_calls).toBeUndefined() + expect(body.input).toEqual([ + { role: 'user', content: [{ type: 'input_text', text: 'Locate the implementation.' }] }, + ]) + }) }) // --------------------------------------------------------------------------- diff --git a/packages/opencode-llm-vendor/src/protocols/openai-responses.ts b/packages/opencode-llm-vendor/src/protocols/openai-responses.ts index c874ae3..9ce9da3 100644 --- a/packages/opencode-llm-vendor/src/protocols/openai-responses.ts +++ b/packages/opencode-llm-vendor/src/protocols/openai-responses.ts @@ -437,10 +437,10 @@ const lowerOptions = Effect.fn('OpenAIResponses.lowerOptions')(function* (reques } }) -const GPT_5_6_MODELS = new Set(['gpt-5.6-sol', 'gpt-5.6-terra', 'gpt-5.6-luna']) +const MAX_REASONING_MODELS = new Set(['gpt-5.6-sol', 'gpt-5.6-terra', 'gpt-5.6-luna', 'gpt-6-sol', 'gpt-6-luna']) export const isReasoningEffortForModel = (modelId: string, effort: ReasoningEffort): boolean => - OpenAIOptions.isReasoningEffort(effort) || (effort === 'max' && GPT_5_6_MODELS.has(modelId)) + OpenAIOptions.isReasoningEffort(effort) || (effort === 'max' && MAX_REASONING_MODELS.has(modelId)) const fromRequest = Effect.fn('OpenAIResponses.fromRequest')(function* (request: LLMRequest) { const generation = request.generation From 5e0fc8b06be8b33981309fe2ed5ce9ee1266ee24 Mon Sep 17 00:00:00 2001 From: Kyle Mistele Date: Wed, 23 Sep 2026 11:13:30 -0700 Subject: [PATCH 2/2] fix: narrow model catalog update HumanLayer-Session: https://app.dev.codelayer.gg/sessions/01a0cf31-76ac-77af-a656-2e0e74a69323 --- package.json | 2 +- packages/agentlayer-core/models.json | 338432 +++++----------- packages/agentlayer-core/test/models.test.ts | 10 +- 3 files changed, 104783 insertions(+), 233661 deletions(-) diff --git a/package.json b/package.json index 94056a8..2c7100e 100644 --- a/package.json +++ b/package.json @@ -16,7 +16,7 @@ "hook:prepush": "env -i HOME=\"$HOME\" PATH=\"$PATH\" TMPDIR=\"$TMPDIR\" bun check", "models:refresh": "bun -e \"const data = await fetch('https://models.dev/api.json').then((response) => response.json()); await Bun.write('packages/agentlayer-core/models.json', JSON.stringify(data, null, 2) + '\\n')\"", "test": "bun test --concurrency 16 **/*", - "test:ci": "bun test --concurrency 8 --conditions=source ./packages/*/test/*.test.ts ./agents/*/test/*.test.ts", + "test:ci": "bun test --conditions=source ./packages/*/test/*.test.ts ./agents/*/test/*.test.ts", "typecheck": "bun --filter '*' --parallel typecheck", "release:prepare": "bun run scripts/release/prepare.ts", "release:publish": "bun run scripts/release/publish.ts" diff --git a/packages/agentlayer-core/models.json b/packages/agentlayer-core/models.json index 175878f..f81552f 100644 --- a/packages/agentlayer-core/models.json +++ b/packages/agentlayer-core/models.json @@ -1,57 +1,61 @@ { - "deepinfra": { - "id": "deepinfra", + "requesty": { + "id": "requesty", "env": [ - "DEEPINFRA_API_KEY" + "REQUESTY_API_KEY" ], - "npm": "@ai-sdk/deepinfra", - "name": "Deep Infra", - "doc": "https://deepinfra.com/models", + "npm": "@ai-sdk/openai-compatible", + "api": "https://router.requesty.ai/v1", + "name": "Requesty", + "doc": "https://requesty.ai/solution/llm-routing/models", "models": { - "tencent/Hy3": { - "id": "tencent/Hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "xai/grok-4": { + "id": "xai/grok-4", + "name": "Grok 4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "knowledge": "2025-01", + "release_date": "2025-09-09", + "last_updated": "2025-09-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 192000, - "output": 128000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.13, - "output": 0.53, - "cache_read": 0.033 + "input": 3, + "output": 15, + "cache_read": 0.75, + "cache_write": 3 } }, - "meta-llama/Llama-3.3-70B-Instruct-Turbo": { - "id": "meta-llama/Llama-3.3-70B-Instruct-Turbo", - "name": "Llama 3.3 70B Turbo", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", - "attachment": false, - "reasoning": false, + "xai/grok-4-fast": { + "id": "xai/grok-4-fast", + "name": "Grok 4 Fast", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-09-19", + "last_updated": "2025-09-19", "modalities": { "input": [ "text" @@ -60,272 +64,411 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 2000000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.32 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05, + "cache_write": 0.2 } }, - "meta-llama/Llama-4-Scout-17B-16E-Instruct": { - "id": "meta-llama/Llama-4-Scout-17B-16E-Instruct", - "name": "Llama 4 Scout 17B", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 327680, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 1.25, + "output": 10, + "cache_read": 0.31, + "cache_write": 2.375, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { - "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", - "name": "Llama 4 Maverick 17B FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 16384 + "output": 65536 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0.3, + "output": 2.5, + "cache_read": 0.075, + "cache_write": 0.55 } }, - "XiaomiMiMo/MiMo-V2.5-Pro": { - "id": "XiaomiMiMo/MiMo-V2.5-Pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "google/gemini-3-pro-preview": { + "id": "google/gemini-3-pro-preview", + "name": "Gemini 3 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", - "audio" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 16384 + "output": 65536 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 4.5 } }, - "XiaomiMiMo/MiMo-V2.5": { - "id": "XiaomiMiMo/MiMo-V2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", "audio", - "video" + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 1 } }, - "thinkingmachines/Inkling-Small": { - "id": "thinkingmachines/Inkling-Small", - "name": "Inkling Small", - "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", - "family": "ling", + "openai/gpt-5.2-chat": { + "id": "openai/gpt-5.2-chat", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 1048576 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.45, - "output": 1.2, - "cache_read": 0.1 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "thinkingmachines/Inkling": { - "id": "thinkingmachines/Inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 1048576 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4.05, - "cache_read": 0.16 + "input": 21, + "output": 168 } }, - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma 3 27B IT", - "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", - "family": "gemma", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image" + "audio", + "image", + "video" ], "output": [ - "text" + "text", + "audio", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.08, - "output": 0.16 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "google/gemma-3-12b-it": { - "id": "google/gemma-3-12b-it", - "name": "Gemma 3 12B IT", - "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", - "family": "gemma", + "openai/gpt-5-chat": { + "id": "openai/gpt-5-chat", + "name": "GPT-5 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -335,29 +478,44 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.15 + "input": 1.25, + "output": 10 } }, - "google/gemma-3-4b-it": { - "id": "google/gemma-3-4b-it", - "name": "Gemma 3 4B IT", - "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", - "family": "gemma", + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", @@ -367,70 +525,91 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 400000, + "output": 272000 }, "cost": { - "input": 0.05, - "output": 0.1 + "input": 15, + "output": 120 } }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4 Mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-06", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.13, - "output": 0.38 + "input": 1.1, + "output": 4.4, + "cache_read": 0.28 } }, - "google/gemma-4-26B-A4B-it": { - "id": "google/gemma-4-26B-A4B-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "openai/gpt-5.1-chat": { + "id": "openai/gpt-5.1-chat", + "name": "GPT-5.1 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -440,66 +619,93 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.07, - "output": 0.34 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "google/gemma-4-E4B-it": { - "id": "google/gemma-4-E4B-it", - "name": "Gemma 4 E4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.02, - "output": 0.1 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "Qwen/Qwen3-VL-235B-A22B-Instruct": { - "id": "Qwen/Qwen3-VL-235B-A22B-Instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT-5.1-Codex-Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -509,109 +715,146 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.88, + "input": 1.1, + "output": 9, "cache_read": 0.11 } }, - "Qwen/Qwen3.8-27B": { - "id": "Qwen/Qwen3.8-27B", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "xhigh" + "high", + "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 2.5, - "cache_read": 0.05 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "Qwen/Qwen3-Next-80B-A3B-Instruct": { - "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.09, - "output": 1.1 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "Qwen/Qwen3.7-Max": { - "id": "Qwen/Qwen3.7-Max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1-Codex-Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -619,266 +862,226 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 400000, + "output": 100000 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "tiers": [ - { - "input": 5, - "output": 15, - "cache_read": 1, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 6.25, - "output": 18.5, - "cache_read": 1.25, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "Qwen/Qwen3.8-Max": { - "id": "Qwen/Qwen3.8-Max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "openai/gpt-5-image": { + "id": "openai/gpt-5-image", + "name": "GPT-5 Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "knowledge": "2024-10-01", + "release_date": "2025-10-14", + "last_updated": "2025-10-14", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 256000, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 1.65, - "output": 4.951, - "cache_read": 0.206 + "input": 5, + "output": 10, + "cache_read": 1.25 } }, - "Qwen/Qwen3.5-27B": { - "id": "Qwen/Qwen3.5-27B", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.26, - "output": 2.6 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 81920 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.95 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo", - "name": "Qwen3 Coder 480B A35B Instruct Turbo", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 66536 + "context": 128000, + "output": 32000 }, "cost": { - "input": 0.3, - "output": 1, - "cache_read": 0.1 + "input": 0.25, + "output": 2, + "cache_read": 0.03 } }, - "Qwen/Qwen3-Max": { - "id": "Qwen/Qwen3-Max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 65536 - }, - "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24, - "tiers": [ - { - "input": 2.4, - "output": 12, - "cache_read": 0.48, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 3, - "output": 15, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 128000 - } - } - ] - } - }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B-A22B Instruct 2507", - "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 16384 - }, - "cost": { - "input": 0.09, - "output": 0.55 - } - }, - "Qwen/Qwen3.8-Flash": { - "id": "Qwen/Qwen3.8-Flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -886,29 +1089,42 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.113, - "output": 0.382, - "cache_read": 0.0141 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text" @@ -917,380 +1133,493 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 16384 + "context": 16000, + "output": 4000 }, "cost": { - "input": 0.08, - "output": 0.28 + "input": 0.05, + "output": 0.4, + "cache_read": 0.01 } }, - "Qwen/Qwen3.5-35B-A3B": { - "id": "Qwen/Qwen3.5-35B-A3B", - "name": "Qwen 3.5 35B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-01", - "last_updated": "2026-04-20", + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 81920 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 1, - "cache_read": 0.05 + "input": 30, + "output": 180, + "cache_read": 30 } }, - "Qwen/Qwen3.5-122B-A10B": { - "id": "Qwen/Qwen3.5-122B-A10B", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2024-10", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.29, - "output": 2.4 + "input": 0.15, + "output": 0.6, + "cache_read": 0.08 } }, - "Qwen/Qwen3.5-9B": { - "id": "Qwen/Qwen3.5-9B", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "GPT-5 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2024-10-01", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.15 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen 3.5 397B A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-01", - "last_updated": "2026-04-20", + "knowledge": "2025-08-31", + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 81920 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.45, - "output": 3, - "cache_read": 0.22 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "Qwen/Qwen3.8-2.4T-A95B": { - "id": "Qwen/Qwen3.8-2.4T-A95B", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "xhigh" + "high", + "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "Qwen/Qwen3-30B-A3B": { - "id": "Qwen/Qwen3-30B-A3B", - "name": "Qwen3 30B A3B", - "description": "Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning", - "family": "qwen", - "attachment": false, + "anthropic/claude-3-7-sonnet": { + "id": "anthropic/claude-3-7-sonnet", + "name": "Claude Sonnet 3.7", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "knowledge": "2024-01", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 16384 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.12, - "output": 0.5 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.32, - "output": 3.2 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek-V3.1", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes", - "family": "deepseek", - "attachment": false, + "anthropic/claude-opus-4-5": { + "id": "anthropic/claude-opus-4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 0.95, - "cache_read": 0.13 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "anthropic/claude-sonnet-4-5": { + "id": "anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.015 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek-ai/DeepSeek-V4-Pro-0813": { - "id": "deepseek-ai/DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, + "anthropic/claude-opus-4-1": { + "id": "anthropic/claude-opus-4-1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "low", + "medium", "high", "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1.3, - "output": 2.6, - "cache_read": 0.1 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "deepseek-ai/DeepSeek-V4.1-Flash": { - "id": "deepseek-ai/DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -1299,159 +1628,209 @@ "values": [ "none", "low", + "medium", "high", - "xhigh", "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-02-01", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 200000, + "output": 62000 }, "cost": { - "input": 0.2, - "output": 0.6, - "cache_read": 0.006 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.09, - "output": 0.18, - "cache_read": 0.018 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.3, - "output": 2.6, - "cache_read": 0.1 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -1459,31 +1838,40 @@ }, "open_weights": false, "limit": { - "context": 163840, - "output": 64000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.26, - "output": 0.38, - "cache_read": 0.13 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } - }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", + } + } + }, + "qiniu-ai": { + "id": "qiniu-ai", + "env": [ + "QINIU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.qnaigc.com/v1", + "name": "Qiniu", + "doc": "https://developer.qiniu.com/aitokenapi", + "models": { + "deepseek-r1-0528": { + "id": "deepseek-r1-0528", "name": "DeepSeek-R1-0528", "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -1494,27 +1882,22 @@ }, "open_weights": false, "limit": { - "context": 163840, - "output": 64000 - }, - "cost": { - "input": 0.5, - "output": 2.15, - "cache_read": 0.35 + "context": 128000, + "output": 32000 } }, - "deepseek-ai/DeepSeek-V3": { - "id": "deepseek-ai/DeepSeek-V3", - "name": "DeepSeek-V3", - "description": "Open DeepSeek MoE chat model for coding, math, and general reasoning", - "family": "deepseek", + "doubao-1.5-thinking-pro": { + "id": "doubao-1.5-thinking-pro", + "name": "Doubao 1.5 Thinking Pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2024-12-26", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -1523,70 +1906,50 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 8192 - }, - "cost": { - "input": 0.32, - "output": 0.89 + "context": 128000, + "output": 16000 } }, - "deepseek-ai/DeepSeek-V3-0324": { - "id": "deepseek-ai/DeepSeek-V3-0324", - "name": "DeepSeek V3 0324", - "description": "March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding", - "family": "deepseek", - "attachment": false, + "qwen3-vl-30b-a3b-thinking": { + "id": "qwen3-vl-30b-a3b-thinking", + "name": "Qwen3-Vl 30b A3b Thinking", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "release_date": "2026-02-09", + "last_updated": "2026-02-09", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 163840 - }, - "cost": { - "input": 0.24, - "output": 0.9, - "cache_read": 0.135 + "context": 128000, + "output": 32000 } }, - "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp": { - "id": "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", + "claude-3.5-haiku": { + "id": "claude-3.5-haiku", + "name": "Claude 3.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text", @@ -1596,100 +1959,73 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.014 + "context": 200000, + "output": 8192 } }, - "stepfun-ai/Step-3.7-Flash": { - "id": "stepfun-ai/Step-3.7-Flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "deepseek-v3-0324": { + "id": "deepseek-v3-0324", + "name": "DeepSeek-V3-0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 256000 - }, - "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "context": 128000, + "output": 16000 } }, - "MiniMaxAI/MiniMax-M3": { - "id": "MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235b A22B Instruct 2507", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2025-08-12", + "last_updated": "2025-08-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 512000 - }, - "cost": { - "input": 0.28, - "output": 1.1, - "cache_read": 0.056 + "context": 262144, + "output": 64000 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "deepseek-v3": { + "id": "deepseek-v3", + "name": "DeepSeek-V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2025-08-13", + "last_updated": "2025-08-13", "modalities": { "input": [ "text" @@ -1698,30 +2034,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0.15, - "output": 1.15, - "cache_read": 0.03 + "context": 128000, + "output": 16000 } }, - "MiniMaxAI/MiniMax-M2.7": { - "id": "MiniMaxAI/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "kimi-k2": { + "id": "kimi-k2", + "name": "Kimi K2", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -1730,24 +2059,17 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0.25, - "output": 1, - "cache_read": 0.05 + "context": 128000, + "output": 128000 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -1755,295 +2077,188 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 - }, - "cost": { - "input": 0.75, - "output": 3.5, - "cache_read": 0.15 + "context": 40000, + "output": 4096 } }, - "moonshotai/Kimi-K3": { - "id": "moonshotai/Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "qwen3-max-preview": { + "id": "qwen3-max-preview", + "name": "Qwen3 Max Preview", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "structured_output": false, + "temperature": true, + "release_date": "2025-09-06", + "last_updated": "2025-09-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 2.85, - "output": 14.25, - "cache_read": 0.285 + "context": 256000, + "output": 64000 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "claude-3.5-sonnet": { + "id": "claude-3.5-sonnet", + "name": "Claude 3.5 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "structured_output": false, + "temperature": true, + "release_date": "2025-09-09", + "last_updated": "2025-09-09", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.68, - "output": 3.4, - "cache_read": 0.136 + "context": 200000, + "output": 8200 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2025-09-12", + "last_updated": "2025-09-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 131072, "output": 32768 - }, - "status": "deprecated", - "cost": { - "input": 0.45, - "output": 2.25, - "cache_read": 0.07 } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 16384 - }, - "cost": { - "input": 1.05, - "output": 3.5, - "cache_read": 0.205 + "context": 1048576, + "output": 65536 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "claude-4.5-haiku": { + "id": "claude-4.5-haiku", + "name": "Claude 4.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2025-10-16", + "last_updated": "2025-10-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 32768 - }, - "cost": { - "input": 0.75, - "output": 2.4, - "cache_read": 0.14 + "context": 200000, + "output": 64000 } }, - "zai-org/GLM-4.7": { - "id": "zai-org/GLM-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "kling-v2-6": { + "id": "kling-v2-6", + "name": "Kling-V2 6", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-01-13", + "last_updated": "2026-01-13", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 16384 - }, - "cost": { - "input": 0.4, - "output": 1.75, - "cache_read": 0.08 + "context": 99999999, + "output": 99999999 } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM-5", + "glm-4.5": { + "id": "glm-4.5", + "name": "GLM 4.5", "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -2052,14 +2267,10 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -2068,119 +2279,79 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 16384 - }, - "status": "deprecated", - "cost": { - "input": 0.6, - "output": 2.08, - "cache_read": 0.12 + "context": 131072, + "output": 98304 } }, - "zai-org/GLM-5.3-Flash": { - "id": "zai-org/GLM-5.3-Flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "claude-4.1-opus": { + "id": "claude-4.1-opus", + "name": "Claude 4.1 Opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "context": 200000, + "output": 32000 } }, - "zai-org/GLM-4.6": { - "id": "zai-org/GLM-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 0.5, - "output": 2, - "cache_read": 0.1 + "context": 1048576, + "output": 64000 } }, - "zai-org/GLM-4.7-Flash": { - "id": "zai-org/GLM-4.7-Flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ "text" @@ -2189,65 +2360,49 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 16384 - }, - "status": "deprecated", - "cost": { - "input": 0.06, - "output": 0.4, - "cache_read": 0.01 + "context": 262144, + "output": 65536 } }, - "zai-org/GLM-5.3": { - "id": "zai-org/GLM-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "doubao-seed-2.0-pro": { + "id": "doubao-seed-2.0-pro", + "name": "Doubao Seed 2.0 Pro", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.2 + "context": 256000, + "output": 128000 } }, - "nvidia/Nemotron-3-Nano-30B-A3B": { - "id": "nvidia/Nemotron-3-Nano-30B-A3B", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, + "doubao-seed-1.6": { + "id": "doubao-seed-1.6", + "name": "Doubao-Seed 1.6", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -2255,122 +2410,96 @@ } ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "release_date": "2025-08-15", + "last_updated": "2025-08-15", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.025 + "context": 256000, + "output": 32000 } }, - "nvidia/Llama-3.3-Nemotron-Super-49B-v1.5": { - "id": "nvidia/Llama-3.3-Nemotron-Super-49B-v1.5", - "name": "Llama 3.3 Nemotron Super 49B v1.5", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, + "doubao-seed-1.6-thinking": { + "id": "doubao-seed-1.6-thinking", + "name": "Doubao-Seed 1.6 Thinking", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "release_date": "2025-08-15", + "last_updated": "2025-08-15", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 0.4 + "context": 256000, + "output": 32000 } }, - "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning": { - "id": "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning", - "name": "Nemotron 3 Nano Omni 30B A3B Reasoning", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", + "gemini-2.0-flash": { + "id": "gemini-2.0-flash", + "name": "Gemini 2.0 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", - "video", - "audio" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "status": "deprecated", - "cost": { - "input": 0.2, - "output": 0.8 + "context": 1048576, + "output": 8192 } }, - "ByteDance/Seed-2.0-code": { - "id": "ByteDance/Seed-2.0-code", - "name": "Seed 2.0 Code", - "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "qwen-max-2025-01-25": { + "id": "qwen-max-2025-01-25", + "name": "Qwen2.5-Max-2025-01-25", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -2378,39 +2507,22 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 131072 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1, - "tiers": [ - { - "input": 1, - "output": 6, - "cache_read": 0.2, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "context": 128000, + "output": 4096 } }, - "ByteDance/Seed-2.0-pro": { - "id": "ByteDance/Seed-2.0-pro", - "name": "Seed 2.0 Pro", - "description": "Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows", - "family": "seed", + "claude-4.0-sonnet": { + "id": "claude-4.0-sonnet", + "name": "Claude 4.0 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -2422,43 +2534,49 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1, - "tiers": [ - { - "input": 1, - "output": 6, - "cache_read": 0.2, - "tier": { - "type": "context", - "size": 128000 - } - } + "context": 200000, + "output": 64000 + } + }, + "doubao-1.5-pro-32k": { + "id": "doubao-1.5-pro-32k", + "name": "Doubao 1.5 Pro 32k", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 12000 } }, - "ByteDance/Seed-2.0-mini": { - "id": "ByteDance/Seed-2.0-mini", - "name": "Seed 2.0 Mini", - "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen3-30b-a3b-instruct-2507": { + "id": "qwen3-30b-a3b-instruct-2507", + "name": "Qwen3 30b A3b Instruct 2507", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-02-04", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -2466,48 +2584,26 @@ }, "open_weights": false, "limit": { - "context": 256000, + "context": 128000, "output": 32000 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.02, - "tiers": [ - { - "input": 0.2, - "output": 0.8, - "cache_read": 0.2, - "tier": { - "type": "context", - "size": 128000 - } - } - ] } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "qwen3-next-80b-a3b-thinking": { + "id": "qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B Thinking", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-09-12", + "last_updated": "2025-09-12", "modalities": { "input": [ "text" @@ -2516,35 +2612,51 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.03, - "output": 0.14 + "output": 32768 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "qwen3-235b-a22b-thinking-2507": { + "id": "qwen3-235b-a22b-thinking-2507", + "name": "Qwen3 235B A22B Thinking 2507", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-12", + "last_updated": "2025-08-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 4096 + } + }, + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, "temperature": true, "release_date": "2025-08-05", "last_updated": "2025-08-05", @@ -2556,57 +2668,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.037, - "output": 0.17 + "context": 128000, + "output": 32000 } - } - } - }, - "perplexity-agent": { - "id": "perplexity-agent", - "env": [ - "PERPLEXITY_API_KEY" - ], - "npm": "@ai-sdk/openai", - "api": "https://api.perplexity.ai/v1", - "name": "Perplexity Agent", - "doc": "https://docs.perplexity.ai/docs/agent-api/models", - "models": { - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + }, + "doubao-1.5-vision-pro": { + "id": "doubao-1.5-vision-pro", + "name": "Doubao 1.5 Vision Pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -2614,89 +2697,80 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1 + "context": 128000, + "output": 16000 } }, - "anthropic/claude-opus-4-5": { - "id": "anthropic/claude-opus-4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gemini-3.0-pro-image-preview": { + "id": "gemini-3.0-pro-image-preview", + "name": "Gemini 3.0 Pro Image Preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 32768, + "output": 8192 + } + }, + "gemini-2.5-flash-image": { + "id": "gemini-2.5-flash-image", + "name": "Gemini 2.5 Flash Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-10-22", + "last_updated": "2025-10-22", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "open_weights": false, + "limit": { + "context": 32768, + "output": 8192 } }, - "anthropic/claude-sonnet-4-5": { - "id": "anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -2704,44 +2778,26 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 1048576, "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 } }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "claude-3.7-sonnet": { + "id": "claude-3.7-sonnet", + "name": "Claude 3.7 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -2751,42 +2807,27 @@ "limit": { "context": 200000, "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "qwen3-30b-a3b-thinking-2507": { + "id": "qwen3-30b-a3b-thinking-2507", + "name": "Qwen3 30b A3b Thinking 2507", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "release_date": "2026-02-04", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -2794,44 +2835,27 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "context": 126000, + "output": 32000 } }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen2.5-vl-72b-instruct": { + "id": "qwen2.5-vl-72b-instruct", + "name": "Qwen 2.5 VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -2839,69 +2863,80 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "context": 128000, + "output": 8192 } }, - "deepseek/deepseek-v4-flash-0731": { - "id": "deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "attachment": false, "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-06", + "last_updated": "2025-08-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + } + }, + "doubao-seed-1.6-flash": { + "id": "doubao-seed-1.6-flash", + "name": "Doubao-Seed 1.6 Flash", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, + "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2025-08-15", + "last_updated": "2025-08-15", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.13, - "output": 0.26, - "cache_read": 0.028 + "context": 256000, + "output": 32000 } }, - "perplexity/sonar": { - "id": "perplexity/sonar", - "name": "Sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar", + "deepseek-v3.1": { + "id": "deepseek-v3.1", + "name": "DeepSeek-V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "release_date": "2025-08-19", + "last_updated": "2025-08-19", "modalities": { "input": [ "text" @@ -2913,45 +2948,23 @@ "open_weights": false, "limit": { "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.25, - "output": 2.5, - "cache_read": 0.0625 + "output": 32000 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "qwen3-235b-a22b": { + "id": "qwen3-235b-a22b", + "name": "Qwen 3 235B A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -2959,62 +2972,24 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "context": 128000, + "output": 32000 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "qwen3-coder-480b-a35b-instruct": { + "id": "qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-05", + "release_date": "2025-08-14", + "last_updated": "2025-08-14", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -3022,62 +2997,30 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "context": 262000, + "output": 4096 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-05", + "release_date": "2026-02-22", + "last_updated": "2026-02-22", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -3085,94 +3028,59 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "context": 256000, + "output": 64000 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "mimo-v2-flash": { + "id": "mimo-v2-flash", + "name": "Mimo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2024-12-01", + "release_date": "2025-12-16", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "tiers": [ - { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 - } + "input": 0.1, + "output": 0.3, + "cache_read": 0.01 } }, - "xai/grok-4-1-fast-non-reasoning": { - "id": "xai/grok-4-1-fast-non-reasoning", - "name": "Grok 4.1 Fast (Non-Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "qwen-vl-max-2025-01-25": { + "id": "qwen-vl-max-2025-01-25", + "name": "Qwen VL-MAX-2025-01-25", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -3180,43 +3088,27 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "context": 128000, + "output": 4096 } }, - "xai/grok-4.6": { - "id": "xai/grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", + "qwen2.5-vl-7b-instruct": { + "id": "qwen2.5-vl-7b-instruct", + "name": "Qwen 2.5 VL 7B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -3224,55 +3116,26 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "context": 128000, + "output": 8192 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "Nemotron 3 Super 120B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "GLM 4.5 Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -3281,131 +3144,90 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 32000 - }, - "cost": { - "input": 0.25, - "output": 2.5 + "context": 131000, + "output": 4096 } }, - "moonshot-ai/kimi-k3": { - "id": "moonshot-ai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "claude-4.5-opus": { + "id": "claude-4.5-opus", + "name": "Claude 4.5 Opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-30", + "structured_output": false, + "temperature": true, + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "context": 200000, + "output": 200000 } }, - "moonshot-ai/kimi-k2.7-code": { - "id": "moonshot-ai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "gemini-3.0-pro-preview": { + "id": "gemini-3.0-pro-preview", + "name": "Gemini 3.0 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-07-30", + "structured_output": false, + "temperature": true, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ "text", "image", - "video" + "video", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "context": 1000000, + "output": 64000 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "doubao-seed-2.0-mini": { + "id": "doubao-seed-2.0-mini", + "name": "Doubao Seed 2.0 Mini", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -3413,40 +3235,22 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "context": 256000, + "output": 32000 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "claude-4.0-opus": { + "id": "claude-4.0-opus", + "name": "Claude 4.0 Opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -3458,44 +3262,25 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "context": 200000, + "output": 32000 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "gpt-oss-20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": false, + "temperature": true, + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -3503,44 +3288,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "context": 128000, + "output": 4096 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "gemini-3.0-flash-preview": { + "id": "gemini-3.0-flash-preview", + "name": "Gemini 3.0 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -3548,45 +3318,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "context": 1000000, + "output": 64000 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, + "MiniMax-M1": { + "id": "MiniMax-M1", + "name": "MiniMax M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -3594,41 +3348,26 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "context": 1000000, + "output": 80000 } - } - } - }, - "bailing": { - "id": "bailing", - "env": [ - "BAILING_API_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.tbox.cn/api/llm/v1/chat/completions", - "name": "Bailing", - "doc": "https://alipaytbox.yuque.com/sxs0ba/ling/intro", - "models": { - "Ring-1T": { - "id": "Ring-1T", - "name": "Ring-1T", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "ring", + }, + "qwen-turbo": { + "id": "qwen-turbo", + "name": "Qwen-Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-10", - "last_updated": "2025-10", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -3637,28 +3376,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32000 - }, - "cost": { - "input": 0.57, - "output": 2.29 + "context": 1000000, + "output": 4096 } }, - "Ling-1T": { - "id": "Ling-1T", - "name": "Ling-1T", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "ling", + "qwen3-30b-a3b": { + "id": "qwen3-30b-a3b", + "name": "Qwen3 30B A3B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-10", - "last_updated": "2025-10", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -3667,42 +3406,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32000 - }, - "cost": { - "input": 0.57, - "output": 2.29 + "context": 40000, + "output": 4096 } - } - } - }, - "poe": { - "id": "poe", - "env": [ - "POE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.poe.com/v1", - "name": "Poe", - "doc": "https://creator.poe.com/docs/external-applications/openai-compatible-api", - "models": { - "cerebras/qwen3-32b-cs": { - "id": "cerebras/qwen3-32b-cs", - "name": "qwen3-32b-cs", - "description": "Legacy model retained for compatibility with older integrations", + }, + "claude-4.5-sonnet": { + "id": "claude-4.5-sonnet", + "name": "Claude 4.5 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-05-15", - "last_updated": "2025-05-15", + "structured_output": false, + "temperature": true, + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -3710,24 +3435,31 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 - }, - "status": "deprecated" + "context": 200000, + "output": 64000 + } }, - "cerebras/llama-3.1-8b-cs": { - "id": "cerebras/llama-3.1-8b-cs", - "name": "Llama-3.1-8B-CS", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "doubao-seed-2.0-lite": { + "id": "doubao-seed-2.0-lite", + "name": "Doubao Seed 2.0 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-05-13", - "last_updated": "2025-05-13", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -3735,27 +3467,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 0 - }, - "cost": { - "input": 0.1, - "output": 0.1 + "context": 256000, + "output": 32000 } }, - "cerebras/llama-3.3-70b-cs": { - "id": "cerebras/llama-3.3-70b-cs", - "name": "llama-3.3-70b-cs", - "description": "Legacy model retained for compatibility with older integrations", + "gemini-2.0-flash-lite": { + "id": "gemini-2.0-flash-lite", + "name": "Gemini 2.0 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-05-13", - "last_updated": "2025-05-13", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -3763,34 +3496,31 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 - }, - "status": "deprecated" + "context": 1048576, + "output": 8192 + } }, - "cerebras/gpt-oss-120b-cs": { - "id": "cerebras/gpt-oss-120b-cs", - "name": "GPT-OSS-120B-CS", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "doubao-seed-2.0-code": { + "id": "doubao-seed-2.0-code", + "name": "Doubao Seed 2.0 Code", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -3798,25 +3528,21 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 0 - }, - "cost": { - "input": 0.35, - "output": 0.75 + "context": 256000, + "output": 128000 } }, - "cerebras/qwen3-235b-2507-cs": { - "id": "cerebras/qwen3-235b-2507-cs", - "name": "qwen3-235b-2507-cs", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "structured_output": false, + "temperature": true, + "release_date": "2025-11-07", + "last_updated": "2025-11-07", "modalities": { "input": [ "text" @@ -3827,50 +3553,48 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 - }, - "status": "deprecated" + "context": 256000, + "output": 100000 + } }, - "empiriolabs/deepseek-v4-pro-el": { - "id": "empiriolabs/deepseek-v4-pro-el", - "name": "DeepSeek-V4-Pro-EL", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Moonshotai/Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "release_date": "2026-04-24", - "last_updated": "2026-05-02", + "structured_output": false, + "temperature": true, + "release_date": "2026-01-28", + "last_updated": "2026-01-28", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.67, - "output": 3.33 + "context": 256000, + "output": 256000 } }, - "empiriolabs/deepseek-v4-flash-el": { - "id": "empiriolabs/deepseek-v4-flash-el", - "name": "DeepSeek-V4-Flash-EL", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "attachment": false, + "reasoning": false, "tool_call": true, - "release_date": "2026-04-24", - "last_updated": "2026-05-02", + "structured_output": false, + "temperature": true, + "release_date": "2025-09-08", + "last_updated": "2025-09-08", "modalities": { "input": [ "text" @@ -3879,33 +3603,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.14, - "output": 0.28 + "context": 256000, + "output": 100000 } }, - "anthropic/claude-haiku-3.5": { - "id": "anthropic/claude-haiku-3.5", - "name": "Claude-Haiku-3.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "stepfun-ai/gelab-zero-4b-preview": { + "id": "stepfun-ai/gelab-zero-4b-preview", + "name": "Stepfun-Ai/Gelab Zero 4b Preview", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -3913,39 +3631,25 @@ }, "open_weights": false, "limit": { - "context": 189096, - "output": 8192 - }, - "cost": { - "input": 0.68, - "output": 3.4, - "cache_read": 0.068, - "cache_write": 0.85 + "context": 8192, + "output": 4096 } }, - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude-Opus-4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "x-ai/grok-code-fast-1": { + "id": "x-ai/grok-code-fast-1", + "name": "x-AI/Grok-Code-Fast 1", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 0, - "max": 31999 - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": false, + "temperature": true, + "release_date": "2025-09-02", + "last_updated": "2025-09-02", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -3953,47 +3657,28 @@ }, "open_weights": false, "limit": { - "context": 196608, - "output": 32000 - }, - "cost": { - "input": 13, - "output": 64, - "cache_read": 1.3, - "cache_write": 16 + "context": 256000, + "output": 10000 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude-Opus-4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "x-ai/grok-4.1-fast-reasoning": { + "id": "x-ai/grok-4.1-fast-reasoning", + "name": "X-Ai/Grok 4.1 Fast Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 63999 - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-11-21", - "last_updated": "2025-11-21", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-19", + "last_updated": "2025-12-19", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -4001,43 +3686,27 @@ }, "open_weights": false, "limit": { - "context": 196608, - "output": 64000 - }, - "cost": { - "input": 4.3, - "output": 21, - "cache_read": 0.43, - "cache_write": 5.3 + "context": 20000000, + "output": 2000000 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude-Opus-4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "x-ai/grok-4.1-fast-non-reasoning": { + "id": "x-ai/grok-4.1-fast-non-reasoning", + "name": "X-Ai/Grok 4.1 Fast Non Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2026-02-04", - "last_updated": "2026-02-04", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-19", + "last_updated": "2025-12-19", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -4045,32 +3714,28 @@ }, "open_weights": false, "limit": { - "context": 983040, - "output": 128000 - }, - "cost": { - "input": 4.3, - "output": 21, - "cache_read": 0.43, - "cache_write": 5.3 + "context": 2000000, + "output": 2000000 } }, - "anthropic/claude-sonnet-3.5-june": { - "id": "anthropic/claude-sonnet-3.5-june", - "name": "Claude-Sonnet-3.5-June", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-sonnet", + "x-ai/grok-4-fast-reasoning": { + "id": "x-ai/grok-4-fast-reasoning", + "name": "X-Ai/Grok-4-Fast-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2024-11-18", - "last_updated": "2024-11-18", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -4078,45 +3743,28 @@ }, "open_weights": false, "limit": { - "context": 189096, - "output": 8192 - }, - "status": "deprecated", - "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "context": 2000000, + "output": 2000000 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude-Opus-4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "x-ai/grok-4-fast": { + "id": "x-ai/grok-4-fast", + "name": "x-AI/Grok-4-Fast", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2026-04-15", - "last_updated": "2026-04-15", + "structured_output": false, + "temperature": true, + "release_date": "2025-09-20", + "last_updated": "2025-09-20", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -4124,32 +3772,27 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 128000 - }, - "cost": { - "input": 4.3, - "output": 21, - "cache_read": 0.43, - "cache_write": 5.4 + "context": 2000000, + "output": 2000000 } }, - "anthropic/claude-haiku-3": { - "id": "anthropic/claude-haiku-3", - "name": "Claude-Haiku-3", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "x-ai/grok-4-fast-non-reasoning": { + "id": "x-ai/grok-4-fast-non-reasoning", + "name": "X-Ai/Grok-4-Fast-Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-03-09", - "last_updated": "2024-03-09", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" @@ -4157,46 +3800,25 @@ }, "open_weights": false, "limit": { - "context": 189096, - "output": 8192 - }, - "cost": { - "input": 0.21, - "output": 1.1, - "cache_read": 0.021, - "cache_write": 0.26 + "context": 2000000, + "output": 2000000 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude-Opus-4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "x-ai/grok-4.1-fast": { + "id": "x-ai/grok-4.1-fast", + "name": "x-AI/Grok-4.1-Fast", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": false, + "temperature": true, + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -4204,31 +3826,25 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 128000 - }, - "cost": { - "input": 4.2929, - "output": 21.4646 + "context": 2000000, + "output": 2000000 } }, - "anthropic/claude-sonnet-3.7": { - "id": "anthropic/claude-sonnet-3.7", - "name": "Claude-Sonnet-3.7", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "z-ai/autoglm-phone-9b": { + "id": "z-ai/autoglm-phone-9b", + "name": "Z-Ai/Autoglm Phone 9b", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -4236,39 +3852,29 @@ }, "open_weights": false, "limit": { - "context": 196608, - "output": 128000 - }, - "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "context": 12800, + "output": 4096 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude-Haiku-4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "Z-Ai/GLM 4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 0, - "max": 63999 + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -4276,39 +3882,54 @@ }, "open_weights": false, "limit": { - "context": 192000, - "output": 64000 + "context": 200000, + "output": 200000 + } + }, + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "Z-AI/GLM 4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": false, + "temperature": true, + "release_date": "2025-10-11", + "last_updated": "2025-10-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "cost": { - "input": 0.85, - "output": 4.3, - "cache_read": 0.085, - "cache_write": 1.1 + "open_weights": false, + "limit": { + "context": 200000, + "output": 200000 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude-Sonnet-4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "Z-Ai/GLM 5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 0, - "max": 31999 + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-09-26", - "last_updated": "2025-09-26", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -4316,33 +3937,24 @@ }, "open_weights": false, "limit": { - "context": 983040, - "output": 32768 - }, - "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "context": 200000, + "output": 128000 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude-Sonnet-4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "OpenAI/GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-05-21", - "last_updated": "2025-05-21", + "structured_output": false, + "temperature": true, + "release_date": "2025-09-19", + "last_updated": "2025-09-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -4350,32 +3962,26 @@ }, "open_weights": false, "limit": { - "context": 983040, - "output": 64000 - }, - "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "context": 400000, + "output": 128000 } }, - "anthropic/claude-sonnet-3.5": { - "id": "anthropic/claude-sonnet-3.5", - "name": "Claude-Sonnet-3.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-sonnet", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "OpenAI/GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2024-06-05", - "last_updated": "2024-06-05", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -4383,78 +3989,57 @@ }, "open_weights": false, "limit": { - "context": 189096, - "output": 8192 - }, - "status": "deprecated", - "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "context": 400000, + "output": 128000 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude-Sonnet-4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "attachment": true, + "xiaomi/mimo-v2-flash": { + "id": "xiaomi/mimo-v2-flash", + "name": "Xiaomi/Mimo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "knowledge": "2024-12-01", + "release_date": "2025-12-16", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 983040, - "output": 128000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 2.6, - "output": 13, - "cache_read": 0.26, - "cache_write": 3.2 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01 } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Claude-Opus-4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "stepfun/step-3.5-flash": { + "id": "stepfun/step-3.5-flash", + "name": "Stepfun/Step-3.5 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "release_date": "2025-05-21", - "last_updated": "2025-05-21", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-02-02", + "last_updated": "2026-02-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -4462,110 +4047,100 @@ }, "open_weights": false, "limit": { - "context": 192512, - "output": 28672 - }, - "cost": { - "input": 13, - "output": 64, - "cache_read": 1.3, - "cache_write": 16 + "context": 64000, + "output": 4096 } }, - "elevenlabs/elevenlabs-v2.5-turbo": { - "id": "elevenlabs/elevenlabs-v2.5-turbo", - "name": "ElevenLabs-v2.5-Turbo", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "elevenlabs", - "attachment": true, + "meituan/longcat-flash-lite": { + "id": "meituan/longcat-flash-lite", + "name": "Meituan/Longcat-Flash-Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-10-28", - "last_updated": "2024-10-28", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-06", + "last_updated": "2026-02-06", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 0 + "context": 256000, + "output": 320000 } }, - "elevenlabs/elevenlabs-v3": { - "id": "elevenlabs/elevenlabs-v3", - "name": "ElevenLabs-v3", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "elevenlabs", - "attachment": true, + "meituan/longcat-flash-chat": { + "id": "meituan/longcat-flash-chat", + "name": "Meituan/Longcat-Flash-Chat", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-11-05", + "last_updated": "2025-11-05", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 0 + "context": 131072, + "output": 131072 } }, - "elevenlabs/elevenlabs-music": { - "id": "elevenlabs/elevenlabs-music", - "name": "ElevenLabs-Music", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "elevenlabs", - "attachment": true, + "deepseek/deepseek-v3.1-terminus": { + "id": "deepseek/deepseek-v3.1-terminus", + "name": "DeepSeek/DeepSeek-V3.1-Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-08-29", - "last_updated": "2025-08-29", + "structured_output": false, + "temperature": true, + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 2000, - "output": 0 + "context": 128000, + "output": 32000 } }, - "novita/glm-4.6v": { - "id": "novita/glm-4.6v", - "name": "glm-4.6v", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "attachment": true, + "deepseek/deepseek-v3.1-terminus-thinking": { + "id": "deepseek/deepseek-v3.1-terminus-thinking", + "name": "DeepSeek/DeepSeek-V3.1-Terminus-Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": false, - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -4573,21 +4148,22 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 32768 + "context": 128000, + "output": 32000 } }, - "novita/glm-4.6": { - "id": "novita/glm-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "deepseek/deepseek-v3.2-exp-thinking": { + "id": "deepseek/deepseek-v3.2-exp-thinking", + "name": "DeepSeek/DeepSeek-V3.2-Exp-Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text" @@ -4598,25 +4174,21 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "output": 32000 } }, - "novita/glm-5": { - "id": "novita/glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "deepseek/deepseek-v3.2-exp": { + "id": "deepseek/deepseek-v3.2-exp", + "name": "DeepSeek/DeepSeek-V3.2-Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text" @@ -4627,69 +4199,48 @@ }, "open_weights": false, "limit": { - "context": 205000, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "context": 128000, + "output": 32000 } }, - "novita/kimi-k2.6": { - "id": "novita/kimi-k2.6", - "name": "Kimi-K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "attachment": true, + "deepseek/deepseek-v3.2-251201": { + "id": "deepseek/deepseek-v3.2-251201", + "name": "Deepseek/DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-05-02", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 262144 - }, - "cost": { - "input": 0.96, - "output": 4.04, - "cache_read": 0.16 + "context": 128000, + "output": 32000 } }, - "novita/kimi-k2-thinking": { - "id": "novita/kimi-k2-thinking", - "name": "kimi-k2-thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": true, + "deepseek/deepseek-math-v2": { + "id": "deepseek/deepseek-math-v2", + "name": "Deepseek/Deepseek-Math-V2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": false, - "release_date": "2025-11-07", - "last_updated": "2025-11-07", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-12-04", + "last_updated": "2025-12-04", "modalities": { "input": [ "text" @@ -4700,15 +4251,15 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 0 + "context": 160000, + "output": 160000 } }, - "novita/kimi-k2.5": { - "id": "novita/kimi-k2.5", - "name": "Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "attachment": true, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "Minimax/Minimax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -4716,14 +4267,13 @@ } ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -4731,20 +4281,15 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "context": 204800, + "output": 128000 } }, - "novita/glm-4.7-flash": { - "id": "novita/glm-4.7-flash", - "name": "glm-4.7-flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "attachment": true, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "Minimax/Minimax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -4752,9 +4297,10 @@ } ], "tool_call": true, - "temperature": false, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -4765,21 +4311,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 65500 + "context": 204800, + "output": 128000 } }, - "novita/glm-4.7": { - "id": "novita/glm-4.7", - "name": "glm-4.7", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": true, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "Minimax/Minimax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2025-10-28", + "last_updated": "2025-10-28", "modalities": { "input": [ "text" @@ -4790,16 +4341,15 @@ }, "open_weights": false, "limit": { - "context": 205000, - "output": 131072 - }, - "status": "deprecated" + "context": 200000, + "output": 128000 + } }, - "novita/glm-4.7-n": { - "id": "novita/glm-4.7-n", - "name": "glm-4.7-n", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "attachment": true, + "minimax/minimax-m2.5-highspeed": { + "id": "minimax/minimax-m2.5-highspeed", + "name": "Minimax/Minimax-M2.5 Highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -4807,9 +4357,10 @@ } ], "tool_call": true, - "temperature": false, - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text" @@ -4820,25 +4371,34 @@ }, "open_weights": false, "limit": { - "context": 205000, - "output": 131072 + "context": 204800, + "output": 128000 } - }, - "novita/deepseek-v3.2": { - "id": "novita/deepseek-v3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + } + } + }, + "alibaba-cn": { + "id": "alibaba-cn", + "env": [ + "DASHSCOPE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://dashscope.aliyuncs.com/compatible-mode/v1", + "name": "Alibaba (China)", + "doc": "https://www.alibabacloud.com/help/en/model-studio/models", + "models": { + "qwen2-5-math-72b-instruct": { + "id": "qwen2-5-math-72b-instruct", + "name": "Qwen2.5-Math 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -4849,30 +4409,26 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 0 + "context": 4096, + "output": 3072 }, "cost": { - "input": 0.27, - "output": 0.4, - "cache_read": 0.13 + "input": 0.574, + "output": 1.721 } }, - "novita/minimax-m2.1": { - "id": "novita/minimax-m2.1", - "name": "minimax-m2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": true, + "deepseek-r1-0528": { + "id": "deepseek-r1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-12-26", - "last_updated": "2025-12-26", + "temperature": true, + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -4883,127 +4439,126 @@ }, "open_weights": false, "limit": { - "context": 205000, - "output": 131072 + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0.574, + "output": 2.294 } }, - "lumalabs/ray2": { - "id": "lumalabs/ray2", - "name": "Ray2", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ray", - "attachment": true, - "reasoning": false, + "qwen3-omni-flash": { + "id": "qwen3-omni-flash", + "name": "Qwen3-Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-02-20", - "last_updated": "2025-02-20", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ - "video" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 5000, - "output": 0 + "context": 65536, + "output": 16384 + }, + "cost": { + "input": 0.058, + "output": 0.23, + "input_audio": 3.584, + "output_audio": 7.168 } }, - "poetools/claude-code": { - "id": "poetools/claude-code", - "name": "claude-code", - "description": "Claude model for careful reasoning, writing, coding, and tool use", - "attachment": true, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-11-27", - "last_updated": "2025-11-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "interleaved": { + "field": "reasoning_content" }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "google/imagen-3-fast": { - "id": "google/imagen-3-fast", - "name": "Imagen-3-Fast", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2024-10-17", - "last_updated": "2024-10-17", + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" ], "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 480, - "output": 0 - } - }, - "google/imagen-4-ultra": { - "id": "google/imagen-4-ultra", - "name": "Imagen-4-Ultra", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2025-05-24", - "last_updated": "2025-05-24", - "modalities": { - "input": [ "text" - ], - "output": [ - "image" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "google/gemini-deep-research": { - "id": "google/gemini-deep-research", - "name": "gemini-deep-research", - "description": "Legacy model retained for compatibility with older integrations", - "attachment": true, + "qwen-plus": { + "id": "qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -5011,377 +4566,329 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 0 + "context": 1000000, + "output": 32768 }, - "status": "deprecated", "cost": { - "input": 1.6, - "output": 9.6 + "input": 0.115, + "output": 0.287, + "reasoning": 1.147 } }, - "google/imagen-4": { - "id": "google/imagen-4", - "name": "Imagen-4", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": true, + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.216, + "output": 0.861 } }, - "google/imagen-4-fast": { - "id": "google/imagen-4-fast", - "name": "Imagen-4-Fast", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": true, + "qwen2-5-coder-7b-instruct": { + "id": "qwen2-5-coder-7b-instruct", + "name": "Qwen2.5-Coder 7B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-06-25", - "last_updated": "2025-06-25", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-11", + "last_updated": "2024-11", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.144, + "output": 0.287 } }, - "google/lyria": { - "id": "google/lyria", - "name": "Lyria", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "lyria", - "attachment": true, + "deepseek-v3": { + "id": "deepseek-v3", + "name": "DeepSeek V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-06-04", - "last_updated": "2025-06-04", + "temperature": true, + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 65536, + "output": 8192 + }, + "cost": { + "input": 0.287, + "output": 1.147 } }, - "google/nano-banana": { - "id": "google/nano-banana", - "name": "Nano-Banana", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "nano-banana", - "attachment": true, + "qwen3-omni-flash-realtime": { + "id": "qwen3-omni-flash-realtime", + "name": "Qwen3-Omni Flash Realtime", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ "text", - "image" + "audio" ] }, "open_weights": false, "limit": { "context": 65536, - "output": 0 + "output": 16384 }, "cost": { - "input": 0.21, - "output": 1.8, - "cache_read": 0.021 + "input": 0.23, + "output": 0.918, + "input_audio": 3.584, + "output_audio": 7.168 } }, - "google/veo-3-fast": { - "id": "google/veo-3-fast", - "name": "Veo-3-Fast", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, - "reasoning": false, + "deepseek-r1-distill-llama-70b": { + "id": "deepseek-r1-distill-llama-70b", + "name": "DeepSeek R1 Distill Llama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-10-13", - "last_updated": "2025-10-13", + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" ], - "output": [ - "video" - ] - }, - "open_weights": false, - "limit": { - "context": 480, - "output": 0 - } - }, - "google/gemini-2.0-flash": { - "id": "google/gemini-2.0-flash", - "name": "Gemini-2.0-Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2024-12-11", - "last_updated": "2024-12-11", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], "output": [ "text" ] }, "open_weights": false, "limit": { - "context": 990000, - "output": 8192 + "context": 32768, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.42 + "input": 0.287, + "output": 0.861 } }, - "google/veo-3.1": { - "id": "google/veo-3.1", - "name": "Veo-3.1", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, - "reasoning": false, + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 38912 + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" ], "output": [ - "video" - ] - }, - "open_weights": false, - "limit": { - "context": 480, - "output": 0 - } - }, - "google/nano-banana-pro": { - "id": "google/nano-banana-pro", - "name": "Nano-Banana-Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "nano-banana", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 0 + "context": 131072, + "output": 16384 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 0.287, + "output": 1.147, + "reasoning": 2.868 } }, - "google/gemini-2.0-flash-lite": { - "id": "google/gemini-2.0-flash-lite", - "name": "Gemini-2.0-Flash-Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "qwen-omni-turbo-realtime": { + "id": "qwen-omni-turbo-realtime", + "name": "Qwen-Omni Turbo Realtime", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-02-05", - "last_updated": "2025-02-05", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-05-08", + "last_updated": "2025-05-08", "modalities": { "input": [ "text", "image", - "video", "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 990000, - "output": 8192 + "context": 32768, + "output": 2048 }, "cost": { - "input": 0.052, - "output": 0.21 + "input": 0.23, + "output": 0.918, + "input_audio": 3.584, + "output_audio": 7.168 } }, - "google/gemini-3-flash": { - "id": "google/gemini-3-flash", - "name": "Gemini-3-Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "high" - ] - } - ], + "qwen2-5-math-7b-instruct": { + "id": "qwen2-5-math-7b-instruct", + "name": "Qwen2.5-Math 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-10-07", - "last_updated": "2025-10-07", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 4096, + "output": 3072 }, "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.04 + "input": 0.144, + "output": 0.287 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini-3.5-Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "high" - ] - } - ], + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.5152, - "output": 9.0909, - "cache_read": 0.1515 + "input": 0.144, + "output": 0.574 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini-2.5-Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "budget_tokens", - "min": 0, - "max": 32768 + "max": 262144 } ], "tool_call": true, - "temperature": false, - "release_date": "2025-02-05", - "last_updated": "2025-02-05", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" @@ -5389,39 +4896,51 @@ }, "open_weights": false, "limit": { - "context": 1065535, - "output": 65535 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.87, - "output": 7, - "cache_read": 0.087 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini-2.5-Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "budget_tokens", - "min": 0, - "max": 24576 + "max": 262144 } ], "tool_call": true, - "temperature": false, - "release_date": "2025-04-26", - "last_updated": "2025-04-26", + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -5429,58 +4948,61 @@ }, "open_weights": false, "limit": { - "context": 1065535, - "output": 65535 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.21, - "output": 1.8, - "cache_read": 0.021 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "google/imagen-3": { - "id": "google/imagen-3", - "name": "Imagen-3", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", - "attachment": true, + "qwen-long": { + "id": "qwen-long", + "name": "Qwen Long", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-10-15", - "last_updated": "2024-10-15", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-01-25", + "last_updated": "2025-01-25", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 480, - "output": 0 + "context": 10000000, + "output": 8192 + }, + "cost": { + "input": 0.072, + "output": 0.287 } }, - "google/gemini-3-pro": { - "id": "google/gemini-3-pro", - "name": "Gemini-3-Pro", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen-math-turbo": { + "id": "qwen-math-turbo", + "name": "Qwen Math Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-10-22", - "last_updated": "2025-10-22", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -5488,256 +5010,291 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 4096, + "output": 3072 }, - "status": "deprecated", "cost": { - "input": 1.6, - "output": 9.6, - "cache_read": 0.16 + "input": 0.287, + "output": 0.861 } }, - "google/veo-2": { - "id": "google/veo-2", - "name": "Veo-2", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-12-02", - "last_updated": "2024-12-02", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 480, - "output": 0 + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.861, + "output": 3.441 } }, - "google/veo-3.1-fast": { - "id": "google/veo-3.1-fast", - "name": "Veo-3.1-Fast", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, + "qwen2-5-omni-7b": { + "id": "qwen2-5-omni-7b", + "name": "Qwen2.5-Omni 7B", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-12", + "last_updated": "2024-12", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ - "video" + "text", + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 32768, + "output": 2048 + }, + "cost": { + "input": 0.087, + "output": 0.345, + "input_audio": 5.448 } }, - "google/gemma-4-31b": { - "id": "google/gemma-4-31b", - "name": "Gemma-4-31B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": true, - "reasoning": false, + "qwen3-8b": { + "id": "qwen3-8b", + "name": "Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 38912 + } + ], "tool_call": true, - "temperature": false, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, + "context": 131072, "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.072, + "output": 0.287, + "reasoning": 0.717 } }, - "google/veo-3": { - "id": "google/veo-3", - "name": "Veo-3", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, + "qwen2-5-14b-instruct": { + "id": "qwen2-5-14b-instruct", + "name": "Qwen2.5 14B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-05-21", - "last_updated": "2025-05-21", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 0 + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.144, + "output": 0.431 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini-2.5-Flash-Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "budget_tokens", - "min": 0, - "max": 24576 + "max": 131072 } ], "tool_call": true, - "temperature": false, - "release_date": "2025-06-19", - "last_updated": "2025-06-19", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-14", + "last_updated": "2026-04-14", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1024000, - "output": 64000 + "context": 202752, + "output": 128000 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 0.87, + "output": 3.48, + "cache_read": 0.17 } }, - "google/gemini-3.1-pro": { - "id": "google/gemini-3.1-pro", - "name": "Gemini-3.1-Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "attachment": true, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "high" + "high", + "max" ] } ], "tool_call": true, - "temperature": false, - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini-3.1-Flash-Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "attachment": true, + "qwen3-next-80b-a3b-thinking": { + "id": "qwen3-next-80b-a3b-thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "high" - ] + "type": "budget_tokens" } ], "tool_call": true, - "temperature": false, - "release_date": "2026-02-18", - "last_updated": "2026-02-18", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 1.5 + "input": 0.144, + "output": 1.434 } }, - "xai/grok-3": { - "id": "xai/grok-3", - "name": "Grok 3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, + "qvq-max": { + "id": "qvq-max", + "name": "QVQ Max", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qvq", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-04-11", - "last_updated": "2025-04-11", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -5749,25 +5306,25 @@ "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 + "input": 1.147, + "output": 4.588 } }, - "xai/grok-4.20-multi-agent": { - "id": "xai/grok-4.20-multi-agent", - "name": "Grok-4.20-Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2026-03-13", - "last_updated": "2026-03-13", + "qwen-plus-character": { + "id": "qwen-plus-character", + "name": "Qwen Plus Character", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01", + "last_updated": "2024-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -5775,63 +5332,67 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 0 + "context": 32768, + "output": 4096 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 0.115, + "output": 0.287 } }, - "xai/grok-4": { - "id": "xai/grok-4", - "name": "Grok-4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Moonshot Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-07-10", - "last_updated": "2025-07-10", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 128000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 + "input": 0.574, + "output": 2.294 } }, - "xai/grok-4-fast-reasoning": { - "id": "xai/grok-4-fast-reasoning", - "name": "Grok-4-Fast-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-09-16", - "last_updated": "2025-09-16", + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -5839,31 +5400,41 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 128000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.574, + "output": 2.294 } }, - "xai/grok-4.1-fast-reasoning": { - "id": "xai/grok-4.1-fast-reasoning", - "name": "Grok-4.1-Fast-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "qwen3.5-flash": { + "id": "qwen3.5-flash", + "name": "Qwen3.5 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -5871,65 +5442,77 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.172, + "output": 1.72, + "reasoning": 1.72 } }, - "xai/grok-4-fast-non-reasoning": { - "id": "xai/grok-4-fast-non-reasoning", - "name": "Grok-4-Fast-Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": false, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-09-16", - "last_updated": "2025-09-16", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 1.1, + "output": 3.851, + "cache_read": 0.275, + "cache_write": 0 } }, - "xai/grok-3-mini": { - "id": "xai/grok-3-mini", - "name": "Grok 3 Mini", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 131072 } ], "tool_call": true, - "temperature": false, - "release_date": "2025-04-11", - "last_updated": "2025-04-11", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -5937,57 +5520,67 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 0.5, - "cache_read": 0.075 + "input": 0.1875, + "output": 1.125, + "cache_write": 0.234375 } }, - "xai/grok-code-fast-1": { - "id": "xai/grok-code-fast-1", - "name": "Grok Code Fast 1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen2-5-vl-72b-instruct": { + "id": "qwen2-5-vl-72b-instruct", + "name": "Qwen2.5-VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-08-22", - "last_updated": "2025-08-22", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 1.5, - "cache_read": 0.02 + "input": 2.294, + "output": 6.881 } }, - "xai/grok-4.1-fast-non-reasoning": { - "id": "xai/grok-4.1-fast-non-reasoning", - "name": "Grok-4.1-Fast-Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": false, + "qwen3-vl-plus": { + "id": "qwen3-vl-plus", + "name": "Qwen3-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", @@ -5999,126 +5592,162 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.143353, + "output": 1.433525, + "reasoning": 4.300576 } }, - "ideogramai/ideogram-v2": { - "id": "ideogramai/ideogram-v2", - "name": "Ideogram-v2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ideogram", - "attachment": true, + "qwen-vl-ocr": { + "id": "qwen-vl-ocr", + "name": "Qwen-VL OCR", + "description": "OCR model for extracting structured text from documents and screenshots", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2024-08-21", - "last_updated": "2024-08-21", + "tool_call": false, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-10-28", + "last_updated": "2025-04-13", "modalities": { "input": [ "text", "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 150, - "output": 0 + "context": 34096, + "output": 4096 + }, + "cost": { + "input": 0.717, + "output": 0.717 } }, - "ideogramai/ideogram-v2a-turbo": { - "id": "ideogramai/ideogram-v2a-turbo", - "name": "Ideogram-v2a-Turbo", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ideogram", - "attachment": true, + "qwen-mt-turbo": { + "id": "qwen-mt-turbo", + "name": "Qwen-MT Turbo", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "tool_call": false, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-01", + "last_updated": "2025-01", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 150, - "output": 0 + "context": 16384, + "output": 8192 + }, + "cost": { + "input": 0.101, + "output": 0.28 } }, - "ideogramai/ideogram-v2a": { - "id": "ideogramai/ideogram-v2a", - "name": "Ideogram-v2a", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ideogram", - "attachment": true, + "qwen-math-plus": { + "id": "qwen-math-plus", + "name": "Qwen Math Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-08-16", + "last_updated": "2024-09-19", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 150, - "output": 0 + "context": 4096, + "output": 3072 + }, + "cost": { + "input": 0.574, + "output": 1.721 } }, - "ideogramai/ideogram": { - "id": "ideogramai/ideogram", - "name": "Ideogram", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ideogram", - "attachment": true, + "qwen-mt-plus": { + "id": "qwen-mt-plus", + "name": "Qwen-MT Plus", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2024-04-03", - "last_updated": "2024-04-03", + "tool_call": false, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-01", + "last_updated": "2025-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 150, - "output": 0 + "context": 16384, + "output": 8192 + }, + "cost": { + "input": 0.259, + "output": 0.775 } }, - "fireworks-ai/kimi-k2.5-fw": { - "id": "fireworks-ai/kimi-k2.5-fw", - "name": "Kimi-K2.5-FW", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "attachment": true, - "reasoning": false, + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], "tool_call": true, - "temperature": false, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -6126,78 +5755,93 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 245760, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.573, + "output": 3.44, + "reasoning": 3.44 } }, - "runwayml/runway": { - "id": "runwayml/runway", - "name": "Runway", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "runway", - "attachment": true, + "qwen-omni-turbo": { + "id": "qwen-omni-turbo", + "name": "Qwen-Omni Turbo", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-10-11", - "last_updated": "2024-10-11", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-01-19", + "last_updated": "2025-03-26", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ - "video" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 256, - "output": 0 + "context": 32768, + "output": 2048 + }, + "cost": { + "input": 0.058, + "output": 0.23, + "input_audio": 3.584, + "output_audio": 7.168 } }, - "runwayml/runway-gen-4-turbo": { - "id": "runwayml/runway-gen-4-turbo", - "name": "Runway-Gen-4-Turbo", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "runway", - "attachment": true, + "qwen2-5-72b-instruct": { + "id": "qwen2-5-72b-instruct", + "name": "Qwen2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-05-09", - "last_updated": "2025-05-09", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256, - "output": 0 + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.574, + "output": 1.721 } }, - "trytako/tako": { - "id": "trytako/tako", - "name": "Tako", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "tako", - "attachment": true, - "reasoning": false, + "deepseek-r1-distill-qwen-7b": { + "id": "deepseek-r1-distill-qwen-7b", + "name": "DeepSeek R1 Distill Qwen 7B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -6208,186 +5852,186 @@ }, "open_weights": false, "limit": { - "context": 2048, - "output": 0 + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 0.072, + "output": 0.144 } }, - "stabilityai/stablediffusionxl": { - "id": "stabilityai/stablediffusionxl", - "name": "StableDiffusionXL", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "stable-diffusion", - "attachment": true, + "deepseek-v3-1": { + "id": "deepseek-v3-1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2023-07-09", - "last_updated": "2023-07-09", + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 200, - "output": 0 + "context": 131072, + "output": 65536 + }, + "cost": { + "input": 0.574, + "output": 1.721 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen2-5-coder-32b-instruct": { + "id": "qwen2-5-coder-32b-instruct", + "name": "Qwen2.5-Coder 32B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-11", + "last_updated": "2024-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 2.2, - "output": 14, - "cache_read": 0.22 + "input": 0.287, + "output": 0.861 } }, - "openai/dall-e-3": { - "id": "openai/dall-e-3", - "name": "DALL-E-3", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "dall-e", - "attachment": true, - "reasoning": false, + "qwen-flash": { + "id": "qwen-flash", + "name": "Qwen Flash", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], "tool_call": true, - "temperature": false, - "release_date": "2023-11-06", - "last_updated": "2023-11-06", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 800, - "output": 0 + "context": 1000000, + "output": 32768 + }, + "cost": { + "input": 0.022, + "output": 0.216 } }, - "openai/sora-2": { - "id": "openai/sora-2", - "name": "Sora-2", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "sora", - "attachment": true, + "deepseek-v3-2-exp": { + "id": "deepseek-v3-2-exp", + "name": "DeepSeek V3.2 Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "output": 65536 + }, + "cost": { + "input": 0.287, + "output": 0.431 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4-Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "moonshot-kimi-k2-instruct": { + "id": "moonshot-kimi-k2-instruct", + "name": "Moonshot Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": false, + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 27, - "output": 160 + "input": 0.574, + "output": 2.294 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5-Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, + "deepseek-r1-distill-qwen-14b": { + "id": "deepseek-r1-distill-qwen-14b", + "name": "DeepSeek R1 Distill Qwen 14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2023-09-13", - "last_updated": "2023-09-13", + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -6395,83 +6039,62 @@ }, "open_weights": false, "limit": { - "context": 16384, - "output": 2048 + "context": 32768, + "output": 16384 }, "cost": { - "input": 0.45, - "output": 1.4 + "input": 0.144, + "output": 0.431 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5-Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, + "qwen3-vl-235b-a22b": { + "id": "qwen3-vl-235b-a22b", + "name": "Qwen3-VL 235B-A22B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 27.2727, - "output": 163.6364 + "input": 0.286705, + "output": 1.14682, + "reasoning": 2.867051 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4-Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen-deep-research": { + "id": "qwen-deep-research", + "name": "Qwen Deep Research", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01", + "last_updated": "2024-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -6479,78 +6102,72 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.18, - "output": 1.1, - "cache_read": 0.018 + "input": 7.742, + "output": 23.367 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Moonshot Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, - "temperature": false, - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.6, - "output": 13, - "cache_read": 0.16 + "input": 0.574, + "output": 2.411 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "qwen3-vl-30b-a3b": { + "id": "qwen3-vl-30b-a3b", + "name": "Qwen3-VL 30B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-11-12", - "last_updated": "2025-11-12", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text", @@ -6560,28 +6177,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 0.108, + "output": 0.431, + "reasoning": 1.076 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "qwen-vl-max": { + "id": "qwen-vl-max", + "name": "Qwen-VL Max", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-08", + "last_updated": "2025-08-13", "modalities": { "input": [ "text", @@ -6593,21 +6211,26 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 131072, "output": 8192 + }, + "cost": { + "input": 0.23, + "output": 0.574 } }, - "openai/gpt-4o-search": { - "id": "openai/gpt-4o-search", - "name": "GPT-4o-Search", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "deepseek-r1-distill-qwen-1-5b": { + "id": "deepseek-r1-distill-qwen-1-5b", + "name": "DeepSeek R1 Distill Qwen 1.5B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-03-11", - "last_updated": "2025-03-11", + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -6618,257 +6241,238 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 32768, + "output": 16384 }, "cost": { - "input": 2.2, - "output": 9 + "input": 0, + "output": 0 } }, - "openai/sora-2-pro": { - "id": "openai/sora-2-pro", - "name": "Sora-2-Pro", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "sora", - "attachment": true, + "qwen-max": { + "id": "qwen-max", + "name": "Qwen Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.345, + "output": 1.377 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.1, - "output": 9 + "input": 0.3, + "output": 1.2 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "qwen3-235b-a22b": { + "id": "qwen3-235b-a22b", + "name": "Qwen3 235B-A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 38912 } ], "tool_call": true, - "temperature": false, - "release_date": "2025-06-25", - "last_updated": "2025-06-25", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.22, - "output": 1.8, - "cache_read": 0.022 + "input": 0.287, + "output": 1.147, + "reasoning": 2.868 } }, - "openai/gpt-image-2": { - "id": "openai/gpt-image-2", - "name": "GPT-Image-2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "deepseek-r1-distill-llama-8b": { + "id": "deepseek-r1-distill-llama-8b", + "name": "DeepSeek R1 Distill Llama 8B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 32768, + "output": 16384 }, "cost": { - "input": 5.0505, - "output": 32.3232, - "cache_read": 1.2626 + "input": 0, + "output": 0 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2-Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Moonshot Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, - "temperature": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 19, - "output": 150 + "input": 0.929, + "output": 3.858 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "qwen3-coder-480b-a35b-instruct": { + "id": "qwen3-coder-480b-a35b-instruct", + "name": "Qwen3-Coder 480B-A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.99, - "output": 4, - "cache_read": 0.25 + "input": 0.861, + "output": 3.441 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": true, + "qwq-plus": { + "id": "qwq-plus", + "name": "QwQ Plus", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-03-05", + "last_updated": "2025-03-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -6876,98 +6480,98 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.99, - "output": 4 + "input": 0.23, + "output": 0.574 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen2-5-32b-instruct": { + "id": "qwen2-5-32b-instruct", + "name": "Qwen2.5 32B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2026-02-10", - "last_updated": "2026-02-10", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 1.6, - "output": 13, - "cache_read": 0.16 + "input": 0.287, + "output": 0.861 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], "tool_call": true, - "temperature": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.09, - "output": 0.36, - "cache_read": 0.022 + "input": 0.43, + "output": 2.58, + "reasoning": 2.58 } }, - "openai/o3-deep-research": { - "id": "openai/o3-deep-research", - "name": "o3-deep-research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "release_date": "2025-06-27", - "last_updated": "2025-06-27", + "tongyi-intent-detect-v3": { + "id": "tongyi-intent-detect-v3", + "name": "Tongyi Intent Detect V3", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "yi", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01", + "last_updated": "2024-01", "modalities": { "input": [ "text" @@ -6978,41 +6582,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 8192, + "output": 1024 }, "cost": { - "input": 9, - "output": 36, - "cache_read": 2.2 + "input": 0.058, + "output": 0.144 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "qwen3-coder-flash": { + "id": "qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7020,29 +6612,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.045, - "output": 0.36, - "cache_read": 0.0045 + "input": 0.144, + "output": 0.574 } }, - "openai/gpt-5.2-instant": { - "id": "openai/gpt-5.2-instant", - "name": "GPT-5.2-Instant", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, + "deepseek-r1-distill-qwen-32b": { + "id": "deepseek-r1-distill-qwen-32b", + "name": "DeepSeek R1 Distill Qwen 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7050,102 +6642,100 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 32768, "output": 16384 }, "cost": { - "input": 1.6, - "output": 13, - "cache_read": 0.16 + "input": 0.287, + "output": 0.861 } }, - "openai/o1": { - "id": "openai/o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, + "qwq-32b": { + "id": "qwq-32b", + "name": "QwQ 32B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2024-12-18", - "last_updated": "2024-12-18", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-12", + "last_updated": "2024-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 14, - "output": 54 + "input": 0.287, + "output": 0.861 } }, - "openai/chatgpt-4o-latest": { - "id": "openai/chatgpt-4o-latest", - "name": "ChatGPT-4o-Latest", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gpt", - "attachment": true, - "reasoning": false, + "qwen3-14b": { + "id": "qwen3-14b", + "name": "Qwen3 14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 38912 + } + ], "tool_call": true, - "temperature": false, - "release_date": "2024-08-14", - "last_updated": "2024-08-14", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 131072, "output": 8192 }, - "status": "deprecated", "cost": { - "input": 4.5, - "output": 14 + "input": 0.144, + "output": 0.574, + "reasoning": 1.434 } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5-Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "qwen3-asr-flash": { + "id": "qwen3-asr-flash", + "name": "Qwen3-ASR Flash", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "knowledge": "2024-04", + "release_date": "2025-09-08", + "last_updated": "2025-09-08", "modalities": { "input": [ - "text", - "image" + "audio" ], "output": [ "text" @@ -7153,29 +6743,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 53248, + "output": 4096 }, "cost": { - "input": 14, - "output": 110 + "input": 0.032, + "output": 0.032 } }, - "openai/gpt-3.5-turbo-instruct": { - "id": "openai/gpt-3.5-turbo-instruct", - "name": "GPT-3.5-Turbo-Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "qwen-doc-turbo": { + "id": "qwen-doc-turbo", + "name": "Qwen Doc Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2023-09-20", - "last_updated": "2023-09-20", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01", + "last_updated": "2024-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7183,25 +6773,37 @@ }, "open_weights": false, "limit": { - "context": 3500, - "output": 1024 + "context": 131072, + "output": 8192 }, "cost": { - "input": 1.4, - "output": 1.8 + "input": 0.087, + "output": 0.144 } }, - "openai/gpt-5.3-codex-spark": { - "id": "openai/gpt-5.3-codex-spark", - "name": "GPT-5.3-Codex-Spark", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 32768 + } + ], "tool_call": true, - "temperature": false, - "release_date": "2026-03-04", - "last_updated": "2026-03-04", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -7212,40 +6814,38 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 202752, "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.86, + "output": 3.15 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "qwen-turbo": { + "id": "qwen-turbo", + "name": "Qwen Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 38912 } ], "tool_call": true, - "temperature": false, - "release_date": "2025-11-12", - "last_updated": "2025-11-12", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-11-01", + "last_updated": "2025-07-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7253,57 +6853,57 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 16384 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 0.044, + "output": 0.087, + "reasoning": 0.431 } }, - "openai/gpt-5-chat": { - "id": "openai/gpt-5-chat", - "name": "GPT-5-Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, + "qwen2-5-7b-instruct": { + "id": "qwen2-5-7b-instruct", + "name": "Qwen2.5 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 8192 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 0.072, + "output": 0.144 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "qwen2-5-vl-7b-instruct": { + "id": "qwen2-5-vl-7b-instruct", + "name": "Qwen2.5-VL 7B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text", @@ -7313,42 +6913,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 124096, - "output": 4096 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.14, - "output": 0.54, - "cache_read": 0.068 + "input": 0.287, + "output": 0.717 } }, - "openai/o3-pro": { - "id": "openai/o3-pro", - "name": "o3-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": true, + "qwen3.6-max-preview": { + "id": "qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 131072 } ], "tool_call": true, - "temperature": false, - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-20", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7356,24 +6954,27 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 245800, + "output": 65536 }, "cost": { - "input": 18, - "output": 72 + "input": 1.32, + "output": 7.9, + "cache_read": 0.132 } }, - "openai/gpt-5.3-instant": { - "id": "openai/gpt-5.3-instant", - "name": "GPT-5.3-Instant", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "qwen-vl-plus": { + "id": "qwen-vl-plus", + "name": "Qwen-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-08-15", "modalities": { "input": [ "text", @@ -7385,31 +6986,40 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 111616, - "output": 16384 + "context": 131072, + "output": 8192 }, "cost": { - "input": 1.6, - "output": 13, - "cache_read": 0.16 + "input": 0.115, + "output": 0.287 } }, - "openai/gpt-4o-aug": { - "id": "openai/gpt-4o-aug", - "name": "GPT-4o-Aug", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], "tool_call": true, - "temperature": false, - "release_date": "2024-11-21", - "last_updated": "2024-11-21", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -7417,172 +7027,159 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 2.2, - "output": 9, - "cache_read": 1.1 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "openai/o3-mini-high": { - "id": "openai/o3-mini-high", - "name": "o3-mini-high", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": true, + "MiniMax/MiniMax-M2.7": { + "id": "MiniMax/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-01-31", - "last_updated": "2025-01-31", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.99, - "output": 4 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "openai/gpt-5.1-instant": { - "id": "openai/gpt-5.1-instant", - "name": "GPT-5.1-Instant", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "kimi/kimi-k2.5": { + "id": "kimi/kimi-k2.5", + "name": "kimi/kimi-k2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": false, - "release_date": "2025-11-12", - "last_updated": "2025-11-12", + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "openai/gpt-image-1": { - "id": "openai/gpt-image-1", - "name": "GPT-Image-1", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", - "attachment": true, - "reasoning": false, + "siliconflow/deepseek-r1-0528": { + "id": "siliconflow/deepseek-r1-0528", + "name": "siliconflow/deepseek-r1-0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-03-31", - "last_updated": "2025-03-31", + "structured_output": true, + "temperature": true, + "release_date": "2025-05-28", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 0 + "context": 163840, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 2.18 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4-Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "temperature": false, - "release_date": "2026-03-12", - "last_updated": "2026-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.68, - "output": 4, - "cache_read": 0.068 - } - }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT-5.1-Codex-Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, + "siliconflow/deepseek-v3.1-terminus": { + "id": "siliconflow/deepseek-v3.1-terminus", + "name": "siliconflow/deepseek-v3.1-terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "structured_output": true, + "temperature": true, + "release_date": "2025-09-29", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7590,41 +7187,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 163840, + "output": 65536 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 0.27, + "output": 1 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "siliconflow/deepseek-v3-0324": { + "id": "siliconflow/deepseek-v3-0324", + "name": "siliconflow/deepseek-v3-0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "structured_output": true, + "temperature": true, + "release_date": "2024-12-26", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7632,73 +7217,61 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 163840, + "output": 163840 }, "cost": { - "input": 1.6, - "output": 13, - "cache_read": 0.16 + "input": 0.25, + "output": 1 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "siliconflow/deepseek-v3.2": { + "id": "siliconflow/deepseek-v3.2", + "name": "siliconflow/deepseek-v3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "temperature": true, + "release_date": "2025-12-03", + "last_updated": "2025-12-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 163840, + "output": 65536 }, "cost": { - "input": 4.5455, - "output": 27.2727, - "cache_read": 0.4545 + "input": 0.27, + "output": 0.42 } }, - "openai/o4-mini-deep-research": { - "id": "openai/o4-mini-deep-research", - "name": "o4-mini-deep-research", - "description": "Research model for long-horizon investigation, synthesis, and analytical reports", - "family": "o-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-06-27", - "last_updated": "2025-06-27", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -7707,32 +7280,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.8, - "output": 7.2, - "cache_read": 0.45 + "input": 1, + "output": 5 } - }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + } + } + }, + "regolo-ai": { + "id": "regolo-ai", + "env": [ + "REGOLO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.regolo.ai/v1", + "name": "Regolo AI", + "doc": "https://docs.regolo.ai/", + "models": { + "llama-3.1-8b-instruct": { + "id": "llama-3.1-8b-instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "temperature": true, + "release_date": "2025-04-07", + "last_updated": "2025-04-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7740,30 +7323,29 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 120000, + "output": 120000 }, "cost": { - "input": 1.8, - "output": 7.2, - "cache_read": 0.45 + "input": 0.05, + "output": 0.25 } }, - "openai/gpt-4-classic-0314": { - "id": "openai/gpt-4-classic-0314", - "name": "GPT-4-Classic-0314", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gpt", - "attachment": true, - "reasoning": false, + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax 2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2024-08-26", - "last_updated": "2024-08-26", + "temperature": true, + "release_date": "2026-03-10", + "last_updated": "2026-03-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7771,30 +7353,29 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 190000, + "output": 64000 }, - "status": "deprecated", "cost": { - "input": 27, - "output": 54 + "input": 0.8, + "output": 3.5 } }, - "openai/gpt-4-classic": { - "id": "openai/gpt-4-classic", - "name": "GPT-4-Classic", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gpt", - "attachment": true, - "reasoning": false, + "mistral-small3.2": { + "id": "mistral-small3.2", + "name": "Mistral Small 3.2", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2024-03-25", - "last_updated": "2024-03-25", + "temperature": true, + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7802,26 +7383,25 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 120000, + "output": 120000 }, - "status": "deprecated", "cost": { - "input": 27, - "output": 54 + "input": 0.5, + "output": 2.2 } }, - "openai/gpt-4o-mini-search": { - "id": "openai/gpt-4o-mini-search", - "name": "GPT-4o-mini-Search", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "qwen3-reranker-4b": { + "id": "qwen3-reranker-4b", + "name": "Qwen3-Reranker-4B", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": false, - "release_date": "2025-03-11", - "last_updated": "2025-03-11", + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ "text" @@ -7830,61 +7410,59 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 32768, "output": 8192 }, "cost": { - "input": 0.14, - "output": 0.54 + "input": 0.12, + "output": 0.12 } }, - "openai/gpt-3.5-turbo-raw": { - "id": "openai/gpt-3.5-turbo-raw", - "name": "GPT-3.5-Turbo-Raw", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "qwen3-embedding-8b": { + "id": "qwen3-embedding-8b", + "name": "Qwen3-Embedding-8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": false, - "release_date": "2023-09-27", - "last_updated": "2023-09-27", + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4524, - "output": 2048 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.45, - "output": 1.4 + "input": 0.1, + "output": 0.1 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "temperature": true, + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -7892,55 +7470,28 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.36, - "output": 1.4, - "cache_read": 0.09 - } - }, - "openai/gpt-image-1-mini": { - "id": "openai/gpt-image-1-mini", - "name": "GPT-Image-1-Mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "input": 0.6, + "output": 2.7 } }, - "openai/gpt-image-1.5": { - "id": "openai/gpt-image-1.5", - "name": "gpt-image-1.5", + "qwen-image": { + "id": "qwen-image", + "name": "Qwen-Image", "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "temperature": true, + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "image" @@ -7948,86 +7499,51 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 0 - } - }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4-Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2023-09-13", - "last_updated": "2023-09-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, + "context": 8192, "output": 4096 }, "cost": { - "input": 9, - "output": 27 + "input": 0.5, + "output": 2 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1-Codex-Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "qwen3.5-122b": { + "id": "qwen3.5-122b", + "name": "Qwen3.5-122B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-11-12", - "last_updated": "2025-11-12", + "temperature": true, + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.22, - "output": 1.8, - "cache_read": 0.022 + "input": 0.9, + "output": 3.6 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT-OSS-120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -8040,13 +7556,12 @@ } ], "tool_call": true, - "temperature": false, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -8054,78 +7569,56 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.8, - "output": 7.2, - "cache_read": 0.45 + "input": 1, + "output": 4.2 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "Qwen3-Coder-Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": true, + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 1.1, - "output": 9, - "cache_read": 0.11 + "input": 0.3, + "output": 1.2 } }, - "openai/o1-pro": { - "id": "openai/o1-pro", - "name": "o1-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5-9B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2025-03-19", - "last_updated": "2025-03-19", + "temperature": true, + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ "text", @@ -8135,75 +7628,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 8192 }, "cost": { - "input": 140, - "output": 540 + "input": 0.15, + "output": 0.6 } }, - "topazlabs-co/topazlabs": { - "id": "topazlabs-co/topazlabs", - "name": "TopazLabs", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "topazlabs", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": false, - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 204, - "output": 0 - } - } - } - }, - "cerebras": { - "id": "cerebras", - "env": [ - "CEREBRAS_API_KEY" - ], - "npm": "@ai-sdk/cerebras", - "name": "Cerebras", - "doc": "https://inference-docs.cerebras.ai/models/overview", - "models": { - "qwen-3.8-27b": { - "id": "qwen-3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", + "mistral-small-4-119b": { + "id": "mistral-small-4-119b", + "name": "Mistral Small 4 119B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-09-03", + "release_date": "2026-03-15", + "last_updated": "2026-03-15", "modalities": { "input": [ "text", @@ -8213,20 +7659,20 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 32768 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0.99, - "output": 1.49 + "input": 0.75, + "output": 3 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT-OSS-20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "family": "gpt-oss", "attachment": false, "reasoning": true, @@ -8241,10 +7687,9 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2026-06-10", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ "text" @@ -8255,39 +7700,41 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 40960 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.35, - "output": 0.75 + "input": 0.4, + "output": 1.8 } } } }, - "groq": { - "id": "groq", + "stackit": { + "id": "stackit", "env": [ - "GROQ_API_KEY" + "STACKIT_API_KEY" ], - "npm": "@ai-sdk/groq", - "name": "Groq", - "doc": "https://console.groq.com/docs/models", + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.openai-compat.model-serving.eu01.onstackit.cloud/v1", + "name": "STACKIT", + "doc": "https://docs.stackit.cloud/products/data-and-ai/ai-model-serving/basics/available-shared-models", "models": { - "whisper-large-v3": { - "id": "whisper-large-v3", - "name": "Whisper", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", + "cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic": { + "id": "cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic", + "name": "Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2023-09-01", - "last_updated": "2025-09-05", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" @@ -8295,25 +7742,30 @@ }, "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.53, + "output": 0.76 } }, - "llama-3.3-70b-versatile": { - "id": "llama-3.3-70b-versatile", - "name": "Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma 3 27B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2025-05-17", + "last_updated": "2025-05-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -8321,24 +7773,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 37000, + "output": 4096 }, "cost": { - "input": 0.59, - "output": 0.79 + "input": 0.53, + "output": 0.76 } }, - "allam-2-7b": { - "id": "allam-2-7b", - "name": "ALLaM-2-7b", - "description": "ALLaM-2-7b instruction tuned model by SDAIA", + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-01-23", - "last_updated": "2025-01-23", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -8349,28 +7803,30 @@ }, "open_weights": true, "limit": { - "context": 4096, - "output": 4096 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.53, + "output": 0.76 } }, - "whisper-large-v3-turbo": { - "id": "whisper-large-v3-turbo", - "name": "Whisper Large V3 Turbo", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, + "Qwen/Qwen3-VL-Embedding-8B": { + "id": "Qwen/Qwen3-VL-Embedding-8B", + "name": "Qwen3-VL Embedding 8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", + "structured_output": false, + "temperature": false, + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "audio" + "text", + "image" ], "output": [ "text" @@ -8378,25 +7834,30 @@ }, "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 32000, + "output": 4096 + }, + "cost": { + "input": 0.09, + "output": 0.09 } }, - "llama-3.1-8b-instant": { - "id": "llama-3.1-8b-instant", - "name": "Llama 3.1 8B", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", - "attachment": false, + "Qwen/Qwen3-VL-235B-A22B-Instruct-FP8": { + "id": "Qwen/Qwen3-VL-235B-A22B-Instruct-FP8", + "name": "Qwen3-VL 235B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -8404,25 +7865,36 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 218000, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.08 + "input": 1.76, + "output": 2.05 } }, - "groq/compound-mini": { - "id": "groq/compound-mini", - "name": "Compound Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "groq", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-09-04", - "last_updated": "2025-09-04", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -8431,23 +7903,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, + "context": 131000, "output": 8192 + }, + "cost": { + "input": 0.53, + "output": 0.76 } }, - "groq/compound": { - "id": "groq/compound", - "name": "Compound", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "groq", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-09-04", - "last_updated": "2025-09-04", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -8456,23 +7934,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, "output": 8192 + }, + "cost": { + "input": 0.18, + "output": 0.29 } }, - "meta-llama/llama-prompt-guard-2-86m": { - "id": "meta-llama/llama-prompt-guard-2-86m", - "name": "Prompt Guard 2 86M", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", + "intfloat/e5-mistral-7b-instruct": { + "id": "intfloat/e5-mistral-7b-instruct", + "name": "E5 Mistral 7B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "mistral", "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": false, "temperature": false, - "release_date": "2025-05-29", - "last_updated": "2025-05-29", + "release_date": "2023-12-11", + "last_updated": "2023-12-11", "modalities": { "input": [ "text" @@ -8483,102 +7966,121 @@ }, "open_weights": true, "limit": { - "context": 512, - "output": 512 + "context": 4096, + "output": 4096 }, - "status": "beta", "cost": { - "input": 0.04, - "output": 0.04 + "input": 0.02, + "output": 0.02 } - }, - "meta-llama/llama-prompt-guard-2-22m": { - "id": "meta-llama/llama-prompt-guard-2-22m", - "name": "Llama Prompt Guard 2 22M", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", + } + } + }, + "vercel": { + "id": "vercel", + "env": [ + "AI_GATEWAY_API_KEY" + ], + "npm": "@ai-sdk/gateway", + "name": "Vercel AI Gateway", + "doc": "https://github.com/vercel/ai/tree/5eb85cc45a259553501f535b8ac79a77d0e79223/packages/gateway", + "models": { + "xai/grok-imagine-video-1.5": { + "id": "xai/grok-imagine-video-1.5", + "name": "Grok Imagine Video 1.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-05-29", - "last_updated": "2025-05-29", + "temperature": true, + "release_date": "2026-06-22", + "last_updated": "2026-06-22", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512, - "output": 512 - }, - "status": "beta", - "cost": { - "input": 0.03, - "output": 0.03 + "context": 0, + "output": 0 } }, - "canopylabs/orpheus-v1-english": { - "id": "canopylabs/orpheus-v1-english", - "name": "Canopy Labs Orpheus V1 English", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "canopylabs", + "xai/grok-4.1-fast-reasoning": { + "id": "xai/grok-4.1-fast-reasoning", + "name": "Grok 4.1 Fast Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-12-19", - "last_updated": "2025-12-19", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-11-19", + "last_updated": "2025-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 4000, - "output": 50000 + "context": 1000000, + "output": 1000000 }, - "status": "beta" + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } }, - "canopylabs/orpheus-arabic-saudi": { - "id": "canopylabs/orpheus-arabic-saudi", - "name": "Canopy Labs Orpheus Arabic Saudi", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "canopylabs", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "xai/grok-4.20-non-reasoning-beta": { + "id": "xai/grok-4.20-non-reasoning-beta", + "name": "Grok 4.20 Beta Non-Reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 4000, - "output": 50000 + "context": 2000000, + "output": 2000000 }, - "status": "beta" + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.4 + } }, - "qwen/qwen3.8-27b": { - "id": "qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", + "xai/grok-4.3": { + "id": "xai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -8586,7 +8088,6 @@ "type": "effort", "values": [ "none", - "default", "low", "medium", "high" @@ -8596,114 +8097,144 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-04-30", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131042, - "output": 16384 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.8, - "output": 4 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "qwen/qwen3.6-27b": { - "id": "qwen/qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "default" - ] - } - ], + "xai/grok-tts": { + "id": "xai/grok-tts", + "name": "Grok TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "xai/grok-4.1-fast-non-reasoning": { + "id": "xai/grok-4.1-fast-non-reasoning", + "name": "Grok 4.1 Fast Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2024-10", + "release_date": "2025-11-19", + "last_updated": "2025-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.3 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "xai/grok-voice-think-fast-1.0": { + "id": "xai/grok-voice-think-fast-1.0", + "name": "Grok Voice Think Fast 1.0", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "grok", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-09-25", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 0, + "output": 0 + } + }, + "xai/grok-imagine-video": { + "id": "xai/grok-imagine-video", + "name": "Grok Imagine", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-01-28", + "last_updated": "2026-01-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "video" + ] }, - "cost": { - "input": 0.075, - "output": 0.3, - "cache_read": 0.0375 + "open_weights": false, + "limit": { + "context": 0, + "output": 0 } }, - "openai/gpt-oss-safeguard-20b": { - "id": "openai/gpt-oss-safeguard-20b", - "name": "Safety GPT OSS 20B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", + "xai/grok-4.20-multi-agent-beta": { + "id": "xai/grok-4.20-multi-agent-beta", + "name": "Grok 4.20 Multi Agent Beta", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -8712,40 +8243,67 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2026-06-29", + "release_date": "2026-03-11", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 2000000, + "output": 2000000 }, - "status": "beta", "cost": { - "input": 0.075, - "output": 0.3 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "xai/grok-stt": { + "id": "xai/grok-stt", + "name": "Grok STT", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "grok", "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "xai/grok-4.5": { + "id": "xai/grok-4.5", + "name": "Grok 4.5", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -8760,205 +8318,135 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-10-21", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 2, + "output": 6, + "cache_read": 0.5 } - } - } - }, - "blueclaw": { - "id": "blueclaw", - "env": [ - "BLUECLAW_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://openai.blueclaw.network/v1", - "name": "Blue Claw", - "doc": "https://blueclaw.network", - "models": { - "Qwen3.6-27B": { - "id": "Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 65536 - }, - "status": "beta" }, - "Qwen/Qwen3.6-35B-A3B-FP8": { - "id": "Qwen/Qwen3.6-35B-A3B-FP8", - "name": "Qwen3.6 35B A3B FP8", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": false, + "xai/grok-4.20-reasoning": { + "id": "xai/grok-4.20-reasoning", + "name": "Grok 4.20 Reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-03-10", + "last_updated": "2026-03-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 2000000, + "output": 2000000 }, - "status": "beta" - } - } - }, - "zai": { - "id": "zai", - "env": [ - "ZHIPU_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.z.ai/api/paas/v4", - "name": "Z.AI", - "doc": "https://docs.z.ai/guides/overview/pricing", - "models": { - "glm-4.6v": { - "id": "glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "cost": { + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 + } + }, + "xai/grok-4.20-reasoning-beta": { + "id": "xai/grok-4.20-reasoning-beta", + "name": "Grok 4.20 Beta Reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2026-03-11", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "glm-4.5": { - "id": "glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", + "xai/grok-imagine-video-1.5-preview": { + "id": "xai/grok-imagine-video-1.5-preview", + "name": "Grok Imagine Video 1.5 Preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-05-30", + "last_updated": "2026-05-30", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "context": 0, + "output": 0 } }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", + "xai/grok-4.20-non-reasoning": { + "id": "xai/grok-4.20-non-reasoning", + "name": "Grok 4.20 Non-Reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "release_date": "2026-03-10", + "last_updated": "2026-03-23", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ @@ -8967,120 +8455,128 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24, - "cache_write": 0 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + "xai/grok-4.20-multi-agent": { + "id": "xai/grok-4.20-multi-agent", + "name": "Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-03-10", + "last_updated": "2026-03-23", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03, - "cache_write": 0 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "glm-4.6": { - "id": "glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", + "xai/grok-imagine-image": { + "id": "xai/grok-imagine-image", + "name": "Grok Imagine Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-01-28", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "xai/grok-build-0.1": { + "id": "xai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2026-05-20", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 1, + "output": 2, + "cache_read": 0.2 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "moonshotai/kimi-k2": { + "id": "moonshotai/kimi-k2", + "name": "Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2025-07-11", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -9089,124 +8585,107 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, + "context": 131072, "output": 131072 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0 + "input": 0.57, + "output": 2.3 } }, - "glm-4.6v-flash": { - "id": "glm-4.6v-flash", - "name": "GLM-4.6V-Flash", - "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 256000, "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "glm-4.5v": { - "id": "glm-4.5v", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 64000, - "output": 16384 + "context": 216144, + "output": 216144 }, "cost": { - "input": 0.6, - "output": 1.8 + "input": 0.47, + "output": 2, + "cache_read": 0.141 } }, - "glm-5.3-flashx": { - "id": "glm-5.3-flashx", - "name": "GLM-5.3-FlashX", - "description": "High-speed GLM-5.3-Flash serving option for coding and agent workflows", - "family": "glm-flash", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-18", - "last_updated": "2026-09-18", + "knowledge": "2025-01", + "release_date": "2026-01-26", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" @@ -9214,22 +8693,21 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262114, + "output": 262114 }, "cost": { - "input": 0.37, - "output": 1.25, - "cache_read": 0.075, - "cache_write": 0 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "glm-4.7-flash": { - "id": "glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -9237,13 +8715,15 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2025-01", + "release_date": "2026-04-20", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -9251,421 +8731,264 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, + "moonshotai/kimi-k2.7-code-highspeed": { + "id": "moonshotai/kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code High Speed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2025-01", + "release_date": "2026-06-15", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 1.9, + "output": 8, + "cache_read": 0.38 } }, - "glm-4.5-flash": { - "id": "glm-4.5-flash", - "name": "GLM-4.5-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "klingai/kling-v3.0-motion-control": { + "id": "klingai/kling-v3.0-motion-control", + "name": "Kling v3.0 Motion Control", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ling", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-03-04", + "last_updated": "2026-03-04", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "context": 0, + "output": 0 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "klingai/kling-v2.6-i2v": { + "id": "klingai/kling-v2.6-i2v", + "name": "Kling v2.6 Image-to-Video", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ling", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2025-12-03", + "last_updated": "2025-12-21", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "context": 0, + "output": 0 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "klingai/kling-v2.5-turbo-t2v": { + "id": "klingai/kling-v2.5-turbo-t2v", + "name": "Kling v2.5 Turbo Text-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ling", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "context": 0, + "output": 0 } }, - "glm-4.7-flashx": { - "id": "glm-4.7-flashx", - "name": "GLM-4.7-FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "klingai/kling-v3.0-i2v": { + "id": "klingai/kling-v3.0-i2v", + "name": "Kling v3.0 Image-to-Video", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ling", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0 + "context": 0, + "output": 0 } }, - "glm-5-turbo": { - "id": "glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", + "klingai/kling-v2.5-turbo-i2v": { + "id": "klingai/kling-v2.5-turbo-i2v", + "name": "Kling v2.5 Turbo Image-to-Video", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ling", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24, - "cache_write": 0 + "context": 0, + "output": 0 } }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "klingai/kling-v3.0-t2v": { + "id": "klingai/kling-v3.0-t2v", + "name": "Kling v3.0 Text-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ling", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 - }, - "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03, - "cache_write": 0 + "context": 0, + "output": 0 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "klingai/kling-v2.6-motion-control": { + "id": "klingai/kling-v2.6-motion-control", + "name": "Kling v2.6 Motion Control", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ling", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2025-12-18", + "last_updated": "2025-12-21", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "context": 0, + "output": 0 } - } - } - }, - "empiriolabs": { - "id": "empiriolabs", - "env": [ - "EMPIRIOLABS_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.empiriolabs.ai/v1", - "name": "EmpirioLabs AI", - "doc": "https://docs.empiriolabs.ai", - "models": { - "qwen3-8-max-0902": { - "id": "qwen3-8-max-0902", - "name": "Qwen3.8 Max 0902", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 - } - ], - "tool_call": true, - "structured_output": true, + }, + "klingai/kling-v2.6-t2v": { + "id": "klingai/kling-v2.6-t2v", + "name": "Kling v2.6 Text-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ling", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "release_date": "2025-12-03", + "last_updated": "2025-12-21", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 2 + "context": 0, + "output": 0 } }, - "qwen3-6-plus": { - "id": "qwen3-6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "voyage/voyage-4-lite": { + "id": "voyage/voyage-4-lite", + "name": "voyage-4-lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "voyage", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-01-15", + "last_updated": "2026-03-06", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -9673,190 +8996,74 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.5, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 2, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 2 - } + "context": 32000, + "output": 0 } }, - "qwen3-6-27b": { - "id": "qwen3-6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 80000 - } - ], - "tool_call": true, - "structured_output": true, + "voyage/voyage-law-2": { + "id": "voyage/voyage-law-2", + "name": "voyage-law-2", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "voyage", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2024-04-15", + "last_updated": "2024-03", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.412564, - "output": 2.475384, - "cache_read": 0.412564 + "context": 8192, + "output": 1536 } }, - "glm-5-3-flash": { - "id": "glm-5-3-flash", - "name": "GLM 5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "voyage/voyage-4": { + "id": "voyage/voyage-4", + "name": "voyage-4", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "voyage", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-01-15", + "last_updated": "2026-03-06", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.075, - "output": 0.25, - "cache_read": 0.075 - } - }, - "kimi-k2-7-code": { - "id": "kimi-k2-7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 131072 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.95 + "context": 32000, + "output": 0 } }, - "seed-2-0-lite": { - "id": "seed-2-0-lite", - "name": "Seed 2.0 Lite", - "description": "Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "voyage/voyage-code-3": { + "id": "voyage/voyage-code-3", + "name": "voyage-code-3", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "voyage", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2024-12-04", + "last_updated": "2024-09", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -9864,56 +9071,21 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.31, - "output": 2.5, - "cache_read": 0.31, - "tiers": [ - { - "input": 0.62, - "output": 5, - "cache_read": 0.62, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "context": 8192, + "output": 1536 } }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + "voyage/voyage-4-large": { + "id": "voyage/voyage-4-large", + "name": "voyage-4-large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "voyage", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 393216 - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2026-01-15", + "last_updated": "2026-03-06", "modalities": { "input": [ "text" @@ -9922,87 +9094,51 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 393216 - }, - "cost": { - "input": 0.424, - "output": 1.272, - "cache_read": 0.424 + "context": 32000, + "output": 0 } }, - "mimo-v2-5": { - "id": "mimo-v2-5", - "name": "MiMo V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "voyage/rerank-2.5": { + "id": "voyage/rerank-2.5", + "name": "Voyage Rerank 2.5", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "voyage", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0.7, - "output": 1.4, - "cache_read": 0.014 + "context": 32000, + "output": 32000 } }, - "seed-2-0-mini": { - "id": "seed-2-0-mini", - "name": "Seed 2.0 Mini", - "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "voyage/rerank-2.5-lite": { + "id": "voyage/rerank-2.5-lite", + "name": "Voyage Rerank 2.5 Lite", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "voyage", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -10010,132 +9146,46 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.12, - "output": 0.5, - "cache_read": 0.12, - "tiers": [ - { - "input": 0.24, - "output": 1, - "cache_read": 0.24, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "context": 32000, + "output": 32000 } }, - "muse-glimmer-30b": { - "id": "muse-glimmer-30b", - "name": "Muse Glimmer 30B", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "voyage/voyage-code-2": { + "id": "voyage/voyage-code-2", + "name": "voyage-code-2", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "voyage", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "release_date": "2024-01-01", + "last_updated": "2024-01", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.2, - "output": 0.8, - "cache_read": 0.05 - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 3 + "context": 8192, + "output": 1536 } }, - "glm-4-7-flash": { - "id": "glm-4-7-flash", - "name": "GLM 4.7 Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", + "voyage/voyage-3.5-lite": { + "id": "voyage/voyage-3.5-lite", + "name": "voyage-3.5-lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "voyage", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text" @@ -10144,84 +9194,48 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 + "context": 8192, + "output": 1536 } }, - "mimo-v2-6-pro": { - "id": "mimo-v2-6-pro", - "name": "MiMo V2.6 Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "voyage/voyage-3.5": { + "id": "voyage/voyage-3.5", + "name": "voyage-3.5", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "voyage", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2.175, - "output": 4.35, - "cache_read": 2.175 + "context": 8192, + "output": 1536 } }, - "qwen3-7-max": { - "id": "qwen3-7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "voyage/voyage-3-large": { + "id": "voyage/voyage-3-large", + "name": "voyage-3-large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "voyage", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 64000 - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-06-12", + "release_date": "2025-01-07", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -10232,28 +9246,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 2.5 + "context": 8192, + "output": 1536 } }, - "step-3-5-flash": { - "id": "step-3-5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "voyage/voyage-finance-2": { + "id": "voyage/voyage-finance-2", + "name": "voyage-finance-2", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "voyage", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", + "release_date": "2024-06-03", + "last_updated": "2024-03", "modalities": { "input": [ "text" @@ -10262,52 +9269,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 131072 - }, - "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "context": 8192, + "output": 1536 } }, - "qwen3-8-max": { - "id": "qwen3-8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 - } - ], + "mistral/mistral-nemo": { + "id": "mistral/mistral-nemo", + "name": "Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "knowledge": "2024-04", + "release_date": "2024-07-18", + "last_updated": "2024-07-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -10315,45 +9297,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 128000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 2 + "input": 0.15, + "output": 0.15 } }, - "deepseek-v3-2": { - "id": "deepseek-v3-2", - "name": "DeepSeek V3.2", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", + "mistral/codestral-embed": { + "id": "mistral/codestral-embed", + "name": "Codestral Embed", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "codestral-embed", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 393216 - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -10362,30 +9324,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.57, - "output": 1.71, - "cache_read": 0.57 + "context": 8192, + "output": 1536 } }, - "minimax-m2-7-highspeed": { - "id": "minimax-m2-7-highspeed", - "name": "MiniMax M2.7 Highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", + "mistral/mistral-embed": { + "id": "mistral/mistral-embed", + "name": "Mistral Embed", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "mistral-embed", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2023-12-11", + "last_updated": "2023-12-11", "modalities": { "input": [ "text" @@ -10394,35 +9349,24 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "context": 8192, + "output": 1536 } }, - "mimo-v2-5-pro": { - "id": "mimo-v2-5-pro", - "name": "MiMo V2.5 Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "mistral/devstral-small": { + "id": "mistral/devstral-small", + "name": "Devstral Small 1.1", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2024-10", + "release_date": "2025-05-21", + "last_updated": "2025-05-07", "modalities": { "input": [ "text" @@ -10431,34 +9375,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 64000 }, "cost": { - "input": 2.175, - "output": 4.35, - "cache_read": 0.018 + "input": 0.1, + "output": 0.3 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "mistral/mistral-large-3": { + "id": "mistral/mistral-large-3", + "name": "Mistral Large 3", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2024-10", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -10467,64 +9409,37 @@ "open_weights": false, "limit": { "context": 256000, - "output": 65536 + "output": 256000 }, "cost": { - "input": 1.08, - "output": 5.52, - "cache_read": 1.08, - "tiers": [ - { - "input": 2.16, - "output": 11.04, - "cache_read": 2.16, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 2.7, - "output": 13.8, - "cache_read": 2.7, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.5, + "output": 1.5 } }, - "seed-2-0-pro": { - "id": "seed-2-0-pro", - "name": "Seed 2.0 Pro", - "description": "Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows", - "family": "seed", - "attachment": true, + "mistral/mistral-medium-3.5": { + "id": "mistral/mistral-medium-3.5", + "name": "Mistral Medium Latest", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-04-29", + "last_updated": "2026-05-21", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -10533,100 +9448,60 @@ "open_weights": false, "limit": { "context": 256000, - "output": 128000 + "output": 256000 }, "cost": { - "input": 0.63, - "output": 3.79, - "cache_read": 0.63, - "tiers": [ - { - "input": 1.26, - "output": 7.58, - "cache_read": 1.26, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 1.5, + "output": 7.5 } }, - "qwen3-5-4b": { - "id": "qwen3-5-4b", - "name": "Qwen3.5 4B", - "description": "Qwen3.5 4B is a low-cost multimodal reasoning model with 256K context, image and video input, function tools, and structured output.", + "mistral/mistral-medium": { + "id": "mistral/mistral-medium", + "name": "Mistral Medium 3.1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 32768 - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-02", - "last_updated": "2026-03-02", + "knowledge": "2024-10", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "output": 64000 }, "cost": { - "input": 0.04, - "output": 0.07, - "cache_read": 0.02 + "input": 0.4, + "output": 2 } }, - "step-5-preview": { - "id": "step-5-preview", - "name": "Step 5 Preview", - "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "mistral/devstral-small-2": { + "id": "mistral/devstral-small-2", + "name": "Devstral Small 2", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-16", - "last_updated": "2026-09-20", + "knowledge": "2024-10", + "release_date": "2025-12-09", + "last_updated": "2025-05-07", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -10634,106 +9509,61 @@ }, "open_weights": false, "limit": { - "context": 1024000, - "output": 131072 + "context": 256000, + "output": 256000 }, "cost": { - "input": 1, - "output": 2.7, - "cache_read": 0.05 + "input": 0.1, + "output": 0.3 } }, - "qwen3-5-27b": { - "id": "qwen3-5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "mistral/ministral-14b": { + "id": "mistral/ministral-14b", + "name": "Ministral 14B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 80000 - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2024-10", + "release_date": "2025-12-02", + "last_updated": "2025-12-01", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 256000, - "output": 64000 + "output": 256000 }, "cost": { - "input": 0.086, - "output": 0.688, - "cache_read": 0.086, - "tiers": [ - { - "input": 0.258, - "output": 2.064, - "cache_read": 0.258, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.2, + "output": 0.2 } }, - "seed-2-1-turbo": { - "id": "seed-2-1-turbo", - "name": "Seed 2.1 Turbo", - "description": "Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "mistral/devstral-2": { + "id": "mistral/devstral-2", + "name": "Devstral 2", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", + "knowledge": "2024-10", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -10742,36 +9572,25 @@ "open_weights": false, "limit": { "context": 256000, - "output": 65536 + "output": 256000 }, "cost": { - "input": 0.63, - "output": 3.13, - "cache_read": 0.63 + "input": 0.4, + "output": 2 } }, - "fugu-ultra-v1-0": { - "id": "fugu-ultra-v1-0", - "name": "Fugu Ultra v1.0", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", + "mistral/mistral-small": { + "id": "mistral/mistral-small", + "name": "Mistral Small (latest)", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "temperature": true, + "knowledge": "2025-06", + "release_date": "2024-09-17", + "last_updated": "2026-03-16", "modalities": { "input": [ "text", @@ -10781,67 +9600,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 32000, + "output": 4000 }, "cost": { - "input": 7.5, - "output": 45, - "cache_read": 1.5, - "tiers": [ - { - "input": 15, - "output": 67.5, - "cache_read": 3, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 15, - "output": 67.5, - "cache_read": 3 - } + "input": 0.1, + "output": 0.3 } }, - "qwen3-5-122b-a10b": { - "id": "qwen3-5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 80000 - } - ], + "mistral/ministral-8b": { + "id": "mistral/ministral-8b", + "name": "Ministral 8B (latest)", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-04", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -10849,87 +9632,60 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.115, - "output": 0.917, - "cache_read": 0.115, - "tiers": [ - { - "input": 0.287, - "output": 2.294, - "cache_read": 0.287, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.1, + "output": 0.1 } }, - "qwen3-5-flash": { - "id": "qwen3-5-flash", - "name": "Qwen3.5 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "mistral/codestral": { + "id": "mistral/codestral", + "name": "Codestral (latest)", + "description": "Mistral code model for completions, refactors, and developer IDE workflows", + "family": "codestral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2024-10", + "release_date": "2024-05-29", + "last_updated": "2025-01-04", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 256000, + "output": 4096 }, "cost": { - "input": 0.09, - "output": 0.368, - "cache_read": 0.09 + "input": 0.3, + "output": 0.9 } }, - "mimo-v2-6-pro-ultraspeed": { - "id": "mimo-v2-6-pro-ultraspeed", - "name": "MiMo V2.6 Pro UltraSpeed", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", - "family": "mimo", + "mistral/pixtral-12b": { + "id": "mistral/pixtral-12b", + "name": "Pixtral 12B", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "knowledge": "2024-09", + "release_date": "2024-09-01", + "last_updated": "2024-09-01", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" @@ -10937,44 +9693,30 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 128000 }, "cost": { - "input": 21.75, - "output": 43.5, - "cache_read": 21.75 + "input": 0.15, + "output": 0.15 } }, - "glm-5-2": { - "id": "glm-5-2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "mistral/pixtral-large": { + "id": "mistral/pixtral-large", + "name": "Pixtral Large (latest)", + "description": "Mistral's larger vision model for document-heavy image understanding and chat", + "family": "pixtral", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2024-11-04", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -10982,38 +9724,29 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 1.4 + "input": 2, + "output": 6 } }, - "mimo-v2-6-flash": { - "id": "mimo-v2-6-flash", - "name": "MiMo V2.6 Flash", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "mistral/ministral-3b": { + "id": "mistral/ministral-3b", + "name": "Ministral 3B (latest)", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-04", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -11021,145 +9754,87 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.7, - "output": 1.4, - "cache_read": 0.7 + "input": 0.04, + "output": 0.04 } }, - "muse-spark-1-2": { - "id": "muse-spark-1-2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", - "attachment": true, + "mistral/magistral-small": { + "id": "mistral/magistral-small", + "name": "Magistral Small", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-small", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "knowledge": "2025-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-17", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 1 + "input": 0.5, + "output": 1.5 } }, - "qwen3-7-flash": { - "id": "qwen3-7-flash", - "name": "Qwen3.7 Flash", - "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", - "family": "qwen", - "attachment": true, + "mistral/magistral-medium": { + "id": "mistral/magistral-medium", + "name": "Magistral Medium (latest)", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-medium", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 131072 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "knowledge": "2025-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-20", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 991000, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.03, - "output": 0.13, - "cache_read": 0.006, - "tiers": [ - { - "input": 0.1, - "output": 0.4, - "cache_read": 0.02, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.2, - "output": 0.8, - "cache_read": 0.04, - "tier": { - "type": "context", - "size": 256000 - } - } - ] + "input": 2, + "output": 5 } }, - "minimax-m2-7": { - "id": "minimax-m2-7", - "name": "MiniMax M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "google/gemini-embedding-2": { + "id": "google/gemini-embedding-2", + "name": "Gemini Embedding 2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "gemini-embedding", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-03-10", + "last_updated": "2026-03-23", "modalities": { "input": [ "text" @@ -11168,41 +9843,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.03 + "context": 0, + "output": 0 } }, - "qwen3-6-35b-a3b": { - "id": "qwen3-6-35b-a3b", - "name": "Qwen3.6 35B A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "google/text-multilingual-embedding-002": { + "id": "google/text-multilingual-embedding-002", + "name": "Text Multilingual Embedding 002", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2024-03-01", + "last_updated": "2024-03", "modalities": { "input": [ "text" @@ -11211,389 +9868,308 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.07, - "output": 0.42, - "cache_read": 0.035 + "context": 8192, + "output": 1536 } }, - "qwen3-5-9b": { - "id": "qwen3-5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 65000 }, "cost": { - "input": 0.09, - "output": 0.13, - "cache_read": 0.045 + "input": 0.25, + "output": 1.5, + "cache_read": 0.03 } }, - "step-3-5-flash-2603": { - "id": "step-3-5-flash-2603", - "name": "Step 3.5 Flash 2603", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, + "google/gemini-3.1-flash-image": { + "id": "google/gemini-3.1-flash-image", + "name": "Gemini 3.1 Flash Image (Nano Banana 2)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", + "minimal", "high" ] } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "qwen3-5-35b-a3b": { - "id": "qwen3-5-35b-a3b", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "google/gemini-3-flash": { + "id": "google/gemini-3-flash", + "name": "Gemini 3 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 80000 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-03", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 1000000, + "output": 65000 }, "cost": { - "input": 0.057, - "output": 0.459, - "cache_read": 0.057, - "tiers": [ - { - "input": 0.229, - "output": 1.835, - "cache_read": 0.229, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "google/veo-3.1-generate-001": { + "id": "google/veo-3.1-generate-001", + "name": "Veo 3.1", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 393216 - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2025-10-15", + "last_updated": "2026-06-08", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 393216 - }, - "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 1.32 + "context": 0, + "output": 0 } }, - "kimi-k2-6": { - "id": "kimi-k2-6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 81920 } ], "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-06-12", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 16000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.8939, - "output": 3.7131, - "cache_read": 0.1788 + "input": 1.5, + "output": 9, + "cache_read": 0.15 } }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-12", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 524288 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.225, - "output": 0.9, - "cache_read": 0.045, - "tiers": [ - { - "input": 0.45, - "output": 1.8, - "cache_read": 0.09, - "tier": { - "type": "context", - "size": 512000 - } - } - ], - "context_over_200k": { - "input": 0.45, - "output": 1.8, - "cache_read": 0.09 - } + "input": 0.14, + "output": 0.4 } }, - "glm-5-3": { - "id": "glm-5-3", - "name": "GLM 5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "google/veo-3.0-generate-001": { + "id": "google/veo-3.0-generate-001", + "name": "Veo 3.0", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2025-05-20", + "last_updated": "2026-06-08", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 0, + "output": 0 + } + }, + "google/veo-3.0-fast-generate-001": { + "id": "google/veo-3.0-fast-generate-001", + "name": "Veo 3.0 Fast Generate", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-07-31", + "last_updated": "2026-06-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "video" + ] }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 1.4 + "open_weights": false, + "limit": { + "context": 0, + "output": 0 } }, - "fugu-ultra-v1-1": { - "id": "fugu-ultra-v1-1", - "name": "Fugu Ultra v1.1", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "google/text-embedding-005": { + "id": "google/text-embedding-005", + "name": "Text Embedding 005", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-08-01", + "last_updated": "2024-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -11601,406 +10177,297 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "context": 8192, + "output": 1536 } }, - "kimi-k2-7-code-highspeed": { - "id": "kimi-k2-7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "google/gemini-embedding-001": { + "id": "google/gemini-embedding-001", + "name": "Gemini Embedding 001", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "gemini-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 131072 - }, - "cost": { - "input": 1.9, - "output": 8, - "cache_read": 1.9 + "context": 8192, + "output": 1536 } }, - "deepseek-v4-1-flash": { - "id": "deepseek-v4-1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "google/gemini-3.1-flash-lite-image": { + "id": "google/gemini-3.1-flash-lite-image", + "name": "Gemini 3.1 Flash Lite Image (Nano Banana 2 Lite)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 393216 - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 393216 + "context": 65536, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.3 + "input": 0.25, + "output": 1.5, + "cache_read": 0.03 } }, - "qwen3-7-plus": { - "id": "qwen3-7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 256000 - } - ], - "tool_call": true, - "structured_output": true, + "google/gemini-2.5-flash-image": { + "id": "google/gemini-2.5-flash-image", + "name": "Nano Banana (Gemini 2.5 Flash Image)", + "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", + "family": "gemini-flash", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-12", + "knowledge": "2025-01", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, + "context": 32768, "output": 65536 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.4, - "tiers": [ - { - "input": 1.2, - "output": 4.8, - "cache_read": 1.2, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.2, - "output": 4.8, - "cache_read": 1.2 - } + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "glm-5-1": { - "id": "glm-5-1", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "google/imagen-4.0-fast-generate-001": { + "id": "google/imagen-4.0-fast-generate-001", + "name": "Imagen 4 Fast", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-06-01", + "last_updated": "2025-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 480, + "output": 0 + } + }, + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] + "type": "toggle" }, { "type": "budget_tokens", - "min": 1, - "max": 38912 + "min": 512, + "max": 24576 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-06-12", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.825, - "output": 3.301, - "cache_read": 0.165, - "tiers": [ - { - "input": 1.1, - "output": 3.851, - "cache_read": 0.22, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 } }, - "glm-4-6v-flash": { - "id": "glm-4-6v-flash", - "name": "GLM 4.6V Flash", - "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", + "google/gemini-omni-flash-preview": { + "id": "google/gemini-omni-flash-preview", + "name": "Gemini Omni Flash Preview", + "description": "Omni-modal model for text, vision, audio, and multimodal agent tasks", + "family": "gemini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 1000000, + "output": 57920 }, "cost": { - "input": 0, - "output": 0 + "input": 1.5, + "output": 9 } }, - "fugu-ultra-v2-0": { - "id": "fugu-ultra-v2-0", - "name": "Fugu Ultra v2.0", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, + "google/gemini-3.1-flash-image-preview": { + "id": "google/gemini-3.1-flash-image-preview", + "name": "Gemini 3.1 Flash Image Preview (Nano Banana 2)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", - "xhigh", - "max" + "minimal", + "high" ] } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "mistral-small-4": { - "id": "mistral-small-4", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, + "google/imagen-4.0-ultra-generate-001": { + "id": "google/imagen-4.0-ultra-generate-001", + "name": "Imagen 4 Ultra", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "tool_call": false, + "temperature": false, + "release_date": "2025-05-24", + "last_updated": "2025-05-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 65536 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.15 + "context": 480, + "output": 0 } }, - "qwen3-8-omni-flash": { - "id": "qwen3-8-omni-flash", - "name": "Qwen3.8 Omni Flash", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -12009,37 +10476,32 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 64000 }, "cost": { - "input": 0.3, - "output": 0.94, - "cache_read": 0.3 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "qwen3-5-plus": { - "id": "qwen3-5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -12047,60 +10509,42 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.36, - "output": 2.21, - "cache_read": 0.36, - "tiers": [ - { - "input": 1.08, - "output": 6.62, - "cache_read": 1.08, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.08, - "output": 6.62, - "cache_read": 1.08 - } + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "muse-spark-1-3": { - "id": "muse-spark-1-3", - "name": "Muse Spark 1.3", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", - "family": "muse", + "google/gemini-3-pro-preview": { + "id": "google/gemini-3-pro-preview", + "name": "Gemini 3 Pro Preview", + "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", - "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -12108,98 +10552,117 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 1 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "google/veo-3.1-fast-generate-001": { + "id": "google/veo-3.1-fast-generate-001", + "name": "Veo 3.1 Fast Generate", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 393216 - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-06-12", + "release_date": "2025-10-15", + "last_updated": "2026-06-08", "modalities": { "input": [ "text" ], "output": [ + "video" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "google/gemini-3-pro-image": { + "id": "google/gemini-3-pro-image", + "name": "Nano Banana Pro (Gemini 3 Pro Image)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-pro", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-09-01", + "last_updated": "2025-09", + "modalities": { + "input": [ "text" + ], + "output": [ + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 393216 + "context": 65536, + "output": 32768 }, "cost": { - "input": 1.65, - "output": 3.3, - "cache_read": 1.65 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "qwen3-8-flash": { - "id": "qwen3-8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -12208,80 +10671,69 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 65000 }, "cost": { - "input": 0.16, - "output": 0.47, - "cache_read": 0.16 + "input": 0.25, + "output": 1.5, + "cache_read": 0.03 } }, - "gemma-3-27b": { - "id": "gemma-3-27b", - "name": "Gemma 3 27B", - "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", - "family": "gemma", - "attachment": true, + "google/imagen-4.0-generate-001": { + "id": "google/imagen-4.0-generate-001", + "name": "Imagen 4", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "release_date": "2025-05-20", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, + "context": 480, "output": 0 } }, - "qwen3-6-flash": { - "id": "qwen3-6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] + "type": "toggle" }, { "type": "budget_tokens", - "min": 1, - "max": 64000 + "min": 0, + "max": 24576 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -12289,61 +10741,43 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.25, - "tiers": [ - { - "input": 1, - "output": 4, - "cache_read": 1, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1, - "output": 4, - "cache_read": 1 - } + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "seed-2-0-code": { - "id": "seed-2-0-code", - "name": "Seed 2.0 Code", - "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", - "family": "seed", + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -12351,83 +10785,200 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.4, + "input": 1.25, + "output": 10, + "cache_read": 0.125, "tiers": [ { - "input": 0.8, - "output": 4.8, - "cache_read": 0.8, + "input": 2.5, + "output": 15, + "cache_read": 0.25, "tier": { "type": "context", - "size": 128000 + "size": 200000 } } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } + } + }, + "prodia/flux-fast-schnell": { + "id": "prodia/flux-fast-schnell", + "name": "Flux Schnell", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-08-02", + "last_updated": "2026-06-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" ] + }, + "open_weights": false, + "limit": { + "context": 512, + "output": 0 } }, - "qwen3-8-27b": { - "id": "qwen3-8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, + "openai/gpt-oss-safeguard-20b": { + "id": "openai/gpt-oss-safeguard-20b", + "name": "gpt-oss-safeguard-20b", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2024-10", + "release_date": "2025-10-29", + "last_updated": "2024-12-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "input": 65536, + "output": 65536 + }, + "cost": { + "input": 0.075, + "output": 0.3, + "cache_read": 0.037 + } + }, + "openai/gpt-3.5-turbo-instruct": { + "id": "openai/gpt-3.5-turbo-instruct", + "name": "GPT-3.5 Turbo Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2021-09", + "release_date": "2023-09-18", + "last_updated": "2023-03-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "input": 4096, + "output": 4096 + }, + "cost": { + "input": 1.5, + "output": 2 + } + }, + "openai/gpt-5.2-chat": { + "id": "openai/gpt-5.2-chat", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-11", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "input": 111616, + "output": 16384 }, "cost": { - "input": 0.17, - "output": 0.5, - "cache_read": 0.08 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "muse-spark-1-1": { - "id": "muse-spark-1-1", - "name": "Muse Spark 1.1", - "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", - "family": "muse", + "openai/text-embedding-3-large": { + "id": "openai/text-embedding-3-large", + "name": "text-embedding-3-large", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-01-25", + "last_updated": "2024-01-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "input": 6656, + "output": 1536 + } + }, + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT 5.2 ", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", "medium", "high", "xhigh" @@ -12435,16 +10986,16 @@ } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", + "knowledge": "2024-10", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -12452,45 +11003,96 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 1 + "input": 21, + "output": 168 } }, - "qwen3-6-max-preview": { - "id": "qwen3-6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "openai/gpt-4o-mini-search-preview": { + "id": "openai/gpt-4o-mini-search-preview", + "name": "GPT 4o Mini Search Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2025-03-12", + "last_updated": "2025-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 111616, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "openai/gpt-5-chat": { + "id": "openai/gpt-5-chat", + "name": "GPT-5 Chat", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 393216 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "knowledge": "2024-10", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 111616, + "output": 16384 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2021-09", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" @@ -12501,38 +11103,26 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 16385, + "input": 12289, + "output": 4096 }, "cost": { - "input": 1.31, - "output": 7.88, - "cache_read": 1.31, - "tiers": [ - { - "input": 1.97, - "output": 11.82, - "cache_read": 1.97, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.5, + "output": 1.5 } }, - "step-3-7-flash": { - "id": "step-3-7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5 pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", "high" ] } @@ -12540,113 +11130,88 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "knowledge": "2024-10", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 131072 + "context": 400000, + "input": 128000, + "output": 272000 }, "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "input": 15, + "output": 120 } }, - "qwen3-5-397b-a17b": { - "id": "qwen3-5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "openai/o3-pro": { + "id": "openai/o3-pro", + "name": "o3 Pro", + "description": "High-effort o3 tier for difficult technical reasoning and careful answers", + "family": "o-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 80000 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "knowledge": "2024-10", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 200000, + "input": 100000, + "output": 100000 }, "cost": { - "input": 0.172, - "output": 1.032, - "cache_read": 0.172, - "tiers": [ - { - "input": 0.43, - "output": 2.58, - "cache_read": 0.43, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 20, + "output": 80 } }, - "glm-4-5-flash": { - "id": "glm-4-5-flash", - "name": "GLM 4.5 Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "openai/gpt-4o-transcribe": { + "id": "openai/gpt-4o-transcribe", + "name": "GPT-4o Transcribe", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2024-03-13", + "last_updated": "2024-03-13", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" @@ -12654,19 +11219,19 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 98304 + "context": 0, + "output": 0 }, "cost": { - "input": 0, - "output": 0 + "input": 2.5, + "output": 10 } }, - "gemma-4-26b-a4b": { - "id": "gemma-4-26b-a4b", - "name": "Gemma 4 26B-A4B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT 5.4 Nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -12677,158 +11242,128 @@ "low", "medium", "high", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-06-12", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.29, - "cache_read": 0.025 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1, - "max": 393216 - } - ], + "openai/gpt-5.3-chat": { + "id": "openai/gpt-5.3-chat", + "name": "GPT-5.3 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-06-12", + "release_date": "2026-03-03", + "last_updated": "2026-03-06", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 393216 + "context": 128000, + "input": 111616, + "output": 16384 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.14 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } - } - } - }, - "sensenova": { - "id": "sensenova", - "env": [ - "SENSENOVA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://token.sensenova.cn/v1", - "name": "SenseNova (China)", - "doc": "https://platform.sensenova.cn/docs", - "models": { - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + }, + "openai/gpt-5.1-thinking": { + "id": "openai/gpt-5.1-thinking", + "name": "GPT 5.1 Thinking", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-11-12", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "sensenova-6.8-flash-lite": { - "id": "sensenova-6.8-flash-lite", - "name": "SenseNova 6.8 Flash Lite", - "description": "SenseNova lightweight multimodal agent model for real-world complex tasks, data analysis, and complex information presentation", + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1-Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -12838,12 +11373,14 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-28", + "knowledge": "2024-10", + "release_date": "2025-11-12", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -12851,319 +11388,344 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT 5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "high" + "low", + "medium", + "high", + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-10", + "release_date": "2025-11-19", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "openai/gpt-4o-mini-transcribe": { + "id": "openai/gpt-4o-mini-transcribe", + "name": "GPT-4o mini Transcribe", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "o-mini", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-03-13", + "last_updated": "2024-03-13", + "modalities": { + "input": [ + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + }, + "cost": { + "input": 1.25, + "output": 5 + } + }, + "openai/text-embedding-ada-002": { + "id": "openai/text-embedding-ada-002", + "name": "text-embedding-ada-002", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2022-12-15", + "last_updated": "2022-12-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "input": 6656, + "output": 1536 + } + }, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2024-10", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT 5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "high" + "low", + "medium", + "high", + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } - } - } - }, - "alibaba-token-plan": { - "id": "alibaba-token-plan", - "env": [ - "ALIBABA_TOKEN_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://token-plan.ap-southeast-1.maas.aliyuncs.com/compatible-mode/v1", - "name": "Alibaba Token Plan", - "doc": "https://www.alibabacloud.com/help/en/model-studio/token-plan-overview", - "models": { - "happyhorse-1.1-r2v": { - "id": "happyhorse-1.1-r2v", - "name": "HappyHorse 1.1 Reference-to-Video", - "description": "Video model for reference-guided video generation", - "attachment": true, + }, + "openai/text-embedding-3-small": { + "id": "openai/text-embedding-3-small", + "name": "text-embedding-3-small", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-07-17", - "last_updated": "2026-07-17", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ - "image", "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 - }, - "cost": { - "input": 0, - "output": 0 + "context": 8192, + "input": 6656, + "output": 1536 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "openai/gpt-realtime-1.5": { + "id": "openai/gpt-realtime-1.5", + "name": "GPT-Realtime-1.5", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 0, + "output": 0 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 4, + "output": 16, + "cache_read": 0.4 } }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT 5.6 Luna", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1, + "output": 6, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "xhigh" + "high" ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "knowledge": "2024-10", + "release_date": "2025-11-12", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ @@ -13172,55 +11734,76 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "happyhorse-1.1-t2v": { - "id": "happyhorse-1.1-t2v", - "name": "HappyHorse 1.1 Text-to-Video", - "description": "Video model for prompt-driven text-to-video generation", - "attachment": false, - "reasoning": false, - "tool_call": false, + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT 5.6 Terra", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-07-17", - "last_updated": "2026-07-17", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125 } }, - "qwen-image-2.0": { - "id": "qwen-image-2.0", - "name": "Qwen Image 2.0", + "openai/gpt-image-1.5": { + "id": "openai/gpt-image-1.5", + "name": "GPT Image 1.5", "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", + "family": "gpt-image", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2025-12-16", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -13231,34 +11814,38 @@ }, "open_weights": false, "limit": { - "context": 8192, + "context": 0, "output": 0 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 32, + "cache_read": 1.25 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2024-10", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -13267,103 +11854,93 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202752, - "output": 16384 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.1, + "output": 0.5 } }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT 5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "qwen3.8-max-preview": { - "id": "qwen3.8-max-preview", - "name": "Qwen3.8 Max Preview", - "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", - "family": "qwen", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT 5.4 Mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", + "high", "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-07-19", - "last_updated": "2026-07-19", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -13371,211 +11948,182 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "qwen-image-2.0-pro": { - "id": "qwen-image-2.0-pro", - "name": "Qwen Image 2.0 Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", + "openai/gpt-realtime-mini": { + "id": "openai/gpt-realtime-mini", + "name": "GPT-Realtime mini", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2025-10-10", + "last_updated": "2025-10-10", "modalities": { "input": [ - "text" + "text", + "audio" ], "output": [ - "image" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 8192, + "context": 0, "output": 0 }, "cost": { - "input": 0, - "output": 0 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "openai/tts-1": { + "id": "openai/tts-1", + "name": "TTS-1", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "o", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "context": 0, + "output": 0 } }, - "wan2.7-image": { - "id": "wan2.7-image", - "name": "Wan2.7 Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "openai/whisper-1": { + "id": "openai/whisper-1", + "name": "Whisper", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2022-09-21", + "last_updated": "2022-09-21", "modalities": { "input": [ - "text" + "audio" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 8192, - "output": 0 - }, - "cost": { - "input": 0, + "context": 0, "output": 0 } }, - "wan2.7-image-pro": { - "id": "wan2.7-image-pro", - "name": "Wan2.7 Image Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "openai/tts-1-hd": { + "id": "openai/tts-1-hd", + "name": "TTS-1 HD", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "o", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" ], "output": [ - "image" + "audio" ] }, "open_weights": false, "limit": { - "context": 8192, - "output": 0 - }, - "cost": { - "input": 0, + "context": 0, "output": 0 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "openai/o3-deep-research": { + "id": "openai/o3-deep-research", + "name": "o3-deep-research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "medium" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2024-10", + "release_date": "2025-06-26", + "last_updated": "2024-06-26", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 98304 + "context": 200000, + "input": 100000, + "output": 100000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 10, + "output": 40, + "cache_read": 2.5 } }, - "happyhorse-1.1-i2v": { - "id": "happyhorse-1.1-i2v", - "name": "HappyHorse 1.1 Image-to-Video", - "description": "Video model for image-to-video generation", - "attachment": true, + "openai/gpt-image-1": { + "id": "openai/gpt-image-1", + "name": "GPT Image 1", + "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", + "family": "gpt-image", + "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-07-17", - "last_updated": "2026-07-17", + "release_date": "2025-03-25", + "last_updated": "2025-04-24", "modalities": { "input": [ - "image", "text" ], "output": [ - "video" + "image" ] }, "open_weights": false, @@ -13584,175 +12132,145 @@ "output": 0 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 40, + "cache_read": 1.25 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 131072 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "openai/gpt-realtime-2": { + "id": "openai/gpt-realtime-2", + "name": "gpt-realtime-2", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", - "image", - "video" + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 0, + "output": 0 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 4, + "output": 24, + "cache_read": 0.4 } }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "openai/gpt-image-1-mini": { + "id": "openai/gpt-image-1-mini", + "name": "GPT Image 1 Mini", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 0, + "output": 0 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2, + "output": 8, + "cache_read": 0.2 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT 5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "input": 196601, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 30, + "output": 180 } }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT 5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", "medium", + "high", "xhigh" ] - }, - { - "type": "budget_tokens", - "max": 262144 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": true, + "knowledge": "2025-12-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -13761,39 +12279,37 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "input": 872000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 30, + "output": 180 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", - "max" + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-10", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -13804,134 +12320,127 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "input": 122880, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.05, + "output": 0.2 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT 5.6 Sol", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2-Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-10", + "release_date": "2025-12-18", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", + "openai/gpt-5.1-instant": { + "id": "openai/gpt-5.1-instant", + "name": "GPT-5.1 Instant", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 131072 - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-10", + "release_date": "2025-11-12", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -13939,173 +12448,155 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "input": 111616, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "openai/gpt-image-2": { + "id": "openai/gpt-image-2", + "name": "GPT Image 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-03", - "last_updated": "2025-12-05", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 0, + "output": 0 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 30, + "cache_read": 1.25 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT 5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-12-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "input": 872000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -14113,246 +12604,262 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "max" + "minimal", + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } - } - } - }, - "cloudflare-workers-ai": { - "id": "cloudflare-workers-ai", - "env": [ - "CLOUDFLARE_ACCOUNT_ID", - "CLOUDFLARE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.cloudflare.com/client/v4/accounts/${CLOUDFLARE_ACCOUNT_ID}/ai/v1", - "name": "Cloudflare Workers AI", - "doc": "https://developers.cloudflare.com/workers-ai/models/", - "models": { - "@cf/meta/llama-3.3-70b-instruct-fp8-fast": { - "id": "@cf/meta/llama-3.3-70b-instruct-fp8-fast", - "name": "Llama 3.3 70B Instruct fp8 Fast", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, + }, + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": false, "temperature": true, "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 24000, - "output": 24000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.293, - "output": 2.253 + "input": 10, + "output": 30 } }, - "@cf/meta/llama-guard-3-8b": { - "id": "@cf/meta/llama-guard-3-8b", - "name": "Llama Guard 3 8B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": false, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.484, - "output": 0.03 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "@cf/meta/llama-3.1-8b-instruct-fp8": { - "id": "@cf/meta/llama-3.1-8b-instruct-fp8", - "name": "Llama 3.1 8B Instruct fp8", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 32000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.152, - "output": 0.287 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "@cf/meta/llama-3.2-11b-vision-instruct": { - "id": "@cf/meta/llama-3.2-11b-vision-instruct", - "name": "Llama 3.2 11B Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.0485, - "output": 0.676 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "@cf/meta/llama-3.2-1b-instruct": { - "id": "@cf/meta/llama-3.2-1b-instruct", - "name": "Llama 3.2 1B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "openai/o1": { + "id": "openai/o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 60000, - "output": 60000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.027, - "output": 0.201 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "@cf/meta/llama-4-scout-17b-16e-instruct": { - "id": "@cf/meta/llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout 17B 16E Instruct", - "description": "Open Llama with long-context vision for efficient multimodal agents", - "family": "llama", + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -14362,29 +12869,40 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131000, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.27, - "output": 0.85 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "@cf/meta/llama-3.2-3b-instruct": { - "id": "@cf/meta/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ "text" @@ -14393,38 +12911,40 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 80000, - "output": 80000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.0509, - "output": 0.335 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "@cf/google/gemma-4-26b-a4b-it": { - "id": "@cf/google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", @@ -14434,175 +12954,158 @@ "text" ] }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 16384 - }, - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "@cf/aisingapore/gemma-sea-lion-v4-27b-it": { - "id": "@cf/aisingapore/gemma-sea-lion-v4-27b-it", - "name": "Gemma Sea Lion V4 27B It", - "description": "Gemma 3 27B tuned by AI Singapore for Southeast Asian languages and instruction following", - "family": "gemma", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.351, - "output": 0.555 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "@cf/ibm-granite/granite-4.0-h-micro": { - "id": "@cf/ibm-granite/granite-4.0-h-micro", - "name": "Granite 4.0 H Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "granite", - "attachment": false, + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-10-02", - "last_updated": "2025-10-02", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.017, - "output": 0.112 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "@cf/deepseek-ai/deepseek-v4-flash-0731": { - "id": "@cf/deepseek-ai/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1310720, - "output": 1048576 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.014 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "@cf/deepseek-ai/deepseek-v4-pro-0813": { - "id": "@cf/deepseek-ai/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "@cf/deepseek-ai/deepseek-r1-distill-qwen-32b": { - "id": "@cf/deepseek-ai/deepseek-r1-distill-qwen-32b", - "name": "Deepseek R1 Distill Qwen 32B", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", + "zai/glm-4.7": { + "id": "zai/glm-4.7", + "name": "GLM 4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "knowledge": "2024-10", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -14611,32 +13114,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 80000, - "output": 80000 + "context": 200000, + "output": 120000 }, "cost": { - "input": 0.497, - "output": 4.881 + "input": 0.6, + "output": 2.2, + "cache_read": 0.12 } }, - "@cf/mistralai/mistral-small-3.1-24b-instruct": { - "id": "@cf/mistralai/mistral-small-3.1-24b-instruct", - "name": "Mistral Small 3.1 24B Instruct", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, - "reasoning": false, + "zai/glm-4.5v": { + "id": "zai/glm-4.5v", + "name": "GLM 4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-17", + "knowledge": "2025-08", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -14644,43 +13153,36 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 66000, + "output": 16000 }, "cost": { - "input": 0.351, - "output": 0.555 + "input": 0.6, + "output": 1.8, + "cache_read": 0.11 } }, - "@cf/moonshotai/kimi-k2.6": { - "id": "@cf/moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "zai/glm-4.5": { + "id": "zai/glm-4.5", + "name": "GLM 4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "interleaved": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-07", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -14688,33 +13190,36 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 256000 + "context": 128000, + "output": 96000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "@cf/moonshotai/kimi-k2.7-code": { - "id": "@cf/moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "zai/glm-4.7-flashx": { + "id": "zai/glm-4.7-flashx", + "name": "GLM 4.7 FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "interleaved": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -14722,62 +13227,56 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.06, + "output": 0.4, + "cache_read": 0.01 } }, - "@cf/zai-org/glm-5.3-flash": { - "id": "@cf/zai-org/glm-5.3-flash", - "name": "Glm 5.3 Flash", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", + "zai/glm-5.1": { + "id": "zai/glm-5.1", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1310720, - "output": 1048576 + "context": 202000, + "output": 202000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 1.3, + "output": 4.3, + "cache_read": 0.26 } }, - "@cf/zai-org/glm-4.7-flash": { - "id": "@cf/zai-org/glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "zai/glm-4.6": { + "id": "zai/glm-4.6", + "name": "GLM 4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -14786,14 +13285,11 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "interleaved": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -14804,35 +13300,38 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "output": 96000 }, "cost": { - "input": 0.0605, - "output": 0.4 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "@cf/zai-org/glm-5.2": { - "id": "@cf/zai-org/glm-5.2", - "name": "Glm 5.2", + "zai/glm-5.2": { + "id": "zai/glm-5.2", + "name": "GLM 5.2", "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", + "release_date": "2026-06-16", "last_updated": "2026-06-13", "modalities": { "input": [ @@ -14844,8 +13343,8 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 256000 + "context": 1040000, + "output": 128000 }, "cost": { "input": 1.4, @@ -14853,52 +13352,49 @@ "cache_read": 0.26 } }, - "@cf/zai-org/glm-5.3": { - "id": "@cf/zai-org/glm-5.3", - "name": "Glm 5.3", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "zai/glm-4.6v": { + "id": "zai/glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "family": "glm", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2024-10", + "release_date": "2025-09-30", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1310720, - "output": 1048576 + "context": 128000, + "output": 24000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.3, + "output": 0.9, + "cache_read": 0.05 } }, - "@cf/nvidia/nemotron-3-120b-a12b": { - "id": "@cf/nvidia/nemotron-3-120b-a12b", - "name": "Nemotron 3 Super 120B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "zai/glm-5.2-fast": { + "id": "zai/glm-5.2-fast", + "name": "GLM 5.2 Fast", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -14908,18 +13404,15 @@ { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ "text" @@ -14928,106 +13421,104 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 3, + "output": 10.25, + "cache_read": 0.5 } }, - "@cf/qwen/qwen3.8-27b": { - "id": "@cf/qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "zai/glm-5v-turbo": { + "id": "zai/glm-5v-turbo", + "name": "GLM 5V Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.45, - "output": 3.2, - "cache_read": 0.05 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "@cf/qwen/qwq-32b": { - "id": "@cf/qwen/qwq-32b", - "name": "Qwq 32B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "zai/glm-4.6v-flash": { + "id": "zai/glm-4.6v-flash", + "name": "GLM-4.6V-Flash", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-03-05", - "last_updated": "2025-03-05", + "knowledge": "2024-10", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 24000, + "context": 128000, "output": 24000 - }, - "cost": { - "input": 0.66, - "output": 1 } }, - "@cf/qwen/qwen3-30b-a3b-fp8": { - "id": "@cf/qwen/qwen3-30b-a3b-fp8", - "name": "Qwen3 30B A3b fp8", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "zai/glm-4.5-air": { + "id": "zai/glm-4.5-air", + "name": "GLM 4.5 Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -15038,26 +13529,32 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 128000, + "output": 96000 }, "cost": { - "input": 0.0509, - "output": 0.335 + "input": 0.2, + "output": 1.1, + "cache_read": 0.03 } }, - "@cf/qwen/qwen2.5-coder-32b-instruct": { - "id": "@cf/qwen/qwen2.5-coder-32b-instruct", - "name": "Qwen2.5 Coder 32B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "zai/glm-4.7-flash": { + "id": "zai/glm-4.7-flash", + "name": "GLM 4.7 Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "release_date": "2024-11-12", - "last_updated": "2024-11-12", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -15066,38 +13563,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 200000, + "output": 131000 }, "cost": { - "input": 0.66, - "output": 1 + "input": 0.07, + "output": 0.4 } }, - "@cf/openai/gpt-oss-20b": { - "id": "@cf/openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "zai/glm-5": { + "id": "zai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -15108,36 +13599,32 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 202800, + "output": 131100 }, "cost": { - "input": 0.2, - "output": 0.3 + "input": 0.95, + "output": 3.15, + "cache_read": 0.2 } }, - "@cf/openai/gpt-oss-120b": { - "id": "@cf/openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "zai/glm-5-turbo": { + "id": "zai/glm-5-turbo", + "name": "GLM 5 Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-03-15", + "last_updated": "2026-03-16", "modalities": { "input": [ "text" @@ -15146,113 +13633,98 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 202800, + "output": 131100 }, "cost": { - "input": 0.35, - "output": 0.75 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } - } - } - }, - "poolside": { - "id": "poolside", - "env": [ - "POOLSIDE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.poolside.ai/v1", - "name": "Poolside", - "doc": "https://platform.poolside.ai", - "models": { - "poolside/laguna-s-2.1": { - "id": "poolside/laguna-s-2.1", - "name": "Laguna S 2.1", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", + }, + "bytedance/seedream-5.0-lite": { + "id": "bytedance/seedream-5.0-lite", + "name": "Seedream 5.0 Lite", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "seed", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "release_date": "2026-02-13", + "last_updated": "2026-01-28", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 32768 + "context": 0, + "output": 0 + } + }, + "bytedance/seedance-2.0-fast": { + "id": "bytedance/seedance-2.0-fast", + "name": "Seedance 2.0 Fast", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-04-14", + "last_updated": "2026-04-14", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "video" + ] }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "open_weights": false, + "limit": { + "context": 0, + "output": 0 } }, - "poolside/laguna-xs-2.1": { - "id": "poolside/laguna-xs-2.1", - "name": "Laguna XS 2.1", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", + "bytedance/seedance-v1.0-pro": { + "id": "bytedance/seedance-v1.0-pro", + "name": "Seedance v1.0 Pro", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-07-02", - "last_updated": "2026-07-02", + "release_date": "2025-06-11", + "last_updated": "2025-06-11", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "context": 0, + "output": 0 } }, - "poolside/laguna-m.1": { - "id": "poolside/laguna-m.1", - "name": "Laguna M.1", - "description": "Poolside's open-weight model for agentic coding and long-horizon work", - "family": "laguna", + "bytedance/seed-1.6": { + "id": "bytedance/seed-1.6", + "name": "Seed 1.6", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -15261,211 +13733,186 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-06-13", + "knowledge": "2024-10", + "release_date": "2025-09-01", + "last_updated": "2025-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 2, + "cache_read": 0.05 } - } - } - }, - "nano-gpt": { - "id": "nano-gpt", - "env": [ - "NANO_GPT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://nano-gpt.com/api/v1", - "name": "NanoGPT", - "doc": "https://docs.nano-gpt.com", - "models": { - "deepseek-reasoner-cheaper": { - "id": "deepseek-reasoner-cheaper", - "name": "Deepseek R1 Cheaper", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "deepseek", - "attachment": false, + }, + "bytedance/seedance-2.0": { + "id": "bytedance/seedance-2.0", + "name": "Seedance 2.0", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", + "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-01-20", + "temperature": true, + "release_date": "2026-04-14", + "last_updated": "2026-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 - }, - "cost": { - "input": 0.4, - "output": 1.7, - "cache_read": 0.2 + "context": 0, + "output": 0 } }, - "nano-gpt-help": { - "id": "nano-gpt-help", - "name": "NanoGPT Help", - "description": "Text-only NanoGPT support assistant. Questions are processed by the Help inference provider; do not paste secrets or account credentials. Covers the website, models, API, pricing, memory, media generation, and support.", - "family": "gpt", + "bytedance/seedance-v1.5-pro": { + "id": "bytedance/seedance-v1.5-pro", + "name": "Seedance v1.5 Pro", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-06-06", - "last_updated": "2026-06-06", + "temperature": true, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 6000, - "input": 6000, - "output": 512 - }, - "cost": { - "input": 0, + "context": 0, "output": 0 } }, - "doubao-seed-2-0-lite-260215": { - "id": "doubao-seed-2-0-lite-260215", - "name": "Doubao Seed 2.0 Lite", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bytedance/seedance-v1.0-pro-fast": { + "id": "bytedance/seedance-v1.0-pro-fast", + "name": "Seedance v1.0 Pro Fast", + "description": "Video model for prompt-guided generation, editing, and motion workflows", "family": "seed", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "temperature": true, + "release_date": "2025-10-24", + "last_updated": "2025-10-31", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32000 - }, - "cost": { - "input": 0.1462, - "output": 0.8738, - "cache_read": 0.0731 + "context": 0, + "output": 0 } }, - "doubao-seed-2-0-mini-260215": { - "id": "doubao-seed-2-0-mini-260215", - "name": "Doubao Seed 2.0 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bytedance/seedream-4.0": { + "id": "bytedance/seedream-4.0", + "name": "Seedream 4.0", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "family": "seed", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "temperature": true, + "release_date": "2025-09-09", + "last_updated": "2025-08-28", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32000 - }, - "cost": { - "input": 0.0493, - "output": 0.4845, - "cache_read": 0.02465 + "context": 0, + "output": 0 } }, - "fastgpt": { - "id": "fastgpt", - "name": "Web Answer", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bytedance/seedream-4.5": { + "id": "bytedance/seedream-4.5", + "name": "Seedream 4.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "seed", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-12-23", - "last_updated": "2024-01-01", + "temperature": true, + "release_date": "2025-12-03", + "last_updated": "2025-11-28", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 - }, - "cost": { - "input": 7.5, - "output": 7.5 + "context": 0, + "output": 0 } }, - "doubao-seed-2-0-pro-260215": { - "id": "doubao-seed-2-0-pro-260215", - "name": "Doubao Seed 2.0 Pro", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bytedance/seed-1.8": { + "id": "bytedance/seed-1.8", + "name": "Seed 1.8", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "family": "seed", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-01", + "last_updated": "2025-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -15474,31 +13921,28 @@ "open_weights": false, "limit": { "context": 256000, - "input": 256000, - "output": 128000 + "output": 64000 }, "cost": { - "input": 0.782, - "output": 3.876, - "cache_read": 0.391 + "input": 0.25, + "output": 2, + "cache_read": 0.05 } }, - "ernie-5.0-thinking-preview": { - "id": "ernie-5.0-thinking-preview", - "name": "Ernie 5.0 Thinking Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "ernie", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "morph/morph-v3-large": { + "id": "morph/morph-v3-large", + "name": "Morph v3 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "morph", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "temperature": false, + "release_date": "2024-08-15", + "last_updated": "2024-08-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -15506,43 +13950,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 32000, + "output": 32000 }, "cost": { - "input": 1, - "output": 3.5, - "cache_read": 0.5 + "input": 0.9, + "output": 1.9 } }, - "gemini-2.5-flash-lite-preview-09-2025-thinking": { - "id": "gemini-2.5-flash-lite-preview-09-2025-thinking", - "name": "Gemini 2.5 Flash Lite Preview (09/2025) – Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "morph/morph-v3-fast": { + "id": "morph/morph-v3-fast", + "name": "Morph v3 Fast", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "morph", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-08-15", + "last_updated": "2024-08-15", "modalities": { "input": [ - "text", - "image", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -15550,28 +13979,26 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 16000, + "output": 16000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 + "input": 0.8, + "output": 1.2 } }, - "gemma-4-12b-it-station-keeper": { - "id": "gemma-4-12b-it-station-keeper", - "name": "Gemma 4 12B StationKeeper", - "description": "Gemma 4 12B StationKeeper is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 131,072-token context window.", - "family": "gemma", + "sakana/fugu-ultra": { + "id": "sakana/fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "aura", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "release_date": "2026-09-09", - "last_updated": "2026-09-09", + "temperature": true, + "release_date": "2026-06-21", + "last_updated": "2026-06-21", "modalities": { "input": [ "text", @@ -15581,29 +14008,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.05, - "output": 0.25, - "cache_read": 0.025 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "glm-4-air-0111": { - "id": "glm-4-air-0111", - "name": "GLM 4 Air 0111", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "glm", + "nvidia/nemotron-3-ultra-550b-a55b": { + "id": "nvidia/nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-11", - "last_updated": "2025-01-11", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -15614,31 +14045,35 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 1000000, + "output": 65000 }, "cost": { - "input": 0.1394, - "output": 0.1394, - "cache_read": 0.0697 + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 } }, - "glm-4.1v-thinking-flashx": { - "id": "glm-4.1v-thinking-flashx", - "name": "GLM 4.1V Thinking FlashX", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "glm-flash", - "attachment": true, - "reasoning": false, + "nvidia/nemotron-3-nano-30b-a3b": { + "id": "nvidia/nemotron-3-nano-30b-a3b", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": false, - "structured_output": false, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -15646,28 +14081,31 @@ }, "open_weights": false, "limit": { - "context": 64000, - "input": 64000, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 0.3, - "cache_read": 0.15 + "input": 0.05, + "output": 0.24 } }, - "deepseek-v3-0324": { - "id": "deepseek-v3-0324", - "name": "DeepSeek Chat 0324", - "description": "March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding", - "family": "deepseek", + "nvidia/nemotron-nano-9b-v2": { + "id": "nvidia/nemotron-nano-9b-v2", + "name": "Nvidia Nemotron Nano 9B V2", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "knowledge": "2024-10", + "release_date": "2025-08-18", + "last_updated": "2025-08-18", "modalities": { "input": [ "text" @@ -15676,29 +14114,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.77, - "cache_read": 0.135 + "input": 0.06, + "output": 0.23 } }, - "venice-uncensored": { - "id": "venice-uncensored", - "name": "Venice Uncensored", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "venice", + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "NVIDIA Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": false, - "structured_output": false, - "release_date": "2025-10-01", - "last_updated": "2025-02-24", + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text" @@ -15707,30 +14148,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.4 + "input": 0.15, + "output": 0.65 } }, - "Gemma-4-31B-MeroMero-v2:thinking": { - "id": "Gemma-4-31B-MeroMero-v2:thinking", - "name": "Gemma 4 31B MeroMero v2 Thinking", - "description": "Gemma 4 31B MeroMero v2 with thinking enabled for more deliberate emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.", - "family": "gemma", + "nvidia/nemotron-nano-12b-v2-vl": { + "id": "nvidia/nemotron-nano-12b-v2-vl", + "name": "Nvidia Nemotron Nano 12B V2 VL", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-08-24", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-10-28", + "last_updated": "2025-10-28", "modalities": { "input": [ "text", @@ -15740,33 +14184,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.45, - "cache_read": 0.05 + "input": 0.2, + "output": 0.6 } }, - "gemini-2.0-pro-reasoner": { - "id": "gemini-2.0-pro-reasoner", - "name": "Gemini 2.0 Pro Reasoner", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo M2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo-v2.5", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-02-05", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "audio" + "image", + "pdf" ], "output": [ "text" @@ -15774,27 +14223,32 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 1050000, + "output": 131100 }, "cost": { - "input": 1.292, - "output": 4.998, - "cache_read": 0.323 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "Meta-Llama-3-1-8B-Instruct-FP8": { - "id": "Meta-Llama-3-1-8B-Instruct-FP8", - "name": "Llama 3.1 8B (decentralized)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "llama", + "xiaomi/mimo-v2-flash": { + "id": "xiaomi/mimo-v2-flash", + "name": "MiMo V2 Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-07-23", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-16", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -15803,29 +14257,34 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 262144, + "output": 32000 }, "cost": { - "input": 0.02, - "output": 0.03, + "input": 0.1, + "output": 0.3, "cache_read": 0.01 } }, - "glm-4-plus-0111": { - "id": "glm-4-plus-0111", - "name": "GLM 4 Plus 0111", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "glm", + "xiaomi/mimo-v2-pro": { + "id": "xiaomi/mimo-v2-pro", + "name": "MiMo V2 Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -15836,92 +14295,88 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 9.996, - "output": 9.996, - "cache_read": 4.998 + "input": 1, + "output": 3, + "cache_read": 0.2 } }, - "holo3-35b-a3b:thinking": { - "id": "holo3-35b-a3b:thinking", - "name": "Holo3-35B-A3B Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo V2.5 Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo-v2.5-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "input": 65536, - "output": 8192 + "context": 1050000, + "output": 131000 }, "cost": { - "input": 0.25, - "output": 1.8, - "cache_read": 0.125 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } }, - "gemma-4-26b-a4b-it-moonlight": { - "id": "gemma-4-26b-a4b-it-moonlight", - "name": "Moonlight Dusk", - "description": "Moonlight Dusk is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "quiverai/arrow-1.1": { + "id": "quiverai/arrow-1.1", + "name": "Arrow 1.1", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "o", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 - }, - "cost": { - "input": 0.12, - "output": 0.38, - "cache_read": 0.06 + "context": 131072, + "output": 131072 } }, - "phi-4-mini-instruct": { - "id": "phi-4-mini-instruct", - "name": "Phi 4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "phi", + "inception/mercury-coder-small": { + "id": "inception/mercury-coder-small", + "name": "Mercury Coder Small Beta", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "mercury", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "tool_call": true, + "temperature": true, + "release_date": "2025-02-26", + "last_updated": "2025-02-26", "modalities": { "input": [ "text" @@ -15930,29 +14385,37 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, + "context": 32000, "output": 16384 }, "cost": { - "input": 0.17, - "output": 0.68, - "cache_read": 0.085 + "input": 0.25, + "output": 1 } }, - "auto-model-premium": { - "id": "auto-model-premium", - "name": "Auto model (Premium)", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", + "inception/mercury-2": { + "id": "inception/mercury-2", + "name": "Mercury 2", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2025-04-16", - "last_updated": "2024-06-01", + "temperature": true, + "release_date": "2026-02-24", + "last_updated": "2026-03-06", "modalities": { "input": [ "text" @@ -15963,75 +14426,67 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 + "context": 128000, + "output": 128000 }, "cost": { - "input": 9.996, - "output": 19.992, - "cache_read": 4.998 + "input": 0.25, + "output": 0.75, + "cache_read": 0.024999999999999998 } }, - "gemma-4-31b-it-darkidol": { - "id": "gemma-4-31b-it-darkidol", - "name": "DarkIdol", - "description": "DarkIdol is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", - "family": "gemma", + "anthropic/claude-3.5-haiku": { + "id": "anthropic/claude-3.5-haiku", + "name": "Claude 3.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "temperature": true, + "knowledge": "2024-07-31", + "release_date": "2024-11-04", + "last_updated": "2024-10-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.45, - "cache_read": 0.05 + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 } }, - "gemini-2.5-flash-lite-preview-06-17": { - "id": "gemini-2.5-flash-lite-preview-06-17", - "name": "Gemini 2.5 Flash Lite Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" @@ -16039,42 +14494,42 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "gemini-2.5-flash-preview-09-2025": { - "id": "gemini-2.5-flash-preview-09-2025", - "name": "Gemini 2.5 Flash Preview (09/2025)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "interleaved": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "audio", "pdf" ], "output": [ @@ -16083,63 +14538,95 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "gemma-4-26b-a4b-it-opusdistill": { - "id": "gemma-4-26b-a4b-it-opusdistill", - "name": "Opus Distill", - "description": "Opus Distill is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", - "family": "gemma", + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "temperature": true, + "knowledge": "2026-01-31", + "release_date": "2026-06-29", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.12, - "output": 0.38, - "cache_read": 0.06 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "auto-model-standard": { - "id": "auto-model-standard", - "name": "Auto model (Standard)", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2025-04-16", - "last_updated": "2024-06-01", + "temperature": true, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -16148,30 +14635,46 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 1000000, - "output": 1000000 + "output": 128000 }, "cost": { - "input": 9.996, - "output": 19.992, - "cache_read": 4.998 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "longcat-2.0": { - "id": "longcat-2.0", - "name": "LongCat 2.0", - "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", - "family": "longcat", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -16179,42 +14682,47 @@ }, "open_weights": false, "limit": { - "context": 1048756, - "input": 1048756, - "output": 262144 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.75, - "output": 3, - "cache_read": 0.015 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-2.5-pro-preview-03-25": { - "id": "gemini-2.5-pro-preview-03-25", - "name": "Gemini 2.5 Pro Preview 0325", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh" ] } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "tool_call": true, + "temperature": true, + "knowledge": "2026-01-31", + "release_date": "2026-07-01", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" @@ -16222,31 +14730,43 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 0.25 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "agnes-3.0-flash": { - "id": "agnes-3.0-flash", - "name": "Agnes 3.0 Flash", - "description": "Agnes 3.0 Flash is a low-cost model for coding, tool use, and multi-turn agent tasks. It supports text and image input, optional thinking, and a 512K-token context window.", + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-09-09", - "last_updated": "2026-09-09", + "interleaved": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -16254,71 +14774,51 @@ }, "open_weights": false, "limit": { - "context": 524288, - "input": 524288, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.05, - "output": 0.15, - "cache_read": 0.005 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "Gemma-4-31B-Queen": { - "id": "Gemma-4-31B-Queen", - "name": "Gemma 4 31B Queen", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemma", + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "interleaved": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306, - "cache_read": 0.153 - } - }, - "auto-model": { - "id": "auto-model", - "name": "Auto model", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "release_date": "2025-04-16", - "last_updated": "2024-06-01", - "modalities": { - "input": [ - "text" + "image", + "pdf" ], "output": [ "text" @@ -16327,25 +14827,45 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 1000000, - "output": 1000000 + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "qvq-max": { - "id": "qvq-max", - "name": "Qwen: QvQ Max", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "qwen", + "anthropic/claude-3-haiku": { + "id": "anthropic/claude-3-haiku", + "name": "Claude Haiku 3", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-haiku", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-28", - "last_updated": "2025-03-28", + "tool_call": true, + "temperature": true, + "knowledge": "2023-08-31", + "release_date": "2024-03-13", + "last_updated": "2024-03-13", "modalities": { "input": [ "text", @@ -16355,31 +14875,35 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 200000, + "output": 4096 }, "cost": { - "input": 1.2, - "output": 4.8, - "cache_read": 0.6 + "input": 0.25, + "output": 1.25, + "cache_read": 0.03, + "cache_write": 0.3 } }, - "asi1-mini": { - "id": "asi1-mini", - "name": "ASI1 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ @@ -16388,24 +14912,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1, - "output": 1, - "cache_read": 0.5 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "gemini-2.5-pro-exp-03-25": { - "id": "gemini-2.5-pro-exp-03-25", - "name": "Gemini 2.5 Pro Experimental 0325", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -16413,17 +14940,23 @@ "medium", "high" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "tool_call": true, + "interleaved": true, + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" @@ -16431,62 +14964,82 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 0.25 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemma-4-31b-it-isometry": { - "id": "gemma-4-31b-it-isometry", - "name": "Isometry", - "description": "Isometry is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", - "family": "gemma", + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "structured_output": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 200000, + "output": 32000 }, + "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.45, - "cache_read": 0.05 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "celeris-1": { - "id": "celeris-1", - "name": "Celeris 1", - "description": "Celeris 1 is a diffusion language model built for ultra-low-latency classification, extraction, judging, query rewriting, and other short structured responses.", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-07-25", - "last_updated": "2026-07-25", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -16494,33 +15047,30 @@ }, "open_weights": false, "limit": { - "context": 8192, - "input": 8192, - "output": 8192 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 1 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "gemini-3-pro-image-preview": { - "id": "gemini-3-pro-image-preview", - "name": "Gemini 3 Pro Image", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", - "attachment": true, + "cohere/rerank-v3.5": { + "id": "cohere/rerank-v3.5", + "name": "Cohere Rerank 3.5", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "o", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "release_date": "2024-12-02", + "last_updated": "2024-12-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -16528,64 +15078,54 @@ }, "open_weights": false, "limit": { - "context": 65536, - "input": 65536, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "context": 4096, + "output": 4096 } }, - "gemma-4-26b-a4b-it-darksoul": { - "id": "gemma-4-26b-a4b-it-darksoul", - "name": "Dark Soul", - "description": "Dark Soul is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "cohere/command-a": { + "id": "cohere/command-a", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 256000, + "output": 8000 }, "cost": { - "input": 0.12, - "output": 0.38, - "cache_read": 0.06 + "input": 2.5, + "output": 10 } }, - "glm-4.1v-thinking-flash": { - "id": "glm-4.1v-thinking-flash", - "name": "GLM 4.1V Thinking Flash", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "glm-flash", - "attachment": true, + "cohere/rerank-v4-fast": { + "id": "cohere/rerank-v4-fast", + "name": "Cohere Rerank 4 Fast", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "o", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "temperature": true, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -16593,75 +15133,49 @@ }, "open_weights": false, "limit": { - "context": 64000, - "input": 64000, - "output": 8192 - }, - "cost": { - "input": 0.3, - "output": 0.3, - "cache_read": 0.15 + "context": 32000, + "output": 32000 } }, - "Gemma-4-31B-GarnetV2": { - "id": "Gemma-4-31B-GarnetV2", - "name": "Gemma 4 31B Garnet V2", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "cohere/embed-v4.0": { + "id": "cohere/embed-v4.0", + "name": "Embed v4.0", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306, - "cache_read": 0.153 + "context": 128000, + "output": 1536 } }, - "gemini-2.5-flash-nothinking": { - "id": "gemini-2.5-flash-nothinking", - "name": "Gemini 2.5 Flash (No Thinking)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", - "attachment": true, + "cohere/rerank-v4-pro": { + "id": "cohere/rerank-v4-pro", + "name": "Cohere Rerank 4 Pro", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "o", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "temperature": true, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -16669,28 +15183,32 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "context": 32000, + "output": 32000 } }, - "Gemma-4-31B-MeroMero-v2": { - "id": "Gemma-4-31B-MeroMero-v2", - "name": "Gemma 4 31B MeroMero v2", - "description": "Gemma 4 31B MeroMero v2 is a LoRA finetune for emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.", - "family": "gemma", + "stepfun/step-3.7-flash": { + "id": "stepfun/step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "family": "step", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-08-23", + "temperature": true, + "knowledge": "2026-01-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", @@ -16700,58 +15218,63 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.1, - "output": 0.45, - "cache_read": 0.05 + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 } }, - "holo3-35b-a3b": { - "id": "holo3-35b-a3b", - "name": "Holo3-35B-A3B", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "stepfun/step-3.5-flash": { + "id": "stepfun/step-3.5-flash", + "name": "StepFun 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "family": "step", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "input": 65536, - "output": 8192 + "context": 262114, + "output": 262114 }, "cost": { - "input": 0.25, - "output": 1.8, - "cache_read": 0.125 + "input": 0.09, + "output": 0.3, + "cache_read": 0.02 } }, - "claw-medium": { - "id": "claw-medium", - "name": "Claw Medium", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "interfaze/interfaze-beta": { + "id": "interfaze/interfaze-beta", + "name": "Interfaze Beta", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -16759,18 +15282,20 @@ "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "tool_call": false, + "temperature": true, + "release_date": "2025-10-07", + "last_updated": "2026-04-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -16778,607 +15303,476 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 1.5, + "output": 3.5 } }, - "gemma-4-26b-a4b-it-shadowsiren": { - "id": "gemma-4-26b-a4b-it-shadowsiren", - "name": "Shadow Siren", - "description": "Shadow Siren is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "bfl/flux-kontext-max": { + "id": "bfl/flux-kontext-max", + "name": "FLUX.1 Kontext Max", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-05-29", + "last_updated": "2025-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 - }, - "cost": { - "input": 0.12, - "output": 0.38, - "cache_read": 0.06 + "context": 512, + "output": 0 } }, - "gemini-2.0-pro-exp-02-05": { - "id": "gemini-2.0-pro-exp-02-05", - "name": "Gemini 2.0 Pro 0205", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", - "attachment": true, + "bfl/flux-2-flex": { + "id": "bfl/flux-2-flex", + "name": "FLUX.2 [flex]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-02-05", - "last_updated": "2025-02-05", + "temperature": true, + "release_date": "2025-11-25", + "last_updated": "2026-06-08", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 2097152, - "input": 2097152, - "output": 8192 - }, - "cost": { - "input": 1.989, - "output": 7.956, - "cache_read": 0.49725 + "context": 0, + "output": 0 } }, - "claw-high": { - "id": "claw-high", - "name": "Claw High", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bfl/flux-pro-1.1-ultra": { + "id": "bfl/flux-pro-1.1-ultra", + "name": "FLUX1.1 [pro] Ultra", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-11-01", + "last_updated": "2024-11", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "context": 512, + "output": 0 } }, - "deepseek-chat": { - "id": "deepseek-chat", - "name": "DeepSeek V3/Deepseek Chat", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": true, + "bfl/flux-2-max": { + "id": "bfl/flux-2-max", + "name": "FLUX.2 [max]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "release_date": "2025-12-16", + "last_updated": "2026-06-08", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 - }, - "cost": { - "input": 0.1, - "output": 0.425, - "cache_read": 0.05 + "context": 67300, + "output": 67300 } }, - "Gemma-4-26B-A4B-MeroMero:thinking": { - "id": "Gemma-4-26B-A4B-MeroMero:thinking", - "name": "Gemma 4 26B A4B MeroMero Thinking", - "description": "Gemma 4 26B A4B MeroMero with thinking enabled for more deliberate emotive dialogue, relationship scenes, creative writing, and multimodal roleplay.", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-08-26", + "bfl/flux-pro-1.1": { + "id": "bfl/flux-pro-1.1", + "name": "FLUX1.1 [pro]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-10-02", + "last_updated": "2024-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 - }, - "cost": { - "input": 0.12, - "output": 0.38, - "cache_read": 0.06 + "context": 512, + "output": 0 } }, - "claw-low": { - "id": "claw-low", - "name": "Claw Low", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bfl/flux-pro-1.0-fill": { + "id": "bfl/flux-pro-1.0-fill", + "name": "FLUX.1 Fill [pro]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-10-01", + "last_updated": "2024-10", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0.08333 + "context": 512, + "output": 0 } }, - "doubao-1.5-pro-32k": { - "id": "doubao-1.5-pro-32k", - "name": "Doubao 1.5 Pro 32k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "bfl/flux-2-klein-4b": { + "id": "bfl/flux-2-klein-4b", + "name": "FLUX.2 [klein] 4B", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-11-20", - "last_updated": "2025-01-22", + "temperature": true, + "release_date": "2026-01-15", + "last_updated": "2026-06-08", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 8192 - }, - "cost": { - "input": 0.1343, - "output": 0.3349, - "cache_read": 0.06715 + "context": 0, + "output": 0 } }, - "doubao-seed-2-0-code-preview-260215": { - "id": "doubao-seed-2-0-code-preview-260215", - "name": "Doubao Seed 2.0 Code Preview", - "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", - "family": "seed", + "bfl/flux-2-klein-9b": { + "id": "bfl/flux-2-klein-9b", + "name": "FLUX.2 [klein] 9B", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-01-15", + "last_updated": "2026-06-08", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 128000 - }, - "cost": { - "input": 0.782, - "output": 3.893, - "cache_read": 0.391 + "context": 0, + "output": 0 } }, - "hermes-low": { - "id": "hermes-low", - "name": "Hermes Low", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "hermes", + "bfl/flux-kontext-pro": { + "id": "bfl/flux-kontext-pro", + "name": "FLUX.1 Kontext Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-05-29", + "last_updated": "2025-06", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0.08333 + "context": 512, + "output": 0 } }, - "gemma-4-31b-it-gemsicle": { - "id": "gemma-4-31b-it-gemsicle", - "name": "Gemsicle", - "description": "Gemsicle is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "bfl/flux-2-pro": { + "id": "bfl/flux-2-pro", + "name": "FLUX.2 [pro]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-11-25", + "last_updated": "2026-06-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.45, - "cache_read": 0.05 + "context": 67300, + "output": 67300 } }, - "mercury-coder-small": { - "id": "mercury-coder-small", - "name": "Mercury Coder Small", - "description": "Model by Inception AI. A diffusion large language model that runs incredibly quickly (500+ tokens/second) while matching Claude 3.5 Haiku and GPT-4o-mini. 1st in speed on Copilot arena, and matching 2nd in quality.", - "family": "mercury", + "recraft/recraft-v4.1-pro": { + "id": "recraft/recraft-v4.1-pro", + "name": "Recraft V4.1 Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "temperature": true, + "release_date": "2026-05-14", + "last_updated": "2026-05-14", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 - }, - "cost": { - "input": 0.25, - "output": 1, - "cache_read": 0.125 + "context": 0, + "output": 0 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, + "recraft/recraft-v4.1": { + "id": "recraft/recraft-v4.1", + "name": "Recraft V4.1", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-05-14", + "last_updated": "2026-05-14", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "cache_write": 0.375 + "context": 0, + "output": 0 } }, - "ernie-x1.1-preview": { - "id": "ernie-x1.1-preview", - "name": "ERNIE X1.1", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "ernie", - "attachment": true, + "recraft/recraft-v4": { + "id": "recraft/recraft-v4", + "name": "Recraft V4", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-09-10", - "last_updated": "2025-09-10", + "temperature": true, + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 64000, - "input": 64000, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "context": 0, + "output": 0 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, + "recraft/recraft-v4-pro": { + "id": "recraft/recraft-v4-pro", + "name": "Recraft V4 Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "context": 0, + "output": 0 } }, - "hermes-high": { - "id": "hermes-high", - "name": "Hermes High", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "hermes", + "recraft/recraft-v2": { + "id": "recraft/recraft-v2", + "name": "Recraft V2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-03-13", + "last_updated": "2024-03", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "context": 512, + "output": 0 } }, - "gemini-2.5-flash-preview-05-20": { - "id": "gemini-2.5-flash-preview-05-20", - "name": "Gemini 2.5 Flash 0520", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", - "attachment": true, + "recraft/recraft-v3": { + "id": "recraft/recraft-v3", + "name": "Recraft V3", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "temperature": true, + "release_date": "2024-10-30", + "last_updated": "2024-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1048000, - "input": 1048000, - "output": 65536 + "context": 512, + "output": 0 + } + }, + "recraft/recraft-v4.1-utility-pro": { + "id": "recraft/recraft-v4.1-utility-pro", + "name": "Recraft V4.1 Utility Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-05-14", + "last_updated": "2026-05-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "open_weights": false, + "limit": { + "context": 0, + "output": 0 } }, - "glm-4-long": { - "id": "glm-4-long", - "name": "GLM-4 Long", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "glm", + "recraft/recraft-v4.1-utility": { + "id": "recraft/recraft-v4.1-utility", + "name": "Recraft V4.1 Utility", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "recraft", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-08-01", + "temperature": true, + "release_date": "2026-05-14", + "last_updated": "2026-05-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "arcee-ai/trinity-large-preview": { + "id": "arcee-ai/trinity-large-preview", + "name": "Trinity Large Preview", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "trinity", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2026-01-27", + "last_updated": "2025-01", "modalities": { "input": [ "text" @@ -17389,27 +15783,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 4096 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.2006, - "output": 0.2006, - "cache_read": 0.1003 + "input": 0.25, + "output": 1 } }, - "pokee-isaac": { - "id": "pokee-isaac", - "name": "Pokee-Isaac 28B", - "description": "Pokee-Isaac is a 28B agentic model with a roughly 10-million-token context window, function calling, and OpenAI-compatible structured output. Pokee bills in $0.01 increments, rounding each non-zero request up to the next cent.", + "arcee-ai/trinity-large-thinking": { + "id": "arcee-ai/trinity-large-thinking", + "name": "Trinity Large Thinking", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "trinity", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-08-04", - "last_updated": "2026-08-04", + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-03", "modalities": { "input": [ "text" @@ -17418,29 +15811,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 10000000, - "input": 10000000, - "output": 60000 + "context": 262100, + "output": 80000 }, "cost": { - "input": 0.15, - "output": 1, - "cache_read": 0.075 + "input": 0.25, + "output": 0.8999999999999999 } }, - "MiniMax-M1": { - "id": "MiniMax-M1", - "name": "MiniMax M1", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "minimax", + "arcee-ai/trinity-mini": { + "id": "arcee-ai/trinity-mini", + "name": "Trinity Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "trinity", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-01-08", - "last_updated": "2025-06-16", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-01", + "last_updated": "2025-12", "modalities": { "input": [ "text" @@ -17449,79 +15841,65 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, + "context": 131072, "output": 131072 }, "cost": { - "input": 0.1394, - "output": 1.3328, - "cache_read": 0.0697 + "input": 0.045, + "output": 0.15 } }, - "Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled": { - "id": "Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled", - "name": "Gemma 4 31B Claude 4.6 Opus Reasoning Distilled", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "claude", - "attachment": true, + "perplexity/sonar-reasoning-pro": { + "id": "perplexity/sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded reasoning model for multi-step research and cited answers", + "family": "sonar-reasoning", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", + "low", + "medium", "high" ] } ], - "tool_call": true, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "tool_call": false, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306, - "cache_read": 0.0306 + "context": 127000, + "output": 8000 } }, - "Qwen3.5-27B-BlueStar-v3-Derestricted": { - "id": "Qwen3.5-27B-BlueStar-v3-Derestricted", - "name": "Qwen3.5 27B BlueStar v3 Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "qwen3.5", + "perplexity/sonar": { + "id": "perplexity/sonar", + "name": "Sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "temperature": true, + "knowledge": "2025-02", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ "text", @@ -17531,30 +15909,24 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.306, - "output": 0.306, - "cache_read": 0.153 + "context": 127000, + "output": 8000 } }, - "gemma-4-31b-it-novelist": { - "id": "gemma-4-31b-it-novelist", - "name": "Novelist", - "description": "Novelist is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", - "family": "gemma", + "perplexity/sonar-pro": { + "id": "perplexity/sonar-pro", + "name": "Sonar Pro", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ "text", @@ -17564,30 +15936,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.45, - "cache_read": 0.05 + "context": 200000, + "output": 8000 } }, - "ernie-5.1:thinking": { - "id": "ernie-5.1:thinking", - "name": "ERNIE 5.1 Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "ernie", + "amazon/titan-embed-text-v2": { + "id": "amazon/titan-embed-text-v2", + "name": "Titan Text Embeddings V2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "titan-embed", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-05-10", - "last_updated": "2026-05-10", + "temperature": true, + "release_date": "2024-04-30", + "last_updated": "2024-04", "modalities": { "input": [ "text" @@ -17598,63 +15963,73 @@ }, "open_weights": false, "limit": { - "context": 119000, - "input": 119000, - "output": 64000 - }, - "cost": { - "input": 0.75, - "output": 3, - "cache_read": 0.75 + "context": 8192, + "output": 1536 } }, - "gemma-4-12b-it-semancer": { - "id": "gemma-4-12b-it-semancer", - "name": "Gemma 4 12B Semancer", - "description": "Gemma 4 12B Semancer is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 131,072-token context window.", - "family": "gemma", + "amazon/nova-2-lite": { + "id": "amazon/nova-2-lite", + "name": "Nova 2 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "release_date": "2026-09-09", - "last_updated": "2026-09-09", + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-02", + "last_updated": "2024-12-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.05, - "output": 0.25, - "cache_read": 0.025 + "input": 0.3, + "output": 2.5, + "cache_read": 0.075 } }, - "mistral-code-latest": { - "id": "mistral-code-latest", - "name": "Mistral Code Latest", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "mistral", - "attachment": false, + "amazon/nova-lite": { + "id": "amazon/nova-lite", + "name": "Nova Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -17662,106 +16037,103 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 300000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.15 + "input": 0.06, + "output": 0.24, + "cache_read": 0.015 } }, - "Qwen3.5-27B-Queen-Derestricted": { - "id": "Qwen3.5-27B-Queen-Derestricted", - "name": "Qwen3.5 27B Queen Derestricted", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "qwen3.5", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "amazon/nova-micro": { + "id": "amazon/nova-micro", + "name": "Nova Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.306, - "output": 0.306, - "cache_read": 0.153 + "input": 0.035, + "output": 0.14, + "cache_read": 0.00875 } }, - "gemma-4-31b-it-fabled": { - "id": "gemma-4-31b-it-fabled", - "name": "Fabled", - "description": "Fabled is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", - "family": "gemma", + "amazon/nova-pro": { + "id": "amazon/nova-pro", + "name": "Nova Pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 300000, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.45, - "cache_read": 0.05 + "input": 0.8, + "output": 3.2, + "cache_read": 0.2 } }, - "Gemma-4-26B-A4B-MeroMero": { - "id": "Gemma-4-26B-A4B-MeroMero", - "name": "Gemma 4 26B A4B MeroMero", - "description": "Gemma 4 26B A4B MeroMero is an NVFP4 multimodal mixture-of-experts fine-tune for emotive dialogue, relationship scenes, creative writing, and roleplay.", - "family": "gemma", + "alibaba/qwen3-vl-thinking": { + "id": "alibaba/qwen3-vl-thinking", + "name": "Qwen3 VL Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1, + "max": 81920 + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-08-26", + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-09-23", + "last_updated": "2025-09-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -17769,40 +16141,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, + "context": 131072, "output": 32768 }, "cost": { - "input": 0.12, - "output": 0.38, - "cache_read": 0.06 + "input": 0.4, + "output": 4 } }, - "Gemma-4-31B-Cognitive-Unshackled": { - "id": "Gemma-4-31B-Cognitive-Unshackled", - "name": "Gemma 4 31B Cognitive Unshackled", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "alibaba/qwen3-coder-plus": { + "id": "alibaba/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -17810,26 +16171,51 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.306, - "output": 0.306, - "cache_read": 0.153 + "input": 1, + "output": 5, + "cache_read": 0.2 } }, - "universal-summarizer": { - "id": "universal-summarizer", - "name": "Universal Summarizer", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/wan-v2.6-r2v": { + "id": "alibaba/wan-v2.6-r2v", + "name": "Wan v2.6 Reference-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "o", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-12-23", - "last_updated": "2024-01-01", + "temperature": true, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "video" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "alibaba/qwen3-embedding-0.6b": { + "id": "alibaba/qwen3-embedding-0.6b", + "name": "Qwen3 Embedding 0.6B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text" @@ -17841,65 +16227,54 @@ "open_weights": false, "limit": { "context": 32768, - "input": 32768, "output": 32768 - }, - "cost": { - "input": 30, - "output": 30 } }, - "gemma-4-12b-it": { - "id": "gemma-4-12b-it", - "name": "Gemma 4 12B Instruct", - "description": "Compact Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, + "alibaba/qwen3-max-preview": { + "id": "alibaba/qwen3-max-preview", + "name": "Qwen3 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2025-04", + "release_date": "2025-09-05", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 131072, + "context": 262144, "output": 32768 }, "cost": { - "input": 0.05, - "output": 0.25, - "cache_read": 0.025 + "input": 1.2, + "output": 6, + "cache_read": 0.24 } }, - "gemini-exp-1206": { - "id": "gemini-exp-1206", - "name": "Gemini 2.0 Pro 1206", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", - "attachment": true, + "alibaba/qwen3-embedding-8b": { + "id": "alibaba/qwen3-embedding-8b", + "name": "Qwen3 Embedding 8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "temperature": true, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -17907,42 +16282,66 @@ }, "open_weights": false, "limit": { - "context": 2097152, - "input": 2097152, - "output": 8192 + "context": 32768, + "output": 32768 + } + }, + "alibaba/qwen3-next-80b-a3b-instruct": { + "id": "alibaba/qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-11", + "last_updated": "2025-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.258, - "output": 4.998, - "cache_read": 0.629 + "input": 0.15, + "output": 1.2 } }, - "gemini-2.5-flash-preview-09-2025-thinking": { - "id": "gemini-2.5-flash-preview-09-2025-thinking", - "name": "Gemini 2.5 Flash Preview (09/2025) – Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", + "alibaba/qwen3.7-plus": { + "id": "alibaba/qwen3.7-plus", + "name": "Qwen 3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen3.7-plus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 } ], "tool_call": true, - "structured_output": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", "image", - "audio", "pdf" ], "output": [ @@ -17951,58 +16350,53 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "cache_write": 0.5 } }, - "doubao-1.5-pro-256k": { - "id": "doubao-1.5-pro-256k", - "name": "Doubao 1.5 Pro 256k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/wan-v2.7-r2v": { + "id": "alibaba/wan-v2.7-r2v", + "name": "Wan v2.7 Reference-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "o", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 16384 - }, - "cost": { - "input": 0.799, - "output": 1.445, - "cache_read": 0.3995 + "context": 0, + "output": 0 } }, - "gemma-4-31b-it-garnet": { - "id": "gemma-4-31b-it-garnet", - "name": "Garnet", - "description": "Garnet is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", - "family": "gemma", + "alibaba/qwen3-vl-instruct": { + "id": "alibaba/qwen3-vl-instruct", + "name": "Qwen3 VL Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-24", "modalities": { "input": [ "text", @@ -18014,42 +16408,38 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 131072, + "output": 129024 }, "cost": { - "input": 0.1, - "output": 0.45, - "cache_read": 0.05 + "input": 0.4, + "output": 1.6 } }, - "gemini-2.5-pro-preview-05-06": { - "id": "gemini-2.5-pro-preview-05-06", - "name": "Gemini 2.5 Pro Preview 0506", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", + "alibaba/qwen3.7-max": { + "id": "alibaba/qwen3.7-max", + "name": "Qwen 3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-06", - "last_updated": "2025-05-06", + "tool_call": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -18057,61 +16447,53 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 991000, + "output": 64000 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 0.25 + "input": 1.25, + "output": 3.75, + "cache_read": 0.25, + "cache_write": 1.5625 } }, - "mercury-2": { - "id": "mercury-2", - "name": "Mercury 2", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "mercury", + "alibaba/wan-v2.5-t2v-preview": { + "id": "alibaba/wan-v2.5-t2v-preview", + "name": "Wan v2.5 Text-to-Video Preview", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "o", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 50000 - }, - "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.025 + "context": 0, + "output": 0 } }, - "command-a-reasoning-08-2025": { - "id": "command-a-reasoning-08-2025", - "name": "Cohere Command A (08/2025)", - "description": "Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows", - "family": "command-a", + "alibaba/qwen3-max": { + "id": "alibaba/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" @@ -18120,31 +16502,35 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 8192 + "context": 262144, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 1.2, + "output": 6, + "cache_read": 0.24 } }, - "longcat-2.0:thinking": { - "id": "longcat-2.0:thinking", - "name": "LongCat 2.0 Thinking", - "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", - "family": "longcat", + "alibaba/qwen3-next-80b-a3b-thinking": { + "id": "alibaba/qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B Thinking", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "knowledge": "2025-09", + "release_date": "2025-09-11", + "last_updated": "2025-09", "modalities": { "input": [ "text" @@ -18153,29 +16539,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, - "input": 1048756, - "output": 262144 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.75, - "output": 3, - "cache_read": 0.015 + "input": 0.15, + "output": 1.2 } }, - "deepseek-r1-sambanova": { - "id": "deepseek-r1-sambanova", - "name": "DeepSeek R1 Fast", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "deepseek", + "alibaba/qwen3-embedding-4b": { + "id": "alibaba/qwen3-embedding-4b", + "name": "Qwen3 Embedding 4B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-02-20", - "last_updated": "2025-02-20", + "temperature": true, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ "text" @@ -18184,44 +16568,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 - }, - "cost": { - "input": 4.998, - "output": 6.987, - "cache_read": 2.499 + "context": 32768, + "output": 32768 } }, - "gemini-2.5-flash-preview-04-17": { - "id": "gemini-2.5-flash-preview-04-17", - "name": "Gemini 2.5 Flash Preview", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", + "alibaba/qwen3.5-flash": { + "id": "alibaba/qwen3.5-flash", + "name": "Qwen 3.5 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-17", - "last_updated": "2025-04-17", + "tool_call": true, + "temperature": true, + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" @@ -18229,29 +16607,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 0.1, + "output": 0.4, + "cache_read": 0.001, + "cache_write": 0.125 } }, - "deepseek-reasoner": { - "id": "deepseek-reasoner", - "name": "DeepSeek Reasoner", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "alibaba/qwen3-coder": { + "id": "alibaba/qwen3-coder", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "knowledge": "2025-04", + "release_date": "2025-07-22", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -18260,45 +16638,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 64000, - "input": 64000, + "context": 262144, "output": 65536 }, "cost": { - "input": 0.4, - "output": 1.7, - "cache_read": 0.2 + "input": 1.5, + "output": 7.5, + "cache_read": 0.3 } }, - "gemini-2.5-flash-lite-preview-09-2025": { - "id": "gemini-2.5-flash-lite-preview-09-2025", - "name": "Gemini 2.5 Flash Lite Preview (09/2025)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", - "attachment": true, + "alibaba/qwen-3-235b": { + "id": "alibaba/qwen-3-235b", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -18306,42 +16672,41 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 + "input": 0.22, + "output": 0.88 } }, - "gemini-2.5-flash-preview-04-17:thinking": { - "id": "gemini-2.5-flash-preview-04-17:thinking", - "name": "Gemini 2.5 Flash Preview Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", + "alibaba/qwen3.5-plus": { + "id": "alibaba/qwen3.5-plus", + "name": "Qwen 3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-17", - "last_updated": "2025-04-17", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" @@ -18349,253 +16714,268 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 3.5, - "cache_read": 0.015 + "input": 0.4, + "output": 2.4, + "cache_read": 0.04, + "cache_write": 0.5 } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek R1", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", + "alibaba/wan-v2.6-t2v": { + "id": "alibaba/wan-v2.6-t2v", + "name": "Wan v2.6 Text-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "o", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 - }, - "cost": { - "input": 0.4, - "output": 1.7, - "cache_read": 0.2 + "context": 0, + "output": 0 } }, - "gemini-2.5-pro-preview-06-05": { - "id": "gemini-2.5-pro-preview-06-05", - "name": "Gemini 2.5 Pro Preview 0605", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", - "attachment": true, + "alibaba/qwen3-max-thinking": { + "id": "alibaba/qwen3-max-thinking", + "name": "Qwen 3 Max Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-23", + "last_updated": "2025-01", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, + "context": 256000, "output": 65536 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 0.25 + "input": 1.2, + "output": 6, + "cache_read": 0.24 } }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, + "alibaba/wan-v2.6-i2v-flash": { + "id": "alibaba/wan-v2.6-i2v-flash", + "name": "Wan v2.6 Image-to-Video Flash", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "o", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 + "context": 0, + "output": 0 } }, - "deepseek-chat-cheaper": { - "id": "deepseek-chat-cheaper", - "name": "DeepSeek V3/Chat Cheaper", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "deepseek", - "attachment": true, + "alibaba/wan-v2.6-r2v-flash": { + "id": "alibaba/wan-v2.6-r2v-flash", + "name": "Wan v2.6 Reference-to-Video Flash", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "o", + "attachment": false, "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "video" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "alibaba/qwen3-coder-next": { + "id": "alibaba/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "temperature": true, + "release_date": "2025-07-22", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.1, - "output": 0.425, - "cache_read": 0.05 + "input": 0.5, + "output": 1.2 } }, - "GLM-4.6-Derestricted-v5": { - "id": "GLM-4.6-Derestricted-v5", - "name": "GLM 4.6 Derestricted v5", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "glm", - "attachment": false, + "alibaba/qwen3.6-27b": { + "id": "alibaba/qwen3.6-27b", + "name": "Qwen 3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 131072 } ], "tool_call": true, - "structured_output": false, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 8192 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.4, - "output": 1.5, - "cache_read": 0.2 + "input": 0.6, + "output": 3.6 } }, - "doubao-1.5-vision-pro-32k": { - "id": "doubao-1.5-vision-pro-32k", - "name": "Doubao 1.5 Vision Pro 32k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "alibaba/wan-v2.6-i2v": { + "id": "alibaba/wan-v2.6-i2v", + "name": "Wan v2.6 Image-to-Video", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "o", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-11-20", - "last_updated": "2025-01-22", + "temperature": true, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 8192 - }, - "cost": { - "input": 0.459, - "output": 1.377, - "cache_read": 0.2295 + "context": 0, + "output": 0 } }, - "glm-z1-airx": { - "id": "glm-z1-airx", - "name": "GLM Z1 AirX", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "glm", + "alibaba/wan-v2.7-t2v": { + "id": "alibaba/wan-v2.7-t2v", + "name": "Wan v2.7 Text-to-Video", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "o", "attachment": false, "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "video" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + }, + "alibaba/qwen-3-30b": { + "id": "alibaba/qwen-3-30b", + "name": "Qwen3-30B-A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -18606,30 +16986,38 @@ }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, + "context": 40960, "output": 16384 }, "cost": { - "input": 0.7, - "output": 0.7, - "cache_read": 0.35 + "input": 0.12, + "output": 0.5 } }, - "auto-model-basic": { - "id": "auto-model-basic", - "name": "Auto model (Basic)", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": false, - "reasoning": false, + "alibaba/qwen3-235b-a22b-thinking": { + "id": "alibaba/qwen3-235b-a22b-thinking", + "name": "Qwen3 235B A22B Thinking 2507", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1, + "max": 81920 + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2025-04-16", - "last_updated": "2024-06-01", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-04", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -18637,26 +17025,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 9.996, - "output": 19.992, - "cache_read": 4.998 + "input": 0.4, + "output": 4 } }, - "qwen3-vl-235b-a22b-instruct-original": { - "id": "qwen3-vl-235b-a22b-instruct-original", - "name": "Qwen3 VL 235B A22B Instruct Original", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "alibaba/qwen3-vl-235b-a22b-instruct": { + "id": "alibaba/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "temperature": true, + "release_date": "2025-09-23", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", @@ -18668,70 +17055,71 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 131072, + "output": 129024 }, "cost": { - "input": 0.3, - "output": 1.9, - "cache_read": 0.15 + "input": 0.4, + "output": 1.6 } }, - "gemma-4-26b-a4b-it-luminous": { - "id": "gemma-4-26b-a4b-it-luminous", - "name": "Luminous Mirror", - "description": "Luminous Mirror is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", - "family": "gemma", - "attachment": true, - "reasoning": false, + "alibaba/qwen-3-14b": { + "id": "alibaba/qwen-3-14b", + "name": "Qwen3-14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 40960, + "output": 16384 }, "cost": { "input": 0.12, - "output": 0.38, - "cache_read": 0.06 + "output": 0.24 } }, - "sarvam-105b": { - "id": "sarvam-105b", - "name": "Sarvam 105B", - "description": "Flagship Indian-language reasoning model for enterprise multilingual applications", - "family": "sarvam", + "alibaba/qwen-3-32b": { + "id": "alibaba/qwen-3-32b", + "name": "Qwen 3.32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 38912 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-09-01", - "last_updated": "2025-09-01", + "knowledge": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -18740,34 +17128,40 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 4096 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.054, - "output": 0.2124, - "cache_read": 0.0336 + "input": 0.16, + "output": 0.64 } }, - "gemma-4-31b-it-gembrain": { - "id": "gemma-4-31b-it-gembrain", - "name": "Gembrain", - "description": "Gembrain is a multimodal Gemma 4 31B creative finetune for expressive dialogue, long-form storytelling, and roleplay.", - "family": "gemma", + "alibaba/qwen-3.6-max-preview": { + "id": "alibaba/qwen-3.6-max-preview", + "name": "Qwen 3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 131072 + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "temperature": true, + "release_date": "2026-04-20", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -18775,27 +17169,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 240000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.45, - "cache_read": 0.05 + "input": 1.3, + "output": 7.8, + "cache_read": 0.26, + "cache_write": 1.625 } }, - "phi-4-multimodal-instruct": { - "id": "phi-4-multimodal-instruct", - "name": "Phi 4 Multimodal", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "phi", + "alibaba/qwen3-coder-30b-a3b": { + "id": "alibaba/qwen3-coder-30b-a3b", + "name": "Qwen 3 Coder 30B A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-31", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -18804,43 +17200,43 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 262144, + "output": 8192 }, "cost": { - "input": 0.07, - "output": 0.11, - "cache_read": 0.035 + "input": 0.15, + "output": 0.6 } }, - "gemini-2.5-flash-preview-05-20:thinking": { - "id": "gemini-2.5-flash-preview-05-20:thinking", - "name": "Gemini 2.5 Flash 0520 Thinking", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemini", + "alibaba/qwen3.6-plus": { + "id": "alibaba/qwen3.6-plus", + "name": "Qwen 3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 131072 } ], - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -18848,103 +17244,97 @@ }, "open_weights": false, "limit": { - "context": 1048000, - "input": 1048000, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 3.5, - "cache_read": 0.015 + "input": 0.5, + "output": 3, + "cache_read": 0.1, + "cache_write": 0.625 } }, - "gemma-4-26b-a4b-it-musica": { - "id": "gemma-4-26b-a4b-it-musica", - "name": "Musica", - "description": "Musica is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", - "family": "gemma", - "attachment": true, + "meituan/longcat-flash-thinking-2601": { + "id": "meituan/longcat-flash-thinking-2601", + "name": "LongCat Flash Thinking 2601", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "longcat", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "temperature": true, + "release_date": "2026-01-15", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, + "context": 32768, "output": 32768 - }, - "cost": { - "input": 0.12, - "output": 0.38, - "cache_read": 0.06 } }, - "gemma-4-26b-a4b-it-chimerax": { - "id": "gemma-4-26b-a4b-it-chimerax", - "name": "Chimera X", - "description": "Chimera X is a Gemma 4 26B A4B multimodal mixture-of-experts fine-tune for creative writing, expressive dialogue, and roleplay.", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "meituan/longcat-flash-chat": { + "id": "meituan/longcat-flash-chat", + "name": "LongCat Flash Chat", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "longcat", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-01", + "last_updated": "2025-08-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 - }, - "cost": { - "input": 0.12, - "output": 0.38, - "cache_read": 0.06 + "context": 128000, + "output": 100000 } }, - "hermes-medium": { - "id": "hermes-medium", - "name": "Hermes Medium", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "hermes", + "meta/muse-spark-1.1": { + "id": "meta/muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "muse", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "temperature": true, + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text" @@ -18956,26 +17346,26 @@ "open_weights": false, "limit": { "context": 1048576, - "input": 1048576, - "output": 131072 + "output": 1048576 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } }, - "doubao-seed-1-6-250615": { - "id": "doubao-seed-1-6-250615", - "name": "Doubao Seed 1.6", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "seed", + "meta/llama-3.2-1b": { + "id": "meta/llama-3.2-1b", + "name": "Llama 3.2 1B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-06-15", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-18", "modalities": { "input": [ "text" @@ -18986,36 +17376,26 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 16384 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.204, - "output": 0.51, - "cache_read": 0.102 + "input": 0.1, + "output": 0.1 } }, - "Gemma-4-31B-DarkIdol": { - "id": "Gemma-4-31B-DarkIdol", - "name": "Gemma 4 31B DarkIdol", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gemma", + "meta/llama-3.2-11b": { + "id": "meta/llama-3.2-11b", + "name": "Llama 3.2 11B Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text", @@ -19025,29 +17405,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.306, - "output": 0.306, - "cache_read": 0.153 + "input": 0.16, + "output": 0.16 } }, - "ernie-5.1": { - "id": "ernie-5.1", - "name": "ERNIE 5.1", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "ernie", + "meta/llama-3.1-8b": { + "id": "meta/llama-3.1-8b", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-05-10", - "last_updated": "2026-05-10", + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -19058,30 +17437,30 @@ }, "open_weights": false, "limit": { - "context": 119000, - "input": 119000, - "output": 64000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.75, - "output": 3, - "cache_read": 0.75 + "input": 0.22, + "output": 0.22 } }, - "doubao-seed-1-6-flash-250615": { - "id": "doubao-seed-1-6-flash-250615", - "name": "Doubao Seed 1.6 Flash", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "seed", - "attachment": false, + "meta/llama-3.2-90b": { + "id": "meta/llama-3.2-90b", + "name": "Llama 3.2 90B Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-06-15", + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -19089,27 +17468,26 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 16384 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.0374, - "output": 0.374, - "cache_read": 0.0187 + "input": 0.72, + "output": 0.72 } }, - "mistral-code-agent-latest": { - "id": "mistral-code-agent-latest", - "name": "Mistral Code Agent Latest", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "mistral", + "meta/llama-3.1-70b": { + "id": "meta/llama-3.1-70b", + "name": "Llama 3.1 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -19118,61 +17496,28 @@ "text" ] }, - "open_weights": true, - "limit": { - "context": 262144, - "input": 262144, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.2 - } - }, - "kimi-k2-instruct-fast": { - "id": "kimi-k2-instruct-fast", - "name": "Kimi K2 0711 Fast", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "kimi-k2", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-15", - "last_updated": "2025-07-15", - "modalities": { - "input": [ - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 16384 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.2 + "input": 0.72, + "output": 0.72 } }, - "featherless-ai/Qwerky-72B": { - "id": "featherless-ai/Qwerky-72B", - "name": "Qwerky 72B", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "qwerky", + "meta/llama-3.2-3b": { + "id": "meta/llama-3.2-3b", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-03-20", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-18", "modalities": { "input": [ "text" @@ -19181,33 +17526,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "input": 32000, + "context": 128000, "output": 8192 }, "cost": { - "input": 0.5, - "output": 0.5, - "cache_read": 0.25 + "input": 0.15, + "output": 0.15 } }, - "unsloth/gemma-3-27b-it": { - "id": "unsloth/gemma-3-27b-it", - "name": "Gemma 3 27B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "unsloth", + "meta/llama-4-scout": { + "id": "meta/llama-4-scout", + "name": "Llama-4-Scout-17B-16E-Instruct-FP8", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-03-10", + "tool_call": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" @@ -19216,62 +17560,28 @@ "open_weights": true, "limit": { "context": 128000, - "input": 128000, - "output": 96000 + "output": 4096 }, "cost": { - "input": 0.2992, - "output": 0.2992, - "cache_read": 0.1496 + "input": 0, + "output": 0 } }, - "unsloth/gemma-3-12b-it": { - "id": "unsloth/gemma-3-12b-it", - "name": "Gemma 3 12B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "unsloth", + "meta/llama-3.3-70b": { + "id": "meta/llama-3.3-70b", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-03-10", + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "pdf" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "input": 131072, - "output": 16384 - }, - "cost": { - "input": 0.272, - "output": 0.272, - "cache_read": 0.136 - } - }, - "unsloth/gemma-3-4b-it": { - "id": "unsloth/gemma-3-4b-it", - "name": "Gemma 3 4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "unsloth", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-03-10", - "modalities": { - "input": [ - "text", - "pdf" ], "output": [ "text" @@ -19280,141 +17590,105 @@ "open_weights": true, "limit": { "context": 128000, - "input": 128000, - "output": 8192 + "output": 4096 }, "cost": { - "input": 0.2006, - "output": 0.2006, - "cache_read": 0.1003 + "input": 0, + "output": 0 } }, - "bytedance-seed/seed-2.0-lite": { - "id": "bytedance-seed/seed-2.0-lite", - "name": "ByteDance Seed 2.0 Lite", - "description": "Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation", - "family": "seed", + "meta/llama-4-maverick": { + "id": "meta/llama-4-maverick", + "name": "Llama-4-Maverick-17B-128E-Instruct-FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "input": 262144, - "output": 131072 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.125 - } - }, - "bytedance-seed/seed-2-1-turbo": { - "id": "bytedance-seed/seed-2-1-turbo", - "name": "ByteDance Seed 2.1 Turbo", - "description": "ByteDance Seed 2.1 Turbo is a multimodal model for coding and long-horizon agent workflows, including end-to-end software delivery and multi-step task execution. It supports text, image, and video input with a 262k context window.", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 235929 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.25 + "input": 0, + "output": 0 } }, - "bytedance-seed/seed-2.0-code": { - "id": "bytedance-seed/seed-2.0-code", - "name": "ByteDance Seed 2.0 Code", - "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", - "family": "seed", - "attachment": true, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-05", + "release_date": "2026-04-23", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.25 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "poolside/laguna-s-2.1:thinking": { - "id": "poolside/laguna-s-2.1:thinking", - "name": "Laguna S 2.1 Thinking", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", + "deepseek/deepseek-v3.1-terminus": { + "id": "deepseek/deepseek-v3.1-terminus", + "name": "DeepSeek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2025-07", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ "text" @@ -19425,29 +17699,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.01 + "input": 0.27, + "output": 1, + "cache_read": 0.135 } }, - "poolside/laguna-s-2.1": { - "id": "poolside/laguna-s-2.1", - "name": "Laguna S 2.1", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", + "deepseek/deepseek-v3": { + "id": "deepseek/deepseek-v3", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2024-07", + "release_date": "2024-12-26", + "last_updated": "2024-12-26", "modalities": { "input": [ "text" @@ -19456,38 +17728,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.01 + "input": 0.27, + "output": 1.12, + "cache_read": 0.135 } }, - "abliteration-ai/abliterated-model-large-v2": { - "id": "abliteration-ai/abliterated-model-large-v2", - "name": "Abliterated Model Large V2", - "description": "Abliteration.ai's default large text reasoning model is derived from GLM-5.3 for harder reasoning and evaluation workloads, with automatic prompt caching and a one-million-token context window.", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-08-31", - "last_updated": "2026-08-31", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-23", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -19496,46 +17772,34 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "input": 1000000, - "output": 999990 + "output": 384000 }, "cost": { - "input": 3, - "output": 5, - "cache_read": 0.3 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } }, - "abliteration-ai/abliterated-model": { - "id": "abliteration-ai/abliterated-model", - "name": "Abliterated Model", - "description": "Abliteration.ai's multimodal reasoning model supports text and image input, structured output, automatic prompt caching, and a 262K-token context window.", - "attachment": true, + "deepseek/deepseek-v3.2-thinking": { + "id": "deepseek/deepseek-v3.2-thinking", + "name": "DeepSeek V3.2 Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": false, - "structured_output": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "reasoning_options": [], + "tool_call": true, + "interleaved": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -19543,40 +17807,31 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 262134 + "context": 128000, + "output": 8000 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.1 + "input": 0.62, + "output": 1.85 } }, - "abliteration-ai/abliterated-model-large": { - "id": "abliteration-ai/abliterated-model-large", - "name": "Abliterated Model Large", - "description": "Abliteration.ai's large text reasoning model is derived from GLM-5.2 and supports native tool calling, structured output, automatic prompt caching, and a one-million-token context window.", + "deepseek/deepseek-v3.1": { + "id": "deepseek/deepseek-v3.1", + "name": "DeepSeek-V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ "text" @@ -19587,86 +17842,61 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 999990 + "context": 128000, + "output": 128000 }, "cost": { - "input": 3, - "output": 5, - "cache_read": 0.3 + "input": 0.6, + "output": 1.7 } }, - "prism-ml/ternary-bonsai-2-27b": { - "id": "prism-ml/ternary-bonsai-2-27b", - "name": "Ternary Bonsai 2 27B", - "description": "Ternary Bonsai 2 27B is a 27B-parameter reasoning model from PrismML derived from Qwen3.8-27B. It supports coding, mathematics, tool calling, and image understanding with a 262K-token context window.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-09-18", - "last_updated": "2026-09-18", + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 128000, + "output": 8000 }, "cost": { - "input": 0.075, - "output": 0.5, - "cache_read": 0.0375 + "input": 0.28, + "output": 0.42, + "cache_read": 0.028 } }, - "anthropic/claude-opus-4.5:thinking": { - "id": "anthropic/claude-opus-4.5:thinking", - "name": "Claude 4.5 Opus Thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "deepseek/deepseek-r1": { + "id": "deepseek/deepseek-r1", + "name": "DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -19674,82 +17904,61 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 1.35, + "output": 5.4 } }, - "anthropic/claude-sonnet-4.5:thinking": { - "id": "anthropic/claude-sonnet-4.5:thinking", - "name": "Claude Sonnet 4.5 Thinking", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax M2.7 High Speed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 + "context": 204800, + "output": 131100 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "anthropic/claude-opus-4.6:thinking:low": { - "id": "anthropic/claude-opus-4.6:thinking:low", - "name": "Claude 4.6 Opus Thinking Low", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -19757,34 +17966,33 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 204800, + "output": 131000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude 4.1 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": false, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-10", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -19792,41 +18000,32 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude 4.5 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax-m3", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2026-05-31", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", @@ -19837,129 +18036,98 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude 4.6 Opus", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2024-10", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 205000, + "output": 205000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude 4.7 Opus", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "Minimax M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 204800, + "output": 131000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "anthropic/claude-sonnet-4:thinking:64000": { - "id": "anthropic/claude-sonnet-4:thinking:64000", - "name": "Claude 4 Sonnet Thinking (64K)", - "description": "Claude 4 Sonnet with maximum thinking budget (64,000 tokens).", - "family": "claude-sonnet", - "attachment": true, + "minimax/minimax-m2.5-highspeed": { + "id": "minimax/minimax-m2.5-highspeed", + "name": "MiniMax M2.5 High Speed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -19967,44 +18135,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 64000 + "context": 204800, + "output": 131000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.6, + "output": 2.4, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "anthropic/claude-opus-latest": { - "id": "anthropic/claude-opus-latest", - "name": "Claude Opus Latest", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "minimax/minimax-m2.1-lightning": { + "id": "minimax/minimax-m2.1-lightning", + "name": "MiniMax M2.1 Lightning", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-29", - "last_updated": "2026-03-29", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-12-23", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -20012,36 +18168,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.3, + "output": 2.4, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "anthropic/claude-opus-4.1:thinking:1024": { - "id": "anthropic/claude-opus-4.1:thinking:1024", - "name": "Claude 4.1 Opus Thinking (1K)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "kwaipilot/kat-coder-pro-v1": { + "id": "kwaipilot/kat-coder-pro-v1", + "name": "KAT-Coder-Pro V1", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "kat-coder", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-10", + "release_date": "2025-11-09", + "last_updated": "2025-10-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -20049,42 +18201,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, + "context": 256000, "output": 32000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "anthropic/claude-haiku-latest": { - "id": "anthropic/claude-haiku-latest", - "name": "Claude Haiku Latest", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "kwaipilot/kat-coder-pro-v2": { + "id": "kwaipilot/kat-coder-pro-v2", + "name": "Kat Coder Pro V2", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "kat-coder", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-29", - "last_updated": "2026-03-29", + "temperature": true, + "release_date": "2026-03-27", + "last_updated": "2026-03-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -20092,80 +18232,71 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } - }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + } + } + }, + "submodel": { + "id": "submodel", + "env": [ + "SUBMODEL_INSTAGEN_ACCESS_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://llm.submodel.ai/v1", + "name": "submodel", + "doc": "https://submodel.gitbook.io", + "models": { + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen3 235B A22B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.2, + "output": 0.6 } }, - "anthropic/claude-sonnet-4:thinking:8192": { - "id": "anthropic/claude-sonnet-4:thinking:8192", - "name": "Claude 4 Sonnet Thinking (8K)", - "description": "Claude 4 Sonnet with reduced thinking budget (8,192 tokens).", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -20173,197 +18304,146 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.2, + "output": 0.8 } }, - "anthropic/claude-sonnet-4:thinking:1024": { - "id": "anthropic/claude-sonnet-4:thinking:1024", - "name": "Claude 4 Sonnet Thinking (1K)", - "description": "Claude 4 Sonnet with minimal thinking budget (1,024 tokens).", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 64000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.2, + "output": 0.3 } }, - "anthropic/claude-opus-4.6:thinking": { - "id": "anthropic/claude-opus-4.6:thinking", - "name": "Claude 4.6 Opus Thinking", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2025-08-23", + "last_updated": "2025-08-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.1, + "output": 0.5 } }, - "anthropic/claude-sonnet-latest": { - "id": "anthropic/claude-sonnet-latest", - "name": "Claude Sonnet Latest", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "zai-org/GLM-4.5-FP8": { + "id": "zai-org/GLM-4.5-FP8", + "name": "GLM 4.5 FP8", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "temperature": true, + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.2, + "output": 0.8 } }, - "anthropic/claude-opus-4.6:thinking:medium": { - "id": "anthropic/claude-opus-4.6:thinking:medium", - "name": "Claude 4.6 Opus Thinking Medium", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "zai-org/GLM-4.5-Air": { + "id": "zai-org/GLM-4.5-Air", + "name": "GLM 4.5 Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.1, + "output": 0.5 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, + "deepseek-ai/DeepSeek-V3-0324": { + "id": "deepseek-ai/DeepSeek-V3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2025-08-23", + "last_updated": "2025-08-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -20371,46 +18451,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 + "context": 75000, + "output": 163840 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1 + "input": 0.2, + "output": 0.8 } }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -20418,45 +18481,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 75000, + "output": 163840 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.5, + "output": 2.15 } }, - "anthropic/claude-fable-latest": { - "id": "anthropic/claude-fable-latest", - "name": "Claude Fable Latest", - "description": "Compatibility alias for Claude Fable.", - "family": "claude-fable", - "attachment": true, + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "release_date": "2025-08-23", + "last_updated": "2025-08-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -20464,305 +18511,293 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 75000, + "output": 163840 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 0.2, + "output": 0.8 + } + } + } + }, + "huggingface": { + "id": "huggingface", + "env": [ + "HF_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://router.huggingface.co/v1", + "name": "Hugging Face", + "doc": "https://huggingface.co/docs/inference-providers", + "models": { + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 4096 + }, + "cost": { + "input": 0.59, + "output": 0.79 } }, - "anthropic/claude-sonnet-4:thinking:32768": { - "id": "anthropic/claude-sonnet-4:thinking:32768", - "name": "Claude 4 Sonnet Thinking (32K)", - "description": "Claude 4 Sonnet with extended thinking budget (32,768 tokens).", - "family": "claude-sonnet", - "attachment": true, + "moonshotai/Kimi-K2-Thinking": { + "id": "moonshotai/Kimi-K2-Thinking", + "name": "Kimi-K2-Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "anthropic/claude-fable-5.1": { - "id": "anthropic/claude-fable-5.1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "moonshotai/Kimi-K2-Instruct-0905": { + "id": "moonshotai/Kimi-K2-Instruct-0905", + "name": "Kimi-K2-Instruct-0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-04", + "last_updated": "2025-09-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 1, + "output": 3 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "moonshotai/Kimi-K2-Instruct": { + "id": "moonshotai/Kimi-K2-Instruct", + "name": "Kimi-K2-Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2024-10", + "release_date": "2025-07-14", + "last_updated": "2025-07-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 1, + "output": 3 } }, - "anthropic/claude-opus-4.1:thinking:8192": { - "id": "anthropic/claude-opus-4.1:thinking:8192", - "name": "Claude 4.1 Opus Thinking (8K)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi-K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "anthropic/claude-haiku-4.5:thinking": { - "id": "anthropic/claude-haiku-4.5:thinking", - "name": "Claude Haiku 4.5 Thinking", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-01", + "release_date": "2026-01-01", + "last_updated": "2026-01-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1, - "output": 5, + "input": 0.6, + "output": 3, "cache_read": 0.1 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.95, + "output": 4 } }, - "anthropic/claude-opus-4:thinking:1024": { - "id": "anthropic/claude-opus-4:thinking:1024", - "name": "Claude 4 Opus Thinking (1K)", - "description": "Claude 4 Opus with minimal thinking budget (1,024 tokens).", - "family": "claude-opus", - "attachment": true, + "stepfun-ai/Step-3.5-Flash": { + "id": "stepfun-ai/Step-3.5-Flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 262144, + "output": 256000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5 + "input": 0.1, + "output": 0.3 } }, - "anthropic/claude-opus-4.8:thinking": { - "id": "anthropic/claude-opus-4.8:thinking", - "name": "Claude Opus 4.8 Thinking", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "stepfun-ai/Step-3.7-Flash": { + "id": "stepfun-ai/Step-3.7-Flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -20771,636 +18806,559 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "knowledge": "2026-01-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 256000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.2, + "output": 1.15 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude 4 Sonnet", - "description": "Claude 4 Sonnet by Anthropic. A new generation model with improved capabilities, especially on programming and development. NOTE: Inputs > 200k tokens are charged at 2x input, 1.5x output rate.", - "family": "claude-sonnet", + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.14, + "output": 0.4 } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "google/gemma-4-26B-A4B-it": { + "id": "google/gemma-4-26B-A4B-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.13, + "output": 0.4 } }, - "anthropic/claude-sonnet-4.6:thinking": { - "id": "anthropic/claude-sonnet-4.6:thinking", - "name": "Claude Sonnet 4.6 Thinking", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.15, + "output": 0.95 } }, - "anthropic/claude-opus-4:thinking:32768": { - "id": "anthropic/claude-opus-4:thinking:32768", - "name": "Claude 4 Opus Thinking (32K)", - "description": "Claude 4 Opus with extended thinking budget (32,768 tokens).", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen3-Coder-Next": { + "id": "Qwen/Qwen3-Coder-Next", + "name": "Qwen3-Coder-Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5 + "input": 0.2, + "output": 1.5 } }, - "anthropic/claude-opus-4.6:thinking:max": { - "id": "anthropic/claude-opus-4.6:thinking:max", - "name": "Claude 4.6 Opus Thinking Max", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "Qwen/Qwen3-Embedding-8B": { + "id": "Qwen/Qwen3-Embedding-8B", + "name": "Qwen 3 Embedding 8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32000, + "output": 4096 + }, + "cost": { + "input": 0.01, + "output": 0 + } + }, + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.47, + "output": 3.19 } }, - "anthropic/claude-sonnet-5:thinking": { - "id": "anthropic/claude-sonnet-5:thinking", - "name": "Claude Sonnet 5 Thinking", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "Qwen/Qwen3-Next-80B-A3B-Instruct": { + "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", + "name": "Qwen3-Next-80B-A3B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-11", + "last_updated": "2025-09-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 66536 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.25, + "output": 1 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen3.5-397B-A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-04", + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.6, + "output": 3.6 } }, - "anthropic/claude-opus-4.7:thinking": { - "id": "anthropic/claude-opus-4.7:thinking", - "name": "Claude 4.7 Opus Thinking", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen3-235B-A22B-Thinking-2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.3, + "output": 3 } }, - "anthropic/claude-opus-5.5": { - "id": "anthropic/claude-opus-5.5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "Qwen/Qwen3-Embedding-4B": { + "id": "Qwen/Qwen3-Embedding-4B", + "name": "Qwen 3 Embedding 4B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 32000, + "output": 2048 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 + "input": 0.01, + "output": 0 } }, - "anthropic/claude-opus-4.1:thinking": { - "id": "anthropic/claude-opus-4.1:thinking", - "name": "Claude 4.1 Opus Thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen3-Coder-480B-A35B-Instruct": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "name": "Qwen3-Coder-480B-A35B-Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 262144, + "output": 66536 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5 + "input": 2, + "output": 2 } }, - "anthropic/claude-sonnet-4:thinking": { - "id": "anthropic/claude-sonnet-4:thinking", - "name": "Claude 4 Sonnet Thinking", - "description": "Anthropic's Claude 4 Sonnet with the ability to show its thinking process step by step.", - "family": "claude-sonnet", + "Qwen/Qwen3.5-122B-A10B": { + "id": "Qwen/Qwen3.5-122B-A10B", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2025-02-24", - "last_updated": "2025-02-24", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 64000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.4, + "output": 3.2 } }, - "anthropic/claude-opus-4:thinking": { - "id": "anthropic/claude-opus-4:thinking", - "name": "Claude 4 Opus Thinking", - "description": "Anthropic's Claude 4 Opus with the ability to show its thinking process step by step.", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen3-Coder-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5 + "input": 0.07, + "output": 0.26 } }, - "anthropic/claude-opus-4:thinking:8192": { - "id": "anthropic/claude-opus-4:thinking:8192", - "name": "Claude 4 Opus Thinking (8K)", - "description": "Claude 4 Opus with reduced thinking budget (8,192 tokens).", - "family": "claude-opus", + "Qwen/Qwen3.5-27B": { + "id": "Qwen/Qwen3.5-27B", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5 + "input": 0.3, + "output": 2.4 } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Claude 4 Opus", - "description": "Claude 4 Opus by Anthropic. The premium version of the new Claude models. A new generation model with improved capabilities, especially on programming and development.", - "family": "claude-opus", - "attachment": true, + "Qwen/Qwen3-Next-80B-A3B-Thinking": { + "id": "Qwen/Qwen3-Next-80B-A3B-Thinking", + "name": "Qwen3-Next-80B-A3B-Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-11", + "last_updated": "2025-09-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5 + "input": 0.3, + "output": 2 } }, - "anthropic/claude-opus-4.1:thinking:32768": { - "id": "anthropic/claude-opus-4.1:thinking:32768", - "name": "Claude 4.1 Opus Thinking (32K)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "Qwen/Qwen3.5-9B": { + "id": "Qwen/Qwen3.5-9B", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 32000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5 + "input": 0.17, + "output": 0.25 } }, - "cohere/command-r-plus-08-2024": { - "id": "cohere/command-r-plus-08-2024", - "name": "Cohere: Command R+", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -21411,39 +19369,28 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 131072, + "output": 16384 }, "cost": { - "input": 2.856, - "output": 14.246, - "cache_read": 1.428 + "input": 0.29, + "output": 0.59 } }, - "deepseek/deepseek-v4-flash:thinking": { - "id": "deepseek/deepseek-v4-flash:thinking", - "name": "DeepSeek V4 Flash (Thinking)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "Qwen/Qwen3-235B-A22B": { + "id": "Qwen/Qwen3-235B-A22B", + "name": "Qwen3 235B-A22B", + "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -21454,43 +19401,31 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 384000 + "context": 40960, + "output": 16384 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.2, + "output": 0.8 } }, - "deepseek/deepseek-v4-flash-0731": { - "id": "deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "Qwen/Qwen3.5-35B-A3B": { + "id": "Qwen/Qwen3.5-35B-A3B", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -21498,44 +19433,39 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.05, - "output": 0.16, - "cache_read": 0.013 + "input": 0.25, + "output": 2 } }, - "deepseek/deepseek-v4.1-flash": { - "id": "deepseek/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -21543,44 +19473,39 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 384000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.13, - "output": 0.52, - "cache_read": 0.006 + "input": 0.25, + "output": 0.69 } }, - "deepseek/deepseek-v4.1-flash:thinking": { - "id": "deepseek/deepseek-v4.1-flash:thinking", - "name": "DeepSeek V4.1 Flash Thinking", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -21588,43 +19513,31 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 384000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.13, - "output": 0.52, - "cache_read": 0.006 + "input": 0.1, + "output": 0.5 } }, - "deepseek/deepseek-v4-flash-vision-exp": { - "id": "deepseek/deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", - "attachment": true, + "XiaomiMiMo/MiMo-V2.5-Pro": { + "id": "XiaomiMiMo/MiMo-V2.5-Pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -21633,33 +19546,29 @@ "open_weights": true, "limit": { "context": 1048576, - "input": 1048576, - "output": 384000 + "output": 131072 }, "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 + "input": 1, + "output": 3 } }, - "deepseek/deepseek-v3.2:thinking": { - "id": "deepseek/deepseek-v3.2:thinking", - "name": "DeepSeek V3.2 Thinking", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", - "attachment": true, + "XiaomiMiMo/MiMo-V2-Flash": { + "id": "XiaomiMiMo/MiMo-V2-Flash", + "name": "MiMo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "knowledge": "2024-12", + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -21667,38 +19576,30 @@ }, "open_weights": true, "limit": { - "context": 163000, - "input": 163000, - "output": 65536 + "context": 262144, + "output": 4096 }, "cost": { - "input": 0.28, - "output": 0.42, - "cache_read": 0.14 + "input": 0.1, + "output": 0.3 } }, - "deepseek/deepseek-latest": { - "id": "deepseek/deepseek-latest", - "name": "DeepSeek Latest", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "zai-org/GLM-4.7-Flash": { + "id": "zai-org/GLM-4.7-Flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ "text" @@ -21709,39 +19610,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 384000 + "context": 200000, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 2.5, - "cache_read": 0.04 + "input": 0, + "output": 0 } }, - "deepseek/deepseek-v4-pro:thinking": { - "id": "deepseek/deepseek-v4-pro:thinking", - "name": "DeepSeek V4 Pro (Thinking)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "zai-org/GLM-4.6": { + "id": "zai-org/GLM-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -21752,39 +19641,29 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 384000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.1, - "output": 2.2, - "cache_read": 0.11 + "input": 0.55, + "output": 2.2 } }, - "deepseek/deepseek-v4-pro-0813:thinking": { - "id": "deepseek/deepseek-v4-pro-0813:thinking", - "name": "DeepSeek V4 Pro 0813 Thinking", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -21795,39 +19674,31 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 384000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 1.1, - "output": 2.5, - "cache_read": 0.04 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "deepseek/deepseek-v4-pro-0813": { - "id": "deepseek/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "zai-org/GLM-4.7": { + "id": "zai-org/GLM-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -21838,43 +19709,32 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 384000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.1, - "output": 2.5, - "cache_read": 0.04 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "deepseek/deepseek-v4-flash-0731:thinking": { - "id": "deepseek/deepseek-v4-flash-0731:thinking", - "name": "DeepSeek V4 Flash 0731 (Thinking)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "zai-org/GLM-4.5V": { + "id": "zai-org/GLM-4.5V", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -21882,39 +19742,27 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 131072 + "context": 65536, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.16, - "cache_read": 0.013 + "input": 0.6, + "output": 1.8 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -21925,38 +19773,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 384000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 1.1, - "output": 2.2, - "cache_read": 0.11 + "input": 1.4, + "output": 4.4 } }, - "deepseek/deepseek-v4-flash-latest": { - "id": "deepseek/deepseek-v4-flash-latest", - "name": "DeepSeek V4 Flash Latest", - "description": "Compatibility alias that routes to the newest dated DeepSeek V4 Flash release. Currently routes to DeepSeek V4 Flash 0731. ⚠️ This route goes directly to DeepSeek, so privacy and logging guarantees are limited.", - "family": "deepseek", + "zai-org/GLM-4.5-Air": { + "id": "zai-org/GLM-4.5-Air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-08-02", - "last_updated": "2026-08-02", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -21967,33 +19804,32 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 131072 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.05, - "output": 0.16, - "cache_read": 0.013 + "input": 0.13, + "output": 0.85 } }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", - "attachment": true, - "reasoning": false, + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2026-04-03", + "last_updated": "2026-04-03", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -22001,39 +19837,28 @@ }, "open_weights": true, "limit": { - "context": 163000, - "input": 163000, - "output": 65536 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0.28, - "output": 0.42, - "cache_read": 0.14 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "zai-org/GLM-4.5": { + "id": "zai-org/GLM-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -22044,27 +19869,28 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 384000 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.6, + "output": 2.2 } }, - "baseten/Kimi-K2-Instruct-FP4": { - "id": "baseten/Kimi-K2-Instruct-FP4", - "name": "Kimi K2 0711 Instruct FP4", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "deepseek-ai/DeepSeek-R1": { + "id": "deepseek-ai/DeepSeek-R1", + "name": "DeepSeek-R1", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-07-11", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" @@ -22075,62 +19901,62 @@ }, "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 131072 + "context": 64000, + "output": 32768 }, "cost": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.2 + "input": 0.7, + "output": 2.5 } }, - "unbiased/pareto": { - "id": "unbiased/pareto", - "name": "Pareto", - "description": "Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer", - "attachment": true, - "reasoning": false, + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", + "name": "DeepSeek-R1-0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", + "knowledge": "2025-05", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 131072 + "context": 163840, + "output": 163840 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.25 + "input": 3, + "output": 5 } }, - "lightonai/LightOnOCR-2-1B": { - "id": "lightonai/LightOnOCR-2-1B", - "name": "LightOnOCR 2", - "description": "LightOnOCR 2 hosted by IONOS in Berlin, Germany. Zero data retention.", - "attachment": true, - "reasoning": false, + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -22138,26 +19964,38 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0.1785, - "output": 0.3465 + "input": 0.14, + "output": 0.28 } }, - "abacusai/Dracarys-72B-Instruct": { - "id": "abacusai/Dracarys-72B-Instruct", - "name": "Llama 3.1 70B Dracarys 2", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-02", - "last_updated": "2025-08-02", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -22168,27 +20006,28 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "shisa-ai/shisa-v2.1-llama3.3-70b": { - "id": "shisa-ai/shisa-v2.1-llama3.3-70b", - "name": "Shisa V2.1 Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-12-06", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -22199,27 +20038,30 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 4096 + "context": 163840, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 0.5, - "cache_read": 0.25 + "input": 0.28, + "output": 0.4 } }, - "shisa-ai/shisa-v2-llama3.3-70b": { - "id": "shisa-ai/shisa-v2-llama3.3-70b", - "name": "Shisa V2 Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "MiniMaxAI/MiniMax-M2.1": { + "id": "MiniMaxAI/MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-10", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -22230,27 +20072,26 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 0.5, - "cache_read": 0.25 + "input": 0.3, + "output": 1.2 } }, - "Steelskull/L3.3-MS-Nevoria-70b": { - "id": "Steelskull/L3.3-MS-Nevoria-70b", - "name": "Steelskull Nevoria 70b", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "MiniMaxAI/MiniMax-M2": { + "id": "MiniMaxAI/MiniMax-M2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-12-06", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ "text" @@ -22261,27 +20102,29 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 204800, + "output": 128000 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.3, + "output": 1.2 } }, - "Steelskull/L3.3-Cu-Mai-R1-70b": { - "id": "Steelskull/L3.3-Cu-Mai-R1-70b", - "name": "Llama 3.3 70B Cu Mai", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-12-06", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -22292,30 +20135,32 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "Steelskull/L3.3-Nevoria-R1-70b": { - "id": "Steelskull/L3.3-Nevoria-R1-70b", - "name": "Steelskull Nevoria R1 70b", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-12-06", + "MiniMaxAI/MiniMax-M3": { + "id": "MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -22323,27 +20168,30 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 524288, + "output": 128000 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.3, + "output": 1.2 } }, - "Steelskull/L3.3-Electra-R1-70b": { - "id": "Steelskull/L3.3-Electra-R1-70b", - "name": "Steelskull Electra R1 70b", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "MiniMaxAI/MiniMax-M2.7": { + "id": "MiniMaxAI/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-12-06", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -22354,38 +20202,39 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.69989, - "output": 0.69989, - "cache_read": 0.349945 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } - }, - "tencent/hy3": { - "id": "tencent/hy3", - "name": "Tencent Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + } + } + }, + "minimax-coding-plan": { + "id": "minimax-coding-plan", + "env": [ + "MINIMAX_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://api.minimax.io/anthropic/v1", + "name": "MiniMax Token Plan (minimax.io)", + "doc": "https://platform.minimax.io/docs/token-plan/intro", + "models": { + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -22396,38 +20245,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.066, - "output": 0.26, - "cache_read": 0.029 + "input": 0, + "output": 0 } }, - "tencent/hy4-preview": { - "id": "tencent/hy4-preview", - "name": "Tencent Hy4 Preview", - "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", - "family": "Hy", + "MiniMax-M2.5-highspeed": { + "id": "MiniMax-M2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -22438,31 +20275,31 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 64000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.834, - "output": 2.501, - "cache_read": 0.042 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "anthracite-org/magnum-v4-72b": { - "id": "anthracite-org/magnum-v4-72b", - "name": "Magnum v4 72B", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-01-01", + "MiniMax-M2.7-highspeed": { + "id": "MiniMax-M2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -22470,27 +20307,28 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2.006, - "output": 2.992, - "cache_read": 1.003 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "anthracite-org/magnum-v2-72b": { - "id": "anthracite-org/magnum-v2-72b", - "name": "Magnum V2 72B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "MiniMax-M2": { + "id": "MiniMax-M2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-07-01", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ "text" @@ -22501,31 +20339,29 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 196608, + "output": 128000 }, "cost": { - "input": 2.006, - "output": 2.992, - "cache_read": 1.003 + "input": 0, + "output": 0 } }, - "meta-llama/llama-4-scout": { - "id": "meta-llama/llama-4-scout", - "name": "Llama 4 Scout", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": true, - "reasoning": false, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -22533,31 +20369,37 @@ }, "open_weights": true, "limit": { - "context": 328000, - "input": 328000, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.085, - "output": 0.46, - "cache_read": 0.0425 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "meta-llama/llama-4-maverick": { - "id": "meta-llama/llama-4-maverick", - "name": "Llama 4 Maverick", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-25", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -22565,29 +20407,28 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "meta-llama/llama-3.3-70b-instruct": { - "id": "meta-llama/llama-3.3-70b-instruct", - "name": "Llama 3.3 70b Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -22598,29 +20439,40 @@ }, "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.23, - "cache_read": 0.025 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "meta-llama/llama-3.1-8b-instruct": { - "id": "meta-llama/llama-3.1-8b-instruct", - "name": "Llama 3.1 8b Instruct", - "description": "Compact open Llama model for lightweight chat, drafting, and self-hosting", - "family": "llama", + } + } + }, + "novita-ai": { + "id": "novita-ai", + "env": [ + "NOVITA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.novita.ai/openai", + "name": "NovitaAI", + "doc": "https://novita.ai/docs/guides/introduction", + "models": { + "inclusionai/ling-2.6-1t": { + "id": "inclusionai/ling-2.6-1t", + "name": "Ling-2.6-1T", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "ling", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2026-04-23", + "last_updated": "2026-06-29", "modalities": { "input": [ "text" @@ -22631,59 +20483,59 @@ }, "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.0544, - "output": 0.085, - "cache_read": 0.0272 + "input": 0.3, + "output": 2.5, + "cache_read": 0.06 } }, - "meta-llama/llama-3.2-3b-instruct": { - "id": "meta-llama/llama-3.2-3b-instruct", - "name": "Llama 3.2 3b Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-09-25", + "inclusionai/ring-2.6-1t": { + "id": "inclusionai/ring-2.6-1t", + "name": "Ring-2.6-1T", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "ring", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-08", + "last_updated": "2026-05-27", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 8192 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.0306, - "output": 0.0493, - "cache_read": 0.0153 + "input": 0.3, + "output": 2.5, + "cache_read": 0.06 } }, - "LLM360/K2-Think": { - "id": "LLM360/K2-Think", - "name": "K2-Think", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "inclusionai/ling-2.6-flash": { + "id": "inclusionai/ling-2.6-flash", + "name": "Ling-2.6-flash", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -22694,27 +20546,26 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, + "context": 262144, "output": 32768 }, "cost": { - "input": 0.17, - "output": 0.68, - "cache_read": 0.085 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "deepcogito/cogito-v1-preview-qwen-32B": { - "id": "deepcogito/cogito-v1-preview-qwen-32B", - "name": "Cogito v1 Preview Qwen 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "meta-llama/llama-3.1-8b-instruct": { + "id": "meta-llama/llama-3.1-8b-instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-05-10", - "last_updated": "2025-05-10", + "temperature": true, + "release_date": "2024-07-24", + "last_updated": "2024-07-24", "modalities": { "input": [ "text" @@ -22725,27 +20576,26 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 32768 + "context": 16384, + "output": 16384 }, "cost": { - "input": 1.8, - "output": 1.8, - "cache_read": 0.9 + "input": 0.02, + "output": 0.05 } }, - "GalrionSoftworks/MN-LooseCannon-12B-v1": { - "id": "GalrionSoftworks/MN-LooseCannon-12B-v1", - "name": "MN-LooseCannon-12B-v1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + "meta-llama/llama-3-70b-instruct": { + "id": "meta-llama/llama-3-70b-instruct", + "name": "Llama3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-07-01", + "structured_output": true, + "temperature": true, + "release_date": "2024-04-25", + "last_updated": "2024-04-25", "modalities": { "input": [ "text" @@ -22756,58 +20606,55 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 8192, + "output": 8000 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.51, + "output": 0.74 } }, - "meganova-ai/manta-mini-1.0": { - "id": "meganova-ai/manta-mini-1.0", - "name": "Manta Mini 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova", - "attachment": false, + "meta-llama/llama-4-scout-17b-16e-instruct": { + "id": "meta-llama/llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-09-20", - "last_updated": "2025-12-20", + "temperature": true, + "release_date": "2025-04-06", + "last_updated": "2025-04-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "input": 8192, - "output": 8192 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.02, - "output": 0.16, - "cache_read": 0.01 + "input": 0.18, + "output": 0.59 } }, - "meganova-ai/manta-flash-1.0": { - "id": "meganova-ai/manta-flash-1.0", - "name": "Manta Flash 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova", + "meta-llama/llama-3.3-70b-instruct": { + "id": "meta-llama/llama-3.3-70b-instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-20", - "last_updated": "2025-12-20", + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-07", + "last_updated": "2024-12-07", "modalities": { "input": [ "text" @@ -22816,29 +20663,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 16384 + "context": 131072, + "output": 120000 }, "cost": { - "input": 0.02, - "output": 0.16, - "cache_read": 0.01 + "input": 0.135, + "output": 0.4 } }, - "meganova-ai/manta-pro-1.0": { - "id": "meganova-ai/manta-pro-1.0", - "name": "Manta Pro 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova", + "meta-llama/llama-3-8b-instruct": { + "id": "meta-llama/llama-3-8b-instruct", + "name": "Llama 3 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-09-20", - "last_updated": "2025-12-20", + "temperature": true, + "release_date": "2024-04-25", + "last_updated": "2024-04-25", "modalities": { "input": [ "text" @@ -22847,69 +20692,59 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "input": 65536, - "output": 32768 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.06, - "output": 0.5, - "cache_read": 0.03 + "input": 0.04, + "output": 0.04 } }, - "perceptron/perceptron-mk1": { - "id": "perceptron/perceptron-mk1", - "name": "Perceptron Mk1", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "meta-llama/llama-3.2-3b-instruct": { + "id": "meta-llama/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "temperature": true, + "release_date": "2024-09-18", + "last_updated": "2024-09-18", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 32768, - "input": 32768, - "output": 8192 + "output": 32000 }, "cost": { - "input": 0.15, - "output": 1.5, - "cache_read": 0.075 + "input": 0.03, + "output": 0.05 } }, - "z-ai/glm-4.6v": { - "id": "z-ai/glm-4.6v", - "name": "GLM 4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "meta-llama/llama-4-maverick-17b-128e-instruct-fp8": { + "id": "meta-llama/llama-4-maverick-17b-128e-instruct-fp8", + "name": "Llama 4 Maverick Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2025-04-06", + "last_updated": "2025-04-06", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -22917,28 +20752,24 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 24000 + "context": 1048576, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.15 + "input": 0.27, + "output": 0.85 } }, - "z-ai/glm-4.6-original": { - "id": "z-ai/glm-4.6-original", - "name": "GLM 4.6 Original", - "description": "GLM-4.6, Zhipu's flagship text model with 256K context window and advanced reasoning capabilities. Direct via Z-AI (Zhipu).", - "family": "glm", + "moonshotai/kimi-k2-instruct": { + "id": "moonshotai/kimi-k2-instruct", + "name": "Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2025-07-11", + "last_updated": "2025-07-11", "modalities": { "input": [ "text" @@ -22949,29 +20780,30 @@ }, "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 65535 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.35, - "output": 1.4, - "cache_read": 0.175 + "input": 0.57, + "output": 2.3 } }, - "z-ai/glm-4.5": { - "id": "z-ai/glm-4.5", - "name": "GLM 4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2025-11-07", + "last_updated": "2026-06-29", "modalities": { "input": [ "text" @@ -22982,28 +20814,36 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 1.3, + "input": 0.6, + "output": 2.5, "cache_read": 0.15 } }, - "z-ai/glm-5v-turbo": { - "id": "z-ai/glm-5v-turbo", - "name": "GLM 5V Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", @@ -23014,40 +20854,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202800, - "input": 202800, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "z-ai/glm-5.3-flash": { - "id": "z-ai/glm-5.3-flash", - "name": "GLM 5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -23060,40 +20898,31 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.075, - "output": 0.25, - "cache_read": 0.015 + "input": 0.8, + "output": 3.4, + "cache_read": 0.16 } }, - "z-ai/glm-5.3-flash-uncensored": { - "id": "z-ai/glm-5.3-flash-uncensored", - "name": "GLM 5.3 Flash Uncensored", - "description": "GLM 5.3 Flash Uncensored is an uncensored fine-tune of the efficient 320B mixture-of-experts reasoning model, built for unrestricted chat, creative writing, coding, agentic work, tool use, and long-context tasks.", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-07-29", - "last_updated": "2026-08-27", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -23101,39 +20930,26 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.2, - "output": 0.8, - "cache_read": 0.1 + "input": 0.6, + "output": 2.5 } }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "GLM 4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", + "minimaxai/minimax-m1-80k": { + "id": "minimaxai/minimax-m1-80k", + "name": "MiniMax M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text" @@ -23144,33 +20960,29 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 65535 + "context": 1000000, + "output": 40000 }, "cost": { - "input": 0.35, - "output": 1.4, - "cache_read": 0.175 + "input": 0.55, + "output": 2.2 } }, - "z-ai/glm-4.7-flash:thinking": { - "id": "z-ai/glm-4.7-flash:thinking", - "name": "GLM 4.7 Flash Thinking", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, + "baidu/ernie-4.5-vl-28b-a3b": { + "id": "baidu/ernie-4.5-vl-28b-a3b", + "name": "ERNIE 4.5 VL 28B A3B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-06-30", + "last_updated": "2026-06-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -23178,28 +20990,26 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 128000 + "context": 30000, + "output": 8000 }, "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.035 + "input": 0.14, + "output": 0.56 } }, - "z-ai/glm-5-original:thinking": { - "id": "z-ai/glm-5-original:thinking", - "name": "GLM 5 Original Thinking", - "description": "GLM-5 original with extended thinking capabilities for complex reasoning.", - "family": "glm", + "baidu/ernie-4.5-21B-a3b": { + "id": "baidu/ernie-4.5-21B-a3b", + "name": "ERNIE 4.5 21B A3B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "ernie", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ "text" @@ -23210,40 +21020,29 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 128000 + "context": 120000, + "output": 8000 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.07, + "output": 0.28 } }, - "z-ai/glm-5.3:thinking": { - "id": "z-ai/glm-5.3:thinking", - "name": "GLM 5.3 Thinking", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "baidu/ernie-4.5-vl-424b-a47b": { + "id": "baidu/ernie-4.5-vl-424b-a47b", + "name": "ERNIE 4.5 VL 424B A47B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -23251,29 +21050,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 123000, + "output": 16000 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.42, + "output": 1.25 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM 5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "baidu/ernie-4.5-21B-a3b-thinking": { + "id": "baidu/ernie-4.5-21B-a3b-thinking", + "name": "ERNIE-4.5-21B-A3B-Thinking", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "ernie", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-03", + "release_date": "2025-09-19", + "last_updated": "2025-09-19", "modalities": { "input": [ "text" @@ -23284,29 +21081,25 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 128000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 2.55, - "cache_read": 0.13 + "input": 0.07, + "output": 0.28 } }, - "z-ai/GLM-4.5-Air": { - "id": "z-ai/GLM-4.5-Air", - "name": "GLM 4.5 Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "baidu/ernie-4.5-300b-a47b-paddle": { + "id": "baidu/ernie-4.5-300b-a47b-paddle", + "name": "ERNIE 4.5 300B A47B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ "text" @@ -23317,21 +21110,51 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 98304 + "context": 123000, + "output": 12000 }, "cost": { - "input": 0.12, - "output": 0.8, - "cache_read": 0.06 + "input": 0.28, + "output": 1.1 } }, - "z-ai/glm-4.5v": { - "id": "z-ai/glm-4.5v", - "name": "GLM 4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "baidu/ernie-4.5-vl-28b-a3b-thinking": { + "id": "baidu/ernie-4.5-vl-28b-a3b-thinking", + "name": "ERNIE-4.5-VL-28B-A3B-Thinking", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-11-26", + "last_updated": "2025-11-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 65536 + }, + "cost": { + "input": 0.39, + "output": 0.39 + } + }, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -23339,12 +21162,11 @@ "type": "toggle" } ], - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -23356,31 +21178,95 @@ }, "open_weights": true, "limit": { - "context": 65536, - "input": 65536, - "output": 16384 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.3 + "input": 0.14, + "output": 0.4 } }, - "z-ai/GLM-4.6-turbo:thinking": { - "id": "z-ai/GLM-4.6-turbo:thinking", - "name": "GLM 4.6 Turbo (Thinking)", - "description": "GLM 4.6 Turbo with thinking mode enabled for enhanced reasoning; shows internal reasoning and supports long context.", - "family": "glm", - "attachment": false, + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.13, + "output": 0.4 + } + }, + "google/gemma-3-12b-it": { + "id": "google/gemma-3-12b-it", + "name": "Gemma 3 12B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-10-02", - "last_updated": "2025-10-02", + "temperature": true, + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.05, + "output": 0.1 + } + }, + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma 3 27B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-03-25", + "last_updated": "2025-03-25", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" @@ -23388,30 +21274,24 @@ }, "open_weights": true, "limit": { - "context": 204800, - "input": 204800, - "output": 131072 + "context": 98304, + "output": 16384 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.5 + "input": 0.119, + "output": 0.2 } }, - "z-ai/glm-4.7:thinking": { - "id": "z-ai/glm-4.7:thinking", - "name": "GLM 4.7 Thinking", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", + "microsoft/wizardlm-2-8x22b": { + "id": "microsoft/wizardlm-2-8x22b", + "name": "Wizardlm 2 8x22B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2024-04-24", + "last_updated": "2024-04-24", "modalities": { "input": [ "text" @@ -23422,28 +21302,26 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 65535 + "context": 65535, + "output": 8000 }, "cost": { - "input": 0.2, - "output": 0.8, - "cache_read": 0.1 + "input": 0.62, + "output": 0.62 } }, - "z-ai/glm-5.1:thinking": { - "id": "z-ai/glm-5.1:thinking", - "name": "GLM 5.1 Thinking", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "OpenAI GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } @@ -23451,11 +21329,12 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -23463,27 +21342,35 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.75, - "output": 2.6, - "cache_read": 0.15 + "input": 0.05, + "output": 0.25 } }, - "z-ai/glm-4.6v-original": { - "id": "z-ai/glm-4.6v-original", - "name": "GLM 4.6V Original", - "description": "GLM-4.6V scales its context window to 128k tokens in training, and achieves SoTA performance in visual understanding among models of similar parameter scales. Integrates native Function Calling capabilities, bridging 'visual perception' and 'executable action' for multimodal agents. Direct via Z-AI (Zhipu).", - "family": "glm", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "OpenAI: GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": false, - "structured_output": false, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "structured_output": true, + "temperature": true, + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ "text", @@ -23495,62 +21382,53 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 24000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 0.9, - "cache_read": 0.3 + "input": 0.04, + "output": 0.15 } }, - "z-ai/glm-5v-turbo:thinking": { - "id": "z-ai/glm-5v-turbo:thinking", - "name": "GLM 5V Turbo Thinking", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "sao10K/l31-70b-euryale-v2.2": { + "id": "sao10K/l31-70b-euryale-v2.2", + "name": "L31 70B Euryale V2.2", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202800, - "input": 202800, - "output": 131072 + "context": 8192, + "output": 8192 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 1.48, + "output": 1.48 } }, - "z-ai/GLM-4.6-turbo": { - "id": "z-ai/GLM-4.6-turbo", - "name": "GLM 4.6 Turbo", - "description": "Fast variant of GLM 4.6 for general chat, coding, and analysis with improved latency and strong reasoning.", - "family": "glm", + "sao10K/L3-8B-stheno-v3.2": { + "id": "sao10K/L3-8B-stheno-v3.2", + "name": "L3 8B Stheno V3.2", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-10-02", - "last_updated": "2025-10-02", + "tool_call": true, + "temperature": true, + "release_date": "2024-11-29", + "last_updated": "2024-11-29", "modalities": { "input": [ "text" @@ -23561,30 +21439,25 @@ }, "open_weights": true, "limit": { - "context": 204800, - "input": 204800, - "output": 131072 + "context": 8192, + "output": 32000 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.5 + "input": 0.05, + "output": 0.05 } }, - "z-ai/GLM-4.5-Air:thinking": { - "id": "z-ai/GLM-4.5-Air:thinking", - "name": "GLM 4.5 Air (Thinking)", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "sao10K/l3-8b-lunaris": { + "id": "sao10K/l3-8b-lunaris", + "name": "Sao10k L3 8B Lunaris\t", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2024-11-28", + "last_updated": "2024-11-28", "modalities": { "input": [ "text" @@ -23595,37 +21468,24 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 98304 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.12, - "output": 0.8, - "cache_read": 0.06 + "input": 0.05, + "output": 0.05 } }, - "z-ai/glm-5.2:thinking": { - "id": "z-ai/glm-5.2:thinking", - "name": "GLM 5.2 Thinking", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "sao10K/l3-70b-euryale-v2.1": { + "id": "sao10K/l3-70b-euryale-v2.1", + "name": "L3 70B Euryale V2.1\t", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2024-06-18", + "last_updated": "2024-06-18", "modalities": { "input": [ "text" @@ -23636,30 +21496,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.42, - "output": 1.32, - "cache_read": 0.078 + "input": 1.48, + "output": 1.48 } }, - "z-ai/glm-4.7-flash": { - "id": "z-ai/glm-4.7-flash", - "name": "GLM 4.7 Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", + "baichuan/baichuan-m2-32b": { + "id": "baichuan/baichuan-m2-32b", + "name": "baichuan-m2-32b", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "baichuan", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2024-12", + "release_date": "2025-08-13", + "last_updated": "2025-08-13", "modalities": { "input": [ "text" @@ -23670,30 +21527,26 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { "input": 0.07, - "output": 0.4, - "cache_read": 0.035 + "output": 0.07 } }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "GLM 4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", + "mistralai/mistral-nemo": { + "id": "mistralai/mistral-nemo", + "name": "Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2024-07-30", + "last_updated": "2024-07-30", "modalities": { "input": [ "text" @@ -23704,29 +21557,28 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 65535 + "context": 60288, + "output": 16000 }, "cost": { - "input": 0.2, - "output": 0.8, - "cache_read": 0.1 + "input": 0.04, + "output": 0.17 } }, - "z-ai/glm-5:thinking": { - "id": "z-ai/glm-5:thinking", - "name": "GLM 5 Thinking", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "xiaomimimo/mimo-v2-flash": { + "id": "xiaomimimo/mimo-v2-flash", + "name": "XiaomiMiMo/MiMo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2024-12", + "release_date": "2025-12-19", + "last_updated": "2025-12-19", "modalities": { "input": [ "text" @@ -23737,38 +21589,32 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 128000 + "context": 262144, + "output": 32000 }, "cost": { - "input": 0.5, - "output": 2.55, - "cache_read": 0.13 + "input": 0.1, + "output": 0.3, + "cache_read": 0.3 } }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "xiaomimimo/mimo-v2-pro": { + "id": "xiaomimimo/mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-05-27", "modalities": { "input": [ "text" @@ -23777,32 +21623,50 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "input": 1048576, "output": 131072 }, "cost": { - "input": 0.42, - "output": 1.32, - "cache_read": 0.078 + "input": 2, + "output": 6, + "cache_read": 0.4, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "z-ai/GLM-4.5:thinking": { - "id": "z-ai/GLM-4.5:thinking", - "name": "GLM 4.5 (Thinking)", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", + "xiaomimimo/mimo-v2.5-pro": { + "id": "xiaomimimo/mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-27", "modalities": { "input": [ "text" @@ -23813,37 +21677,41 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.3, - "cache_read": 0.15 + "input": 0.522, + "output": 1.044, + "cache_read": 0.0043, + "tiers": [ + { + "input": 0.522, + "output": 1.044, + "cache_read": 0.0043, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.522, + "output": 1.044, + "cache_read": 0.0043 + } } }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "gryphe/mythomax-l2-13b": { + "id": "gryphe/mythomax-l2-13b", + "name": "Mythomax L2 13B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2024-04-25", + "last_updated": "2024-04-25", "modalities": { "input": [ "text" @@ -23854,39 +21722,34 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 131072 + "context": 4096, + "output": 3200 }, "cost": { - "input": 0.75, - "output": 2.6, - "cache_read": 0.15 + "input": 0.09, + "output": 0.09 } }, - "z-ai/glm-4.6:thinking": { - "id": "z-ai/glm-4.6:thinking", - "name": "GLM 4.6 Thinking", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "zai-org/glm-4.7": { + "id": "zai-org/glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -23897,26 +21760,29 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 65535 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.35, - "output": 1.4, - "cache_read": 0.175 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "z-ai/glm-4.5v:thinking": { - "id": "z-ai/glm-4.5v:thinking", - "name": "GLM 4.5V Thinking", + "zai-org/glm-4.5v": { + "id": "zai-org/glm-4.5v", + "name": "GLM 4.5V", "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "family": "glmv", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, "knowledge": "2025-04", "release_date": "2025-08-11", @@ -23924,6 +21790,7 @@ "modalities": { "input": [ "text", + "video", "image" ], "output": [ @@ -23933,27 +21800,33 @@ "open_weights": true, "limit": { "context": 65536, - "input": 65536, "output": 16384 }, "cost": { "input": 0.6, "output": 1.8, - "cache_read": 0.3 + "cache_read": 0.11 } }, - "z-ai/glm-4.7-original": { - "id": "z-ai/glm-4.7-original", - "name": "GLM 4.7 Original", - "description": "GLM-4.7 is a next-gen GLM series text model with stronger reasoning, long-context chat, and reliable tool use. Routed directly via Z-AI (Zhipu).", + "zai-org/glm-4.5": { + "id": "zai-org/glm-4.5", + "name": "GLM-4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -23964,9 +21837,8 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 65535 + "context": 131072, + "output": 98304 }, "cost": { "input": 0.6, @@ -23974,21 +21846,20 @@ "cache_read": 0.11 } }, - "z-ai/glm-5-original": { - "id": "z-ai/glm-5-original", - "name": "GLM 5 Original", - "description": "GLM-5 is Zhipu's latest flagship model with advanced reasoning and instruction following. Routed directly via Z-AI (Zhipu).", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "zai-org/autoglm-phone-9b-multilingual": { + "id": "zai-org/autoglm-phone-9b-multilingual", + "name": "AutoGLM-Phone-9B-Multilingual", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-12-10", + "last_updated": "2025-12-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -23996,28 +21867,34 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 128000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.035, + "output": 0.138 } }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "GLM 5 Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", + "zai-org/glm-5.1": { + "id": "zai-org/glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ "text" @@ -24026,30 +21903,37 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202800, - "input": 202800, + "context": 204800, "output": 131072 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 1.38, + "output": 4.4, + "cache_read": 0.26 } }, - "z-ai/glm-4.7-original:thinking": { - "id": "z-ai/glm-4.7-original:thinking", - "name": "GLM 4.7 Original Thinking", - "description": "GLM-4.7 original with extended thinking capabilities for complex reasoning.", + "zai-org/glm-4.6": { + "id": "zai-org/glm-4.6", + "name": "GLM 4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "temperature": true, + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -24060,20 +21944,19 @@ }, "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 65535 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.6, + "input": 0.55, "output": 2.2, "cache_read": 0.11 } }, - "z-ai/glm-5.3": { - "id": "z-ai/glm-5.3", - "name": "GLM 5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "zai-org/glm-5.2": { + "id": "zai-org/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, @@ -24081,17 +21964,20 @@ { "type": "effort", "values": [ - "low", + "none", "high", "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -24103,39 +21989,36 @@ "open_weights": true, "limit": { "context": 1048576, - "input": 1048576, "output": 131072 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "z-ai/glm-5.3-flash-cybersecurity": { - "id": "z-ai/glm-5.3-flash-cybersecurity", - "name": "GLM 5.3 Flash Cybersecurity", - "description": "GLM 5.3 Flash Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports always-on reasoning, image understanding, tool calling, and a 1,048,576-token context window.", - "family": "glm", + "zai-org/glm-4.6v": { + "id": "zai-org/glm-4.6v", + "name": "GLM 4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glmv", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, - "release_date": "2026-09-19", - "last_updated": "2026-09-19", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", + "video", "image" ], "output": [ @@ -24144,28 +22027,32 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, + "context": 131072, "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.075 + "input": 0.3, + "output": 0.9, + "cache_read": 0.055 } }, - "z-ai/glm-latest": { - "id": "z-ai/glm-latest", - "name": "GLM Latest", - "description": "Compatibility alias that routes to the newest thinking GLM model. Currently routes to GLM 5.2 Thinking.", - "family": "glm", + "zai-org/glm-4.5-air": { + "id": "zai-org/glm-4.5-air", + "name": "GLM 4.5 Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-10-13", + "last_updated": "2025-10-13", "modalities": { "input": [ "text" @@ -24176,26 +22063,33 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 131072, + "output": 98304 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.13, + "output": 0.85, + "cache_read": 0.025 } }, - "inference-net/schematron-v2-small": { - "id": "inference-net/schematron-v2-small", - "name": "Schematron V2 Small", - "description": "Inference.net's 3B-parameter HTML-to-JSON extraction model, focused on accuracy for complex schemas and long web pages. It turns HTML into typed, structured data for web scraping and product catalog ingestion, with a 128K-token context window. Supply HTML in the user message and extraction instructions in a JSON schema via response_format; it does not follow ordinary chat or system prompts.", + "zai-org/glm-4.7-flash": { + "id": "zai-org/glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "structured_output": true, - "release_date": "2026-09-12", - "last_updated": "2026-09-12", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -24204,28 +22098,37 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.23, - "cache_read": 0.025 + "input": 0.07, + "output": 0.4, + "cache_read": 0.01 } }, - "inference-net/schematron-v2-turbo": { - "id": "inference-net/schematron-v2-turbo", - "name": "Schematron V2 Turbo", - "description": "Inference.net's 3B-parameter HTML-to-JSON extraction model, optimized for throughput and low cost on high-volume workloads. It turns HTML into typed, structured data for web scraping and product catalog ingestion, with a 128K-token context window. Supply HTML in the user message and extraction instructions in a JSON schema via response_format; it does not follow ordinary chat or system prompts.", + "zai-org/glm-5": { + "id": "zai-org/glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-09-12", - "last_updated": "2026-09-12", + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -24234,47 +22137,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 202800, + "output": 131072 }, "cost": { - "input": 0.03, - "output": 0.15, - "cache_read": 0.015 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "thinkingmachines/inkling:thinking": { - "id": "thinkingmachines/inkling:thinking", - "name": "Inkling Thinking", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", + "paddlepaddle/paddleocr-vl": { + "id": "paddlepaddle/paddleocr-vl", + "name": "PaddleOCR-VL", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "release_date": "2025-10-22", + "last_updated": "2025-10-22", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" @@ -24282,33 +22169,30 @@ }, "open_weights": true, "limit": { - "context": 1048000, - "input": 1048000, - "output": 32768 + "context": 16384, + "output": 16384 }, "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.17 + "input": 0.02, + "output": 0.02 } }, - "thinkingmachines/Inkling-Small": { - "id": "thinkingmachines/Inkling-Small", - "name": "Inkling Small", - "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", - "family": "ling", + "qwen/qwen3-vl-235b-a22b-thinking": { + "id": "qwen/qwen3-vl-235b-a22b-thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ "text", "image", - "audio" + "video" ], "output": [ "text" @@ -24316,33 +22200,30 @@ }, "open_weights": true, "limit": { - "context": 524288, - "input": 524288, + "context": 131072, "output": 32768 }, "cost": { - "input": 0.5, - "output": 1.2, - "cache_read": 0.1 + "input": 0.98, + "output": 3.95 } }, - "thinkingmachines/inkling": { - "id": "thinkingmachines/inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", + "qwen/qwen3-vl-30b-a3b-thinking": { + "id": "qwen/qwen3-vl-30b-a3b-thinking", + "name": "qwen/qwen3-vl-30b-a3b-thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "release_date": "2025-10-11", + "last_updated": "2025-10-11", "modalities": { "input": [ "text", "image", - "audio" + "video" ], "output": [ "text" @@ -24350,45 +22231,30 @@ }, "open_weights": true, "limit": { - "context": 1048000, - "input": 1048000, + "context": 131072, "output": 32768 }, "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.17 + "input": 0.2, + "output": 1 } }, - "thinkingmachines/Inkling-Small:thinking": { - "id": "thinkingmachines/Inkling-Small:thinking", - "name": "Inkling Small Thinking", - "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], + "qwen/qwen3-235b-a22b-instruct-2507": { + "id": "qwen/qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "knowledge": "2025-04", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -24396,93 +22262,57 @@ }, "open_weights": true, "limit": { - "context": 524288, - "input": 524288, - "output": 32768 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.5, - "output": 1.2, - "cache_read": 0.1 + "input": 0.09, + "output": 0.58 } }, - "meta/muse-spark-1.3": { - "id": "meta/muse-spark-1.3", - "name": "Muse Spark 1.3", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen/qwen3-coder-30b-a3b-instruct": { + "id": "qwen/qwen3-coder-30b-a3b-instruct", + "name": "Qwen3 Coder 30b A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "release_date": "2025-10-09", + "last_updated": "2025-10-09", "modalities": { "input": [ - "text", - "image", - "video", - "pdf", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 943718 + "context": 160000, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 0.07, + "output": 0.27 } }, - "meta/muse-glimmer-30b": { - "id": "meta/muse-glimmer-30b", - "name": "Muse Glimmer 30B", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, + "qwen/qwen3-8b-fp8": { + "id": "qwen/qwen3-8b-fp8", + "name": "Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -24490,188 +22320,129 @@ }, "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 117964 + "context": 128000, + "output": 20000 }, "cost": { - "input": 0.35, - "output": 1.5, - "cache_read": 0.04 + "input": 0.035, + "output": 0.138 } }, - "meta/muse-spark-1.1": { - "id": "meta/muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": false, - "structured_output": false, + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", + "release_date": "2025-09-10", + "last_updated": "2025-09-10", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 0.15, + "output": 1.5 } }, - "meta/muse-spark-1.2": { - "id": "meta/muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", + "qwen/qwen3-vl-8b-instruct": { + "id": "qwen/qwen3-vl-8b-instruct", + "name": "qwen/qwen3-vl-8b-instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "release_date": "2025-10-17", + "last_updated": "2025-10-17", "modalities": { "input": [ "text", "image", - "video", - "pdf", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 0.08, + "output": 0.5 } }, - "meta/muse-spark-1.2-contributor": { - "id": "meta/muse-spark-1.2-contributor", - "name": "Muse Spark 1.2 Contributor (Data Used for Training)", - "description": "A much cheaper opt-in version of Muse Spark 1.2 with the same multimodal coding and agentic capabilities. Prompts and outputs sent to this Contributor model may be used by Meta for training and to improve its products; use the standard Muse Spark 1.2 model if you do not want your data used for training.", - "family": "muse", + "qwen/qwen3-omni-30b-a3b-thinking": { + "id": "qwen/qwen3-omni-30b-a3b-thinking", + "name": "Qwen3 Omni 30B A3B Thinking", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "temperature": true, + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ "text", - "image", - "video", "audio", - "pdf" + "video", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 65536 + "context": 65536, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 + "input": 0.25, + "output": 0.97, + "input_audio": 2.2, + "output_audio": 1.788 } }, - "meta/muse-spark-1.3-contributor": { - "id": "meta/muse-spark-1.3-contributor", - "name": "Muse Spark 1.3 Contributor", - "description": "Meta's Muse Spark 1.3 Contributor is a frontier multimodal reasoning model for long-horizon coding and agentic workflows, with strong gains in computer use, browsing, professional tool use, codebase understanding, and million-token retrieval. It accepts text, images, audio, video, and files, supports tool calling and structured output, and always reasons before answering. Prompts and outputs may be used by Meta for training and to improve its products.", - "family": "muse", - "attachment": true, + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7-Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-27", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -24679,27 +22450,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 943718 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 + "input": 1.25, + "output": 3.75, + "cache_read": 0.25, + "cache_write": 1.5625 } }, - "mlabonne/NeuralDaredevil-8B-abliterated": { - "id": "mlabonne/NeuralDaredevil-8B-abliterated", - "name": "Neural Daredevil 8B abliterated", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-12-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ "text" @@ -24708,29 +22481,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "input": 8192, - "output": 8192 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.44, - "output": 0.44, - "cache_read": 0.22 + "input": 2.11, + "output": 8.45 } }, - "NeverSleep/Lumimaid-v0.2-70B": { - "id": "NeverSleep/Lumimaid-v0.2-70B", - "name": "Lumimaid v0.2", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "qwen/qwen2.5-7b-instruct": { + "id": "qwen/qwen2.5-7b-instruct", + "name": "Qwen2.5 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-07-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text" @@ -24741,27 +22512,26 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 32000, + "output": 32000 }, "cost": { - "input": 1, - "output": 1.5, - "cache_read": 0.5 + "input": 0.07, + "output": 0.07 } }, - "nanogpt/coding-router:low": { - "id": "nanogpt/coding-router:low", - "name": "Coding Router Low", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "temperature": true, + "release_date": "2025-09-10", + "last_updated": "2025-09-10", "modalities": { "input": [ "text" @@ -24770,29 +22540,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0.15, + "output": 1.5 } }, - "nanogpt/coding-router:medium": { - "id": "nanogpt/coding-router:medium", - "name": "Coding Router Medium", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "qwen/qwen3-235b-a22b-thinking-2507": { + "id": "qwen/qwen3-235b-a22b-thinking-2507", + "name": "Qwen3 235B A22b Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ "text" @@ -24801,29 +22571,26 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0.3, + "output": 3 } }, - "nanogpt/coding-router": { - "id": "nanogpt/coding-router", - "name": "Coding Router", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "qwen/qwen-mt-plus": { + "id": "qwen/qwen-mt-plus", + "name": "Qwen MT Plus", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-09-03", + "last_updated": "2025-09-03", "modalities": { "input": [ "text" @@ -24832,29 +22599,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 16384, + "output": 8192 }, "cost": { - "input": 1.1, - "output": 2.2, - "cache_read": 0.11 + "input": 0.25, + "output": 0.75 } }, - "nanogpt/coding-router:max": { - "id": "nanogpt/coding-router:max", - "name": "Coding Router Max", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "qwen/qwen3-32b-fp8": { + "id": "qwen/qwen3-32b-fp8", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "tool_call": false, + "temperature": true, + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -24863,60 +22628,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 40960, + "output": 20000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.1, + "output": 0.45 } }, - "nanogpt/coding-router:high": { - "id": "nanogpt/coding-router:high", - "name": "Coding Router High", - "description": "Automatic model router for matching prompts to suitable backends and budgets", + "qwen/qwen3-4b-fp8": { + "id": "qwen/qwen3-4b-fp8", + "name": "Qwen3 4B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 - }, - "cost": { - "input": 1.1, - "output": 2.2, - "cache_read": 0.11 - } - }, - "LatitudeGames/Wayfarer-Large-70B-Llama-3.3": { - "id": "LatitudeGames/Wayfarer-Large-70B-Llama-3.3", - "name": "Llama 3.3 70B Wayfarer", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-02-20", - "last_updated": "2025-02-20", + "temperature": true, + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -24927,30 +22659,30 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 128000, + "output": 20000 }, "cost": { - "input": 0.7, - "output": 0.7, - "cache_read": 0.35 + "input": 0.03, + "output": 0.03 } }, - "VongolaChouko/Starcannon-Unleashed-12B-v1.0": { - "id": "VongolaChouko/Starcannon-Unleashed-12B-v1.0", - "name": "Mistral Nemo Starcannon 12b v1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", - "attachment": false, + "qwen/qwen2.5-vl-72b-instruct": { + "id": "qwen/qwen2.5-vl-72b-instruct", + "name": "Qwen2.5 VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-07-01", + "temperature": true, + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -24958,29 +22690,25 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.8, + "output": 0.8 } }, - "arcee-ai/trinity-large-thinking": { - "id": "arcee-ai/trinity-large-thinking", - "name": "Trinity Large Thinking", - "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", - "family": "trinity", + "qwen/qwen3-30b-a3b-fp8": { + "id": "qwen/qwen3-30b-a3b-fp8", + "name": "Qwen3 30B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-05-28", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -24991,377 +22719,301 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 80000 + "context": 40960, + "output": 20000 }, "cost": { - "input": 0.25, - "output": 0.9, - "cache_read": 0.125 + "input": 0.09, + "output": 0.45 } }, - "x-ai/grok-4.20-multi-agent": { - "id": "x-ai/grok-4.20-multi-agent", - "name": "Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5-27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "temperature": true, + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "input": 2000000, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 1 + "input": 0.3, + "output": 2.4 } }, - "x-ai/grok-latest": { - "id": "x-ai/grok-latest", - "name": "Grok Latest", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen/qwen-2.5-72b-instruct": { + "id": "qwen/qwen-2.5-72b-instruct", + "name": "Qwen 2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-10-15", + "last_updated": "2024-10-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "input": 500000, - "output": 450000 + "context": 32000, + "output": 8192 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 0.38, + "output": 0.4 } }, - "x-ai/grok-4.7": { - "id": "x-ai/grok-4.7", - "name": "Grok 4.7", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "input": 500000, - "output": 500000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.6, - "output": 4.8, - "cache_read": 0.4 + "input": 0.2, + "output": 1.5 } }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5-35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 900000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.25, + "output": 2 } }, - "x-ai/grok-4.5": { - "id": "x-ai/grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "qwen/qwen3-coder-480b-a35b-instruct": { + "id": "qwen/qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "input": 500000, - "output": 450000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 0.38, + "output": 1.55 } }, - "x-ai/grok-build-0.1": { - "id": "x-ai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5-397B-A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 230400 + "context": 262144, + "output": 64000 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 + "input": 0.6, + "output": 3.6 } }, - "x-ai/grok-4.20": { - "id": "x-ai/grok-4.20", - "name": "Grok 4.20", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "qwen/qwen3-vl-30b-a3b-instruct": { + "id": "qwen/qwen3-vl-30b-a3b-instruct", + "name": "qwen/qwen3-vl-30b-a3b-instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "temperature": true, + "release_date": "2025-10-11", + "last_updated": "2025-10-11", "modalities": { "input": [ "text", + "video", "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "input": 2000000, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 1 + "input": 0.2, + "output": 0.7 } }, - "x-ai/grok-4.6": { - "id": "x-ai/grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", + "qwen/qwen3-omni-30b-a3b-instruct": { + "id": "qwen/qwen3-omni-30b-a3b-instruct", + "name": "Qwen3 Omni 30B A3B Instruct", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2024-04", + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ "text", + "video", + "audio", "image" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "input": 500000, - "output": 450000 + "context": 65536, + "output": 16384 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 0.25, + "output": 0.97, + "input_audio": 2.2, + "output_audio": 1.788 } }, - "Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B": { - "id": "Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B", - "name": "Nemotron Tenyxchat Storybreaker 70b", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, + "qwen/qwen3-vl-235b-a22b-instruct": { + "id": "qwen/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-12-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -25369,30 +23021,36 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.3, + "output": 1.5 } }, - "Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B": { - "id": "Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B", - "name": "Llama 3.05 Storybreaker Ministral 70b", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-12-01", + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5-122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -25400,27 +23058,25 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.4, + "output": 3.2 } }, - "upstage/solar-pro-3": { - "id": "upstage/solar-pro-3", - "name": "Solar Pro 3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "solar-pro", + "qwen/qwen3-235b-a22b-fp8": { + "id": "qwen/qwen3-235b-a22b-fp8", + "name": "Qwen3 235B A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "structured_output": false, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "temperature": true, + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -25429,31 +23085,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 117964 + "context": 40960, + "output": 20000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 0.2, + "output": 0.8 } }, - "upstage/solar-pro4": { - "id": "upstage/solar-pro4", - "name": "Solar Pro 4", - "description": "Upstage's flagship model, specialized for agentic use", - "family": "solar-pro", + "deepseek/deepseek-r1-0528": { + "id": "deepseek/deepseek-r1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-08-06", - "last_updated": "2026-08-06", + "knowledge": "2024-07", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -25462,40 +23117,48 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 524288, - "input": 524288, - "output": 131072 + "context": 163840, + "output": 32768 }, "cost": { - "input": 0.03, - "output": 0.12, - "cache_read": 0.006 + "input": 0.7, + "output": 2.5, + "cache_read": 0.35 } }, - "upstage/solar-mini4:thinking": { - "id": "upstage/solar-mini4:thinking", - "name": "Solar Mini 4 Thinking", - "description": "Solar Mini 4 with reasoning enabled for agentic tasks and harder analysis across a 524K-token context window.", - "family": "solar", + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -25504,42 +23167,34 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 524288, - "input": 524288, - "output": 131072 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.005 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "upstage/solar-pro4:thinking": { - "id": "upstage/solar-pro4:thinking", - "name": "Solar Pro 4 Thinking", - "description": "Upstage's flagship model, specialized for agentic use", - "family": "solar-pro", + "deepseek/deepseek-v3.1-terminus": { + "id": "deepseek/deepseek-v3.1-terminus", + "name": "Deepseek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-08-06", - "last_updated": "2026-08-06", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ "text" @@ -25548,29 +23203,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 524288, - "input": 524288, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.03, - "output": 0.12, - "cache_read": 0.006 + "input": 0.27, + "output": 1, + "cache_read": 0.135 } }, - "upstage/solar-mini4": { - "id": "upstage/solar-mini4", - "name": "Solar Mini 4", - "description": "Upstage's compact 35B-parameter mixture-of-experts model with 3B active parameters and a 524K context window. Built for fast, cost-efficient agentic tasks, with strong Korean and English support.", - "family": "solar", + "deepseek/deepseek-v3-0324": { + "id": "deepseek/deepseek-v3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ "text" @@ -25579,29 +23235,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 524288, - "input": 524288, - "output": 131072 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.005 + "input": 0.27, + "output": 1.12, + "cache_read": 0.135 } }, - "nano/lumen-stealth": { - "id": "nano/lumen-stealth", - "name": "Lumen Stealth", - "description": "Experimental multimodal model focused on reasoning, creative writing, roleplay, and agentic workflows. Available temporarily for evaluation ahead of public release. During this evaluation, prompts and responses are logged and may be reviewed to evaluate and improve the model. Do not send sensitive or confidential information.", + "deepseek/deepseek-ocr": { + "id": "deepseek/deepseek-ocr", + "name": "DeepSeek-OCR", + "description": "OCR model for extracting structured text from documents and screenshots", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "structured_output": true, + "temperature": true, + "release_date": "2025-10-24", + "last_updated": "2025-10-24", "modalities": { "input": [ "text", @@ -25611,28 +23266,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.05, - "output": 0 + "input": 0.03, + "output": 0.03 } }, - "bytedance/doubao-seed-2.1-turbo": { - "id": "bytedance/doubao-seed-2.1-turbo", - "name": "Doubao Seed 2.1 Turbo", - "description": "Fast, lower-cost model in the Doubao Seed 2.1 family for everyday chat, coding assistance, document work, and high-throughput productivity tasks. Supports a 256k context window and up to 128k output tokens. Note: privacy and logging guarantees may be limited.", - "family": "seed", + "deepseek/deepseek-r1-distill-llama-70b": { + "id": "deepseek/deepseek-r1-distill-llama-70b", + "name": "DeepSeek R1 Distill LLama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "structured_output": false, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", + "structured_output": true, + "temperature": true, + "release_date": "2025-01-27", + "last_updated": "2025-01-27", "modalities": { "input": [ "text" @@ -25641,29 +23297,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 128000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.25 + "input": 0.8, + "output": 0.8 } }, - "bytedance/doubao-seed-2.1-pro": { - "id": "bytedance/doubao-seed-2.1-pro", - "name": "Doubao Seed 2.1 Pro", - "description": "Higher-capability model in the Doubao Seed 2.1 family for agentic coding, long-context analysis, complex instruction following, and productivity workflows. Supports a 256k context window and up to 128k output tokens. Note: privacy and logging guarantees may be limited.", - "family": "seed", + "deepseek/deepseek-r1-turbo": { + "id": "deepseek/deepseek-r1-turbo", + "name": "DeepSeek R1 (Turbo)\t", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-03-05", + "last_updated": "2025-03-05", "modalities": { "input": [ "text" @@ -25672,72 +23326,58 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 128000 + "context": 64000, + "output": 16000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.5 + "input": 0.7, + "output": 2.5 } }, - "bytedance/doubao-seed-character": { - "id": "bytedance/doubao-seed-character", - "name": "Doubao Seed Character", - "description": "ByteDance's character-focused Doubao Seed model for roleplay, persona consistency, dialogue, and creative character interactions. It supports text and image input with a 128k context window. Requests route through ZenMux to ByteDance; ZenMux does not publish a model-API zero-retention or training guarantee, so avoid sensitive data.", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "deepseek/deepseek-prover-v2-671b": { + "id": "deepseek/deepseek-prover-v2-671b", + "name": "Deepseek Prover V2 671B", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2026-07-18", - "last_updated": "2026-07-18", + "temperature": true, + "release_date": "2025-04-30", + "last_updated": "2025-04-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 32768 + "context": 160000, + "output": 160000 }, "cost": { - "input": 0.1179, - "output": 0.2947, - "cache_read": 0.0236, - "cache_write": 0.0025 + "input": 0.7, + "output": 2.5 } }, - "liquid/lfm-2.5-2.6b": { - "id": "liquid/lfm-2.5-2.6b", - "name": "LFM2.5 2.6B", - "description": "Liquid AI's compact 2.6B reasoning model for agent workflows, data extraction, RAG, and long-context processing. It supports tool calling and structured output, but Liquid advises against using it for agentic coding. Warning: prompts and responses may be logged and used for model training or service improvement; do not send sensitive data.", - "family": "liquid", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", "low", "medium", "high", @@ -25746,9 +23386,14 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -25759,87 +23404,62 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 32768 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.05 + "input": 1.6, + "output": 3.2, + "cache_read": 0.135 } }, - "google/gemini-flash-latest": { - "id": "google/gemini-flash-latest", - "name": "Gemini Flash Latest", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, + "deepseek/deepseek-v3.2-exp": { + "id": "deepseek/deepseek-v3.2-exp", + "name": "Deepseek V3.2 Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, + "context": 163840, "output": 65536 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667 + "input": 0.27, + "output": 0.41 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "deepseek/deepseek-ocr-2": { + "id": "deepseek/deepseek-ocr-2", + "name": "deepseek/deepseek-ocr-2", + "description": "OCR model for extracting structured text from documents and screenshots", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "reasoning": false, + "tool_call": false, + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -25847,117 +23467,93 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 131072 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.45, - "cache_read": 0.05 + "input": 0.03, + "output": 0.03 } }, - "google/gemini-3.1-pro-preview-low": { - "id": "google/gemini-3.1-pro-preview-low", - "name": "Gemini 3.1 Pro (Preview Low)", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-02-21", - "last_updated": "2026-02-21", + "deepseek/deepseek-r1-distill-qwen-14b": { + "id": "deepseek/deepseek-r1-distill-qwen-14b", + "name": "DeepSeek R1 Distill Qwen 14B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 32768, + "output": 16384 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "input": 0.15, + "output": 0.15 } }, - "google/gemini-3.5-flash-thinking": { - "id": "google/gemini-3.5-flash-thinking", - "name": "Gemini 3.5 Flash Thinking", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "deepseek/deepseek-v3.1": { + "id": "deepseek/deepseek-v3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 0.083333 + "input": 0.27, + "output": 1, + "cache_read": 0.135 } }, - "google/gemma-4-26b-a4b-it:thinking": { - "id": "google/gemma-4-26b-a4b-it:thinking", - "name": "Gemma 4 26B A4B Thinking", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "deepseek/deepseek-r1-0528-qwen3-8b": { + "id": "deepseek/deepseek-r1-0528-qwen3-8b", + "name": "DeepSeek R1 0528 Qwen3 8B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-05-29", + "last_updated": "2025-05-29", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -25965,133 +23561,91 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 131072 + "context": 128000, + "output": 32000 }, "cost": { - "input": 0.13, - "output": 0.4, - "cache_read": 0.065 + "input": 0.06, + "output": 0.09 } }, - "google/gemini-flash-lite-latest": { - "id": "google/gemini-flash-lite-latest", - "name": "Gemini Flash Lite Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "deepseek/deepseek-r1-distill-qwen-32b": { + "id": "deepseek/deepseek-r1-distill-qwen-32b", + "name": "DeepSeek R1 Distill Qwen 32B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 64000, + "output": 32000 }, "cost": { "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.08333 + "output": 0.3 } }, - "google/gemini-3.6-flash": { - "id": "google/gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "deepseek/deepseek-v3-turbo": { + "id": "deepseek/deepseek-v3-turbo", + "name": "DeepSeek V3 (Turbo)\t", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "release_date": "2025-03-05", + "last_updated": "2025-03-05", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 64000, + "output": 16000 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667 + "input": 0.4, + "output": 1.3 } }, - "google/diffusiongemma": { - "id": "google/diffusiongemma", - "name": "DiffusionGemma", - "description": "DiffusionGemma is a high-speed diffusion-based version of Gemma 4 26B A4B. It supports optional reasoning and a 262,144-token context window.", + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "Deepseek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "xhigh" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-09-19", - "last_updated": "2026-09-19", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -26102,38 +23656,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 163840, + "output": 65536 }, "cost": { - "input": 0.05, - "output": 0.15, - "cache_read": 0.025 + "input": 0.269, + "output": 0.4, + "cache_read": 0.1345 } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-03-18", + "last_updated": "2026-05-27", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -26141,47 +23688,35 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, + "context": 204800, "output": 131072 }, "cost": { - "input": 0.12, - "output": 0.38, - "cache_read": 0.06 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "google/gemini-3.5-flash-lite": { - "id": "google/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -26189,35 +23724,33 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 204800, + "output": 131100 }, "cost": { "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.08333 + "output": 1.2, + "cache_read": 0.03 } }, - "google/gemma-4-31b-it:thinking": { - "id": "google/gemma-4-31b-it:thinking", - "name": "Gemma 4 31B Thinking", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "Minimax M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": false, "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -26225,110 +23758,100 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, + "context": 204800, "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.35, - "cache_read": 0.05 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro (Preview)", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "google/gemini-3.1-pro-preview-high": { - "id": "google/gemini-3.1-pro-preview-high", - "name": "Gemini 3.1 Pro (Preview High)", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini", - "attachment": true, + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax-m2.7", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-02-21", - "last_updated": "2026-02-21", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "google/gemma4-31b-splituntied": { - "id": "google/gemma4-31b-splituntied", - "name": "Gemma 4 31B Split-Untied", - "description": "Blazed-Forge's Split-Untied is a text-only Gemma 4 31B community finetune with an untied BF16 output head, built for creative writing, roleplay, expressive dialogue, and tool use.", - "family": "gemma", + "minimax/minimax-m2.5-highspeed": { + "id": "minimax/minimax-m2.5-highspeed", + "name": "MiniMax M2.5 Highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax-m2.5", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -26337,82 +23860,61 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 204800, + "output": 131100 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.05 + "input": 0.6, + "output": 2.4, + "cache_read": 0.03 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "kwaipilot/kat-coder-pro": { + "id": "kwaipilot/kat-coder-pro", + "name": "Kat Coder Pro", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-01-05", + "last_updated": "2026-01-05", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 256000, + "output": 128000 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 0.083333 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "google/gemma-4-26b-a4b-it-cybersecurity": { - "id": "google/gemma-4-26b-a4b-it-cybersecurity", - "name": "Gemma 4 26B A4B Cybersecurity", - "description": "Gemma 4 26B A4B Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports optional reasoning, image understanding, tool calling, and a 262,144-token context window.", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "release_date": "2026-09-19", - "last_updated": "2026-09-19", + "nousresearch/hermes-2-pro-llama-3-8b": { + "id": "nousresearch/hermes-2-pro-llama-3-8b", + "name": "Hermes 2 Pro Llama 3 8B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2024-06-27", + "last_updated": "2024-06-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -26420,46 +23922,52 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.1056, - "output": 0.3344, - "cache_read": 0.0528 + "input": 0.14, + "output": 0.14 } - }, - "google/gemini-3.7-flash": { - "id": "google/gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", + } + } + }, + "xai": { + "id": "xai", + "env": [ + "XAI_API_KEY" + ], + "npm": "@ai-sdk/xai", + "name": "xAI", + "doc": "https://docs.x.ai/docs/models", + "models": { + "grok-4.20-multi-agent-0309": { + "id": "grok-4.20-multi-agent-0309", + "name": "Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], - "tool_call": true, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -26468,46 +23976,48 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro (Preview Custom Tools)", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "grok-4.20-0309-non-reasoning": { + "id": "grok-4.20-0309-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -26515,29 +24025,43 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 2, - "output": 12, + "input": 1.25, + "output": 2.5, "cache_read": 0.2, - "cache_write": 0.375 + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash (Preview)", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "grok-4.3": { + "id": "grok-4.3", + "name": "Grok 4.3", + "description": "xAI's Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", "high" @@ -26547,15 +24071,13 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -26563,73 +24085,98 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "google/gemini-pro-latest": { - "id": "google/gemini-pro-latest", - "name": "Gemini Pro Latest", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "grok-imagine-image-quality": { + "id": "grok-imagine-image-quality", + "name": "Grok Imagine Image Quality", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-03-29", - "last_updated": "2026-03-29", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-03", + "last_updated": "2026-04-03", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ - "text" + "image", + "pdf" ] }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 8000, + "output": 0 + } + }, + "grok-imagine-video": { + "id": "grok-imagine-video", + "name": "Grok Imagine Video", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-01-28", + "last_updated": "2026-01-28", + "modalities": { + "input": [ + "text", + "image", + "video", + "pdf" + ], + "output": [ + "video" + ] }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "open_weights": false, + "limit": { + "context": 1024, + "output": 0 } }, - "google/gemini-3.8-flash": { - "id": "google/gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's latest Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -26639,14 +24186,12 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -26655,47 +24200,49 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667 - } - }, - "google/gemini-3-flash-preview-thinking": { - "id": "google/gemini-3-flash-preview-thinking", - "name": "Gemini 3 Flash Thinking", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 } - ], - "tool_call": false, - "structured_output": false, + } + }, + "grok-4.20-0309-reasoning": { + "id": "grok-4.20-0309-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -26703,46 +24250,76 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "grok-imagine-image": { + "id": "grok-imagine-image", + "name": "Grok Imagine Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-01-28", + "last_updated": "2026-01-28", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "image", + "pdf" + ] + }, + "open_weights": false, + "limit": { + "context": 8000, + "output": 0 + } + }, + "grok-build-0.1": { + "id": "grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -26751,34 +24328,57 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65536 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 0.08333 + "input": 1, + "output": 2, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 4, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 4, + "cache_read": 0.4 + } } - }, - "TEE/gemma-4-31b-it": { - "id": "TEE/gemma-4-31b-it", - "name": "Gemma 4 31B IT TEE", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + } + } + }, + "privatemode-ai": { + "id": "privatemode-ai", + "env": [ + "PRIVATEMODE_API_KEY", + "PRIVATEMODE_ENDPOINT" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "http://localhost:8080/v1", + "name": "Privatemode AI", + "doc": "https://docs.privatemode.ai/api/overview", + "models": { + "qwen3-embedding-4b": { + "id": "qwen3-embedding-4b", + "name": "Qwen3-Embedding 4B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-06", + "release_date": "2025-06-06", + "last_updated": "2025-06-06", "modalities": { "input": [ "text" @@ -26789,34 +24389,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 262144 + "context": 32000, + "output": 2560 }, "cost": { - "input": 0.15, - "output": 0.46, - "cache_read": 0.075 + "input": 0, + "output": 0 } }, - "TEE/qwen3.8-27b": { - "id": "TEE/qwen3.8-27b", - "name": "Qwen3.8 27B TEE", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", + "gemma-3-27b": { + "id": "gemma-3-27b", + "name": "Gemma 3 27B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2024-08", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -26824,28 +24421,25 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 262144 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 3, - "cache_read": 0.15 + "input": 0, + "output": 0 } }, - "TEE/nemotron-3.5-lightning": { - "id": "TEE/nemotron-3.5-lightning", - "name": "Nvidia Nemotron 3.5 Lightning TEE", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -26855,8 +24449,9 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "knowledge": "2025-08", + "release_date": "2025-08-04", + "last_updated": "2025-08-14", "modalities": { "input": [ "text" @@ -26867,42 +24462,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 65536 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.08, - "output": 0.2, - "cache_read": 0.04 + "input": 0, + "output": 0 } }, - "TEE/glm-5.3-flash": { - "id": "TEE/glm-5.3-flash", - "name": "GLM 5.3 Flash TEE", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "whisper-large-v3": { + "id": "whisper-large-v3", + "name": "Whisper large-v3", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2023-09", + "release_date": "2023-09-01", + "last_updated": "2023-09-01", "modalities": { "input": [ - "text", - "image" + "audio" ], "output": [ "text" @@ -26910,33 +24493,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 0, + "output": 4096 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 0, + "output": 0 } }, - "TEE/qwen3.5-27b": { - "id": "TEE/qwen3.5-27b", - "name": "Qwen3.5 27B TEE", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen3-coder-30b-a3b": { + "id": "qwen3-coder-30b-a3b", + "name": "Qwen3-Coder 30B-A3B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -26944,45 +24524,41 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 65536 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 2.4, - "cache_read": 0.15 + "input": 0, + "output": 0 } - }, - "TEE/muse-glimmer-30b": { - "id": "TEE/muse-glimmer-30b", - "name": "Muse Glimmer 30B TEE", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + } + } + }, + "drun": { + "id": "drun", + "env": [ + "DRUN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://chat.d.run/v1", + "name": "D.Run (China)", + "doc": "https://www.d.run", + "models": { + "public/deepseek-v3": { + "id": "public/deepseek-v3", + "name": "DeepSeek V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "knowledge": "2024-07", + "release_date": "2024-12-26", + "last_updated": "2024-12-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -26991,41 +24567,32 @@ "open_weights": true, "limit": { "context": 131072, - "input": 131072, - "output": 131072 + "output": 8192 }, "cost": { - "input": 0.35, - "output": 1.5, - "cache_read": 0.04 + "input": 0.28, + "output": 1.1 } }, - "TEE/kimi-k3": { - "id": "TEE/kimi-k3", - "name": "Kimi K3 TEE", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "public/deepseek-r1": { + "id": "public/deepseek-r1", + "name": "DeepSeek R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -27033,73 +24600,80 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 65535 + "context": 131072, + "output": 32000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 1.5 + "input": 0.55, + "output": 2.2 } }, - "TEE/deepseek-v4.1-flash": { - "id": "TEE/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash TEE", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "public/minimax-m25": { + "id": "public/minimax-m25", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_details" + }, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2025-03-01", + "last_updated": "2025-03-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 384000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.65, - "output": 1.45, - "cache_read": 0.13 + "input": 0.29, + "output": 1.16 } - }, - "TEE/llama3-3-70b": { - "id": "TEE/llama3-3-70b", - "name": "Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + } + } + }, + "alibaba-token-plan-cn": { + "id": "alibaba-token-plan-cn", + "env": [ + "ALIBABA_TOKEN_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://token-plan.cn-beijing.maas.aliyuncs.com/compatible-mode/v1", + "name": "Alibaba Token Plan (China)", + "doc": "https://www.alibabacloud.com/help/zh/model-studio/token-plan-overview", + "models": { + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-03", - "last_updated": "2025-07-03", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -27110,29 +24684,36 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1.75, - "output": 2.75, - "cache_read": 1.75 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "TEE/kimi-k2.6": { - "id": "TEE/kimi-k2.6", - "name": "Kimi K2.6 TEE", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": false, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", @@ -27142,29 +24723,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1.5, - "output": 5.25, - "cache_read": 0.375 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "TEE/gemma-4-26b-a4b-uncensored": { - "id": "TEE/gemma-4-26b-a4b-uncensored", - "name": "Gemma 4 26B A4B Uncensored TEE", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-05-23", - "last_updated": "2026-05-23", + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -27173,66 +24763,46 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "input": 65536, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.7, - "cache_read": 0.075 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "TEE/gemma4-31b:thinking": { - "id": "TEE/gemma4-31b:thinking", - "name": "Gemma 4 31B Thinking TEE", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-05-02", - "last_updated": "2026-05-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "input": 262144, - "output": 131072 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.4, - "output": 1, - "cache_read": 0.4 - } - }, - "TEE/gemma4-31b": { - "id": "TEE/gemma4-31b", - "name": "Gemma 4 31B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2026-04-04", - "last_updated": "2026-04-04", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -27241,27 +24811,35 @@ "open_weights": true, "limit": { "context": 262144, - "input": 262144, - "output": 131072 + "output": 262144 }, "cost": { - "input": 0.4, - "output": 1, - "cache_read": 0.4 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "TEE/qwen3.6-35b-a3b": { - "id": "TEE/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B A3B TEE", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -27272,40 +24850,37 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 262144 + "context": 202752, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.27, - "cache_read": 0.1 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "TEE/glm-5.2:thinking": { - "id": "TEE/glm-5.2:thinking", - "name": "GLM 5.2 Thinking TEE", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -27316,57 +24891,63 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.7 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "TEE/qwen3.5-397b-a17b": { - "id": "TEE/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B TEE", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "wan2.7-image-pro": { + "id": "wan2.7-image-pro", + "name": "Wan2.7 Image Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 65536 + "context": 8192, + "output": 0 }, "cost": { - "input": 0.55, - "output": 3.5, - "cache_read": 0.275 + "input": 0, + "output": 0 } }, - "TEE/glm-5.2": { - "id": "TEE/glm-5.2", - "name": "GLM 5.2 TEE", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, "release_date": "2026-06-13", @@ -27381,81 +24962,84 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "input": 1048576, + "context": 1000000, "output": 131072 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.7 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "TEE/glm-5.1-thinking": { - "id": "TEE/glm-5.1-thinking", - "name": "GLM 5.1 Thinking TEE", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "input": 202752, - "output": 65535 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.5, - "output": 5.25, - "cache_read": 0.3 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "TEE/glm-5.1": { - "id": "TEE/glm-5.1", - "name": "GLM 5.1 TEE", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -27463,31 +25047,63 @@ }, "open_weights": true, "limit": { - "context": 202752, - "input": 202752, - "output": 65535 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 5.25, - "cache_read": 0.3 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "TEE/qwen2.5-vl-72b-instruct": { - "id": "TEE/qwen2.5-vl-72b-instruct", - "name": "Qwen2.5 VL 72B TEE", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen-image-2.0": { + "id": "qwen-image-2.0", + "name": "Qwen Image 2.0", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-02-01", + "temperature": true, + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ - "text", + "text" + ], + "output": [ "image" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 0 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -27495,33 +25111,42 @@ }, "open_weights": true, "limit": { - "context": 65536, - "input": 65536, - "output": 8192 + "context": 196608, + "input": 196601, + "output": 24576 }, "cost": { - "input": 0.7, - "output": 0.7, - "cache_read": 0.35 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "TEE/qwen3.6-27b": { - "id": "TEE/qwen3.6-27b", - "name": "Qwen3.6 27B TEE", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -27530,102 +25155,91 @@ "open_weights": true, "limit": { "context": 262144, - "input": 262144, - "output": 65536 + "output": 16384 }, "cost": { - "input": 0.32, - "output": 2.7, - "cache_read": 0.16 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "TEE/deepseek-v3.2": { - "id": "TEE/deepseek-v3.2", - "name": "DeepSeek V3.2 TEE", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", + "qwen-image-2.0-pro": { + "id": "qwen-image-2.0-pro", + "name": "Qwen Image 2.0 Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 164000, - "input": 164000, - "output": 65536 + "context": 8192, + "output": 0 }, "cost": { - "input": 0.5, - "output": 1, - "cache_read": 0.25 + "input": 0, + "output": 0 } }, - "TEE/gpt-oss-120b": { - "id": "TEE/gpt-oss-120b", - "name": "GPT-OSS 120B TEE", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "wan2.7-image": { + "id": "wan2.7-image", + "name": "Wan2.7 Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 16384 + "context": 8192, + "output": 0 }, "cost": { - "input": 2, - "output": 2, - "cache_read": 2 + "input": 0, + "output": 0 } }, - "TEE/glm-5.3": { - "id": "TEE/glm-5.3", - "name": "GLM 5.3 TEE", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -27634,36 +25248,39 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 202752, + "output": 16384 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "TEE/kimi-k2.7-code": { - "id": "TEE/kimi-k2.7-code", - "name": "Kimi K2.7 Code TEE", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, - "temperature": false, + "structured_output": true, + "temperature": true, "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2025-12-03", + "last_updated": "2025-12-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -27671,58 +25288,81 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, + "context": 131072, "output": 65536 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0, + "output": 0 } }, - "EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2": { - "id": "EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2", - "name": "EVA-Qwen2.5-72B-v0.2", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.799, - "output": 0.799, - "cache_read": 0.3995 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0": { - "id": "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0", - "name": "EVA Llama 3.33 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + } + } + }, + "moonshotai": { + "id": "moonshotai", + "env": [ + "MOONSHOT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.moonshot.ai/v1", + "name": "Moonshot AI", + "doc": "https://platform.moonshot.ai/docs/api/chat", + "models": { + "kimi-k2-0905-preview": { + "id": "kimi-k2-0905-preview", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -27733,27 +25373,31 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2.006, - "output": 2.006, - "cache_read": 1.003 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2": { - "id": "EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2", - "name": "EVA-Qwen2.5-32B-v0.2", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "kimi-k2-thinking-turbo": { + "id": "kimi-k2-thinking-turbo", + "name": "Kimi K2 Thinking Turbo", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ "text" @@ -27764,30 +25408,37 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.799, - "output": 0.799, - "cache_read": 0.3995 + "input": 1.15, + "output": 8, + "cache_read": 0.15 } }, - "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1": { - "id": "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1", - "name": "EVA-LLaMA-3.33-70B-v0.1", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -27795,27 +25446,31 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2.006, - "output": 2.006, - "cache_read": 1.003 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "nothingiisreal/L3.1-70B-Celeste-V0.1-BF16": { - "id": "nothingiisreal/L3.1-70B-Celeste-V0.1-BF16", - "name": "Llama 3.1 70B Celeste v0.1", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-07-23", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ "text" @@ -27826,27 +25481,27 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "MarinaraSpaghetti/NemoMix-Unleashed-12B": { - "id": "MarinaraSpaghetti/NemoMix-Unleashed-12B", - "name": "NemoMix 12B Unleashed", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + "kimi-k2-0711-preview": { + "id": "kimi-k2-0711-preview", + "name": "Kimi K2 0711", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-07-01", + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-07-14", + "last_updated": "2025-07-14", "modalities": { "input": [ "text" @@ -27857,28 +25512,27 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "pamanseau/OpenReasoning-Nemotron-32B": { - "id": "pamanseau/OpenReasoning-Nemotron-32B", - "name": "OpenReasoning Nemotron 32B", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", + "kimi-k2-turbo-preview": { + "id": "kimi-k2-turbo-preview", + "name": "Kimi K2 Turbo", + "description": "Fast Kimi model for responsive chat, coding help, and agent loops", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -27889,31 +25543,41 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.05 + "input": 2.4, + "output": 10, + "cache_read": 0.6 } }, - "huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated": { - "id": "huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated", - "name": "DeepSeek R1 Qwen Abliterated", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-01-20", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -27921,30 +25585,41 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.4, - "output": 1.4, - "cache_read": 0.7 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "huihui-ai/Qwen2.5-32B-Instruct-abliterated": { - "id": "huihui-ai/Qwen2.5-32B-Instruct-abliterated", - "name": "Qwen 2.5 32B Abliterated", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-06", - "last_updated": "2025-01-06", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -27952,30 +25627,37 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.7, - "output": 0.7, - "cache_read": 0.35 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "huihui-ai/Llama-3.3-70B-Instruct-abliterated": { - "id": "huihui-ai/Llama-3.3-70B-Instruct-abliterated", - "name": "Llama 3.3 70B Instruct abliterated", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "kimi-k2.7-code-highspeed": { + "id": "kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code HighSpeed", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -27983,31 +25665,50 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.7, - "output": 0.7, - "cache_read": 0.35 + "input": 1.9, + "output": 8, + "cache_read": 0.38 } - }, - "huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated": { - "id": "huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated", - "name": "DeepSeek R1 Llama 70B Abliterated", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek", - "attachment": false, + } + } + }, + "fireworks-ai": { + "id": "fireworks-ai", + "env": [ + "FIREWORKS_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.fireworks.ai/inference/v1/", + "name": "Fireworks AI", + "doc": "https://fireworks.ai/docs/", + "models": { + "accounts/fireworks/routers/kimi-k2p6-turbo": { + "id": "accounts/fireworks/routers/kimi-k2p6-turbo", + "name": "Kimi K2.6 Turbo", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-01-20", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -28015,27 +25716,41 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.7, - "output": 0.7, - "cache_read": 0.35 + "input": 2, + "output": 8, + "cache_read": 0.3 } }, - "Gryphe/MythoMax-L2-13b": { - "id": "Gryphe/MythoMax-L2-13b", - "name": "MythoMax 13B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "accounts/fireworks/routers/glm-5p2-fast": { + "id": "accounts/fireworks/routers/glm-5p2-fast", + "name": "GLM 5.2 Fast", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-26", + "last_updated": "2026-06-26", "modalities": { "input": [ "text" @@ -28046,40 +25761,38 @@ }, "open_weights": true, "limit": { - "context": 4096, - "input": 4096, - "output": 3686 + "context": 1048575, + "output": 131072 }, "cost": { - "input": 0.1003, - "output": 0.1003, - "cache_read": 0.05015 + "input": 2.1, + "output": 6.6, + "cache_read": 0.21 } }, - "ibm-granite/granite-4.2-8b": { - "id": "ibm-granite/granite-4.2-8b", - "name": "Granite 4.2 8B", - "description": "IBM Granite 4.2 8B is an Apache 2.0-licensed dense model with native step-by-step reasoning and specialized training for agentic work. It can plan before acting, sequence tools, navigate codebases, work in terminals, and verify results across coding, search, mathematics, science, and complex instruction-following tasks.", - "family": "granite", - "attachment": false, + "accounts/fireworks/routers/kimi-k2p7-code-fast": { + "id": "accounts/fireworks/routers/kimi-k2p7-code-fast", + "name": "Kimi K2.7 Code Fast", + "description": "Kimi coding model for software agents, refactors, and repository reasoning", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-08-31", - "last_updated": "2026-08-31", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-12", + "last_updated": "2026-06-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -28087,30 +25800,37 @@ }, "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 117964 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.1, - "output": 0.15, - "cache_read": 0.05 + "input": 1.9, + "output": 8, + "cache_read": 0.38 } }, - "TheDrummer/skyfall-36b-v2": { - "id": "TheDrummer/skyfall-36b-v2", - "name": "TheDrummer Skyfall 36B V2", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-03-10", - "last_updated": "2025-03-10", + "accounts/fireworks/routers/glm-5p1-fast": { + "id": "accounts/fireworks/routers/glm-5p1-fast", + "name": "GLM 5.1 Fast", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -28118,27 +25838,34 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 29491 + "context": 202800, + "output": 131072 }, "cost": { - "input": 0.55, - "output": 0.8, - "cache_read": 0.25 + "input": 2.8, + "output": 8.8, + "cache_read": 0.52 } }, - "TheDrummer/Artemis-v1.1": { - "id": "TheDrummer/Artemis-v1.1", - "name": "TheDrummer/Artemis v1.1", - "description": "TheDrummer's Artemis v1.1 is a Gemma 4 31B fine-tune for creative writing, expressive dialogue, and roleplay, with optional thinking and a 262K context window.", + "accounts/fireworks/routers/kimi-k2p6-fast": { + "id": "accounts/fireworks/routers/kimi-k2p6-fast", + "name": "Kimi K2.6 Fast", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2026-09-06", - "last_updated": "2026-09-06", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-06-05", "modalities": { "input": [ "text", @@ -28150,26 +25877,43 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.1, - "output": 0.45, - "cache_read": 0.05 + "input": 2, + "output": 8, + "cache_read": 0.3 } }, - "TheDrummer/Anubis-70B-v1": { - "id": "TheDrummer/Anubis-70B-v1", - "name": "Anubis 70B v1", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "accounts/fireworks/models/deepseek-v4-flash": { + "id": "accounts/fireworks/models/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-06-16", "modalities": { "input": [ "text" @@ -28180,26 +25924,43 @@ }, "open_weights": true, "limit": { - "context": 65536, - "input": 65536, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.31, - "output": 0.31, - "cache_read": 0.155 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "TheDrummer/Anubis-70B-v1.1": { - "id": "TheDrummer/Anubis-70B-v1.1", - "name": "Anubis 70B v1.1", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "accounts/fireworks/models/deepseek-v4-pro": { + "id": "accounts/fireworks/models/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -28210,26 +25971,39 @@ }, "open_weights": true, "limit": { - "context": 32000, - "input": 32000, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.31, - "output": 0.31, - "cache_read": 0.155 + "input": 1.74, + "output": 3.48, + "cache_read": 0.145 } }, - "TheDrummer/Cydonia-24B-v4.3": { - "id": "TheDrummer/Cydonia-24B-v4.3", - "name": "The Drummer Cydonia 24B v4.3", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "accounts/fireworks/models/minimax-m2p7": { + "id": "accounts/fireworks/models/minimax-m2p7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-25", - "last_updated": "2025-12-25", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-12", + "last_updated": "2026-04-12", "modalities": { "input": [ "text" @@ -28240,26 +26014,39 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 196608, + "output": 196608 }, "cost": { - "input": 0.12, - "output": 0.15, + "input": 0.3, + "output": 1.2, "cache_read": 0.06 } }, - "TheDrummer/Magidonia-24B-v4.3": { - "id": "TheDrummer/Magidonia-24B-v4.3", - "name": "The Drummer Magidonia 24B v4.3", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-25", - "last_updated": "2025-12-25", + "accounts/fireworks/models/minimax-m3": { + "id": "accounts/fireworks/models/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text" @@ -28270,29 +26057,38 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 512000, + "output": 512000 }, "cost": { - "input": 0.1003, - "output": 0.1207, - "cache_read": 0.05015 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "TheDrummer/Cydonia-24B-v4.1": { - "id": "TheDrummer/Cydonia-24B-v4.1", - "name": "The Drummer Cydonia 24B v4.1", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, + "accounts/fireworks/models/kimi-k2p6": { + "id": "accounts/fireworks/models/kimi-k2p6", + "name": "Kimi K2.6", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2025-08-19", - "last_updated": "2025-08-19", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -28300,57 +26096,82 @@ }, "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 117964 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.35, - "output": 0.55, + "input": 0.95, + "output": 4, "cache_read": 0.16 } }, - "TheDrummer/UnslopNemo-12B-v4.1": { - "id": "TheDrummer/UnslopNemo-12B-v4.1", - "name": "UnslopNemo 12b v4", - "description": "Multimodal model for analyzing text, images, documents, and rich media", + "accounts/fireworks/models/qwen3p7-plus": { + "id": "accounts/fireworks/models/qwen3p7-plus", + "name": "Qwen 3.7 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1 + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "input": 8192, - "output": 26214 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.4, + "output": 1.6, + "cache_read": 0.08 } }, - "TheDrummer/Cydonia-24B-v2": { - "id": "TheDrummer/Cydonia-24B-v2", - "name": "The Drummer Cydonia 24B v2", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "accounts/fireworks/models/glm-5p1": { + "id": "accounts/fireworks/models/glm-5p1", + "name": "GLM 5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-17", - "last_updated": "2025-02-17", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text" @@ -28361,26 +26182,41 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 202800, + "output": 131072 }, "cost": { - "input": 0.1003, - "output": 0.1207, - "cache_read": 0.05015 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "TheDrummer/Rocinante-12B-v1.1": { - "id": "TheDrummer/Rocinante-12B-v1.1", - "name": "Rocinante 12b", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "accounts/fireworks/models/glm-5p2": { + "id": "accounts/fireworks/models/glm-5p2", + "name": "GLM 5.2", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ "text" @@ -28391,26 +26227,36 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1048575, + "output": 131072 }, "cost": { - "input": 0.408, - "output": 0.595, - "cache_read": 0.204 + "input": 1.4, + "output": 4.4, + "cache_read": 0.14 } }, - "TheDrummer/Cydonia-24B-v4": { - "id": "TheDrummer/Cydonia-24B-v4", - "name": "The Drummer Cydonia 24B v4", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "accounts/fireworks/models/gpt-oss-120b": { + "id": "accounts/fireworks/models/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2026-06-16", "modalities": { "input": [ "text" @@ -28421,31 +26267,78 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, + "context": 131072, "output": 32768 }, "cost": { - "input": 0.2006, - "output": 0.2414, - "cache_read": 0.1003 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "accounts/fireworks/models/gpt-oss-20b": { + "id": "accounts/fireworks/models/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.07, + "output": 0.3, + "cache_read": 0.035 + } + }, + "accounts/fireworks/models/kimi-k2p7-code": { + "id": "accounts/fireworks/models/kimi-k2p7-code", + "name": "Kimi K2.7 Code", + "description": "Kimi coding model for software agents, refactors, and repository reasoning", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-12", + "last_updated": "2026-06-16", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" @@ -28453,28 +26346,44 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.2, - "output": 0.7, - "cache_read": 0.1 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } - }, - "deepseek-ai/deepseek-v3.2-exp-thinking": { - "id": "deepseek-ai/deepseek-v3.2-exp-thinking", - "name": "DeepSeek V3.2 Exp Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek-thinking", - "attachment": false, + } + } + }, + "vultr": { + "id": "vultr", + "env": [ + "VULTR_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.vultrinference.com/v1", + "name": "Vultr", + "doc": "https://api.vultrinference.com/", + "models": { + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-09-29", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text" @@ -28485,31 +26394,33 @@ }, "open_weights": true, "limit": { - "context": 163840, - "input": 163840, - "output": 65536 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.28, - "output": 0.42, - "cache_read": 0.14 + "input": 0.3, + "output": 1.2 } }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek", - "attachment": false, + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -28517,30 +26428,33 @@ }, "open_weights": true, "limit": { - "context": 163840, - "input": 163840, - "output": 32768 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 1.7, - "cache_read": 0.2 + "input": 0.3, + "output": 2 } }, - "deepseek-ai/DeepSeek-V3.1-Terminus": { - "id": "deepseek-ai/DeepSeek-V3.1-Terminus", - "name": "DeepSeek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2025-08-02", - "last_updated": "2025-08-02", + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -28548,27 +26462,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, + "context": 262144, "output": 65536 }, "cost": { - "input": 0.25, - "output": 0.7, - "cache_read": 0.125 + "input": 0.3, + "output": 2 } }, - "deepseek-ai/deepseek-v3.2-exp": { - "id": "deepseek-ai/deepseek-v3.2-exp", - "name": "DeepSeek V3.2 Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "XiaomiMiMo/MiMo-V2.5-Pro": { + "id": "XiaomiMiMo/MiMo-V2.5-Pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-09-29", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -28579,27 +26493,27 @@ }, "open_weights": true, "limit": { - "context": 163840, - "input": 163840, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.28, - "output": 0.42, - "cache_read": 0.14 + "input": 0.55, + "output": 1.65 } }, - "deepseek-ai/DeepSeek-V3.1:thinking": { - "id": "deepseek-ai/DeepSeek-V3.1:thinking", - "name": "DeepSeek V3.1 Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek-thinking", + "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16": { + "id": "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16", + "name": "NVIDIA Nemotron 3 Nano Omni", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-08-21", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ "text" @@ -28610,27 +26524,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.7, - "cache_read": 0.1 + "input": 0.13, + "output": 0.38 } }, - "deepseek-ai/DeepSeek-V3.1-Terminus:thinking": { - "id": "deepseek-ai/DeepSeek-V3.1-Terminus:thinking", - "name": "DeepSeek V3.1 Terminus (Thinking)", + "nvidia/DeepSeek-V3.2-NVFP4": { + "id": "nvidia/DeepSeek-V3.2-NVFP4", + "name": "DeepSeek V3.2", "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek-thinking", + "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2025-09-22", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -28641,38 +26555,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 0.7, - "cache_read": 0.125 + "input": 0.55, + "output": 1.65 } }, - "stepfun-ai/step-3.5-flash-2603": { - "id": "stepfun-ai/step-3.5-flash-2603", - "name": "Step 3.5 Flash 2603", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "nvidia/Nemotron-Cascade-2-30B-A3B": { + "id": "nvidia/Nemotron-Cascade-2-30B-A3B", + "name": "NVIDIA Nemotron Cascade 2", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -28684,37 +26587,29 @@ "open_weights": true, "limit": { "context": 262144, - "input": 262144, - "output": 65536 + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.05 + "input": 0.15, + "output": 0.6 } }, - "stepfun-ai/step-3.5-flash": { - "id": "stepfun-ai/step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "zai-org/GLM-5.2-FP8": { + "id": "zai-org/GLM-5.2-FP8", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -28725,27 +26620,28 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 65536 + "context": 393216, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.05 + "input": 0.85, + "output": 3.1 } }, - "MiniMaxAI/MiniMax-M1-80k": { - "id": "MiniMaxAI/MiniMax-M1-80k", - "name": "MiniMax M1 80K", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-01-08", - "last_updated": "2025-06-16", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -28754,29 +26650,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "input": 1000000, - "output": 131072 + "output": 384000 }, "cost": { - "input": 0.6052, - "output": 2.4225, - "cache_read": 0.3026 + "input": 0.3, + "output": 1 } }, - "mistralai/ministral-8b-2512": { - "id": "mistralai/ministral-8b-2512", - "name": "Ministral 8B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "MiniMaxAI/MiniMax-M2.7": { + "id": "MiniMaxAI/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-04", - "last_updated": "2025-12-04", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -28787,21 +26682,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.075 + "input": 0.3, + "output": 1.2 } - }, - "mistralai/mistral-small-4-119b-2603:thinking": { - "id": "mistralai/mistral-small-4-119b-2603:thinking", - "name": "Mistral Small 4 119B Thinking", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + } + } + }, + "302ai": { + "id": "302ai", + "env": [ + "302AI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.302.ai/v1", + "name": "302.AI", + "doc": "https://doc.302.ai", + "models": { + "gpt-5.4-mini-2026-03-17": { + "id": "gpt-5.4-mini-2026-03-17", + "name": "gpt-5.4-mini-2026-03-17", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -28809,14 +26714,19 @@ "type": "effort", "values": [ "none", - "high" + "low", + "medium", + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-19", + "last_updated": "2026-03-19", "modalities": { "input": [ "text", @@ -28826,23 +26736,53 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, + "context": 400000, + "input": 272000, + "output": 128000 + }, + "cost": { + "input": 0.75, + "output": 4.5 + } + }, + "chatgpt-4o-latest": { + "id": "chatgpt-4o-latest", + "name": "chatgpt-4o-latest", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-08-08", + "last_updated": "2024-08-08", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, "output": 16384 }, "cost": { - "input": 0.4, - "output": 1.4, - "cache_read": 0.2 + "input": 5, + "output": 15 } }, - "mistralai/mistral-medium-3.5": { - "id": "mistralai/mistral-medium-3.5", - "name": "Mistral Medium 3.5", - "description": "Mistral Medium 3.5 is a 128B dense open-weights flagship model for instruction-following, reasoning, coding, long-horizon agentic work, tool use, structured output, and multimodal prompts. It supports a 256k context window and configurable reasoning effort.", - "family": "mistral-medium", + "gpt-5.4-nano-2026-03-17": { + "id": "gpt-5.4-nano-2026-03-17", + "name": "gpt-5.4-nano-2026-03-17", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -28850,14 +26790,19 @@ "type": "effort", "values": [ "none", - "high" + "low", + "medium", + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-19", + "last_updated": "2026-03-19", "modalities": { "input": [ "text", @@ -28867,29 +26812,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.75 + "input": 0.2, + "output": 1.25 } }, - "mistralai/mistral-small-24b-instruct-2501": { - "id": "mistralai/mistral-small-24b-instruct-2501", - "name": "Mistral Small 24B", - "description": "Mistral Small 24B hosted by IONOS in Berlin, Germany. Zero data retention.", - "family": "mistral-small", + "kimi-k2-0905-preview": { + "id": "kimi-k2-0905-preview", + "name": "kimi-k2-0905-preview", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -28898,90 +26842,88 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.1155, - "output": 0.3465 + "input": 0.632, + "output": 2.53 } }, - "mistralai/mixtral-8x22b-instruct-v0.1": { - "id": "mistralai/mixtral-8x22b-instruct-v0.1", - "name": "Mixtral 8x22B", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mixtral", - "attachment": false, + "grok-4.20-beta-0309-non-reasoning": { + "id": "grok-4.20-beta-0309-non-reasoning", + "name": "grok-4.20-beta-0309-non-reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "tool_call": true, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "input": 65536, - "output": 52428 + "context": 2000000, + "output": 30000 }, "cost": { "input": 2, - "output": 6, - "cache_read": 0.2 + "output": 6 } }, - "mistralai/mistral-saba": { - "id": "mistralai/mistral-saba", - "name": "Mistral Saba", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": false, + "gemini-2.5-flash-nothink": { + "id": "gemini-2.5-flash-nothink", + "name": "gemini-2.5-flash-nothink", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-02-17", + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-24", + "last_updated": "2025-06-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 26214 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.1989, - "output": 0.595, - "cache_read": 0.09945 + "input": 0.3, + "output": 2.5 } }, - "mistralai/mistral-medium-3.1": { - "id": "mistralai/mistral-medium-3.1", - "name": "Mistral Medium 3.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "qwen-plus": { + "id": "qwen-plus", + "name": "Qwen-Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -28992,53 +26934,20 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, + "context": 1000000, "output": 32768 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.2 + "input": 0.12, + "output": 1.2 } }, - "mistralai/mistral-small-3.2-24b-instruct": { - "id": "mistralai/mistral-small-3.2-24b-instruct", - "name": "Mistral Small 3.2 24B (2506)", - "description": "The latest iteration of Mistral Small, version 3.2 (2506) brings enhanced performance and capabilities. With 24 billion parameters, this model delivers state-of-the-art results across text generation tasks with improved efficiency.", - "family": "mistral-small", + "glm-4.7": { + "id": "glm-4.7", + "name": "glm-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "input": 128000, - "output": 16384 - }, - "cost": { - "input": 0.2, - "output": 0.4, - "cache_read": 0.1 - } - }, - "mistralai/mistral-small-4-119b-2603": { - "id": "mistralai/mistral-small-4-119b-2603", - "name": "Mistral Small 4 119B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -29046,13 +26955,16 @@ } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -29060,27 +26972,25 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.4, - "output": 1.4, - "cache_read": 0.2 + "input": 0.286, + "output": 1.142 } }, - "mistralai/mistral-large": { - "id": "mistralai/mistral-large", - "name": "Mistral Large 2411", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "qwen3-235b-a22b-instruct-2507", + "description": "Tool-capable chat model for instruction following and agentic application workflows", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-02-26", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ "text" @@ -29092,39 +27002,35 @@ "open_weights": false, "limit": { "context": 128000, - "input": 128000, - "output": 102400 + "output": 65536 }, "cost": { - "input": 2.006, - "output": 6.001, - "cache_read": 0.2 + "input": 0.29, + "output": 1.143 } }, - "mistralai/mistral-medium-3.5:thinking": { - "id": "mistralai/mistral-medium-3.5:thinking", - "name": "Mistral Medium 3.5 Thinking", - "description": "Mistral Medium 3.5 with reasoning enabled by default (reasoning_effort=high), for complex coding, agentic, and multi-step reasoning prompts.", - "family": "mistral-medium", + "glm-4.5v": { + "id": "glm-4.5v", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-12", + "last_updated": "2025-08-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -29132,89 +27038,119 @@ }, "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 64000, + "output": 16384 }, "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.75 + "input": 0.29, + "output": 0.86 } }, - "mistralai/devstral-small-2505": { - "id": "mistralai/devstral-small-2505", - "name": "Mistral Devstral Small 2505", - "description": "OpenHands+Devstral is 100% local 100% open, and is SOTA for the category on SWE-Bench Verified: 46.8% accuracy.", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-02", - "last_updated": "2025-08-02", + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "claude-opus-4-5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.06, - "output": 0.06, - "cache_read": 0.03 + "input": 5, + "output": 25 } }, - "mistralai/devstral-2-123b-instruct-2512": { - "id": "mistralai/devstral-2-123b-instruct-2512", - "name": "Devstral 2 123B", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "gemini-2.5-pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.4, - "output": 1.4, - "cache_read": 0.2 + "input": 1.25, + "output": 10 } }, - "mistralai/mistral-medium-3": { - "id": "mistralai/mistral-medium-3", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "gpt-5": { + "id": "gpt-5", + "name": "gpt-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ "text", @@ -29226,59 +27162,69 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.2 + "input": 1.25, + "output": 10 } }, - "mistralai/ministral-14b-instruct-2512": { - "id": "mistralai/ministral-14b-instruct-2512", - "name": "Ministral 3 14B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "claude-haiku-4-5-20251001", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-16", + "last_updated": "2025-10-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.05 + "input": 1, + "output": 5 } }, - "mistralai/ministral-14b-2512": { - "id": "mistralai/ministral-14b-2512", - "name": "Ministral 14B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "kimi-k2-thinking-turbo": { + "id": "kimi-k2-thinking-turbo", + "name": "kimi-k2-thinking-turbo", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-04", - "last_updated": "2025-12-04", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -29287,61 +27233,65 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "input": 262144, - "output": 32768 + "output": 262144 }, "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.1 + "input": 1.265, + "output": 9.119 } }, - "mistralai/mistral-small-3.1-24b-instruct": { - "id": "mistralai/mistral-small-3.1-24b-instruct", - "name": "Mistral Small 3.1 24B (2503)", - "description": "Building upon Mistral Small 3 (2501), Mistral Small 3.1 (2503) adds state-of-the-art vision understanding and enhances long context capabilities up to 128k tokens without compromising text performance. With 24 billion parameters, this model achieves top-tier capabilities in both text and vision tasks.", - "family": "mistral-small", + "claude-3-5-haiku-20241022": { + "id": "claude-3-5-haiku-20241022", + "name": "claude-3-5-haiku-20241022", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "tool_call": true, + "temperature": true, + "knowledge": "2024-07-31", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 102400 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.05 + "input": 0.8, + "output": 4 } }, - "mistralai/ministral-3b-2512": { - "id": "mistralai/ministral-3b-2512", - "name": "Ministral 3B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "glm-4.5": { + "id": "glm-4.5", + "name": "GLM-4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-12-04", - "last_updated": "2025-12-04", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ "text" @@ -29353,60 +27303,38 @@ "open_weights": true, "limit": { "context": 131072, - "input": 131072, - "output": 32768 + "output": 98304 }, "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.05 + "input": 0.286, + "output": 1.142 } }, - "mistralai/mistral-nemo-instruct-2407": { - "id": "mistralai/mistral-nemo-instruct-2407", - "name": "Mistral Nemo", - "description": "12B parameter model with multilingual support.", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 16384, - "input": 16384, - "output": 8192 - }, - "cost": { - "input": 0.1003, - "output": 0.1207, - "cache_read": 0.05015 - } - }, - "mistralai/codestral-2508": { - "id": "mistralai/codestral-2508", - "name": "Codestral 2508", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "codestral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-08-01", - "last_updated": "2025-08-01", + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "gpt-5-pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-08", + "last_updated": "2025-10-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -29414,42 +27342,30 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 272000 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.15 + "input": 15, + "output": 120 } }, - "sakana/fugu-max": { - "id": "sakana/fugu-max", - "name": "Fugu Max", - "description": "Sakana AI's cost-performance Fugu model uses learned multi-agent orchestration to route tasks across expert models for reasoning, coding, and tool use.", - "family": "fugu", + "grok-4.20-beta-0309-reasoning": { + "id": "grok-4.20-beta-0309-reasoning", + "name": "grok-4.20-beta-0309-reasoning", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -29457,38 +27373,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 128000 + "context": 2000000, + "output": 30000 }, "cost": { "input": 2, - "output": 6, - "cache_read": 0.25 + "output": 6 } }, - "sakana/fugu-ultra": { - "id": "sakana/fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "gemini-2.5-flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", @@ -29501,40 +27405,31 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 1000000, - "output": 16384 + "output": 65536 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.3, + "output": 2.5 } }, - "sakana/fugu-ultra-v1.1": { - "id": "sakana/fugu-ultra-v1.1", - "name": "Fugu Ultra v1.1", - "description": "Sakana AI's upgraded Fugu Ultra release with stronger coding, agentic task execution, and advanced reasoning through dynamic orchestration of frontier models.", - "family": "fugu", + "gpt-4o": { + "id": "gpt-4o", + "name": "gpt-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -29542,27 +27437,24 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, + "context": 128000, "output": 16384 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 2.5, + "output": 10 } }, - "inclusionai/ling-3.0-flash": { - "id": "inclusionai/ling-3.0-flash", - "name": "Ling 3.0 Flash", - "description": "Ling-3.0-flash is a 124B-parameter Mixture-of-Experts model with approximately 5.1B parameters active per token. It prioritizes token efficiency and production-scale agentic inference, helping coding and tool-using agents complete more work within constrained latency and serving budgets.", - "family": "ling", + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-07-23", - "last_updated": "2026-07-23", + "temperature": true, + "release_date": "2025-12-19", + "last_updated": "2025-12-19", "modalities": { "input": [ "text" @@ -29573,87 +27465,58 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.075, - "output": 0.22, - "cache_read": 0.015 + "input": 0.3, + "output": 1.2 } }, - "inclusionai/ling-3.0-flash-vl": { - "id": "inclusionai/ling-3.0-flash-vl", - "name": "Ling 3.0 Flash VL", - "description": "Ling 3.0 Flash VL is inclusionAI's native multimodal Mixture-of-Experts model with 124B total parameters and 5.5B active parameters per token. It combines image and video understanding with reasoning and tool use for document analysis, charts, visual verification, and interface-based agent tasks. Thinking is enabled by default and can be turned off in settings.", - "family": "ling", + "gemini-2.5-flash-lite-preview-09-2025": { + "id": "gemini-2.5-flash-lite-preview-09-2025", + "name": "gemini-2.5-flash-lite-preview-09-2025", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-09-09", - "last_updated": "2026-09-09", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-09-26", + "last_updated": "2025-09-26", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.012 + "input": 0.1, + "output": 0.4 } }, - "inclusionai/ling-3.0-flash:thinking": { - "id": "inclusionai/ling-3.0-flash:thinking", - "name": "Ling 3.0 Flash Thinking", - "description": "Ling-3.0-flash Thinking enables visible reasoning on inclusionAI's token-efficient 124B-parameter Mixture-of-Experts model for harder coding, tool use, planning, and production-scale agent workflows.", - "family": "ling", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "doubao-seed-1-6-vision-250815": { + "id": "doubao-seed-1-6-vision-250815", + "name": "doubao-seed-1-6-vision-250815", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-07-23", - "last_updated": "2026-07-23", + "temperature": true, + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -29661,112 +27524,105 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.075, - "output": 0.22, - "cache_read": 0.015 + "input": 0.114, + "output": 1.143 } }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "claude-opus-4-1-20250805", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 943718 + "context": 200000, + "output": 32000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2 + "input": 15, + "output": 75 } }, - "moonshotai/kimi-k2.7-code-highspeed": { - "id": "moonshotai/kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code High-Speed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen3-max-2025-09-23": { + "id": "qwen3-max-2025-09-23", + "name": "qwen3-max-2025-09-23", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-24", + "last_updated": "2025-09-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, + "context": 258048, "output": 65536 }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.32 + "input": 0.86, + "output": 3.43 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": false, + "glm-4.7-flashx": { + "id": "glm-4.7-flashx", + "name": "glm-4.7-flashx", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-04", + "release_date": "2026-01-20", + "last_updated": "2026-01-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -29774,34 +27630,72 @@ }, "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 65536 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 2.6, - "cache_read": 0.125 + "input": 0.0715, + "output": 0.429 } }, - "moonshotai/kimi-k2.5:thinking": { - "id": "moonshotai/kimi-k2.5:thinking", - "name": "Kimi K2.5 Thinking", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, + "glm-5.1": { + "id": "glm-5.1", + "name": "glm-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-10", + "last_updated": "2026-04-10", "modalities": { "input": [ - "text", - "image" + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0.86, + "output": 3.5 + } + }, + "glm-4.6": { + "id": "glm-4.6", + "name": "glm-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -29809,30 +27703,26 @@ }, "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.9, - "cache_read": 0.15 + "input": 0.286, + "output": 1.142 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "kimi-k2-thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "knowledge": "2025-06", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -29841,63 +27731,70 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "input": 262144, - "output": 98304 + "output": 262144 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 0.575, + "output": 2.3 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "claude-sonnet-4-5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 1.9, - "cache_read": 0.15 + "input": 3, + "output": 15 } }, - "moonshotai/kimi-k2-instruct": { - "id": "moonshotai/kimi-k2-instruct", - "name": "Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "glm-4.5-x": { + "id": "glm-4.5-x", + "name": "glm-4.5-x", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2025-07-01", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ "text" @@ -29906,141 +27803,156 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 8192 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.2 + "input": 1.143, + "output": 2.29 } }, - "moonshotai/kimi-k2.6:thinking": { - "id": "moonshotai/kimi-k2.6:thinking", - "name": "Kimi K2.6 Thinking", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "deepseek-v3.2-thinking": { + "id": "deepseek-v3.2-thinking", + "name": "DeepSeek-V3.2-Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2024-12", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 65536 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 2.6, - "cache_read": 0.125 + "input": 0.29, + "output": 0.43 } }, - "moonshotai/kimi-k2-instruct-0711": { - "id": "moonshotai/kimi-k2-instruct-0711", - "name": "Kimi K2 0711", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "claude-sonnet-4-6-thinking": { + "id": "claude-sonnet-4-6-thinking", + "name": "claude-sonnet-4-6-thinking", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2025-07-11", + "temperature": true, + "knowledge": "2025-08", + "release_date": "2026-02-18", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.2 + "input": 3, + "output": 15 } }, - "moonshotai/kimi-latest": { - "id": "moonshotai/kimi-latest", - "name": "Kimi Latest", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "claude-opus-4-7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "temperature": true, + "knowledge": "2026-01-31", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 943718 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "grok-4-1-fast-non-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ "text", @@ -30050,91 +27962,104 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 65536 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.2, + "output": 0.5 } }, - "moonshotai/Kimi-K2-Instruct-0905": { - "id": "moonshotai/Kimi-K2-Instruct-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "gpt-5.4-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-19", + "last_updated": "2026-03-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 100352 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.2 + "input": 0.2, + "output": 1.25 } }, - "THUDM/GLM-Z1-9B-0414": { - "id": "THUDM/GLM-Z1-9B-0414", - "name": "GLM Z1 9B 0414", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm-z", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-04-14", + "claude-opus-4-1-20250805-thinking": { + "id": "claude-opus-4-1-20250805-thinking", + "name": "claude-opus-4-1-20250805-thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-05-27", + "last_updated": "2025-05-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 8000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.1 + "input": 15, + "output": 75 } }, - "THUDM/GLM-4-9B-0414": { - "id": "THUDM/GLM-4-9B-0414", - "name": "GLM 4 9B 0414", + "glm-4.5-airx": { + "id": "glm-4.5-airx", + "name": "glm-4.5-airx", "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-04-14", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ "text" @@ -30143,237 +28068,240 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 8000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.1 + "input": 0.572, + "output": 1.714 } }, - "THUDM/GLM-4-32B-0414": { - "id": "THUDM/GLM-4-32B-0414", - "name": "GLM 4 32B 0414", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "grok-4.1": { + "id": "grok-4.1", + "name": "grok-4.1", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-04-14", + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.1 + "input": 2, + "output": 10 } }, - "nvidia/nemotron-3-nano-30b-a3b": { - "id": "nvidia/nemotron-3-nano-30b-a3b", - "name": "Nvidia Nemotron 3 Nano 30B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, + "gemini-2.5-flash-preview-09-2025": { + "id": "gemini-2.5-flash-preview-09-2025", + "name": "gemini-2.5-flash-preview-09-2025", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "knowledge": "2025-01", + "release_date": "2025-09-26", + "last_updated": "2025-09-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 235929 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.17, - "output": 0.68, - "cache_read": 0.085 + "input": 0.3, + "output": 2.5 } }, - "nvidia/nemotron-3.5-lightning": { - "id": "nvidia/nemotron-3.5-lightning", - "name": "Nvidia Nemotron 3.5 Lightning", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron", - "attachment": false, + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "claude-opus-4-5-20251101", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "knowledge": "2025-03-31", + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.01 + "input": 5, + "output": 25 } }, - "nvidia/Llama-3.3-Nemotron-Super-49B-v1": { - "id": "nvidia/Llama-3.3-Nemotron-Super-49B-v1", - "name": "Nvidia Nemotron Super 49B", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "claude-opus-4-20250514": { + "id": "claude-opus-4-20250514", + "name": "claude-opus-4-20250514", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.075 + "input": 15, + "output": 75 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "Nvidia Nemotron 3 Super 120B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "gemini-3-pro-image-preview": { + "id": "gemini-3-pro-image-preview", + "name": "gemini-3-pro-image-preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "knowledge": "2025-06", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 32768, + "output": 64000 }, "cost": { - "input": 0.05, - "output": 0.25, - "cache_read": 0.025 + "input": 2, + "output": 120 } }, - "nvidia/nemotron-3-ultra-550b-a55b": { - "id": "nvidia/nemotron-3-ultra-550b-a55b", - "name": "Nvidia Nemotron 3 Ultra 550B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, + "gemini-2.5-flash-image": { + "id": "gemini-2.5-flash-image", + "name": "gemini-2.5-flash-image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2025-01", + "release_date": "2025-10-08", + "last_updated": "2025-10-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 65536 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.25 + "input": 0.3, + "output": 30 } }, - "nvidia/nemotron-3-super-120b-a12b:thinking": { - "id": "nvidia/nemotron-3-super-120b-a12b:thinking", - "name": "Nvidia Nemotron 3 Super 120B Thinking", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "glm-for-coding": { + "id": "glm-for-coding", + "name": "glm-for-coding", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -30382,64 +28310,71 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.25, - "cache_read": 0.025 + "input": 0.086, + "output": 0.343 } }, - "nvidia/nemotron-3-ultra-550b-a55b:thinking": { - "id": "nvidia/nemotron-3-ultra-550b-a55b:thinking", - "name": "Nvidia Nemotron 3 Ultra 550B Thinking", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "gpt-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-12", + "last_updated": "2025-12-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.25 + "input": 1.75, + "output": 14 } }, - "nvidia/nemotron-3.5-content-safety": { - "id": "nvidia/nemotron-3.5-content-safety", - "name": "Nemotron 3.5 Content Safety", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "nemotron", + "qwen-flash": { + "id": "qwen-flash", + "name": "Qwen-Flash", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -30448,338 +28383,275 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 131072, + "context": 1000000, "output": 32768 }, "cost": { - "input": 0.05, - "output": 0.15, - "cache_read": 0.025 + "input": 0.022, + "output": 0.22 } }, - "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF": { - "id": "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF", - "name": "Nvidia Nemotron 70b", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "claude-opus-4-6-thinking": { + "id": "claude-opus-4-6-thinking", + "name": "claude-opus-4-6-thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "knowledge": "2025-05", + "release_date": "2026-02-06", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.357, - "output": 0.408, - "cache_read": 0.1785 + "input": 5, + "output": 25 } }, - "nvidia/nemotron-3.5-lightning:thinking": { - "id": "nvidia/nemotron-3.5-lightning:thinking", - "name": "Nvidia Nemotron 3.5 Lightning Thinking", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron", - "attachment": false, + "claude-sonnet-4-20250514": { + "id": "claude-sonnet-4-20250514", + "name": "claude-sonnet-4-20250514", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.01 + "input": 3, + "output": 15 } }, - "xiaomi/mimo-v2.5-pro:thinking": { - "id": "xiaomi/mimo-v2.5-pro:thinking", - "name": "MiMo V2.5 Pro Thinking", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "gpt-5.1-chat-latest": { + "id": "gpt-5.1-chat-latest", + "name": "gpt-5.1-chat-latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "high" + "medium" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036, - "cache_write": 0 + "input": 1.25, + "output": 10 } }, - "xiaomi/mimo-v2.6-pro": { - "id": "xiaomi/mimo-v2.6-pro", - "name": "MiMo V2.6 Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "gpt-5.2-chat-latest": { + "id": "gpt-5.2-chat-latest", + "name": "gpt-5.2-chat-latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "high" + "medium" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-12", + "last_updated": "2025-12-12", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036, - "cache_write": 0 + "input": 1.75, + "output": 14 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "grok-4-fast-reasoning": { + "id": "grok-4-fast-reasoning", + "name": "grok-4-fast-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-06", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028, - "cache_write": 0 + "input": 0.2, + "output": 0.5 } }, - "xiaomi/mimo-v2.6-pro-ultraspeed": { - "id": "xiaomi/mimo-v2.6-pro-ultraspeed", - "name": "MiMo V2.6 Pro UltraSpeed", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", - "family": "mimo", + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "gpt-4.1-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 4.35, - "output": 8.7, - "cache_read": 0.036, - "cache_write": 0 + "input": 0.1, + "output": 0.4 } }, - "xiaomi/mimo-v2.5:thinking": { - "id": "xiaomi/mimo-v2.5:thinking", - "name": "MiMo V2.5 Thinking", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "claude-sonnet-4-5-20250929-thinking": { + "id": "claude-sonnet-4-5-20250929-thinking", + "name": "claude-sonnet-4-5-20250929-thinking", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-03", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028, - "cache_write": 0 + "input": 3, + "output": 15 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo V2.5 Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "MiniMax-M2.7-highspeed": { + "id": "MiniMax-M2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-03-19", + "last_updated": "2026-03-19", "modalities": { "input": [ "text" @@ -30788,107 +28660,88 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, + "context": 204800, "output": 131072 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036, - "cache_write": 0 + "input": 0.6, + "output": 4.8 } }, - "xiaomi/mimo-v2.6-flash": { - "id": "xiaomi/mimo-v2.6-flash", - "name": "MiMo V2.6 Flash", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "MiniMax-M2": { + "id": "MiniMax-M2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "release_date": "2025-10-26", + "last_updated": "2025-10-26", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "input": 1048576, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028, - "cache_write": 0 + "input": 0.33, + "output": 1.32 } }, - "undi95/remm-slerp-l2-13b": { - "id": "undi95/remm-slerp-l2-13b", - "name": "ReMM SLERP 13B", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", + "gemini-3.1-flash-image-preview": { + "id": "gemini-3.1-flash-image-preview", + "name": "gemini-3.1-flash-image-preview", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-01-01", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-27", + "last_updated": "2026-02-27", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 6144, - "input": 6144, - "output": 4096 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.799, - "output": 1.207, - "cache_read": 0.3995 + "input": 0.5, + "output": 60 } }, - "Salesforce/Llama-xLAM-2-70b-fc-r": { - "id": "Salesforce/Llama-xLAM-2-70b-fc-r", - "name": "Llama-xLAM-2 70B fc-r", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "qwen3-235b-a22b": { + "id": "qwen3-235b-a22b", + "name": "Qwen3-235B-A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-13", - "last_updated": "2025-04-13", + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -30897,34 +28750,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "input": 128000, "output": 16384 }, "cost": { - "input": 2.5, - "output": 2.5, - "cache_read": 1.25 + "input": 0.29, + "output": 2.86 } }, - "minimax/minimax-latest": { - "id": "minimax/minimax-latest", - "name": "MiniMax Latest", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + "ministral-14b-2512": { + "id": "ministral-14b-2512", + "name": "ministral-14b-2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ "text", @@ -30934,34 +28780,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512000, - "input": 512000, - "output": 80000 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.33, + "output": 0.33 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, + "glm-4.6v": { + "id": "glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -30969,62 +28819,85 @@ }, "open_weights": true, "limit": { - "context": 204800, - "input": 204800, - "output": 131072 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.15 + "input": 0.145, + "output": 0.43 } }, - "minimax/minimax-m2.7-turbo": { - "id": "minimax/minimax-m2.7-turbo", - "name": "MiniMax M2.7 Turbo", - "description": "Efficient MiniMax model for quick assistance, coding, and routine automation", - "family": "minimax-m2.7", - "attachment": false, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "claude-haiku-4-5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-16", + "last_updated": "2025-10-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "input": 204800, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.3 + "input": 1, + "output": 5 } }, - "minimax/minimax-01": { - "id": "minimax/minimax-01", - "name": "MiniMax 01", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + "gpt-5.4": { + "id": "gpt-5.4", + "name": "gpt-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2025-01-15", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ @@ -31033,31 +28906,46 @@ }, "open_weights": false, "limit": { - "context": 1000192, - "input": 1000192, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.1394, - "output": 1.122, - "cache_read": 0.0697 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 0, + "tiers": [ + { + "input": 5, + "output": 22.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5 + } } }, - "minimax/minimax-m2-her": { - "id": "minimax/minimax-m2-her", - "name": "MiniMax M2-her", - "description": "MiniMax M2 variant tuned for conversational and character-driven agent interactions", - "family": "minimax", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "doubao-seed-1-6-thinking-250715": { + "id": "doubao-seed-1-6-thinking-250715", + "name": "doubao-seed-1-6-thinking-250715", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2026-01-23", - "last_updated": "2026-01-23", + "release_date": "2025-07-15", + "last_updated": "2025-07-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -31065,95 +28953,102 @@ }, "open_weights": false, "limit": { - "context": 65532, - "input": 65532, - "output": 2048 + "context": 256000, + "output": 16000 }, "cost": { - "input": 0.302, - "output": 1.207, - "cache_read": 0.151 + "input": 0.121, + "output": 1.21 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "gpt-5.4-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-19", + "last_updated": "2026-03-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "input": 204800, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.315, - "output": 1.26, - "cache_read": 0.1575 + "input": 0.75, + "output": 4.5 } }, - "minimax/minimax-m3:thinking": { - "id": "minimax/minimax-m3:thinking", - "name": "MiniMax M3 Thinking", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "gpt-4.1": { + "id": "gpt-4.1", + "name": "gpt-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512000, - "input": 512000, - "output": 80000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 2, + "output": 8 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "doubao-seed-code-preview-251028": { + "id": "doubao-seed-code-preview-251028", + "name": "doubao-seed-code-preview-251028", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2025-11-11", + "last_updated": "2025-11-11", "modalities": { "input": [ "text", @@ -31163,64 +29058,78 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512000, - "input": 512000, - "output": 80000 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.17, + "output": 1.14 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "claude-opus-4-6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2025-05-31", + "release_date": "2026-02-06", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.33, - "output": 1.32, - "cache_read": 0.165 + "input": 5, + "output": 25 } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", + "qwen3-coder-480b-a35b-instruct": { + "id": "qwen3-coder-480b-a35b-instruct", + "name": "qwen3-coder-480b-a35b-instruct", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -31229,60 +29138,71 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.17, - "output": 1.53, - "cache_read": 0.085 + "input": 0.86, + "output": 3.43 } }, - "inflatebot/MN-12B-Mag-Mell-R1": { - "id": "inflatebot/MN-12B-Mag-Mell-R1", - "name": "Mag Mell R1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-07-01", + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "claude-sonnet-4-5-20250929", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 3, + "output": 15 } }, - "soob3123/GrayLine-Qwen3-8B": { - "id": "soob3123/GrayLine-Qwen3-8B", - "name": "Grayline Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "deepseek-reasoner": { + "id": "deepseek-reasoner", + "name": "Deepseek-Reasoner", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -31291,105 +29211,93 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 0.3, - "cache_read": 0.15 + "input": 0.29, + "output": 0.43 } }, - "soob3123/Veiled-Calla-12B": { - "id": "soob3123/Veiled-Calla-12B", - "name": "Veiled Calla 12B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-13", - "last_updated": "2025-04-13", + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "grok-4-1-fast-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.3, - "output": 0.3, - "cache_read": 0.15 + "input": 0.2, + "output": 0.5 } }, - "soob3123/amoral-gemma3-27B-v2": { - "id": "soob3123/amoral-gemma3-27B-v2", - "name": "Amoral Gemma3 27B v2", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, + "gemini-3-pro-preview": { + "id": "gemini-3-pro-preview", + "name": "gemini-3-pro-preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-05-23", - "last_updated": "2025-05-23", + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 8192 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 0.3, - "cache_read": 0.15 + "input": 2, + "output": 12 } }, - "stepfun/step-5-preview": { - "id": "stepfun/step-5-preview", - "name": "Step 5 Preview", - "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", + "gpt-5-thinking": { + "id": "gpt-5-thinking", + "name": "gpt-5-thinking", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-09-16", - "last_updated": "2026-09-20", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -31397,27 +29305,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 1, - "output": 2.7, - "cache_read": 0.05 + "input": 1.25, + "output": 10 } }, - "stepfun/step-3.7-flash:thinking": { - "id": "stepfun/step-3.7-flash:thinking", - "name": "Step 3.7 Flash Thinking", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "gpt-5-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", "high" @@ -31426,136 +29333,180 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 256000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "input": 0.25, + "output": 2 } }, - "Sao10K/L3.1-70B-Hanami-x1": { - "id": "Sao10K/L3.1-70B-Hanami-x1", - "name": "Llama 3.1 70B Hanami", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "gpt-4.1-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-07-23", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.4, + "output": 1.6 } }, - "Sao10K/L3.1-70B-Euryale-v2.2": { - "id": "Sao10K/L3.1-70B-Euryale-v2.2", - "name": "Llama 3.1 70B Euryale", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-07-23", + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 20480, - "input": 20480, - "output": 16384 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.306, - "output": 0.357, - "cache_read": 0.153 + "input": 0.72, + "output": 3.2 } }, - "Sao10K/L3.3-70B-Euryale-v2.3": { - "id": "Sao10K/L3.3-70B-Euryale-v2.3", - "name": "Llama 3.3 70B Euryale", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "gpt-5.4-pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-12-06", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 20480, - "input": 20480, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 30, + "output": 180, + "cache_read": 0, + "cache_write": 0, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "Sao10K/L3-8B-Stheno-v3.2": { - "id": "Sao10K/L3-8B-Stheno-v3.2", - "name": "Sao10K Stheno 8b", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "glm-4.5-air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-11-29", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ "text" @@ -31566,41 +29517,41 @@ }, "open_weights": true, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.2006, - "output": 0.2006, - "cache_read": 0.1003 + "input": 0.1143, + "output": 0.286 } }, - "inception/mercury-2.5-preview": { - "id": "inception/mercury-2.5-preview", - "name": "Mercury 2.5 Preview", - "description": "Mercury 2.5 Preview is Inception's latest and most intelligent diffusion language model. Instead of generating tokens strictly one at a time, it produces and refines multiple tokens in parallel, reaching up to 1,107 tokens per second on standard GPUs. It delivers a 10+ point intelligence gain over Mercury 2, with tunable reasoning, parallel tool calls, schema-aligned JSON output, and a 260K context window. It is built for latency-sensitive production work such as search agents, voice pipelines, customer support, rapid coding iteration, and coding subagents.", - "family": "mercury", - "attachment": false, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "claude-sonnet-4-6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-18", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -31608,28 +29559,25 @@ }, "open_weights": false, "limit": { - "context": 260000, - "input": 260000, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.04, - "output": 0.15, - "cache_read": 0.004 + "input": 3, + "output": 15 } }, - "ornith-ai/ornith-1.5-35b-a3b": { - "id": "ornith-ai/ornith-1.5-35b-a3b", - "name": "Ornith 1.5 35B", - "description": "Ornith 1.5 35B A3B is an open-weight mixture-of-experts model for agentic coding, tool use, image understanding, and long-context work. This variant disables thinking for faster direct responses.", - "family": "ornith", + "grok-4-fast-non-reasoning": { + "id": "grok-4-fast-non-reasoning", + "name": "grok-4-fast-non-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-07-29", - "last_updated": "2026-08-20", + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", @@ -31639,30 +29587,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.05 + "input": 0.2, + "output": 0.5 } }, - "ornith-ai/ornith-1.5-35b-a3b:thinking": { - "id": "ornith-ai/ornith-1.5-35b-a3b:thinking", - "name": "Ornith 1.5 35B Thinking", - "description": "Ornith 1.5 35B A3B is an open-weight mixture-of-experts model for agentic coding, reasoning, tool use, image understanding, and long-context work. This variant enables thinking by default.", - "family": "ornith", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "gemini-3-flash-preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-07-29", - "last_updated": "2026-08-20", + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ "text", @@ -31672,29 +29617,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.05 + "input": 0.5, + "output": 3 } }, - "amazon/nova-lite-v1": { - "id": "amazon/nova-lite-v1", - "name": "Amazon Nova Lite 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", + "deepseek-chat": { + "id": "deepseek-chat", + "name": "Deepseek-Chat", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-11-29", + "last_updated": "2024-11-29", "modalities": { "input": [ "text" @@ -31705,27 +29649,25 @@ }, "open_weights": false, "limit": { - "context": 300000, - "input": 300000, - "output": 5120 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.0595, - "output": 0.238, - "cache_read": 0.02975 + "input": 0.29, + "output": 0.43 } }, - "amazon/nova-2-lite-v1": { - "id": "amazon/nova-2-lite-v1", - "name": "Amazon Nova 2 Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova", + "MiniMax-M1": { + "id": "MiniMax-M1", + "name": "MiniMax-M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-12-03", + "tool_call": true, + "temperature": true, + "release_date": "2025-06-16", + "last_updated": "2025-06-16", "modalities": { "input": [ "text" @@ -31737,29 +29679,38 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 1000000, - "output": 65535 + "output": 128000 }, "cost": { - "input": 0.51, - "output": 4.25, - "cache_read": 0.255 + "input": 0.132, + "output": 1.254 } }, - "amazon/nova-pro-v1": { - "id": "amazon/nova-pro-v1", - "name": "Amazon Nova Pro 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "grok-4.20-multi-agent-beta-0309": { + "id": "grok-4.20-multi-agent-beta-0309", + "name": "grok-4.20-multi-agent-beta-0309", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -31767,27 +29718,33 @@ }, "open_weights": false, "limit": { - "context": 300000, - "input": 300000, - "output": 32000 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.799, - "output": 3.196, - "cache_read": 0.3995 + "input": 2, + "output": 6 } }, - "Doctor-Shotgun/MS3.2-24B-Magnum-Diamond": { - "id": "Doctor-Shotgun/MS3.2-24B-Magnum-Diamond", - "name": "MS3.2 24B Magnum Diamond", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", + "glm-5": { + "id": "glm-5", + "name": "glm-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -31798,26 +29755,26 @@ }, "open_weights": true, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 0.6, + "output": 2.6 } }, - "aion-labs/aion-2.0": { - "id": "aion-labs/aion-2.0", - "name": "AionLabs: Aion-2.0", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "qwen-max-latest": { + "id": "qwen-max-latest", + "name": "Qwen-Max-Latest", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", "modalities": { "input": [ "text" @@ -31829,29 +29786,28 @@ "open_weights": false, "limit": { "context": 131072, - "input": 131072, - "output": 32768 + "output": 8192 }, "cost": { - "input": 0.8, - "output": 1.6, - "cache_read": 0.2 + "input": 0.343, + "output": 1.372 } }, - "aion-labs/aion-rp-llama-3.1-8b": { - "id": "aion-labs/aion-rp-llama-3.1-8b", - "name": "Llama 3.1 8b (uncensored)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "mistral-large-2512": { + "id": "mistral-large-2512", + "name": "mistral-large-2512", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-07-23", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -31859,27 +29815,24 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 16384 + "context": 128000, + "output": 262144 }, "cost": { - "input": 0.8, - "output": 1.6, - "cache_read": 0.4 + "input": 1.1, + "output": 3.3 } }, - "aion-labs/aion-3.0": { - "id": "aion-labs/aion-3.0", - "name": "AionLabs: Aion 3.0", - "description": "Aion 3.0 is a GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.", + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-07-07", - "last_updated": "2026-07-07", + "temperature": true, + "release_date": "2026-03-19", + "last_updated": "2026-03-19", "modalities": { "input": [ "text" @@ -31890,30 +29843,31 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 3, - "output": 6, - "cache_read": 0.75 + "input": 0.3, + "output": 1.2 } }, - "aion-labs/aion-3.5": { - "id": "aion-labs/aion-3.5", - "name": "AionLabs: Aion 3.5", - "description": "A GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "claude-3-5-haiku-latest": { + "id": "claude-3-5-haiku-latest", + "name": "claude-3-5-haiku-latest", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "temperature": true, + "knowledge": "2024-07-31", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -31921,27 +29875,26 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 200000, + "output": 8192 }, "cost": { - "input": 3, - "output": 6, - "cache_read": 0.75 + "input": 0.8, + "output": 4 } }, - "aion-labs/aion-3.5-mini": { - "id": "aion-labs/aion-3.5-mini", - "name": "AionLabs: Aion 3.5 Mini", - "description": "A GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.", + "qwen3-30b-a3b": { + "id": "qwen3-30b-a3b", + "name": "Qwen3-30B-A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" @@ -31952,27 +29905,25 @@ }, "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.7, - "output": 1.4, - "cache_read": 0.18 + "input": 0.11, + "output": 1.08 } }, - "aion-labs/aion-3.0-mini": { - "id": "aion-labs/aion-3.0-mini", - "name": "AionLabs: Aion 3.0 Mini", - "description": "Aion 3.0 Mini is a DeepSeek-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.", + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "deepseek-v3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-07-07", - "last_updated": "2026-07-07", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -31983,33 +29934,28 @@ }, "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.7, - "output": 1.4, - "cache_read": 0.18 + "input": 0.29, + "output": 0.43 } }, - "qwen/qwen3.5-flash": { - "id": "qwen/qwen3.5-flash", - "name": "Qwen3.5 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "doubao-seed-1-8-251215": { + "id": "doubao-seed-1-8-251215", + "name": "doubao-seed-1-8-251215", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -32017,45 +29963,30 @@ }, "open_weights": false, "limit": { - "context": 991808, - "input": 991808, - "output": 65536 + "context": 224000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.05 + "input": 0.114, + "output": 0.286 } }, - "qwen/qwen3.8-max:thinking": { - "id": "qwen/qwen3.8-max:thinking", - "name": "Qwen3.8 Max Thinking", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "claude-opus-4-5-20251101-thinking": { + "id": "claude-opus-4-5-20251101-thinking", + "name": "claude-opus-4-5-20251101-thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "knowledge": "2025-03", + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" @@ -32063,30 +29994,26 @@ }, "open_weights": false, "limit": { - "context": 991000, - "input": 991000, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 + "input": 5, + "output": 25 } }, - "qwen/qwen3-vl-235b-a22b-instruct": { - "id": "qwen/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.0-flash-lite": { + "id": "gemini-2.0-flash-lite", + "name": "gemini-2.0-flash-lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2024-11", + "release_date": "2025-06-16", + "last_updated": "2025-06-16", "modalities": { "input": [ "text", @@ -32096,39 +30023,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 2000000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.9, - "cache_read": 0.15 + "input": 0.075, + "output": 0.3 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "glm-5-turbo": { + "id": "glm-5-turbo", + "name": "glm-5-turbo", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text" @@ -32139,137 +30063,173 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 65536 + "context": 200000, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 0.72, + "output": 3.2 } }, - "qwen/qwen3.8-27b": { - "id": "qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", + "gpt-5.1": { + "id": "gpt-5.1", + "name": "gpt-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.7, - "cache_read": 0.04 + "input": 1.25, + "output": 10 } - }, - "qwen/qwen3-vl-235b-a22b-thinking": { - "id": "qwen/qwen3-vl-235b-a22b-thinking", - "name": "Qwen3 VL 235B A22B Thinking", - "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", - "family": "qwen", + } + } + }, + "trustedrouter": { + "id": "trustedrouter", + "env": [ + "TRUSTEDROUTER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.trustedrouter.com/v1", + "name": "TrustedRouter", + "doc": "https://trustedrouter.com/docs", + "models": { + "zdr": { + "id": "zdr", + "name": "Zero Data Retention", + "description": "TrustedRouter privacy routing alias that prefers zero data retention model endpoints.", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-06-01", + "last_updated": "2026-06-27", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 6, - "cache_read": 0.25 + "context": 1000000, + "output": 131072 } }, - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "e2e": { + "id": "e2e", + "name": "End-to-End Encrypted", + "description": "TrustedRouter privacy routing alias for end-to-end encrypted provider routes where available.", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-06-01", + "last_updated": "2026-06-27", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 260096, - "input": 260096, - "output": 65536 - }, - "cost": { - "input": 0.27, - "output": 2.16, - "cache_read": 0.135 + "context": 1000000, + "output": 131072 } }, - "qwen/qwen3.5-omni-plus": { - "id": "qwen/qwen3.5-omni-plus", - "name": "Qwen3.5 Omni Plus", - "description": "Qwen3.5 Omni Plus is Qwen's stronger general multimodal model. We verified live support for text prompts, images, audio files, and direct video URLs on Alibaba's chat-completions-compatible API. Alibaba describes Plus as a comprehensive evolution of Qwen3 Omni with support for over 10 hours of audio input.", - "family": "qwen3.5", + "synth-code": { + "id": "synth-code", + "name": "Synth Code", + "description": "TrustedRouter code synthesis orchestration alias that combines multiple model responses into one answer.", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-20", + "last_updated": "2026-06-27", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -32277,16 +30237,14 @@ }, "open_weights": false, "limit": { - "context": 983616, - "input": 983616, - "output": 65536 + "context": 1000000, + "output": 131072 } }, - "qwen/qwen3.7-flash:thinking": { - "id": "qwen/qwen3.7-flash:thinking", - "name": "Qwen3.7 Flash Thinking", - "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", - "family": "qwen", + "fast": { + "id": "fast", + "name": "Fast", + "description": "TrustedRouter speed routing alias that prefers low-latency healthy model endpoints.", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -32294,6 +30252,8 @@ "type": "effort", "values": [ "none", + "low", + "medium", "high" ] } @@ -32301,13 +30261,13 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "release_date": "2026-06-01", + "last_updated": "2026-06-27", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -32315,65 +30275,75 @@ }, "open_weights": false, "limit": { - "context": 983616, - "input": 983616, - "output": 65536 - }, - "cost": { - "input": 0.03, - "output": 0.13, - "cache_read": 0.006, - "cache_write": 0.038 + "context": 1000000, + "output": 131072 } }, - "qwen/qwen3-235b-a22b-instruct-2507": { - "id": "qwen/qwen3-235b-a22b-instruct-2507", - "name": "Qwen 3 235b A22B 2507", - "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "synth": { + "id": "synth", + "name": "Synth", + "description": "TrustedRouter synthesis orchestration alias that combines multiple model responses into one answer.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", + "release_date": "2026-06-20", + "last_updated": "2026-06-27", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 - }, - "cost": { - "input": 0.13, - "output": 0.5, - "cache_read": 0.065 + "context": 1000000, + "output": 131072 } }, - "qwen/qwen-max": { - "id": "qwen/qwen-max", - "name": "Qwen 2.5 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "auto": { + "id": "auto", + "name": "Auto", + "description": "TrustedRouter automatic routing alias that chooses a healthy supported model endpoint for the request.", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", + "release_date": "2026-05-01", + "last_updated": "2026-06-27", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -32381,66 +30351,36 @@ }, "open_weights": false, "limit": { - "context": 32000, - "input": 32000, - "output": 8192 - }, - "cost": { - "input": 1.5997, - "output": 6.392, - "cache_read": 0.79985 + "context": 1000000, + "output": 131072 } }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B (Thinking)", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 256000, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.65, - "cache_read": 0.075 - } - }, - "qwen/qwen3.8-max": { - "id": "qwen/qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "cheap": { + "id": "cheap", + "name": "Cheap", + "description": "TrustedRouter low-cost routing alias that prefers inexpensive healthy model endpoints.", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "release_date": "2026-05-01", + "last_updated": "2026-06-27", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ @@ -32449,30 +30389,42 @@ }, "open_weights": false, "limit": { - "context": 991000, - "input": 991000, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 + "context": 1000000, + "output": 131072 } - }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", - "family": "qwen", + } + } + }, + "zhipuai": { + "id": "zhipuai", + "env": [ + "ZHIPU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://open.bigmodel.cn/api/paas/v4", + "name": "Zhipu AI", + "doc": "https://docs.z.ai/guides/overview/pricing", + "models": { + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ "text" @@ -32481,29 +30433,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 65536 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 1.5, - "cache_read": 0.1 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "qwen/qwen3-coder": { - "id": "qwen/qwen3-coder", - "name": "Qwen 3 Coder 480B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -32512,71 +30477,43 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "input": 262000, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.13, - "output": 0.5, - "cache_read": 0.065 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "qwen/qwen3.5-35b-a3b:thinking": { - "id": "qwen/qwen3.5-35b-a3b:thinking", - "name": "Qwen3.5 35B A3B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 260096, - "input": 260096, - "output": 65536 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.225, - "output": 1.8, - "cache_read": 0.1125 - } - }, - "qwen/qwen3.5-plus": { - "id": "qwen/qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", "image", - "video" + "video", + "pdf" ], "output": [ "text" @@ -32584,33 +30521,38 @@ }, "open_weights": false, "limit": { - "context": 983616, - "input": 983616, - "output": 65536 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.04 + "input": 5, + "output": 22, + "cache_read": 1.2, + "cache_write": 0 } }, - "qwen/qwen3-32b": { - "id": "qwen/qwen3-32b", - "name": "Qwen 3 32b", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": true, - "reasoning": false, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -32618,30 +30560,33 @@ }, "open_weights": true, "limit": { - "context": 41000, - "input": 41000, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.05 + "input": 1, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0 } }, - "qwen/qwen-plus": { - "id": "qwen/qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "glm-4.5-flash": { + "id": "glm-4.5-flash", + "name": "GLM-4.5-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -32650,75 +30595,72 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 995904, - "input": 995904, - "output": 32768 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.3995, - "output": 1.2002, - "cache_read": 0.19975 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "qwen/qwen3.7-flash": { - "id": "qwen/qwen3.7-flash", - "name": "Qwen3.7 Flash", - "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", - "family": "qwen", - "attachment": true, + "glm-4.7-flash": { + "id": "glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 991808, - "input": 991808, - "output": 65536 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.03, - "output": 0.13, - "cache_read": 0.006, - "cache_write": 0.038 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -32727,34 +30669,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 131072, + "output": 98304 }, "cost": { - "input": 1.2002, - "output": 6.001, - "cache_read": 0.6001 + "input": 0.2, + "output": 1.1, + "cache_read": 0.03, + "cache_write": 0 } }, - "qwen/qwen3.8-27b-obliterated:thinking": { - "id": "qwen/qwen3.8-27b-obliterated:thinking", - "name": "Qwen 3.8 27B Obliterated Thinking", - "description": "Qwen 3.8 27B Obliterated with thinking enabled for more deliberate creative work, coding, multimodal analysis, tool use, and long-context problem solving.", - "family": "qwen", + "glm-4.6v": { + "id": "glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-07-29", - "last_updated": "2026-08-24", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -32762,29 +30710,31 @@ }, "open_weights": true, "limit": { - "context": 524288, - "input": 524288, + "context": 128000, "output": 32768 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.2 + "input": 0.3, + "output": 0.9 } }, - "qwen/qwen-turbo": { - "id": "qwen/qwen-turbo", - "name": "Qwen Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", + "glm-4.6": { + "id": "glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11-01", - "last_updated": "2025-04-28", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -32793,31 +30743,35 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 8192 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.04998, - "output": 0.2006, - "cache_read": 0.02499 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "qwen/qwen3-coder-30b-a3b-instruct": { - "id": "qwen/qwen3-coder-30b-a3b-instruct", - "name": "Qwen3 Coder 30B A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", + "glm-4.7-flashx": { + "id": "glm-4.7-flashx", + "name": "GLM-4.7-FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -32828,34 +30782,36 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.1, + "input": 0.07, "output": 0.4, - "cache_read": 0.05 + "cache_read": 0.01, + "cache_write": 0 } }, - "qwen/qwen3.8-27b:thinking": { - "id": "qwen/qwen3.8-27b:thinking", - "name": "Qwen3.8 27B Thinking", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, + "glm-4.5": { + "id": "glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -32863,32 +30819,38 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.15, - "output": 0.7, - "cache_read": 0.04 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "qwen/qwen3.8-27b-obliterated": { - "id": "qwen/qwen3.8-27b-obliterated", - "name": "Qwen 3.8 27B Obliterated", - "description": "Qwen 3.8 27B Obliterated is an open-weight multimodal model LoRA-tuned for fewer refusals across chat, coding, reasoning, tool use, and long-context work.", - "family": "qwen", + "glm-4.5v": { + "id": "glm-4.5v", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-07-29", - "last_updated": "2026-08-24", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -32896,75 +30858,82 @@ }, "open_weights": true, "limit": { - "context": 524288, - "input": 524288, - "output": 32768 + "context": 64000, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.2 + "input": 0.6, + "output": 1.8 } }, - "qwen/qwen3.8-omni-flash": { - "id": "qwen/qwen3.8-omni-flash", - "name": "Qwen3.8 Omni Flash", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": true, + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 991808, - "input": 991808, + "context": 204800, "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } - }, - "qwen/qwen3.8-27b-uncensored": { - "id": "qwen/qwen3.8-27b-uncensored", - "name": "Qwen 3.8 27B Uncensored", - "description": "Qwen 3.8 27B Uncensored is an NVFP4 open-weight multimodal model LoRA-tuned for fewer refusals across chat, coding, reasoning, tool use, and long-context work.", - "family": "qwen", - "attachment": true, + } + } + }, + "cortecs": { + "id": "cortecs", + "env": [ + "CORTECS_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.cortecs.ai/v1", + "name": "Cortecs", + "doc": "https://api.cortecs.ai/v1/models", + "models": { + "deepseek-r1-0528": { + "id": "deepseek-r1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-07-29", - "last_updated": "2026-08-21", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -32972,37 +30941,39 @@ }, "open_weights": true, "limit": { - "context": 524288, - "input": 524288, - "output": 32768 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.15, - "output": 1.2, - "cache_read": 0.125 + "input": 0.585, + "output": 2.307 } }, - "qwen/qwen3.7-max:thinking": { - "id": "qwen/qwen3.7-max:thinking", - "name": "Qwen3.7 Max Thinking", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -33011,36 +30982,35 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 65536 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 0.133, + "output": 0.266, + "cache_read": 0.0028 } }, - "qwen/qwen3-235b-a22b": { - "id": "qwen/qwen3-235b-a22b", - "name": "Qwen 3 235b A22B", - "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", - "family": "qwen", - "attachment": true, - "reasoning": false, + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -33048,28 +31018,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 196608, + "output": 196608 }, "cost": { - "input": 0.3, - "output": 0.5, - "cache_read": 0.15 + "input": 0.32, + "output": 1.18 } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "deepseek-v3-0324": { + "id": "deepseek-v3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2024-07", + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ "text" @@ -33080,66 +31048,41 @@ }, "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.437, - "output": 3.496, - "cache_read": 0.103788 + "input": 0.551, + "output": 1.654 } }, - "qwen/qwen3.8-27b-uncensored:thinking": { - "id": "qwen/qwen3.8-27b-uncensored:thinking", - "name": "Qwen 3.8 27B Uncensored Thinking", - "description": "Qwen 3.8 27B Uncensored with thinking enabled for more deliberate creative work, coding, multimodal analysis, tool use, and long-context problem solving.", - "family": "qwen", + "claude-opus4-7": { + "id": "claude-opus4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-07-29", - "last_updated": "2026-08-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "input": 524288, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 1.2, - "cache_read": 0.125 - } - }, - "qwen/qwen3.5-omni-flash": { - "id": "qwen/qwen3.5-omni-flash", - "name": "Qwen3.5 Omni Flash", - "description": "Qwen3.5 Omni Flash is Qwen's fast multimodal model. We verified live support for text prompts, images, audio files, and direct video URLs on Alibaba's chat-completions-compatible API. Alibaba describes Flash as a fully evolved version of Qwen3 Omni with audio input support across 60+ languages.", - "family": "qwen3.5", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -33147,28 +31090,35 @@ }, "open_weights": false, "limit": { - "context": 49152, - "input": 49152, - "output": 16384 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5.6, + "output": 27.99, + "cache_read": 0.56, + "cache_write": 6.99 } }, - "qwen/qwen3.6-35b-a3b": { - "id": "qwen/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": false, + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM 4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -33176,29 +31126,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 198000, + "output": 198000 }, "cost": { - "input": 0.112, - "output": 0.8, - "cache_read": 0.056 + "input": 0.45, + "output": 2.23 } }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B (Instruct)", + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -33209,27 +31157,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.15, - "output": 0.65, - "cache_read": 0.075 + "input": 0.062, + "output": 0.408 } }, - "qwen/qwen3-max-2026-01-23": { - "id": "qwen/qwen3-max-2026-01-23", - "name": "Qwen3 Max 2026-01-23", - "description": "Qwen3 Max is Alibaba's flagship Qwen 3 reasoning model with native tool use (web search, web extractor, code interpreter) and a 256K context window.", + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3 Coder 30B A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2026-01-26", - "last_updated": "2026-01-26", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-31", + "last_updated": "2025-07-31", "modalities": { "input": [ "text" @@ -33238,34 +31186,34 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 262000, + "output": 262000 }, "cost": { - "input": 1.2002, - "output": 6.001, - "cache_read": 0.6001 + "input": 0.053, + "output": 0.222 } }, - "qwen/qwen3.8-27b-fable": { - "id": "qwen/qwen3.8-27b-fable", - "name": "Qwen 3.8 27B Fable", - "description": "Qwen 3.8 27B Fable is an open-weight multimodal creative finetune for expressive dialogue, long-form storytelling, character work, and roleplay.", - "family": "qwen", - "attachment": true, + "minimax-m2.1": { + "id": "minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "release_date": "2026-07-29", - "last_updated": "2026-08-28", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -33273,32 +31221,29 @@ }, "open_weights": true, "limit": { - "context": 524288, - "input": 524288, - "output": 32768 + "context": 196000, + "output": 196000 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.125 + "input": 0.34, + "output": 1.34 } }, - "qwen/qwen3.8-27b-hemmingway": { - "id": "qwen/qwen3.8-27b-hemmingway", - "name": "Qwen 3.8 27B Hemingway", - "description": "Qwen 3.8 27B Hemingway is an open-weight NVFP4 multimodal creative finetune for long-form prose, character dialogue, storytelling, and roleplay.", + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -33306,33 +31251,45 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.125 + "input": 0.099, + "output": 0.33 } }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "claude-4-6-sonnet": { + "id": "claude-4-6-sonnet", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -33340,33 +31297,31 @@ }, "open_weights": false, "limit": { - "context": 991808, - "input": 991808, - "output": 65536 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.19, - "output": 1.16, - "cache_read": 0.02, - "cache_write": 0.24 + "input": 3.59, + "output": 17.92 } }, - "qwen/qwen3-coder-flash": { - "id": "qwen/qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "claude-sonnet-4": { + "id": "claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-03", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -33374,68 +31329,63 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 1.5, - "cache_read": 0.15 + "input": 3.307, + "output": 16.536 } }, - "qwen/qwen3.5-plus:thinking": { - "id": "qwen/qwen3.5-plus:thinking", - "name": "Qwen3.5 Plus Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "llama-4-maverick": { + "id": "llama-4-maverick", + "name": "Llama 4 Maverick 17B Instruct", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 983616, - "input": 983616, - "output": 65536 + "context": 1000000, + "output": 16384 }, "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.04 + "input": 0.124, + "output": 0.603, + "cache_read": 0.03, + "cache_write": 0.151 } }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -33443,73 +31393,72 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 65536 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.5 + "input": 1.654, + "output": 11.024 } }, - "qwen/qwen3.5-9b": { - "id": "qwen/qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "claude-4-5-sonnet": { + "id": "claude-4-5-sonnet", + "name": "Claude 4.5 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] }, { "type": "budget_tokens", - "min": 1024, - "max": 81920 + "min": 1024 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 65536 + "context": 200000, + "output": 200000 }, "cost": { - "input": 0.05, - "output": 0.15, - "cache_read": 0.025 + "input": 3.259, + "output": 16.296 } }, - "qwen/qwen3.5-122b-a10b:thinking": { - "id": "qwen/qwen3.5-122b-a10b:thinking", - "name": "Qwen3.5 122B A10B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "mixtral-8x7B-instruct-v0.1": { + "id": "mixtral-8x7B-instruct-v0.1", + "name": "Mixtral 8x7B Instruct v0.1", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2023-09", + "release_date": "2023-12-11", + "last_updated": "2023-12-11", "modalities": { "input": [ "text" @@ -33520,34 +31469,33 @@ }, "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 32768 + "context": 32000, + "output": 32000 }, "cost": { - "input": 0.437, - "output": 3.496, - "cache_read": 0.103788 + "input": 0.438, + "output": 0.68 } }, - "qwen/qwen3.5-397b-a17b:thinking": { - "id": "qwen/qwen3.5-397b-a17b:thinking", - "name": "Qwen3.5 397B A17B Thinking", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, + "glm-4.5": { + "id": "glm-4.5", + "name": "GLM 4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "knowledge": "2025-04", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -33555,75 +31503,60 @@ }, "open_weights": true, "limit": { - "context": 258048, - "input": 258048, - "output": 65536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 3.6, - "cache_read": 0.3 + "input": 0.67, + "output": 2.46 } }, - "qwen/qwen3.8-flash": { - "id": "qwen/qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "kimi-k2-instruct": { + "id": "kimi-k2-instruct", + "name": "Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-07-11", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 991808, - "input": 991808, - "output": 131072 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.14, - "output": 0.42, - "cache_read": 0.016, - "cache_write": 0.2 + "input": 0.551, + "output": 2.646 } }, - "qwen/qwen3.8-27b-cybersecurity": { - "id": "qwen/qwen3.8-27b-cybersecurity", - "name": "Qwen 3.8 27B Cybersecurity", - "description": "Qwen 3.8 27B Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports optional reasoning, image understanding, tool calling, and a 262,144-token context window.", - "family": "qwen", - "attachment": true, + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "release_date": "2026-09-19", - "last_updated": "2026-09-19", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -33631,30 +31564,40 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 32768 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.1, - "output": 0.6, - "cache_read": 0.05 + "input": 0.089, + "output": 0.275 } }, - "qwen/qwen-long": { - "id": "qwen/qwen-long", - "name": "Qwen Long 10M", - "description": "Alibaba's huge context window model. Takes in up to 10 million tokens, which is equivalent to dozens of books.", - "family": "qwen", + "claude-opus4-8": { + "id": "claude-opus4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2024-08-01", - "last_updated": "2024-08-01", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ @@ -33663,28 +31606,33 @@ }, "open_weights": false, "limit": { - "context": 10000000, - "input": 10000000, - "output": 8192 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.1003, - "output": 0.408, - "cache_read": 0.05015 + "input": 5.64, + "output": 28.198, + "cache_read": 0.563, + "cache_write": 7.049 } }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5 35B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", @@ -33697,29 +31645,31 @@ }, "open_weights": true, "limit": { - "context": 260096, - "input": 260096, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.225, - "output": 1.8, - "cache_read": 0.1125 + "input": 1.28, + "output": 4.63, + "cache_read": 0.32 } }, - "qwen/qwen3.6-max-preview": { - "id": "qwen/qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "release_date": "2026-04-14", + "last_updated": "2026-04-14", "modalities": { "input": [ "text" @@ -33728,30 +31678,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 245760, - "input": 245760, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.04, - "output": 6.24, - "cache_read": 0.52 + "input": 1.31, + "output": 4.1, + "cache_read": 0.24 } }, - "qwen/qwen3-30b-a3b": { - "id": "qwen/qwen3-30b-a3b", - "name": "Qwen3 30B A3B", - "description": "Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning", - "family": "qwen", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -33762,33 +31724,63 @@ }, "open_weights": true, "limit": { - "context": 41000, - "input": 41000, - "output": 32768 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.05 + "input": 1.553, + "output": 3.106, + "cache_read": 0.003625 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "qwen3-next-80b-a3b-thinking": { + "id": "qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B Thinking", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-11", + "last_updated": "2025-09-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.164, + "output": 1.311 + } + }, + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "knowledge": "2025-12", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -33796,27 +31788,38 @@ }, "open_weights": true, "limit": { - "context": 258048, - "input": 258048, - "output": 65536 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.6, - "output": 3.6, - "cache_read": 0.3 + "input": 0.656, + "output": 2.731 } }, - "qwen/qwen3-8b": { - "id": "qwen/qwen3-8b", - "name": "Qwen 3 8B", - "description": "Qwen 3 8B is a 8B model. Supports switching between thinking and non thinking: trigger thinking with /think and /no_think anywhere in a prompt or system message to toggle chain-of-thought reasoning.", - "family": "qwen", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -33827,29 +31830,27 @@ }, "open_weights": true, "limit": { - "context": 41000, - "input": 41000, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.47, - "output": 0.47, - "cache_read": 0.235 + "input": 1.44, + "output": 4.53, + "cache_read": 0.39 } }, - "qwen/qwen3.5-27b:thinking": { - "id": "qwen/qwen3.5-27b:thinking", - "name": "Qwen3.5 27B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", @@ -33862,32 +31863,34 @@ }, "open_weights": true, "limit": { - "context": 260096, - "input": 260096, - "output": 65536 + "context": 512000, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 2.16, - "cache_read": 0.135 + "input": 0.355, + "output": 1.775, + "cache_read": 0.089 } }, - "qwen/qwen3.8-27b-queen": { - "id": "qwen/qwen3.8-27b-queen", - "name": "Qwen 3.8 27B Queen", - "description": "Qwen 3.8 27B Queen is an open-weight roleplay finetune with image understanding, tool calling, optional reasoning, and a 262,144-token context window.", - "family": "qwen", - "attachment": true, + "minimax-m2": { + "id": "minimax-m2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "release_date": "2026-09-09", - "last_updated": "2026-09-09", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -33895,27 +31898,27 @@ }, "open_weights": true, "limit": { - "context": 524288, - "input": 524288, - "output": 32768 + "context": 400000, + "output": 400000 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.125 + "input": 0.39, + "output": 1.57 } }, - "qwen/qwen3-30b-a3b-instruct-2507": { - "id": "qwen/qwen3-30b-a3b-instruct-2507", - "name": "Qwen3 30B A3B Instruct 2507", - "description": "Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference. Significant improvements in general capabilities, including instruction following, logical reasoning, text comprehension, mathematics, science, coding and tool usage.", - "family": "qwen", + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-m2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-02-20", - "last_updated": "2025-02-20", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -33926,31 +31929,39 @@ }, "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 32768 + "context": 202752, + "output": 196072 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.1 + "input": 0.47, + "output": 1.4 } }, - "qwen/qwen2.5-coder-32b-instruct": { - "id": "qwen/qwen2.5-coder-32b-instruct", - "name": "Qwen 2.5 Coder 32b", - "description": "Open coding-focused Qwen model for code generation, repair, and repository reasoning", - "family": "qwen", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2024-11-12", - "last_updated": "2024-11-12", + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -33958,72 +31969,66 @@ }, "open_weights": true, "limit": { - "context": 32000, - "input": 32000, - "output": 8192 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.2006, - "output": 0.2006, - "cache_read": 0.1003 + "input": 0.55, + "output": 2.76 } }, - "qwen/qwen3.7-plus:thinking": { - "id": "qwen/qwen3.7-plus:thinking", - "name": "Qwen3.7 Plus Thinking", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT Oss 120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2024-01", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 983616, - "input": 983616, - "output": 65536 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "cache_write": 0.5 + "input": 0, + "output": 0 } }, - "qwen/qwen-2.5-72b-instruct": { - "id": "qwen/qwen-2.5-72b-instruct", - "name": "Qwen2.5 72B", + "qwen-2.5-72b-instruct": { + "id": "qwen-2.5-72b-instruct", + "name": "Qwen2.5 72B Instruct", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2025-07-03", - "last_updated": "2025-07-03", + "temperature": true, + "knowledge": "2024-06", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", "modalities": { "input": [ "text" @@ -34034,34 +32039,29 @@ }, "open_weights": true, "limit": { - "context": 131072, - "input": 131072, - "output": 8192 + "context": 33000, + "output": 33000 }, "cost": { - "input": 0.357, - "output": 0.408, - "cache_read": 0.1785 + "input": 0.062, + "output": 0.231 } }, - "qwen/qwen3.6-35b-a3b:thinking": { - "id": "qwen/qwen3.6-35b-a3b:thinking", - "name": "Qwen3.6 35B A3B Thinking", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, + "hermes-4-70b": { + "id": "hermes-4-70b", + "name": "Hermes 4 70B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2023-12", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -34069,68 +32069,88 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 262144, - "output": 16384 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.112, - "output": 0.8, - "cache_read": 0.056 + "input": 0.116, + "output": 0.358 } }, - "qwen/qwen3.6-27b": { - "id": "qwen/qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 260096, - "input": 260096, - "output": 65536 + "context": 200000, + "output": 200000 }, "cost": { - "input": 0.203, - "output": 2.24, - "cache_read": 0.1015 + "input": 1.09, + "output": 5.43 } }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen 3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -34138,35 +32158,38 @@ }, "open_weights": false, "limit": { - "context": 991808, - "input": 991808, - "output": 65536 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0.325, - "output": 1.95, - "cache_read": 0.0325, - "cache_write": 0.40625 + "input": 3, + "output": 16.13, + "cache_read": 0.25 } }, - "qwen/qwen3.6-27b:thinking": { - "id": "qwen/qwen3.6-27b:thinking", - "name": "Qwen3.6 27B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -34174,74 +32197,62 @@ }, "open_weights": true, "limit": { - "context": 260096, - "input": 260096, - "output": 65536 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.203, - "output": 2.24, - "cache_read": 0.1015 + "input": 0.81, + "output": 3.54, + "cache_read": 0.2 } }, - "qwen/qwen3-14b": { - "id": "qwen/qwen3-14b", - "name": "Qwen 3 14b", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT 4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 41000, - "input": 41000, + "context": 1047576, "output": 32768 }, "cost": { - "input": 0.08, - "output": 0.24, - "cache_read": 0.04 + "input": 2.354, + "output": 9.417 } }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "nova-pro-v1": { + "id": "nova-pro-v1", + "name": "Nova Pro 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2024-04", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -34249,81 +32260,75 @@ }, "open_weights": false, "limit": { - "context": 991808, - "input": 991808, - "output": 65536 + "context": 300000, + "output": 5000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "cache_write": 0.5 + "input": 1.016, + "output": 4.061 } }, - "qwen/qwen3.5-flash:thinking": { - "id": "qwen/qwen3.5-flash:thinking", - "name": "Qwen3.5 Flash Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "Qwen3 Coder Next 80B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-04", + "release_date": "2026-02-04", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 991808, - "input": 991808, + "context": 256000, "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.05 + "input": 0.158, + "output": 0.84 } }, - "qwen/qwen3.8-max-0902": { - "id": "qwen/qwen3.8-max-0902", - "name": "Qwen3.8 Max 0902", - "description": "2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding", - "family": "qwen", + "claude-opus4-5": { + "id": "claude-opus4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "max" + "high" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ @@ -34332,28 +32337,26 @@ }, "open_weights": false, "limit": { - "context": 991808, - "input": 991808, - "output": 131072 + "context": 200000, + "output": 200000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.17, - "cache_write": 2.5 + "input": 5.98, + "output": 29.89 } }, - "failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5": { - "id": "failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5", - "name": "Llama 3 70B abliterated", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "qwen3-coder-480b-a35b-instruct": { + "id": "qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ "text" @@ -34364,76 +32367,56 @@ }, "open_weights": true, "limit": { - "context": 8192, - "input": 8192, - "output": 8192 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.7, - "output": 0.7, - "cache_read": 0.35 + "input": 0.441, + "output": 1.984 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT 5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "devstral-2512": { + "id": "devstral-2512", + "name": "Devstral 2 2512", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 0, + "output": 0 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "knowledge": "2026-01", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text" @@ -34442,45 +32425,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16385, - "input": 16385, - "output": 4096 + "context": 250000, + "output": 250000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 0.6, + "output": 3.6 } }, - "openai/gpt-terra-latest": { - "id": "openai/gpt-terra-latest", - "name": "GPT Terra Latest", - "description": "Compatibility alias that routes to GPT 5.6 Terra, the latest supported GPT Terra model.", - "family": "gpt-terra", + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ @@ -34489,84 +32460,63 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 200000, + "output": 131072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 1.235, + "output": 4.118, + "cache_read": 0.308, + "cache_write": 1.544 } }, - "openai/gpt-6-astra-pro": { - "id": "openai/gpt-6-astra-pro", - "name": "GPT 6 Astra Pro", - "description": "GPT 6 Astra in Pro reasoning mode. Uses additional model work for difficult tasks, with higher latency and token usage at the same per-token rates. Reasoning effort remains independently configurable.", - "family": "gpt", - "attachment": true, + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "GLM 4.5 Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-01", + "last_updated": "2025-08-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.22, + "output": 1.34 } }, - "openai/o3-pro-2025-06-10": { - "id": "openai/o3-pro-2025-06-10", - "name": "OpenAI o3-pro (2025-06-10)", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", + "glm-4.7-flash": { + "id": "glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2025-06-10", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ "text" @@ -34575,168 +32525,121 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 203000, + "output": 203000 }, "cost": { - "input": 22, - "output": 88, - "cache_read": 11 + "input": 0.09, + "output": 0.53 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT 5.4 Nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "intellect-3": { + "id": "intellect-3", + "name": "INTELLECT 3", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "knowledge": "2025-11", + "release_date": "2025-11-26", + "last_updated": "2025-11-26", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, + "context": 128000, "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.219, + "output": 1.202 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT 5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "llama-3.1-405b-instruct": { + "id": "llama-3.1-405b-instruct", + "name": "Llama 3.1 405B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, + "context": 128000, "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0, + "output": 0 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT 5.1 Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", - "attachment": true, + "glm-5": { + "id": "glm-5", + "name": "GLM 5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 202752, + "output": 202752 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.08, + "output": 3.44 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "mistral-large-2512": { + "id": "mistral-large-2512", + "name": "Mistral Large 3 2512", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "knowledge": "2025-12", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text", @@ -34746,182 +32649,123 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 256000, + "output": 256000 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.5, + "output": 1.5, + "cache_read": 0.05 } }, - "openai/gpt-chat-latest": { - "id": "openai/gpt-chat-latest", - "name": "GPT Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-05-03", - "last_updated": "2026-05-03", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 163840, + "output": 163840 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.266, + "output": 0.444 } }, - "openai/gpt-5.6-sol-pro": { - "id": "openai/gpt-5.6-sol-pro", - "name": "GPT 5.6 Sol Pro", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "codestral-2508": { + "id": "codestral-2508", + "name": "Codestral 2508", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "mistral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.3, + "output": 0.9, + "cache_read": 0.03 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT 5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, + "qwen3.5-122b-a10b": { + "id": "qwen3.5-122b-a10b", + "name": "Qwen3.5 122B A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2026-01", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.444, + "output": 3.106 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "OpenAI o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "nemotron-3-super-120b-a12b": { + "id": "nemotron-3-super-120b-a12b", + "name": "Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-12", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text" @@ -34930,41 +32774,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.266, + "output": 0.799 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "OpenAI o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "glm-5-turbo": { + "id": "glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text" @@ -34976,20 +32808,20 @@ "open_weights": false, "limit": { "context": 200000, - "input": 200000, - "output": 100000 + "output": 131072 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 1.235, + "output": 4.118, + "cache_read": 0.308, + "cache_write": 1.544 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT 5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "claude-opus4-6": { + "id": "claude-opus4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -34998,17 +32830,19 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", + "knowledge": "2025-05-31", "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -35021,75 +32855,80 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 5.98, + "output": 29.89 } - }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT 4.1 Nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, + } + } + }, + "nebius": { + "id": "nebius", + "env": [ + "NEBIUS_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.tokenfactory.nebius.com/v1", + "name": "Nebius Token Factory", + "doc": "https://docs.tokenfactory.nebius.com/", + "models": { + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-08", + "release_date": "2025-12-05", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "input": 1047576, - "output": 32768 + "context": 128000, + "input": 120000, + "output": 8192 }, "cost": { - "input": 0.1, + "input": 0.13, "output": 0.4, - "cache_read": 0.025 + "cache_read": 0.013, + "cache_write": 0.16 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT 5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "moonshotai/Kimi-K2.5-fast": { + "id": "moonshotai/Kimi-K2.5-fast", + "name": "Kimi-K2.5-fast", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-12-15", + "last_updated": "2026-02-04", "modalities": { "input": [ "text", @@ -35099,162 +32938,146 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 256000, + "input": 256000, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 0.5, + "output": 2.5, + "cache_read": 0.05, + "cache_write": 0.625 } }, - "openai/o1": { - "id": "openai/o1", - "name": "OpenAI o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": false, + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi-K2.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-12-15", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 256000, + "input": 256000, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 0.5, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.05, + "cache_write": 0.625 } }, - "openai/gpt-latest": { - "id": "openai/gpt-latest", - "name": "GPT Latest", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma-3-27b-it", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-03-29", - "last_updated": "2026-03-29", + "temperature": true, + "knowledge": "2025-10", + "release_date": "2026-01-20", + "last_updated": "2026-02-04", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 110000, + "input": 100000, + "output": 8192 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01, + "cache_write": 0.125 } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT 5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", - "attachment": true, + "Qwen/Qwen3-Next-80B-A3B-Thinking-fast": { + "id": "Qwen/Qwen3-Next-80B-A3B-Thinking-fast", + "name": "Qwen3-Next-80B-A3B-Thinking-fast", + "description": "Legacy model retained for compatibility with older integrations", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-25", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 8000, + "input": 7000, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 15, - "output": 120, - "cache_read": 1.5 + "input": 0.15, + "output": 1.2, + "cache_read": 0.015, + "cache_write": 0.1875 } }, - "openai/gpt-4o-2024-08-06": { - "id": "openai/gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "Qwen/Qwen2.5-VL-72B-Instruct": { + "id": "Qwen/Qwen2.5-VL-72B-Instruct", + "name": "Qwen2.5-VL-72B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-08-06", - "last_updated": "2024-08-06", + "knowledge": "2024-12", + "release_date": "2025-01-20", + "last_updated": "2026-02-04", "modalities": { "input": [ "text", @@ -35264,264 +33087,211 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "input": 128000, - "output": 16384 + "input": 120000, + "output": 8192 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.25, + "output": 0.75, + "cache_read": 0.025, + "cache_write": 0.31 } }, - "openai/gpt-6-astra": { - "id": "openai/gpt-6-astra", - "name": "GPT 6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "Qwen/Qwen3-Embedding-8B": { + "id": "Qwen/Qwen3-Embedding-8B", + "name": "Qwen3-Embedding-8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "knowledge": "2025-10", + "release_date": "2026-01-10", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 32768, + "input": 32768, + "output": 0 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.01, + "output": 0 } }, - "openai/gpt-luna-latest": { - "id": "openai/gpt-luna-latest", - "name": "GPT Luna Latest", - "description": "Compatibility alias that routes to GPT 5.6 Luna, the latest supported GPT Luna model.", - "family": "gpt-luna", - "attachment": true, + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen3.5-397B-A17B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-15", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 262144, + "input": 250000, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 + "input": 0.6, + "output": 3.6, + "cache_read": 0.06, + "cache_write": 0.75 } }, - "openai/gpt-5.6-luna-pro": { - "id": "openai/gpt-5.6-luna-pro", - "name": "GPT 5.6 Luna Pro", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, + "Qwen/Qwen3.5-397B-A17B-fast": { + "id": "Qwen/Qwen3.5-397B-A17B-fast", + "name": "Qwen3.5-397B-A17B-fast", + "description": "Legacy model retained for compatibility with older integrations", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-15", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 8000, + "input": 7000, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 + "input": 0.6, + "output": 3.6, + "cache_read": 0.06, + "cache_write": 0.75 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT 5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "Qwen/Qwen3-30B-A3B-Instruct-2507": { + "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen3-30B-A3B-Instruct-2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-12", + "release_date": "2026-01-28", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 128000, + "input": 120000, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01, + "cache_write": 0.125 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "Qwen/Qwen3-Next-80B-A3B-Thinking": { + "id": "Qwen/Qwen3-Next-80B-A3B-Thinking", + "name": "Qwen3-Next-80B-A3B-Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2025-12", + "release_date": "2026-01-28", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "input": 128000, + "input": 120000, "output": 16384 }, "cost": { "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "output": 1.2, + "reasoning": 1.2, + "cache_read": 0.015, + "cache_write": 0.18 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-07", + "release_date": "2025-07-25", + "last_updated": "2025-10-04", "modalities": { "input": [ "text" @@ -35530,39 +33300,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 262144, + "output": 8192 }, "cost": { "input": 0.2, - "output": 0.3 + "output": 0.6 } }, - "openai/gpt-oss-safeguard-20b": { - "id": "openai/gpt-oss-safeguard-20b", - "name": "GPT OSS Safeguard 20B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen3-32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "knowledge": "2025-12", + "release_date": "2026-01-28", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -35574,19 +33333,55 @@ "open_weights": true, "limit": { "context": 128000, - "input": 128000, - "output": 16384 + "input": 120000, + "output": 8192 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01, + "cache_write": 0.125 } }, - "openai/o3-mini-high": { - "id": "openai/o3-mini-high", - "name": "OpenAI o3-mini (High)", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "Qwen/Qwen3-235B-A22B-Thinking-2507-fast": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507-fast", + "name": "Qwen3-235B-A22B-Thinking-2507-fast", + "description": "Legacy model retained for compatibility with older integrations", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-25", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8000, + "input": 7000, + "output": 8192 + }, + "status": "deprecated", + "cost": { + "input": 0.5, + "output": 2, + "cache_read": 0.05, + "cache_write": 0.625 + } + }, + "openai/gpt-oss-120b-fast": { + "id": "openai/gpt-oss-120b-fast", + "name": "gpt-oss-120b-fast", + "description": "Legacy model retained for compatibility with older integrations", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -35601,8 +33396,10 @@ ], "tool_call": true, "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2025-01-31", + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-06-10", + "last_updated": "2026-05-07", "modalities": { "input": [ "text" @@ -35611,236 +33408,302 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 8000, + "input": 7000, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125 } }, - "openai/gpt-sol-latest": { - "id": "openai/gpt-sol-latest", - "name": "GPT Sol Latest", - "description": "Compatibility alias that routes to GPT 5.6 Sol, the latest supported GPT Sol model.", - "family": "gpt-sol", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "temperature": true, + "knowledge": "2025-09", + "release_date": "2026-01-10", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 128000, + "input": 124000, + "output": 8192 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.15, + "output": 0.6, + "reasoning": 0.6, + "cache_read": 0.015, + "cache_write": 0.18 } }, - "openai/gpt-6-sol-pro": { - "id": "openai/gpt-6-sol-pro", - "name": "GPT 6 Sol Pro", - "description": "GPT-6 Sol Pro uses the same underlying model as GPT-6 Sol with Pro reasoning mode enabled for higher-quality responses on complex tasks.", - "family": "gpt", - "attachment": true, + "NousResearch/Hermes-4-405B": { + "id": "NousResearch/Hermes-4-405B", + "name": "Hermes-4-405B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "knowledge": "2025-11", + "release_date": "2026-01-30", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 128000, + "input": 120000, + "output": 8192 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 1, + "output": 3, + "reasoning": 3, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "openai/gpt-astra-latest": { - "id": "openai/gpt-astra-latest", - "name": "GPT Astra Latest", - "description": "Compatibility alias that routes to GPT 6 Astra, the latest supported GPT Astra model.", - "family": "gpt-astra", - "attachment": true, + "NousResearch/Hermes-4-70B": { + "id": "NousResearch/Hermes-4-70B", + "name": "Hermes-4-70B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "temperature": true, + "knowledge": "2025-11", + "release_date": "2026-01-30", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 128000, + "input": 120000, + "output": 8192 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.13, + "output": 0.4, + "reasoning": 0.4, + "cache_read": 0.013, + "cache_write": 0.16 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT 5.4 Mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, + "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B": { + "id": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", + "name": "Nemotron-3-Nano-30B-A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-08-10", + "last_updated": "2026-02-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32000, + "input": 30000, + "output": 4096 + }, + "cost": { + "input": 0.06, + "output": 0.24, + "cache_read": 0.006, + "cache_write": 0.075 + } + }, + "nvidia/Nemotron-3-Nano-Omni": { + "id": "nvidia/Nemotron-3-Nano-Omni", + "name": "Nemotron-3-Nano-Omni", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "knowledge": "2025-01", + "release_date": "2025-01-20", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 65536, + "input": 60000, + "output": 8192 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.06, + "output": 0.24, + "cache_read": 0.006, + "cache_write": 0.075 } }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT 5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, + "nvidia/Llama-3_1-Nemotron-Ultra-253B-v1": { + "id": "nvidia/Llama-3_1-Nemotron-Ultra-253B-v1", + "name": "Llama-3.1-Nemotron-Ultra-253B-v1", + "description": "Flagship Nemotron model for high-throughput reasoning and complex agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-15", + "last_updated": "2026-02-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "input": 120000, + "output": 4096 + }, + "cost": { + "input": 0.6, + "output": 1.8, + "cache_read": 0.06, + "cache_write": 0.75 + } + }, + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "Nemotron-3-Super-120B-A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2026-02", + "release_date": "2026-03-11", + "last_updated": "2026-03-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "input": 256000, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 0.9 + } + }, + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM-5", + "description": "Legacy model retained for compatibility with older integrations", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2026-01", + "release_date": "2026-03-01", + "last_updated": "2026-03-10", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -35848,249 +33711,238 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 200000, + "input": 200000, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 + "input": 1, + "output": 3.2, + "cache_read": 0.1, + "cache_write": 1 } }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT 5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 432000, + "output": 432000 }, "cost": { - "input": 2.5, - "output": 20, - "cache_read": 0.25 + "input": 1.4, + "output": 4.4 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT 5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, + "deepseek-ai/DeepSeek-V3.2-fast": { + "id": "deepseek-ai/DeepSeek-V3.2-fast", + "name": "DeepSeek-V3.2-fast", + "description": "Legacy model retained for compatibility with older integrations", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2025-01", + "release_date": "2025-01-27", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 8000, + "input": 7000, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.4, + "output": 2, + "cache_read": 0.04, + "cache_write": 0.5 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT 5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 1.75, + "output": 3.5, + "cache_read": 0.15 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT 4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "DeepSeek-V3.2", + "description": "Legacy model retained for compatibility with older integrations", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-11", + "release_date": "2026-01-20", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "input": 1047576, - "output": 32768 + "context": 163000, + "input": 160000, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.3, + "output": 0.45, + "reasoning": 0.45, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "openai/gpt-4o-2024-11-20": { - "id": "openai/gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", + "knowledge": "2025-01", + "release_date": "2025-01-20", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 196608, + "input": 190000, + "output": 8192 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "openai/o3-mini-low": { - "id": "openai/o3-mini-low", - "name": "OpenAI o3-mini (Low)", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "MiniMaxAI/MiniMax-M2.5-fast": { + "id": "MiniMaxAI/MiniMax-M2.5-fast", + "name": "MiniMax-M2.5-fast", + "description": "Legacy model retained for compatibility with older integrations", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2026-01-31", - "last_updated": "2025-01-31", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-20", + "last_updated": "2026-05-07", "modalities": { "input": [ "text" @@ -36099,64 +33951,80 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 8000, + "input": 7000, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT 4.1 Mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, + "PrimeIntellect/INTELLECT-3": { + "id": "PrimeIntellect/INTELLECT-3", + "name": "INTELLECT-3", + "description": "Legacy model retained for compatibility with older integrations", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-10", + "release_date": "2026-01-25", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "input": 1047576, - "output": 32768 + "context": 128000, + "input": 120000, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.2, + "output": 1.1, + "cache_read": 0.02, + "cache_write": 0.25 } - }, - "openai/gpt-5.6-terra-pro": { - "id": "openai/gpt-5.6-terra-pro", - "name": "GPT 5.6 Terra Pro", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, + } + } + }, + "auriko": { + "id": "auriko", + "env": [ + "AURIKO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.auriko.ai/v1", + "name": "Auriko", + "doc": "https://docs.auriko.ai", + "models": { + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -36166,46 +34034,44 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "openai/gpt-6-luna": { - "id": "openai/gpt-6-luna", - "name": "GPT 6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -36216,14 +34082,16 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -36232,82 +34100,56 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.05, - "output": 0.25, - "cache_read": 0.005, - "cache_write": 0.0625 - } - }, - "openai/o4-mini-high": { - "id": "openai/o4-mini-high", - "name": "OpenAI o4-mini high", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2025-12-04", - "last_updated": "2025-04-16", - "modalities": { - "input": [ - "text" + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "input": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "openai/gpt-6-luna-pro": { - "id": "openai/gpt-6-luna-pro", - "name": "GPT 6 Luna Pro", - "description": "GPT-6 Luna Pro uses the same underlying model as GPT-6 Luna with Pro reasoning mode enabled for higher-quality responses on complex tasks.", - "family": "gpt", + "grok-4.3": { + "id": "grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", @@ -36320,29 +34162,45 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.05, - "output": 0.25, - "cache_read": 0.005, - "cache_write": 0.0625 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT 5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -36353,14 +34211,16 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -36369,72 +34229,81 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 4096 + "context": 200000, + "output": 131072 }, "cost": { - "input": 10, - "output": 30 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -36445,42 +34314,47 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.35, - "output": 0.75 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT 5.1 Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -36488,39 +34362,79 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 400000, + "context": 1000000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } + }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/o3": { - "id": "openai/o3", - "name": "OpenAI o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "minimax-m2-7-highspeed": { + "id": "minimax-m2-7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_write": 0.375 + } + }, + "minimax-m2-7": { + "id": "minimax-m2-7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -36529,76 +34443,134 @@ "text" ] }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_write": 0.375 + } + }, + "qwen-3.6-plus": { + "id": "qwen-3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 1 + "input": 0.5, + "output": 3, + "cache_read": 0.1, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT 5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.5, + "output": 2.8 } }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT 5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -36608,46 +34580,46 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 1050000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "openai/gpt-5.1-2025-11-13": { - "id": "openai/gpt-5.1-2025-11-13", - "name": "GPT-5.1 (2025-11-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -36656,11 +34628,17 @@ ], "tool_call": true, "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2025-11-13", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -36668,42 +34646,57 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "openai/gpt-6-sol": { - "id": "openai/gpt-6-sol", - "name": "GPT 6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -36716,30 +34709,42 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 1050000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/o1-pro": { - "id": "openai/o1-pro", - "name": "OpenAI o1 Pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2025-03-19", - "last_updated": "2025-03-19", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -36752,90 +34757,100 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 100000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 150, - "output": 600, - "cache_read": 75 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } - }, - "microsoft/wizardlm-2-8x22b": { - "id": "microsoft/wizardlm-2-8x22b", - "name": "WizardLM-2 8x22B", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + } + } + }, + "stepfun-ai": { + "id": "stepfun-ai", + "env": [ + "STEPFUN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.stepfun.ai/step_plan/v1", + "name": "StepFun AI", + "doc": "https://platform.stepfun.ai/docs/en/step-plan/integrations/open-code", + "models": { + "step-2-16k": { + "id": "step-2-16k", + "name": "Step 2 (16K)", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-01-01", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "input": 65536, + "context": 16384, + "input": 16384, "output": 8192 }, "cost": { - "input": 0.493, - "output": 0.493, - "cache_read": 0.2465 + "input": 5.21, + "output": 16.44, + "cache_read": 1.04 } }, - "NousResearch/hermes-3-llama-3.1-70b": { - "id": "NousResearch/hermes-3-llama-3.1-70b", - "name": "Hermes 3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "nousresearch", + "step-tts-2": { + "id": "step-tts-2", + "name": "Step TTS 2", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "step", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": false, - "release_date": "2026-01-07", - "last_updated": "2026-01-07", + "tool_call": false, + "temperature": false, + "release_date": "2026-03-01", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "input": 65536, - "output": 8192 - }, - "cost": { - "input": 0.408, - "output": 0.408, - "cache_read": 0.204 + "context": 0, + "output": 0 } }, - "NousResearch/hermes-4-405b": { - "id": "NousResearch/hermes-4-405b", - "name": "Hermes 4 Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nousresearch", + "step-3.5-flash": { + "id": "step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-06-15", "modalities": { "input": [ "text" @@ -36846,94 +34861,70 @@ }, "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.15 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "NousResearch/hermes-4-405b:thinking": { - "id": "NousResearch/hermes-4-405b:thinking", - "name": "Hermes 4 Large (Thinking)", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nousresearch", + "stepaudio-2.5-asr": { + "id": "stepaudio-2.5-asr", + "name": "StepAudio 2.5 ASR", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "step", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", + "temperature": false, + "release_date": "2026-04-24", + "last_updated": "2026-07-02", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.15 + "context": 0, + "output": 0 } }, - "ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0": { - "id": "ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0", - "name": "Omega Directive 24B Unslop v2.0", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "stepaudio-2.5-tts": { + "id": "stepaudio-2.5-tts", + "name": "StepAudio 2.5 TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "step", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "temperature": false, + "release_date": "2026-04-16", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "input": 32768, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 0.5, - "cache_read": 0.25 + "context": 0, + "output": 0 } - } - } - }, - "abliteration-ai": { - "id": "abliteration-ai", - "env": [ - "ABLIT_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.abliteration.ai/v1", - "name": "abliteration.ai", - "doc": "https://docs.abliteration.ai/models", - "models": { - "abliterated-model-large-v2": { - "id": "abliterated-model-large-v2", - "name": "Abliterated Model Large V2", - "description": "GLM-5.3 model abliterated and finetuned for cyber, ML red teaming, and agent testing", + }, + "step-3.5-flash-2603": { + "id": "step-3.5-flash-2603", + "name": "Step 3.5 Flash 2603", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -36941,16 +34932,18 @@ "type": "effort", "values": [ "low", - "high", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-29", - "last_updated": "2026-08-31", + "knowledge": "2025-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" @@ -36959,46 +34952,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 999990 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 5, - "output": 5, - "cache_read": 0.5 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "abliterated-model": { - "id": "abliterated-model", - "name": "Abliterated Model", - "description": "Multimodal model for analyzing text, images, documents, and rich media", + "step-3.7-flash": { + "id": "step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] - }, - { - "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-01-06", - "last_updated": "2026-07-28", + "knowledge": "2026-01-01", + "release_date": "2026-05-29", + "last_updated": "2026-06-29", "modalities": { "input": [ "text", @@ -37010,40 +34999,28 @@ }, "open_weights": true, "limit": { - "context": 150000, - "input": 150000, - "output": 8192 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 3, - "output": 3, - "cache_read": 0.3 + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 } }, - "abliterated-model-large": { - "id": "abliterated-model-large", - "name": "Abliterated Model Large", - "description": "GLM-5.2 model abliterated and finetuned for cyber, ML red teaming, and agent testing", + "step-1-32k": { + "id": "step-1-32k", + "name": "Step 1 (32K)", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - }, - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-25", - "last_updated": "2026-07-28", + "knowledge": "2024-06", + "release_date": "2025-01-01", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -37054,34 +35031,34 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 999990 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 5, - "output": 5, - "cache_read": 0.5 + "input": 2.05, + "output": 9.59, + "cache_read": 0.41 } } } }, - "crof": { - "id": "crof", + "vivgrid": { + "id": "vivgrid", "env": [ - "CROF_API_KEY" + "VIVGRID_API_KEY" ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://crof.ai/v1", - "name": "CrofAI", - "doc": "https://crof.ai/docs", + "npm": "@ai-sdk/openai", + "api": "https://api.vivgrid.com/v1", + "name": "Vivgrid", + "doc": "https://docs.vivgrid.com/models", "models": { - "gemma-4-31b-it": { - "id": "gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -37095,14 +35072,17 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -37110,23 +35090,23 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 384000 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "qwen3.8-27b": { - "id": "qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -37142,46 +35122,46 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { "input": 0.2, - "output": 1.5, - "cache_read": 0.03 + "output": 1.25, + "cache_read": 0.02 } }, - "kimi-k3-eco": { - "id": "kimi-k3-eco", - "name": "Kimi K3 Eco", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", "high", "max" ] @@ -37192,14 +35172,12 @@ "field": "reasoning_content" }, "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -37208,122 +35186,121 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "output": 128000 }, "cost": { - "input": 1, - "output": 4, - "cache_read": 0.1 + "input": 1.2, + "output": 4.2, + "cache_read": 0.3 } }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM 5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.07, - "output": 0.22, - "cache_read": 0.01 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash (New)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + "gpt-5.1-codex-max": { + "id": "gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 400000, + "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.08, - "output": 0.1, - "cache_read": 0.003 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -37332,48 +35309,47 @@ "values": [ "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 2, - "output": 8, + "input": 2.5, + "output": 15, "cache_read": 0.25 } }, - "deepseek-v4-flash-vision-exp": { - "id": "deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -37388,41 +35364,41 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.08, - "output": 0.2, - "cache_read": 0.007 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -37437,51 +35413,80 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 65536 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.5, - "output": 1.99, - "cache_read": 0.05 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "greg-2-super": { - "id": "greg-2-super", - "name": "Greg 2 Super", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-06-14", - "last_updated": "2026-06-14", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -37489,20 +35494,23 @@ }, "open_weights": false, "limit": { - "context": 229376, - "output": 229376 + "context": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 1.5, - "output": 5, - "cache_read": 0.25 + "input": 0.25, + "output": 2, + "cache_read": 0.03 } }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro (0813)", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -37517,13 +35525,10 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -37534,24 +35539,23 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 128000 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.35, - "output": 0.8, - "cache_read": 0.01 + "input": 0.28, + "output": 0.42 } }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "gemini-3.1-flash-lite-preview": { + "id": "gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -37565,64 +35569,55 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 + "output": 65536 }, "provider": { "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.4, - "output": 0.8, - "cache_read": 0.003, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 1 } }, - "greg-rp": { - "id": "greg-rp", - "name": "Greg (Roleplay)", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -37630,20 +35625,20 @@ }, "open_weights": false, "limit": { - "context": 229376, - "output": 229376 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -37658,40 +35653,69 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-03-13", - "last_updated": "2026-03-13", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.04, - "output": 0.15, - "cache_read": 0.008 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } - }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + } + } + }, + "tinfoil": { + "id": "tinfoil", + "env": [ + "TINFOIL_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.tinfoil.sh/v1", + "name": "Tinfoil", + "doc": "https://docs.tinfoil.sh", + "models": { + "kimi-k2-6": { + "id": "kimi-k2-6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -37699,23 +35723,26 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "video", - "audio" + "video" ], "output": [ "text" @@ -37723,30 +35750,55 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 256000, "output": 262144 }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "cost": { + "input": 1.5, + "output": 5.25 + } + }, + "llama3-3-70b": { + "id": "llama3-3-70b", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.35, - "output": 1.75, - "cache_read": 0.07 + "input": 1.75, + "output": 2.75 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "gpt-oss-safeguard-120b": { + "id": "gpt-oss-safeguard-120b", + "name": "gpt-oss-safeguard-120b", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -37754,13 +35806,11 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-06", + "release_date": "2025-10-29", + "last_updated": "2025-10-29", "modalities": { "input": [ "text" @@ -37771,30 +35821,54 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 131000, + "output": 32768 }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "nomic-embed-text": { + "id": "nomic-embed-text", + "name": "Nomic Embed Text v1.5", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": false, + "release_date": "2024-02", + "last_updated": "2024-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 768 }, "cost": { - "input": 0.3, - "output": 1.05, - "cache_read": 0.05 + "input": 0.05, + "output": 0 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -37802,13 +35876,11 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2024-06", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -37819,24 +35891,19 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 202752 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 131000, + "output": 32768 }, "cost": { - "input": 0.45, - "output": 2.15, - "cache_read": 0.08, - "cache_write": 0 + "input": 0.15, + "output": 0.6 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "glm-5-2": { + "id": "glm-5-2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -37844,21 +35911,20 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -37869,23 +35935,19 @@ }, "open_weights": true, "limit": { - "context": 1000000, + "context": 384000, "output": 131072 }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, "cost": { - "input": 0.35, - "output": 0.8, - "cache_read": 0.003 + "input": 1.5, + "output": 5.25 } }, - "qwen3.6-27b": { - "id": "qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemma4-31b": { + "id": "gemma4-31b", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -37893,23 +35955,24 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" @@ -37917,28 +35980,37 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 256000, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 1.5, - "cache_read": 0.04 + "input": 0.4, + "output": 1 } - }, - "greg-2-ultra": { - "id": "greg-2-ultra", - "name": "Greg 2 Ultra", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + } + } + }, + "mistral": { + "id": "mistral", + "env": [ + "MISTRAL_API_KEY" + ], + "npm": "@ai-sdk/mistral", + "name": "Mistral", + "doc": "https://docs.mistral.ai/getting-started/models/", + "models": { + "codestral-latest": { + "id": "codestral-latest", + "name": "Codestral (latest)", + "description": "Mistral code model for completions, refactors, and developer IDE workflows", + "family": "codestral", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-06-14", - "last_updated": "2026-06-14", + "knowledge": "2024-10", + "release_date": "2024-05-29", + "last_updated": "2025-01-04", "modalities": { "input": [ "text" @@ -37947,28 +36019,59 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 229376, - "output": 229376 + "context": 256000, + "output": 4096 }, "cost": { - "input": 3, - "output": 10, - "cache_read": 0.5 + "input": 0.3, + "output": 0.9 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "mistral-large-latest": { + "id": "mistral-large-latest", + "name": "Mistral Large (latest)", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "open-mistral-7b": { + "id": "open-mistral-7b", + "name": "Mistral 7B", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "knowledge": "2023-12", + "release_date": "2023-09-27", + "last_updated": "2023-09-27", "modalities": { "input": [ "text" @@ -37977,43 +36080,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 8000, + "output": 8000 }, "cost": { - "input": 0.18, - "output": 0.35, - "cache_read": 0.04 + "input": 0.25, + "output": 0.25 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "devstral-small-2507": { + "id": "devstral-small-2507", + "name": "Devstral Small", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-05", + "release_date": "2025-07-10", + "last_updated": "2025-07-10", "modalities": { "input": [ "text" @@ -38024,50 +36112,30 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 128000, + "output": 128000 }, + "status": "deprecated", "cost": { - "input": 0.4, - "output": 1.4, - "cache_read": 0.06 + "input": 0.1, + "output": 0.3 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "ministral-3b-latest": { + "id": "ministral-3b-latest", + "name": "Ministral 3B (latest)", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-04", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -38075,28 +36143,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.55, - "output": 2.25, - "cache_read": 0.05 + "input": 0.04, + "output": 0.04 } }, - "greg-1-mini": { - "id": "greg-1-mini", - "name": "Greg 1 Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "pixtral-large-latest": { + "id": "pixtral-large-latest", + "name": "Pixtral Large (latest)", + "description": "Mistral's larger vision model for document-heavy image understanding and chat", + "family": "pixtral", "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2024-11-04", "modalities": { "input": [ "text", @@ -38106,44 +36172,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 229376, - "output": 229376 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.07, - "output": 0.15, - "cache_read": 0.01 + "input": 2, + "output": 6 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "mistral-nemo": { + "id": "mistral-nemo", + "name": "Mistral Nemo", + "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", + "family": "mistral-nemo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ "text" @@ -38154,50 +36204,55 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai-compatible" + "cost": { + "input": 0.15, + "output": 0.15 + } + }, + "mistral-embed": { + "id": "mistral-embed", + "name": "Mistral Embed", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "mistral-embed", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2023-12-11", + "last_updated": "2023-12-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8000, + "output": 3072 }, "cost": { - "input": 0.12, - "output": 0.21, - "cache_read": 0.003 + "input": 0.1, + "output": 0 } - } - } - }, - "standardcompute": { - "id": "standardcompute", - "env": [ - "STANDARDCOMPUTE_API_KEY" - ], - "npm": "@openrouter/ai-sdk-provider", - "api": "https://api.stdcmpt.com/v1", - "name": "Standard Compute", - "doc": "https://standardcompute.com/models", - "models": { - "standardcompute": { - "id": "standardcompute", - "name": "Standard Compute", - "description": "Flat-rate smart-routing gateway: one model id, each request routed across a curated catalog of 1M-context models (DeepSeek, GLM, MiniMax, Qwen, GPT-5.6, Claude 5, Gemini 2.5, Kimi) or pinned to a user-selected model", - "family": "auto", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + }, + "mistral-small-2506": { + "id": "mistral-small-2506", + "name": "Mistral Small 3.2", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-08-24", + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", "modalities": { "input": [ "text", @@ -38207,40 +36262,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 24576 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.3 } - } - } - }, - "cloudferro-sherlock": { - "id": "cloudferro-sherlock", - "env": [ - "CLOUDFERRO_SHERLOCK_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api-sherlock.cloudferro.com/openai/v1/", - "name": "CloudFerro Sherlock", - "doc": "https://docs.sherlock.cloudferro.com/", - "models": { - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + }, + "ministral-8b-latest": { + "id": "ministral-8b-latest", + "name": "Ministral 8B (latest)", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-10-09", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-04", "modalities": { "input": [ "text" @@ -38251,28 +36294,26 @@ }, "open_weights": true, "limit": { - "context": 70000, - "output": 70000 + "context": 128000, + "output": 128000 }, "cost": { - "input": 2.92, - "output": 2.92 + "input": 0.1, + "output": 0.1 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "open-mixtral-8x22b": { + "id": "open-mixtral-8x22b", + "name": "Mixtral 8x22B", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mixtral", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2024-04", + "release_date": "2024-04-17", + "last_updated": "2024-04-17", "modalities": { "input": [ "text" @@ -38283,28 +36324,26 @@ }, "open_weights": true, "limit": { - "context": 196000, - "input": 180000, - "output": 16000 + "context": 64000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 2, + "output": 6 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "OpenAI GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "mistral-medium-latest": { + "id": "mistral-medium-latest", + "name": "Mistral Medium (latest)", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } @@ -38312,11 +36351,12 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -38324,26 +36364,26 @@ }, "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2.92, - "output": 2.92 + "input": 1.5, + "output": 7.5 } }, - "speakleash/Bielik-11B-v2.6-Instruct": { - "id": "speakleash/Bielik-11B-v2.6-Instruct", - "name": "Bielik 11B v2.6 Instruct", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "devstral-small-2505": { + "id": "devstral-small-2505", + "name": "Devstral Small 2505", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "knowledge": "2025-05", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ "text" @@ -38354,26 +36394,28 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 32000 + "context": 128000, + "output": 128000 }, + "status": "deprecated", "cost": { - "input": 0.67, - "output": 0.67 + "input": 0.1, + "output": 0.3 } }, - "speakleash/Bielik-11B-v3.0-Instruct": { - "id": "speakleash/Bielik-11B-v3.0-Instruct", - "name": "Bielik 11B v3.0 Instruct", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "magistral-small": { + "id": "magistral-small", + "name": "Magistral Small", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-small", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "knowledge": "2025-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-17", "modalities": { "input": [ "text" @@ -38384,302 +36426,211 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 32000 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.67, - "output": 0.67 + "input": 0.5, + "output": 1.5 } - } - } - }, - "anthropic": { - "id": "anthropic", - "env": [ - "ANTHROPIC_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "name": "Anthropic", - "doc": "https://docs.anthropic.com/en/docs/about-claude/models", - "models": { - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + }, + "mistral-medium-2604": { + "id": "mistral-medium-2604", + "name": "Mistral Medium 3.5", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 1.5, + "output": 7.5 } }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "mistral-small-latest": { + "id": "mistral-small-latest", + "name": "Mistral Small (latest)", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.15, + "output": 0.6 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "open-mixtral-8x7b": { + "id": "open-mixtral-8x7b", + "name": "Mixtral 8x7B", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mixtral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2024-01", + "release_date": "2023-12-11", + "last_updated": "2023-12-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 32000, + "output": 32000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.7, + "output": 0.7 } }, - "claude-opus-5-5": { - "id": "claude-opus-5-5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "devstral-latest": { + "id": "devstral-latest", + "name": "Devstral 2", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 8, - "output": 40, - "cache_read": 0.4, - "cache_write": 10 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "context": 262144, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 + "input": 0.4, + "output": 2 } }, - "claude-fable-5-1": { - "id": "claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", + "mistral-small-2603": { + "id": "mistral-small-2603", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high", - "xhigh", - "max" + "none", + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 0.15, + "output": 0.6 } }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "mistral-medium-2505": { + "id": "mistral-medium-2505", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-01", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -38687,112 +36638,60 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.4, + "output": 2 } }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "mistral-large-2411": { + "id": "mistral-large-2411", + "name": "Mistral Large 2.1", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-18", + "last_updated": "2024-11-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "context": 131072, + "output": 16384 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2, + "output": 6 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "mistral-medium-2508": { + "id": "mistral-medium-2508", + "name": "Mistral Medium 3.1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-07", - "last_updated": "2026-06-09", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-08-12", + "last_updated": "2025-08-12", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -38800,387 +36699,288 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.4, + "output": 2 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "open-mistral-nemo": { + "id": "open-mistral-nemo", + "name": "Open Mistral Nemo", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 128000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } - }, + "status": "deprecated", "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.15, + "output": 0.15 } }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "magistral-medium-latest": { + "id": "magistral-medium-latest", + "name": "Magistral Medium (latest)", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-medium", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 2, + "output": 5 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "devstral-medium-latest": { + "id": "devstral-medium-latest", + "name": "Devstral 2 (latest)", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-29", - "last_updated": "2026-06-30", + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.4, + "output": 2 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "devstral-2512": { + "id": "devstral-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-04", - "last_updated": "2026-03-13", + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.4, + "output": 2 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "labs-devstral-small-2512": { + "id": "labs-devstral-small-2512", + "name": "Devstral Small 2", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 256000, + "output": 256000 }, + "status": "deprecated", "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0, + "output": 0 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "pixtral-12b": { + "id": "pixtral-12b", + "name": "Pixtral 12B", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2024-09", + "release_date": "2024-09-01", + "last_updated": "2024-09-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 128000, "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.15, + "output": 0.15 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "mistral-large-2512": { + "id": "mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-14", - "last_updated": "2026-04-16", + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "devstral-medium-2507": { + "id": "devstral-medium-2507", + "name": "Devstral Medium", + "description": "Legacy model retained for compatibility with older integrations", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-07-10", + "last_updated": "2025-07-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, "output": 128000 }, + "status": "deprecated", "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.4, + "output": 2 } } } }, - "tinfoil": { - "id": "tinfoil", + "cloudflare-workers-ai": { + "id": "cloudflare-workers-ai", "env": [ - "TINFOIL_API_KEY" + "CLOUDFLARE_ACCOUNT_ID", + "CLOUDFLARE_API_KEY" ], "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.tinfoil.sh/v1", - "name": "Tinfoil", - "doc": "https://docs.tinfoil.sh", + "api": "https://api.cloudflare.com/client/v4/accounts/${CLOUDFLARE_ACCOUNT_ID}/ai/v1", + "name": "Cloudflare Workers AI", + "doc": "https://developers.cloudflare.com/workers-ai/models/", "models": { - "gpt-oss-safeguard-120b": { - "id": "gpt-oss-safeguard-120b", - "name": "gpt-oss-safeguard-120b", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", + "@cf/ibm-granite/granite-4.0-h-micro": { + "id": "@cf/ibm-granite/granite-4.0-h-micro", + "name": "Granite 4.0 H Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "granite", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "release_date": "2025-10-07", + "last_updated": "2025-10-07", "modalities": { "input": [ "text" @@ -39191,42 +36991,44 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.017, + "output": 0.112 } }, - "glm-5-3-flash": { - "id": "glm-5-3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "@cf/moonshotai/kimi-k2.7-code": { + "id": "@cf/moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" @@ -39234,42 +37036,48 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.4, - "output": 1.25, - "cache_read": 0.1 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "@cf/moonshotai/kimi-k2.6": { + "id": "@cf/moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -39278,66 +37086,36 @@ "open_weights": true, "limit": { "context": 262144, - "output": 131072 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.8 - } - }, - "llama3-3-70b": { - "id": "llama3-3-70b", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 4096 + "output": 256000 }, "cost": { - "input": 1.75, - "output": 2.75 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "gemma4-31b": { - "id": "gemma4-31b", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "@cf/google/gemma-4-26b-a4b-it": { + "id": "@cf/google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, "release_date": "2026-04-02", @@ -39353,19 +37131,19 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0.4, - "output": 1 + "input": 0.1, + "output": 0.3 } }, - "glm-5-3": { - "id": "glm-5-3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "@cf/openai/gpt-oss-120b": { + "id": "@cf/openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -39373,16 +37151,16 @@ "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -39393,46 +37171,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.8, - "output": 5.75, - "cache_read": 0.45 + "input": 0.35, + "output": 0.75 } }, - "deepseek-v4-1-flash": { - "id": "deepseek-v4-1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "@cf/openai/gpt-oss-20b": { + "id": "@cf/openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -39440,27 +37202,26 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 384000 + "context": 128000, + "output": 16384 }, - "status": "beta", "cost": { - "input": 0.65, - "output": 1.45, - "cache_read": 0.13 + "input": 0.2, + "output": 0.3 } }, - "nomic-embed-text": { - "id": "nomic-embed-text", - "name": "Nomic Embed Text v1.5", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "@cf/mistralai/mistral-small-3.1-24b-instruct": { + "id": "@cf/mistralai/mistral-small-3.1-24b-instruct", + "name": "Mistral Small 3.1 24B Instruct", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": false, - "temperature": false, - "release_date": "2024-02", - "last_updated": "2024-02", + "temperature": true, + "release_date": "2025-03-18", + "last_updated": "2025-03-18", "modalities": { "input": [ "text" @@ -39471,22 +37232,25 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 768 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0 + "input": 0.351, + "output": 0.555 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "@cf/nvidia/nemotron-3-120b-a12b": { + "id": "@cf/nvidia/nemotron-3-120b-a12b", + "name": "Nemotron 3 Super 120B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -39497,11 +37261,11 @@ } ], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text" @@ -39512,42 +37276,42 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.5, + "output": 1.5 } - } - } - }, - "llama": { - "id": "llama", - "env": [ - "LLAMA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.llama.com/compat/v1/", - "name": "Llama", - "doc": "https://llama.developer.meta.com/docs/models", - "models": { - "llama-4-maverick-17b-128e-instruct-fp8": { - "id": "llama-4-maverick-17b-128e-instruct-fp8", - "name": "Llama-4-Maverick-17B-128E-Instruct-FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": true, - "reasoning": false, + }, + "@cf/zai-org/glm-5.2": { + "id": "@cf/zai-org/glm-5.2", + "name": "Glm 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -39555,26 +37319,44 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "groq-llama-4-maverick-17b-128e-instruct": { - "id": "groq-llama-4-maverick-17b-128e-instruct", - "name": "Groq-Llama-4-Maverick-17B-128E-Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": true, - "reasoning": false, + "@cf/zai-org/glm-4.7-flash": { + "id": "@cf/zai-org/glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -39585,30 +37367,31 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.0605, + "output": 0.4 } }, - "llama-4-scout-17b-16e-instruct-fp8": { - "id": "llama-4-scout-17b-16e-instruct-fp8", - "name": "Llama-4-Scout-17B-16E-Instruct-FP8", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": true, + "@cf/deepseek-ai/deepseek-r1-distill-qwen-32b": { + "id": "@cf/deepseek-ai/deepseek-r1-distill-qwen-32b", + "name": "Deepseek R1 Distill Qwen 32B", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -39616,26 +37399,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 80000, + "output": 80000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.497, + "output": 4.881 } }, - "cerebras-llama-4-maverick-17b-128e-instruct": { - "id": "cerebras-llama-4-maverick-17b-128e-instruct", - "name": "Cerebras-Llama-4-Maverick-17B-128E-Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": true, - "reasoning": false, + "@cf/qwen/qwen3-30b-a3b-fp8": { + "id": "@cf/qwen/qwen3-30b-a3b-fp8", + "name": "Qwen3 30B A3b fp8", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2025-04-30", + "last_updated": "2025-04-30", "modalities": { "input": [ "text" @@ -39646,26 +37430,26 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.0509, + "output": 0.335 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": true, + "@cf/qwen/qwen2.5-coder-32b-instruct": { + "id": "@cf/qwen/qwen2.5-coder-32b-instruct", + "name": "Qwen2.5 Coder 32B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ "text" @@ -39676,26 +37460,57 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.66, + "output": 1 } }, - "llama-3.3-8b-instruct": { - "id": "llama-3.3-8b-instruct", - "name": "Llama-3.3-8B-Instruct", + "@cf/qwen/qwq-32b": { + "id": "@cf/qwen/qwq-32b", + "name": "Qwq 32B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-03-05", + "last_updated": "2025-03-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 24000, + "output": 24000 + }, + "cost": { + "input": 0.66, + "output": 1 + } + }, + "@cf/meta/llama-3.2-1b-instruct": { + "id": "@cf/meta/llama-3.2-1b-instruct", + "name": "Llama 3.2 1B Instruct", "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "family": "llama", - "attachment": true, + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text" @@ -39706,29 +37521,30 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 60000, + "output": 60000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.027, + "output": 0.201 } }, - "cerebras-llama-4-scout-17b-16e-instruct": { - "id": "cerebras-llama-4-scout-17b-16e-instruct", - "name": "Cerebras-Llama-4-Scout-17B-16E-Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "@cf/meta/llama-3.2-11b-vision-instruct": { + "id": "@cf/meta/llama-3.2-11b-vision-instruct", + "name": "Llama 3.2 11B Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", "family": "llama", "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -39737,39 +37553,30 @@ "open_weights": true, "limit": { "context": 128000, - "output": 4096 + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.0485, + "output": 0.676 } - } - } - }, - "cohere": { - "id": "cohere", - "env": [ - "COHERE_API_KEY" - ], - "npm": "@ai-sdk/cohere", - "name": "Cohere", - "doc": "https://docs.cohere.com/docs/models", - "models": { - "command-r-08-2024": { - "id": "command-r-08-2024", - "name": "Command R", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, + }, + "@cf/meta/llama-4-scout-17b-16e-instruct": { + "id": "@cf/meta/llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout 17B 16E Instruct", + "description": "Open Llama with long-context vision for efficient multimodal agents", + "family": "llama", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -39777,39 +37584,26 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 131000, + "output": 16384 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.27, + "output": 0.85 } }, - "north-mini-code-1-0": { - "id": "north-mini-code-1-0", - "name": "North Mini Code", - "description": "Cohere coding model for practical software engineering and agentic edits", - "family": "north", + "@cf/meta/llama-guard-3-8b": { + "id": "@cf/meta/llama-guard-3-8b", + "name": "Llama Guard 3 8B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-09-23", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "release_date": "2025-01-22", + "last_updated": "2025-01-22", "modalities": { "input": [ "text" @@ -39820,30 +37614,27 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.cohere.ai/compatibility/v1" + "context": 131072, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.484, + "output": 0.03 } }, - "command-a-translate-08-2025": { - "id": "command-a-translate-08-2025", - "name": "Command A Translate", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "family": "command-a", + "@cf/meta/llama-3.3-70b-instruct-fp8-fast": { + "id": "@cf/meta/llama-3.3-70b-instruct-fp8-fast", + "name": "Llama 3.3 70B Instruct fp8 Fast", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -39854,24 +37645,26 @@ }, "open_weights": true, "limit": { - "context": 8000, - "output": 8000 + "context": 24000, + "output": 24000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.293, + "output": 2.253 } }, - "c4ai-aya-expanse-8b": { - "id": "c4ai-aya-expanse-8b", - "name": "Aya Expanse 8B", - "description": "Compact open multilingual model optimized for generation across 23 languages", + "@cf/meta/llama-3.1-8b-instruct-fp8": { + "id": "@cf/meta/llama-3.1-8b-instruct-fp8", + "name": "Llama 3.1 8B Instruct fp8", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2024-10-24", - "last_updated": "2024-10-24", + "release_date": "2024-07-25", + "last_updated": "2024-07-25", "modalities": { "input": [ "text" @@ -39882,22 +37675,26 @@ }, "open_weights": true, "limit": { - "context": 8000, - "output": 4000 + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.152, + "output": 0.287 } }, - "command-r7b-arabic-02-2025": { - "id": "command-r7b-arabic-02-2025", - "name": "Command R7B Arabic", - "description": "Open Command R model optimized for Arabic enterprise chat, RAG, and cultural knowledge", - "family": "command-r", + "@cf/meta/llama-3.2-3b-instruct": { + "id": "@cf/meta/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-02-27", - "last_updated": "2025-02-27", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text" @@ -39908,40 +37705,29 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 80000, + "output": 80000 }, "cost": { - "input": 0.0375, - "output": 0.15 + "input": 0.0509, + "output": 0.335 } }, - "command-a-plus-05-2026": { - "id": "command-a-plus-05-2026", - "name": "Command A Plus", - "description": "Cohere's stronger command model for multilingual agents and enterprise workflows", - "family": "command-a", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1 - } - ], - "tool_call": true, - "structured_output": true, + "@cf/aisingapore/gemma-sea-lion-v4-27b-it": { + "id": "@cf/aisingapore/gemma-sea-lion-v4-27b-it", + "name": "Gemma Sea Lion V4 27B It", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04-01", - "release_date": "2026-05-20", - "last_updated": "2026-06-09", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -39950,25 +37736,38 @@ "open_weights": true, "limit": { "context": 128000, - "output": 64000 + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.351, + "output": 0.555 } - }, - "command-a-03-2025": { - "id": "command-a-03-2025", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", + } + } + }, + "bailing": { + "id": "bailing", + "env": [ + "BAILING_API_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.tbox.cn/api/llm/v1/chat/completions", + "name": "Bailing", + "doc": "https://alipaytbox.yuque.com/sxs0ba/ling/intro", + "models": { + "Ring-1T": { + "id": "Ring-1T", + "name": "Ring-1T", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "ring", "attachment": false, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "knowledge": "2024-06", + "release_date": "2025-10", + "last_updated": "2025-10", "modalities": { "input": [ "text" @@ -39979,28 +37778,29 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 8000 + "context": 128000, + "output": 32000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.57, + "output": 2.29 } }, - "c4ai-aya-vision-8b": { - "id": "c4ai-aya-vision-8b", - "name": "Aya Vision 8B", - "description": "Compact open multilingual vision model for OCR and visual question answering", - "attachment": true, + "Ling-1T": { + "id": "Ling-1T", + "name": "Ling-1T", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "ling", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-03-04", - "last_updated": "2025-05-14", + "knowledge": "2024-06", + "release_date": "2025-10", + "last_updated": "2025-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -40008,281 +37808,277 @@ }, "open_weights": true, "limit": { - "context": 16000, - "output": 4000 + "context": 128000, + "output": 32000 + }, + "cost": { + "input": 0.57, + "output": 2.29 } - }, - "c4ai-aya-vision-32b": { - "id": "c4ai-aya-vision-32b", - "name": "Aya Vision 32B", - "description": "Open multilingual vision model for OCR, visual reasoning, and image question answering", + } + } + }, + "anyapi": { + "id": "anyapi", + "env": [ + "ANYAPI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.anyapi.ai/v1", + "name": "AnyAPI", + "doc": "https://docs.anyapi.ai", + "models": { + "xai/grok-4.3": { + "id": "xai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-03-04", - "last_updated": "2025-05-14", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16000, - "output": 4000 + "context": 1000000, + "output": 30000 } }, - "command-r7b-12-2024": { - "id": "command-r7b-12-2024", - "name": "Command R7B", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-12-02", - "last_updated": "2024-12-02", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 0.0375, - "output": 0.15 + "context": 1048576, + "output": 65536 } }, - "command-r-plus-08-2024": { - "id": "command-r-plus-08-2024", - "name": "Command R+", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", - "attachment": false, - "reasoning": false, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 2.5, - "output": 10 + "context": 1048576, + "output": 65536 } }, - "command-a-vision-07-2025": { - "id": "command-a-vision-07-2025", - "name": "Command A Vision", - "description": "Cohere vision model for multilingual document analysis, OCR, and image understanding", - "family": "command-a", + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-07-31", - "last_updated": "2025-07-31", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8000 - }, - "cost": { - "input": 2.5, - "output": 10 + "context": 1048576, + "output": 65536 } }, - "command-a-reasoning-08-2025": { - "id": "command-a-reasoning-08-2025", - "name": "Command A Reasoning", - "description": "Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows", - "family": "command-a", - "attachment": false, + "google/gemini-3-pro-preview": { + "id": "google/gemini-3-pro-preview", + "name": "Gemini 3 Pro Preview", + "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", + "family": "gemini-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 32000 - }, - "cost": { - "input": 2.5, - "output": 10 + "context": 1048576, + "output": 65536 } }, - "c4ai-aya-expanse-32b": { - "id": "c4ai-aya-expanse-32b", - "name": "Aya Expanse 32B", - "description": "Open multilingual model optimized for generation across 23 languages", - "attachment": false, - "reasoning": false, - "tool_call": false, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-10-24", - "last_updated": "2024-10-24", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 + "context": 1048576, + "output": 65536 } - } - } - }, - "deepseek": { - "id": "deepseek", - "env": [ - "DEEPSEEK_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.deepseek.com", - "name": "DeepSeek", - "doc": "https://api-docs.deepseek.com/quick_start/pricing", - "models": { - "deepseek-v4-flash-vision-exp": { - "id": "deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + }, + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "status": "deprecated", - "cost": { - "input": 0.15, - "output": 0.6, - "reasoning": 0.6, - "cache_read": 0.003 + "context": 200000, + "output": 100000 } }, - "deepseek-flash": { - "id": "deepseek-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -40292,158 +38088,111 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "reasoning": 0.6, - "cache_read": 0.003 + "context": 400000, + "input": 272000, + "output": 128000 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "reasoning": 0.87, - "cache_read": 0.003625 + "context": 200000, + "output": 100000 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "status": "deprecated", - "cost": { - "input": 0.15, - "output": 0.6, - "reasoning": 0.6, - "cache_read": 0.003 + "context": 200000, + "output": 100000 } - } - } - }, - "baseten": { - "id": "baseten", - "env": [ - "BASETEN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.baseten.co/v1", - "name": "Baseten", - "doc": "https://docs.baseten.co/inference/model-apis/overview", - "models": { - "thinkingmachines/inkling-small": { - "id": "thinkingmachines/inkling-small", - "name": "Inkling Small", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ling", + }, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -40453,216 +38202,189 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 1.2, - "cache_read": 0.1 + "context": 400000, + "input": 272000, + "output": 128000 } }, - "thinkingmachines/inkling": { - "id": "thinkingmachines/inkling", - "name": "Inkling", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ling", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "cost": { - "input": 1, - "output": 4.05 + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } } }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek V3.1", - "description": "Legacy model retained for compatibility with older integrations", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-25", - "last_updated": "2025-08-25", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 164000, - "output": 131000 - }, - "status": "deprecated", - "cost": { - "input": 0.5, - "output": 1.5 + "context": 1047576, + "output": 32768 } }, - "deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.13, - "output": 0.26, - "cache_read": 0.028 + "context": 400000, + "input": 272000, + "output": 128000 } }, - "deepseek-ai/DeepSeek-V4-Pro-0813": { - "id": "deepseek-ai/DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 262144 - }, - "cost": { - "input": 1.32, - "output": 3.96 + "context": 1047576, + "output": 32768 } }, - "deepseek-ai/DeepSeek-V4.1-Flash": { - "id": "deepseek-ai/DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -40672,47 +38394,25 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "context": 400000, + "input": 272000, + "output": 128000 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "mistralai/devstral-2512": { + "id": "mistralai/devstral-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ "text" @@ -40723,71 +38423,23 @@ }, "open_weights": true, "limit": { - "context": 1048576, + "context": 262144, "output": 262144 }, - "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.145 - } - }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204000, - "output": 204000 - }, - "status": "deprecated", - "cost": { - "input": 0.3, - "output": 1.2 - } + "status": "deprecated" }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "mistralai/mistral-large-2512": { + "id": "mistralai/mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", @@ -40799,247 +38451,287 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "context": 262144, + "output": 262144 } }, - "moonshotai/Kimi-K3": { - "id": "moonshotai/Kimi-K3", - "name": "Kimi K3", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k3", + "anthropic/claude-sonnet-4-5": { + "id": "anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", - "high", - "max" + "medium", + "high" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 262144 - }, - "cost": { - "input": 3, - "output": 15 + "context": 200000, + "output": 64000 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 128000 }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-01-30", - "last_updated": "2026-02-12", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.12 + "context": 200000, + "output": 64000 } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202800, - "output": 202800 + "context": 1000000, + "output": 128000 }, - "cost": { - "input": 1.3, - "output": 4.3, - "cache_read": 0.26 + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "high", - "max" + "low", + "medium", + "high" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 262144 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.3 + "context": 1000000, + "output": 64000 } }, - "zai-org/GLM-4.7": { - "id": "zai-org/GLM-4.7", - "name": "GLM 4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "cohere/command-r-plus-08-2024": { + "id": "cohere/command-r-plus-08-2024", + "name": "Command R+", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -41050,40 +38742,23 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 200000 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.12 + "context": 128000, + "output": 4000 } }, - "zai-org/GLM-5.2-Fast": { - "id": "zai-org/GLM-5.2-Fast", - "name": "GLM 5.2 Fast", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "perplexity/sonar-reasoning-pro": { + "id": "perplexity/sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", + "family": "sonar-reasoning", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text", @@ -41093,40 +38768,24 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 262144 - }, - "cost": { - "input": 2.1, - "output": 6.6, - "cache_read": 0.21 + "context": 128000, + "output": 4096 } }, - "zai-org/GLM-5.3-Fast": { - "id": "zai-org/GLM-5.3-Fast", - "name": "GLM 5.3 Fast", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "perplexity/sonar-pro": { + "id": "perplexity/sonar-pro", + "name": "Sonar Pro", + "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", + "family": "sonar-pro", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text", @@ -41136,37 +38795,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 262144 - }, - "cost": { - "input": 2.1, - "output": 6.6 + "context": 200000, + "output": 8192 } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM 5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -41177,44 +38828,30 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 202800 - }, - "cost": { - "input": 0.95, - "output": 3.15, - "cache_read": 0.2 + "context": 1000000, + "output": 384000 } }, - "zai-org/GLM-5.3-Flash": { - "id": "zai-org/GLM-5.3-Flash", - "name": "GLM 5.3 Flash", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", - "attachment": true, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -41222,43 +38859,29 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.5 + "context": 1000000, + "output": 384000 } }, - "zai-org/GLM-5.3": { - "id": "zai-org/GLM-5.3", - "name": "GLM 5.3", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "deepseek/deepseek-r1": { + "id": "deepseek/deepseek-r1", + "name": "DeepSeek Reasoner", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -41266,36 +38889,22 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 262144 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.14 + "context": 1000000, + "output": 384000 } }, - "nvidia/Nemotron-120B-A12B": { - "id": "nvidia/Nemotron-120B-A12B", - "name": "Nemotron Super", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "deepseek/deepseek-chat": { + "id": "deepseek/deepseek-chat", + "name": "DeepSeek Chat", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ "text" @@ -41306,220 +38915,260 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 202800 - }, - "cost": { - "input": 0.3, - "output": 0.75, - "cache_read": 0.06 + "context": 1000000, + "output": 384000 } - }, - "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B": { - "id": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B", - "name": "Nemotron Ultra", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, + } + } + }, + "google": { + "id": "google", + "env": [ + "GOOGLE_API_KEY", + "GOOGLE_GENERATIVE_AI_API_KEY", + "GEMINI_API_KEY" + ], + "npm": "@ai-sdk/google", + "name": "Google", + "doc": "https://ai.google.dev/gemini-api/docs/models", + "models": { + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202800, - "output": 202800 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "OpenAI GPT 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "gemini-2.5-flash-preview-tts": { + "id": "gemini-2.5-flash-preview-tts", + "name": "Gemini 2.5 Flash Preview TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gemini-flash", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01", + "release_date": "2025-05-01", + "last_updated": "2025-05-01", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128072, - "output": 128072 + "context": 8192, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.5 + "input": 0.5, + "output": 10 } - } - } - }, - "nan": { - "id": "nan", - "env": [ - "NAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.nan.builders/v1", - "name": "NaN", - "doc": "https://nan.builders/docs/models", - "models": { - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + }, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "audio" + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "glm5.3-flash": { - "id": "glm5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "gemma4": { - "id": "gemma4", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 } }, - "qwen3.6": { - "id": "qwen3.6", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "gemma-4-31b-it": { + "id": "gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -41530,8 +39179,8 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -41544,29 +39193,29 @@ "open_weights": true, "limit": { "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 + "output": 32768 } }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.0-flash": { + "id": "gemini-2.0-flash", + "name": "Gemini 2.0 Flash", + "description": "Earlier Gemini Flash workhorse for responsive multimodal apps and tool use", + "family": "gemini-flash", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": true, + "knowledge": "2024-06", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -41574,27 +39223,28 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1048576, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "glm5.3": { - "id": "glm5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "gemini-embedding-001": { + "id": "gemini-embedding-001", + "name": "Gemini Embedding 001", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "gemini", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-05", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text" @@ -41603,65 +39253,21 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 2048, + "output": 1 }, "cost": { - "input": 0, + "input": 0.15, "output": 0 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0 - } - } - } - }, - "stepfun-ai-step-plan": { - "id": "stepfun-ai-step-plan", - "env": [ - "STEPFUN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.stepfun.ai/step_plan/v1", - "name": "StepFun Step Plan (Global)", - "doc": "https://platform.stepfun.ai/docs/en/step-plan/integrations/reasoning-api", - "models": { - "step-5-preview": { - "id": "step-5-preview", - "name": "Step 5 Preview", - "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", + "gemini-3.1-pro-preview-customtools": { + "id": "gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -41675,17 +39281,18 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "release_date": "2026-09-16", - "last_updated": "2026-09-20", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -41693,188 +39300,149 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 - } - }, - "step-3.5-flash-2603": { - "id": "step-3.5-flash-2603", - "name": "Step 3.5 Flash 2603", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "context": 1048576, + "output": 65536 }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text" + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "step-3.5-flash": { - "id": "step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", - "attachment": false, + "gemini-flash-lite-latest": { + "id": "gemini-flash-lite-latest", + "name": "Gemini Flash-Lite Latest", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 512, + "max": 24576 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "step-3.7-flash": { - "id": "step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "gemini-3-pro-image-preview": { + "id": "gemini-3-pro-image-preview", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "knowledge": "2025-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 120 } - } - } - }, - "nearai": { - "id": "nearai", - "env": [ - "NEARAI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://cloud-api.near.ai/v1", - "name": "NEAR AI Cloud", - "doc": "https://docs.near.ai/", - "models": { - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + }, + "gemini-2.5-flash-image": { + "id": "gemini-2.5-flash-image", + "name": "Nano Banana", + "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-06", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.3, + "output": 30, + "cache_read": 0.075 } }, - "anthropic/claude-sonnet-4-5": { - "id": "anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -41883,18 +39451,22 @@ }, { "type": "budget_tokens", - "min": 1024 + "min": 512, + "max": 24576 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -41903,80 +39475,69 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "input_audio": 0.3 } }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "gemini-omni-flash-preview": { + "id": "gemini-omni-flash-preview", + "name": "Gemini Omni Flash Preview", + "description": "Video generation and editing model for fast, conversational text- and image-to-video workflows", + "family": "gemini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.5, + "output": 17.5 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "gemini-3.1-flash-image-preview": { + "id": "gemini-3.1-flash-image-preview", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "minimal", + "high" + ] } ], - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-01", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ "text", @@ -41984,42 +39545,49 @@ "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.5, + "output": 60 } }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -42028,103 +39596,91 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "black-forest-labs/FLUX.2-klein-4B": { - "id": "black-forest-labs/FLUX.2-klein-4B", - "name": "FLUX.2 Klein 4B", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-14", - "last_updated": "2026-01-14", - "modalities": { - "input": [ - "text" + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "image" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 1, - "output": 1 + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "gemma-4-26b-a4b-it": { + "id": "gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 + "context": 262144, + "output": 32768 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "gemini-3-pro-preview": { + "id": "gemini-3-pro-preview", + "name": "Gemini 3 Pro Preview", + "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", "image", - "audio", "video", + "audio", "pdf" ], "output": [ @@ -42136,15 +39692,16 @@ "context": 1048576, "output": 65536 }, + "status": "deprecated", "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, + "input": 2, + "output": 12, + "cache_read": 0.2, "tiers": [ { - "input": 2.5, - "output": 15, - "cache_read": 0.25, + "input": 4, + "output": 18, + "cache_read": 0.4, "tier": { "type": "context", "size": 200000 @@ -42152,32 +39709,42 @@ } ], "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 4, + "output": 18, + "cache_read": 0.4 } } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", - "audio", "video", + "audio", "pdf" ], "output": [ @@ -42190,26 +39757,65 @@ "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, + "input": 0.5, + "output": 3, + "cache_read": 0.05, "input_audio": 1 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", + "gemini-2.5-pro-preview-tts": { + "id": "gemini-2.5-pro-preview-tts", + "name": "Gemini 2.5 Pro Preview TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gemini-flash", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-05-01", + "last_updated": "2025-05-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 8192, + "output": 16384 + }, + "cost": { + "input": 1, + "output": 20 + } + }, + "gemini-flash-latest": { + "id": "gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", @@ -42228,26 +39834,36 @@ "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 + "input": 0.3, + "output": 2.5, + "cache_read": 0.075, + "input_audio": 1 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "gemini-3.1-flash-lite-preview": { + "id": "gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Legacy model retained for compatibility with older integrations", "family": "gemini-flash-lite", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", @@ -42265,6 +39881,7 @@ "context": 1048576, "output": 65536 }, + "status": "deprecated", "cost": { "input": 0.25, "output": 1.5, @@ -42272,112 +39889,79 @@ "input_audio": 0.5 } }, - "Qwen/Qwen3-Reranker-0.6B": { - "id": "Qwen/Qwen3-Reranker-0.6B", - "name": "Qwen3 Reranker 0.6B", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-06-03", - "last_updated": "2025-06-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "output": 1024 - }, - "cost": { - "input": 0.01, - "output": 0.01 - } - }, - "Qwen/Qwen3-VL-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", - "name": "Qwen3-VL 30B-A3B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.0-flash-lite": { + "id": "gemini-2.0-flash-lite", + "name": "Gemini 2.0 Flash-Lite", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gemini-flash-lite", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2024-06", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, + "context": 1048576, "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0.15, - "output": 0.55 + "input": 0.075, + "output": 0.3 } - }, - "Qwen/Qwen3-Embedding-0.6B": { - "id": "Qwen/Qwen3-Embedding-0.6B", - "name": "Qwen3 Embedding 0.6B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + } + } + }, + "opencode-go": { + "id": "opencode-go", + "env": [ + "OPENCODE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://opencode.ai/zen/go/v1", + "name": "OpenCode Go", + "doc": "https://opencode.ai/docs/zen", + "models": { + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-06-03", - "last_updated": "2025-06-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 1024 - }, - "cost": { - "input": 0.01, - "output": 0.01 - } - }, - "Qwen/Qwen3.6-35B-A3B-FP8": { - "id": "Qwen/Qwen3.6-35B-A3B-FP8", - "name": "Qwen 3.6 35B A3B FP8", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -42388,35 +39972,28 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 8192 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.17, - "output": 1.1, - "cache_read": 0.056 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "zai-org/GLM-5.1-FP8": { - "id": "zai-org/GLM-5.1-FP8", - "name": "GLM-5.1 FP8", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax-m2.5", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "knowledge": "2025-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -42427,44 +40004,44 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 16384 + "context": 204800, + "output": 65536 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 1.4, - "output": 4.4 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.7-plus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -42472,61 +40049,60 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, + "input": 0.4, + "output": 1.6, + "cache_read": 0.04, + "cache_write": 0.5, "tiers": [ { - "input": 5, - "output": 22.5, - "cache_read": 0.5, + "input": 1.2, + "output": 4.8, + "cache_read": 0.12, + "cache_write": 1.5, "tier": { "type": "context", - "size": 272000 + "size": 256000 } } ], "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 + "input": 1.2, + "output": 4.8, + "cache_read": 0.12, + "cache_write": 1.5 } } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "qwen3.7-max", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -42534,30 +40110,41 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "openai/whisper-large-v3": { - "id": "openai/whisper-large-v3", - "name": "Whisper Large v3", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": false, - "release_date": "2023-11-06", - "last_updated": "2023-11-06", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "audio" + "text", + "image", + "video" ], "output": [ "text" @@ -42565,125 +40152,118 @@ }, "open_weights": true, "limit": { - "context": 448, - "output": 448 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.01, - "output": 0.01 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 202752, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 1.74, + "output": 3.48, + "cache_read": 0.0145 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -42692,78 +40272,100 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 1.1, + "input": 1.4, "output": 4.4, - "cache_read": 0.55 + "cache_read": 0.26 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax-m3", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-01", + "release_date": "2026-05-31", + "last_updated": "2026-05-31", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "tiers": [ + { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12, + "tier": { + "type": "context", + "size": 512000 + } + } + ], + "context_over_200k": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 + } } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen3.5", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -42771,312 +40373,356 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 65536 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax-m2.7", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 + } + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2024-10", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 65536 }, + "status": "deprecated", "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo V2.5", + "description": "MiMo omni model for text, image, video, audio, and agents", + "family": "mimo-v2.5", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "mimo-v2-omni": { + "id": "mimo-v2-omni", + "name": "MiMo V2 Omni", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo-v2-omni", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, + "context": 262144, "output": 128000 }, + "status": "deprecated", "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.4, + "output": 2, + "cache_read": 0.08 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } + }, + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo V2 Pro", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo-v2-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, "output": 128000 }, + "status": "deprecated", "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, + "input": 1, + "output": 3, + "cache_read": 0.2, "tiers": [ { - "input": 10, - "output": 45, - "cache_read": 1, + "input": 2, + "output": 6, + "cache_read": 0.4, "tier": { "type": "context", - "size": 272000 + "size": 256000 } } ], "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 + "input": 2, + "output": 6, + "cache_read": 0.4 } } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo V2.5 Pro", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "family": "mimo-v2.5-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 1048576, + "output": 128000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 1.74, + "output": 3.48, + "cache_read": 0.0145 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-04", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, + "context": 202752, "output": 32768 }, + "status": "deprecated", "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -43084,27 +40730,60 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } } - }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", + } + } + }, + "digitalocean": { + "id": "digitalocean", + "env": [ + "DIGITALOCEAN_ACCESS_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.do-ai.run/v1", + "name": "DigitalOcean", + "doc": "https://docs.digitalocean.com/products/gradient-ai-platform/details/models/", + "models": { + "anthropic-claude-haiku-4.5": { + "id": "anthropic-claude-haiku-4.5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -43112,15 +40791,15 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -43128,71 +40807,58 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1, + "output": 5, + "cache_read": 1, + "cache_write": 1.25 } - } - } - }, - "wandb": { - "id": "wandb", - "env": [ - "WANDB_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.inference.wandb.ai/v1", - "name": "CoreWeave", - "doc": "https://docs.wandb.ai/inference", - "models": { - "JetBrains/Mellum2-12B-A2.5B-Instruct": { - "id": "JetBrains/Mellum2-12B-A2.5B-Instruct", - "name": "Mellum2 12B A2.5B", - "description": "Mellum2-12B-A2.5B-Instruct is a fast MoE model with 131K context built for coding, tool use, and low-latency AI workflows.", - "attachment": false, + }, + "openai-gpt-image-1": { + "id": "openai-gpt-image-1", + "name": "GPT Image 1", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "tool_call": false, + "temperature": false, + "release_date": "2025-04-24", + "last_updated": "2025-04-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 0, + "output": 0 }, "cost": { - "input": 0.05, - "output": 0.1, - "cache_read": 0.05 + "input": 5, + "output": 40, + "cache_read": 1.25 } }, - "meta-llama/Llama-3.1-70B-Instruct": { - "id": "meta-llama/Llama-3.1-70B-Instruct", - "name": "Llama 3.1 70B", - "description": "Efficient conversational model optimized for responsive multilingual chatbot interactions.", - "family": "llama", + "e5-large-v2": { + "id": "e5-large-v2", + "name": "E5 Large v2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "tool_call": false, + "temperature": false, + "release_date": "2023-05-19", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" @@ -43203,28 +40869,25 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 512, + "output": 1024 }, "cost": { - "input": 0.8, - "output": 0.8, - "cache_read": 0.8 + "input": 0.02, + "output": 0 } }, - "meta-llama/Llama-3.1-8B-Instruct": { - "id": "meta-llama/Llama-3.1-8B-Instruct", - "name": "Llama 3.1 8B", - "description": "Efficient conversational model optimized for responsive multilingual chatbot interactions.", - "family": "llama", + "bge-m3": { + "id": "bge-m3", + "name": "BGE M3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "bge", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "tool_call": false, + "temperature": false, + "release_date": "2024-01-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" @@ -43235,28 +40898,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 8192, + "output": 1024 }, "cost": { - "input": 0.22, - "output": 0.22, - "cache_read": 0.22 + "input": 0.02, + "output": 0 } }, - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama 3.3 70B", - "description": "Multilingual model excelling in conversational tasks, detailed instruction-following, and coding.", - "family": "llama", + "mistral-3-14B": { + "id": "mistral-3-14B", + "name": "Ministral 3 14B Instruct", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "release_date": "2025-12-15", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" @@ -43267,73 +40928,64 @@ }, "open_weights": true, "limit": { - "context": 128000, + "context": 262144, "output": 128000 }, "cost": { - "input": 0.71, - "output": 0.71, - "cache_read": 0.71 + "input": 0.2, + "output": 0.2 } }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B", - "description": "Gemma 4 31B Dense is designed for advanced reasoning, agentic workflows, and longer context and is natively trained on 140+ languages.", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "nemotron-3-ultra-550b": { + "id": "nemotron-3-ultra-550b", + "name": "Nemotron 3 Ultra", + "description": "Flagship Nemotron model for high-throughput reasoning and complex agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-06-04", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.1, - "output": 0.34, - "cache_read": 0.1 + "context": 131072, + "output": 8192 } }, - "google/gemma-4-26B-A4B-it": { - "id": "google/gemma-4-26B-A4B-it", - "name": "Gemma 4 26B A4B", - "description": "Gemma 4 26B A4B is a multimodal MoE model with LoRA support and function calling for agentic workflows.", - "family": "gemma", - "attachment": true, + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax-m2.5", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-08", + "release_date": "2026-02-12", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -43341,32 +40993,40 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 204800, + "output": 128000 }, + "status": "beta", "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.05 + "input": 0.3, + "output": 1.2 } }, - "Qwen/Qwen3.8-27B": { - "id": "Qwen/Qwen3.8-27B", - "name": "Qwen3.8 27B", - "description": "Qwen3.8-27B is a dense multimodal model suited for coding, research, vision, and long-running agent tasks.", - "family": "qwen", + "openai-gpt-5.4-nano": { + "id": "openai-gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -43376,29 +41036,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 3, - "cache_read": 0.15 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "Qwen/Qwen3-30B-A3B-Instruct-2507": { - "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen3 30B A3B Instruct 2507", - "description": "Qwen3-30B-A3B-Instruct-2507 is a 30.5B MoE instruction-tuned model with enhanced reasoning, coding, and long-context understanding.", - "family": "qwen", + "deepseek-v3": { + "id": "deepseek-v3", + "name": "DeepSeek V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "knowledge": "2024-07", + "release_date": "2024-12-26", + "last_updated": "2025-03-24", "modalities": { "input": [ "text" @@ -43409,69 +41069,61 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.1 + "context": 163840, + "output": 131072 } }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B A3B", - "description": "Qwen3.6-35B-A3B is an MoE multimodal model with 262K context optimized for agentic coding workflows.", - "family": "qwen3.6", + "openai-gpt-image-2": { + "id": "openai-gpt-image-2", + "name": "GPT Image 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-15", - "last_updated": "2026-04-15", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-04-24", + "last_updated": "2025-04-24", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.25 + "context": 0, + "output": 0 } }, - "Qwen/Qwen3.5-35B-A3B": { - "id": "Qwen/Qwen3.5-35B-A3B", - "name": "Qwen3.5-35B-A3B", - "description": "Qwen3.5-35B-A3B is an open-weights multimodal MoE model built for efficient, high-throughput inference across chat, reasoning, and agentic tasks.", - "family": "qwen3.5", + "openai-gpt-5.2": { + "id": "openai-gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -43481,38 +41133,44 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.25 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen3.6-27B is a 27B dense multimodal model with 262K context built for flagship-level agentic coding.", - "family": "qwen3.6", - "attachment": true, + "deepseek-r1-distill-llama-70b": { + "id": "deepseek-r1-distill-llama-70b", + "name": "DeepSeek R1 Distill Llama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2025-01-30", + "last_updated": "2025-01-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -43520,32 +41178,25 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 3.6, - "cache_read": 0.12 + "input": 0.99, + "output": 0.99 } }, - "ibm-granite/granite-4.2-8b": { - "id": "ibm-granite/granite-4.2-8b", - "name": "Granite 4.2 8B", - "description": "Granite 4.2 8B is an instruct model capable of enhanced tool calling, instruction following, and chat capabilities.", - "family": "granite", + "qwen3-embedding-0.6b": { + "id": "qwen3-embedding-0.6b", + "name": "Qwen3 Embedding 0.6B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-24", - "last_updated": "2026-08-24", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-03", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -43556,30 +41207,31 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 8000, + "output": 1024 }, + "status": "beta", "cost": { - "input": 0.1, - "output": 0.15, - "cache_read": 0.05 + "input": 0.04, + "output": 0 } }, - "ibm-granite/granite-4.1-8b": { - "id": "ibm-granite/granite-4.1-8b", - "name": "Granite 4.1 8B", - "description": "Granite 4.1 8B is a long-context instruct model capable of enhanced tool calling, instruction following, and chat capabilities.", - "family": "granite", - "attachment": false, + "gemma-4-31B-it": { + "id": "gemma-4-31B-it", + "name": "Gemma 4 31B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "release_date": "2026-04-22", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -43587,30 +41239,31 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 256000, + "output": 8192 }, "cost": { - "input": 0.05, - "output": 0.1, - "cache_read": 0.05 + "input": 0.18, + "output": 0.5 } }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek V3.1", - "description": "A large hybrid model that supports both thinking and non-thinking modes via prompt templates.", - "family": "deepseek", - "attachment": false, + "llama-4-maverick": { + "id": "llama-4-maverick", + "name": "Llama 4 Maverick 17B 128E Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -43618,144 +41271,148 @@ }, "open_weights": true, "limit": { - "context": 161000, - "output": 161000 + "context": 1000000, + "output": 16384 }, "cost": { - "input": 0.55, - "output": 1.65, - "cache_read": 0.55 + "input": 0.25, + "output": 0.87 } }, - "deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "DeepSeek V4-Flash-0731 is an MoE model great for coding, reasoning, and agentic workloads.", - "family": "deepseek", - "attachment": false, + "anthropic-claude-3.7-sonnet": { + "id": "anthropic-claude-3.7-sonnet", + "name": "Claude 3.7 Sonnet", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2024-11", + "release_date": "2025-02-24", + "last_updated": "2025-02-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 64000 }, + "status": "deprecated", "cost": { - "input": 0.13, - "output": 0.28, - "cache_read": 0.07 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek-ai/DeepSeek-V4-Pro-0813": { - "id": "deepseek-ai/DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4-Pro-0813 is a 1.6T-parameter MoE model excelling at advanced reasoning, coding, and complex agentic workloads.", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "openai-gpt-4o-mini": { + "id": "openai-gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.31, - "output": 3.96, - "cache_read": 0.044 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "deepseek-ai/DeepSeek-V4.1-Flash": { - "id": "deepseek-ai/DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash is a multimodal MoE model for coding, reasoning, and agentic workloads with long contexts.", - "family": "deepseek", + "anthropic-claude-opus-4.7": { + "id": "anthropic-claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.65, - "cache_read": 0.03 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "DeepSeek V4-Flash is an MoE model with 1M context length great for coding, reasoning, and agentic workloads.", + "deepseek-4-flash": { + "id": "deepseek-4-flash", + "name": "Deepseek V4 Flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-05-27", + "last_updated": "2026-05-29", "modalities": { "input": [ "text" @@ -43764,135 +41421,190 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.07 + "context": 262144, + "output": 8192 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "DeepSeek V4-Pro is a 1.6T-parameter MoE model with 49B active parameters excelling at advanced reasoning, coding, and complex agentic workloads.", - "family": "deepseek", - "attachment": false, + "anthropic-claude-4.5-haiku": { + "id": "anthropic-claude-4.5-haiku", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.15, - "output": 2.55, - "cache_read": 0.2 + "input": 1, + "output": 5, + "cache_read": 1, + "cache_write": 1.25 } }, - "MiniMaxAI/MiniMax-M3": { - "id": "MiniMaxAI/MiniMax-M3", - "name": "MiniMax M3", - "description": "MiniMax M3 is a multimodal MoE model with 23B active parameters optimized for coding and agentic workflows.", - "family": "minimax-m3", + "anthropic-claude-4.6-sonnet": { + "id": "anthropic-claude-4.6-sonnet", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.23, - "output": 0.96, - "cache_read": 0.05 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.3, + "cache_write": 3.75, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.3, + "cache_write": 3.75 + } } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Kimi K2.6 is a multimodal Mixture-of-Experts language model featuring 32 billion activated parameters and a total of 1 trillion parameters.", - "family": "kimi-k2", + "anthropic-claude-sonnet-4": { + "id": "anthropic-claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.65, - "output": 3.41, - "cache_read": 0.15 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.3, + "cache_write": 3.75, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.3, + "cache_write": 3.75 + } } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Kimi K2.7 Code is a 1T-parameter MoE model with 32B active parameters purpose-built for long-horizon agentic coding and software engineering.", - "family": "kimi-k2", + "ministral-3-8b-instruct-2512": { + "id": "ministral-3-8b-instruct-2512", + "name": "Ministral 3 8B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ "text", @@ -43906,99 +41618,148 @@ "limit": { "context": 262144, "output": 262144 - }, - "cost": { - "input": 0.71, - "output": 3.5, - "cache_read": 0.15 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM 5.2", - "description": "GLM-5.2 is a Mixture-of-Experts language model featuring 40 billion activated parameters and a total of 744 billion parameters.", - "family": "glm", - "attachment": false, + "anthropic-claude-4.5-sonnet": { + "id": "anthropic-claude-4.5-sonnet", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.76, - "output": 2.42, - "cache_read": 0.14 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.3, + "cache_write": 3.75, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.3, + "cache_write": 3.75 + } } }, - "zai-org/GLM-5.3-Flash": { - "id": "zai-org/GLM-5.3-Flash", - "name": "GLM 5.3 Flash", - "description": "GLM-5.3-Flash is a natively multimodal model with 320B total parameters and 18B active parameters.", - "family": "glm", + "anthropic-claude-opus-4.6": { + "id": "anthropic-claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.05 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 0.5, + "cache_write": 6.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 0.5, + "cache_write": 6.25 + } } }, - "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B": { - "id": "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B", - "name": "Nemotron 3.5 Lightning", - "description": "Nemotron 3.5 Lightning is an MoE model built for fast, reliable agentic tasks across use cases such as financial services, cybersecurity, telecom, and retail.", - "family": "nemotron", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -44009,32 +41770,42 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.07, - "output": 0.2, - "cache_read": 0.04 + "input": 0.975, + "output": 4.3, + "cache_read": 0.26 } }, - "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B": { - "id": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B", - "name": "Nemotron 3 Ultra", - "description": "Nemotron 3 Ultra is a powerful MoE model designed for long-running agents across coding, deep research, and enterprise automation.", - "family": "nemotron", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -44045,28 +41816,37 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.5, - "output": 2.15, - "cache_read": 0.1 + "input": 1.74, + "output": 3.48 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", + "openai-gpt-oss-20b": { + "id": "openai-gpt-oss-20b", "name": "gpt-oss-20b", - "description": "Lower latency Mixture-of-Experts model trained on OpenAI's Harmony response format with reasoning capabilities.", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, + "knowledge": "2024-06", "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -44081,24 +41861,22 @@ "output": 131072 }, "cost": { - "input": 0.03, - "output": 0.13, - "cache_read": 0.03 + "input": 0.05, + "output": 0.45 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Efficient Mixture-of-Experts model designed for high-reasoning, agentic and general-purpose use cases.", - "family": "gpt-oss", + "qwen-2.5-14b-instruct": { + "id": "qwen-2.5-14b-instruct", + "name": "Qwen 2.5 14B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-09", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", "modalities": { "input": [ "text" @@ -44111,78 +41889,34 @@ "limit": { "context": 131072, "output": 131072 - }, - "cost": { - "input": 0.03, - "output": 0.17, - "cache_read": 0.03 } }, - "OpenPipe/Qwen3-14B-Instruct": { - "id": "OpenPipe/Qwen3-14B-Instruct", - "name": "Qwen3 14B Instruct", - "description": "An efficient multilingual, dense, instruction-tuned model, optimized by OpenPipe for building agents with finetuning.", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.05, - "output": 0.22, - "cache_read": 0.05 - } - } - } - }, - "subconscious": { - "id": "subconscious", - "env": [ - "SUBCONSCIOUS_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://api.subconscious.dev/v1", - "name": "Subconscious", - "doc": "https://docs.subconscious.dev", - "models": { - "subconscious/tim-qwen3.6-27b": { - "id": "subconscious/tim-qwen3.6-27b", - "name": "TIM-Qwen3.6 27B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "anthropic-claude-opus-4.8": { + "id": "anthropic-claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-05-11", - "last_updated": "2026-05-11", + "release_date": "2026-05-28", + "last_updated": "2026-05-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -44190,78 +41924,30 @@ }, "open_weights": false, "limit": { - "context": 8192, - "input": 8192, - "output": 5000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 3, - "cache_read": 0.15 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "subconscious/glm-5.2": { - "id": "subconscious/glm-5.2", + "glm-5.2": { + "id": "glm-5.2", "name": "GLM-5.2", "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - } - } - }, - "zeldoc": { - "id": "zeldoc", - "env": [ - "ZELDOC_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.zeldoc.ai/v1", - "name": "Zeldoc", - "doc": "https://docs.zeldoc.ai", - "models": { - "zdev": { - "id": "zdev", - "name": "ZDev", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": true, - "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", + "medium", "high", "max" ] @@ -44273,56 +41959,47 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.05, + "output": 4.4, + "cache_read": 0.21 } - } - } - }, - "databricks": { - "id": "databricks", - "env": [ - "DATABRICKS_HOST", - "DATABRICKS_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://${DATABRICKS_HOST}/ai-gateway/mlflow/v1", - "name": "Databricks", - "doc": "https://docs.databricks.com/aws/en/machine-learning/foundation-models/", - "models": { - "databricks-claude-opus-4-5": { - "id": "databricks-claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", + }, + "anthropic-claude-opus-4.5": { + "id": "anthropic-claude-opus-4.5", + "name": "Claude Opus 4.5", "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "knowledge": "2025-05", + "knowledge": "2025-03-31", "release_date": "2025-11-24", "last_updated": "2025-11-24", "modalities": { @@ -44347,11 +42024,11 @@ "cache_write": 6.25 } }, - "databricks-gpt-5-6-terra": { - "id": "databricks-gpt-5-6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", + "anthropic-claude-4.1-opus": { + "id": "anthropic-claude-4.1-opus", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -44365,11 +42042,10 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -44382,60 +42058,31 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "databricks-kimi-k2-7-code": { - "id": "databricks-kimi-k2-7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "llama3-8b-instruct": { + "id": "llama3-8b-instruct", + "name": "Llama 3.1 Instruct (8B)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -44443,86 +42090,69 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.198, + "output": 0.198 } }, - "databricks-glm-5-2": { - "id": "databricks-glm-5-2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "stable-diffusion-3.5-large": { + "id": "stable-diffusion-3.5-large", + "name": "Stable Diffusion 3.5 Large", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "stable-diffusion", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-10-22", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 256, + "output": 1 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.08, + "output": 0 } }, - "databricks-gemini-3-1-flash-lite": { - "id": "databricks-gemini-3-1-flash-lite", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "openai-gpt-5.4-pro": { + "id": "openai-gpt-5.4-pro", + "name": "GPT-5.4 pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -44530,126 +42160,87 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 30, + "output": 180 } }, - "databricks-gpt-5-mini": { - "id": "databricks-gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "all-mini-lm-l6-v2": { + "id": "all-mini-lm-l6-v2", + "name": "All-MiniLM-L6-v2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2021-08-30", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256, + "output": 384 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.009, + "output": 0 } }, - "databricks-gemini-3-flash": { - "id": "databricks-gemini-3-flash", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "openai-gpt-image-1.5": { + "id": "openai-gpt-image-1.5", + "name": "GPT Image 1.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 0, + "output": 0 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 + "input": 5, + "output": 10, + "cache_read": 1 } }, - "databricks-claude-opus-4-7": { - "id": "databricks-claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "openai-gpt-4o": { + "id": "openai-gpt-4o", + "name": "GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", @@ -44662,42 +42253,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "context": 128000, + "output": 16384 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "databricks-gpt-oss-20b": { - "id": "databricks-gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "anthropic-claude-opus-4": { + "id": "anthropic-claude-opus-4", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -44710,10 +42280,43 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "bge-reranker-v2-m3": { + "id": "bge-reranker-v2-m3", + "name": "BGE Reranker v2 M3", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "bge", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-03-12", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" @@ -44724,25 +42327,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 8192, + "output": 1 }, "cost": { - "input": 0.05, - "output": 0.2 + "input": 0.01, + "output": 0 } }, - "databricks-gpt-5-4-mini": { - "id": "databricks-gpt-5-4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "openai-gpt-5": { + "id": "openai-gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -44751,10 +42355,10 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -44767,37 +42371,20 @@ "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "databricks-gemini-3-pro": { - "id": "databricks-gemini-3-pro", - "name": "Gemini 3 Pro Preview", - "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", - "family": "gemini-pro", - "attachment": true, + "openai-o3-mini": { + "id": "openai-o3-mini", + "name": "o3-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -44811,17 +42398,13 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -44829,37 +42412,50 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 + } + }, + "multi-qa-mpnet-base-dot-v1": { + "id": "multi-qa-mpnet-base-dot-v1", + "name": "Multi-QA-mpnet-base-dot-v1", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2021-08-30", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 512, + "output": 768 + }, + "cost": { + "input": 0.009, + "output": 0 } }, - "databricks-gpt-5-2": { - "id": "databricks-gpt-5-2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -44873,152 +42469,120 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.5, + "output": 2.7 } }, - "databricks-claude-haiku-4-5": { - "id": "databricks-claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "llama3.3-70b-instruct": { + "id": "llama3.3-70b-instruct", + "name": "Llama 3.3 Instruct 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.65, + "output": 0.65 } }, - "databricks-gemini-3-1-pro": { - "id": "databricks-gemini-3-1-pro", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "gte-large-en-v1.5": { + "id": "gte-large-en-v1.5", + "name": "GTE Large (v1.5)", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-03-27", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 8192, + "output": 1024 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.09, + "output": 0 } }, - "databricks-claude-opus-4-6": { - "id": "databricks-claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "openai-gpt-5.4-mini": { + "id": "openai-gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -45026,60 +42590,40 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } - }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "databricks-gpt-5-6-sol": { - "id": "databricks-gpt-5-6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + "openai-gpt-5.5": { + "id": "openai-gpt-5.5", + "name": "GPT-5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", @@ -45092,8 +42636,7 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, "cost": { @@ -45118,49 +42661,46 @@ } } }, - "databricks-claude-opus-4-1": { - "id": "databricks-claude-opus-4-1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "nemotron-3-nano-30b": { + "id": "nemotron-3-nano-30b", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "context": 262144, + "output": 262144 } }, - "databricks-gpt-oss-120b": { - "id": "databricks-gpt-oss-120b", - "name": "GPT OSS 120B", + "openai-gpt-oss-120b": { + "id": "openai-gpt-oss-120b", + "name": "gpt-oss-120b", "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "family": "gpt-oss", "attachment": false, @@ -45176,10 +42716,14 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, + "knowledge": "2024-06", "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -45191,40 +42735,41 @@ "open_weights": true, "limit": { "context": 131072, - "output": 32768 + "output": 131072 }, "cost": { - "input": 0.072, - "output": 0.28 + "input": 0.1, + "output": 0.7 } }, - "databricks-gemini-2-5-pro": { - "id": "databricks-gemini-2-5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "openai-gpt-5-nano": { + "id": "openai-gpt-5-nano", + "name": "GPT-5 nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -45232,203 +42777,135 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "databricks-gemini-2-5-flash": { - "id": "databricks-gemini-2-5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "wan2-2-t2v-a14b": { + "id": "wan2-2-t2v-a14b", + "name": "Wan2.2-T2V-A14B", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-07-28", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ - "text" + "video" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 100, + "output": 1 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 + "input": 0.6, + "output": 0 } }, - "databricks-claude-sonnet-4-5": { - "id": "databricks-claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "mistral-7b-instruct-v0.3": { + "id": "mistral-7b-instruct-v0.3", + "name": "Mistral 7B Instruct v0.3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2024-05-22", + "last_updated": "2024-05-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 32768, + "output": 32768 + } + }, + "qwen3-tts-voicedesign": { + "id": "qwen3-tts-voicedesign", + "name": "Qwen3 TTS VoiceDesign", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "open_weights": true, + "limit": { + "context": 32768, + "output": 1 } }, - "databricks-gpt-5-4": { - "id": "databricks-gpt-5-4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "mistral-nemo-instruct-2407": { + "id": "mistral-nemo-instruct-2407", + "name": "Mistral Nemo Instruct", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mistral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 128000, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0.3, + "output": 0.3 } }, - "databricks-gpt-5-4-nano": { - "id": "databricks-gpt-5-4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -45436,11 +42913,14 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -45450,63 +42930,65 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.95, + "output": 4 } }, - "databricks-claude-sonnet-4-6": { - "id": "databricks-claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, + "deepseek-3.2": { + "id": "deepseek-3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2024-12", + "release_date": "2025-12-02", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 128000, "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.5, + "output": 1.6 } }, - "databricks-gpt-5-6-luna": { - "id": "databricks-gpt-5-6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", + "openai-o1": { + "id": "openai-o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -45522,9 +43004,9 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text", @@ -45537,145 +43019,116 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1, - "tiers": [ - { - "input": 2, - "output": 9, - "cache_read": 0.2, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 9, - "cache_read": 0.2 - } + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "databricks-claude-sonnet-4": { - "id": "databricks-claude-sonnet-4", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen 3.5 397B A17B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen3.5", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-02-15", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 81920 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.55, + "output": 3.5 } }, - "databricks-gpt-5-1": { - "id": "databricks-gpt-5-1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "qwen3-coder-flash": { + "id": "qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.45, + "output": 1.7 } }, - "databricks-gpt-5-nano": { - "id": "databricks-gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "openai-gpt-5.3-codex": { + "id": "openai-gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -45684,43 +43137,32 @@ "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "databricks-gpt-5": { - "id": "databricks-gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "openai-gpt-4.1": { + "id": "openai-gpt-4.1", + "name": "GPT-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -45728,20 +43170,52 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "databricks-gpt-5-5": { - "id": "databricks-gpt-5-5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "anthropic-claude-3.5-haiku": { + "id": "anthropic-claude-3.5-haiku", + "name": "Claude 3.5 Haiku", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-haiku", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-11-05", + "last_updated": "2024-11-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "status": "deprecated", + "cost": { + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 + } + }, + "openai-gpt-5.4": { + "id": "openai-gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "family": "gpt", "attachment": true, "reasoning": true, @@ -45749,18 +43223,20 @@ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -45773,90 +43249,39 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 2.5, + "output": 15, + "cache_read": 0.25 } - } - } - }, - "umans-ai-coding-plan": { - "id": "umans-ai-coding-plan", - "env": [ - "UMANS_AI_CODING_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.code.umans.ai/v1", - "name": "Umans AI Coding Plan", - "doc": "https://app.umans.ai/offers/code/docs", - "models": { - "umans-glm-5.3-flash": { - "id": "umans-glm-5.3-flash", - "name": "GLM 5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + }, + "arcee-trinity-large-thinking": { + "id": "arcee-trinity-large-thinking", + "name": "Trinity Large Thinking", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "trinity", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-04-02", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -45864,27 +43289,24 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131071 + "context": 256000, + "output": 128000 }, + "status": "beta", "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 0.9, + "cache_read": 0.06 } }, - "umans-flash": { - "id": "umans-flash", - "name": "Umans Flash", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "openai-o3": { + "id": "openai-o3", + "name": "o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -45895,162 +43317,138 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "umans-deepseek-v4-pro-0813": { - "id": "umans-deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "anthropic-claude-3.5-sonnet": { + "id": "anthropic-claude-3.5-sonnet", + "name": "Claude 3.5 Sonnet", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-sonnet", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2024-04", + "release_date": "2024-06-20", + "last_updated": "2024-10-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 393215 + "context": 200000, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "umans-deepseek-v4-flash-0731": { - "id": "umans-deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "openai-gpt-5-mini": { + "id": "openai-gpt-5-mini", + "name": "GPT-5 mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "minimal", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 393215 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "umans-kimi-k3": { - "id": "umans-kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "nvidia-nemotron-3-super-120b": { + "id": "nvidia-nemotron-3-super-120b", + "name": "Nemotron-3-Super-120B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", + "medium", "high", "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2026-02", + "release_date": "2026-03-11", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -46058,33 +43456,31 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 256000, + "output": 32768 }, + "status": "beta", "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.3, + "output": 0.65 } }, - "umans-qwen3.6-35b-a3b": { - "id": "umans-qwen3.6-35b-a3b", - "name": "Qwen3.6 35B A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, + "glm-5": { + "id": "glm-5", + "name": "GLM 5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "max" ] } ], @@ -46092,14 +43488,11 @@ "interleaved": { "field": "reasoning_content" }, - "structured_output": true, - "temperature": false, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-02-11", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -46107,132 +43500,94 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 202752, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1, + "output": 3.2 } }, - "umans-coder": { - "id": "umans-coder", - "name": "Umans Coder", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "openai-gpt-5.2-pro": { + "id": "openai-gpt-5.2-pro", + "name": "GPT-5.2 pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": false, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131071 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 21, + "output": 168 } - } - } - }, - "kuae-cloud-coding-plan": { - "id": "kuae-cloud-coding-plan", - "env": [ - "KUAE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://coding-plan-endpoint.kuaecloud.net/v1", - "name": "KUAE Cloud Coding Plan", - "doc": "https://docs.mthreads.com/kuaecloud/kuaecloud-doc-online/coding_plan/", - "models": { - "GLM-4.7": { - "id": "GLM-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + }, + "anthropic-claude-3-opus": { + "id": "anthropic-claude-3-opus", + "name": "Claude 3 Opus", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-opus", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2023-08", + "release_date": "2024-02-29", + "last_updated": "2024-02-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } - } - } - }, - "umans-ai": { - "id": "umans-ai", - "env": [ - "UMANS_AI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.code.umans.ai/v1", - "name": "Umans AI", - "doc": "https://app.umans.ai/offers/code/docs/orgs", - "models": { - "umans-glm-5.3-flash": { - "id": "umans-glm-5.3-flash", - "name": "GLM 5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + }, + "anthropic-claude-fable-5": { + "id": "anthropic-claude-fable-5", + "name": "Anthropic Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -46240,72 +43595,61 @@ "type": "effort", "values": [ "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-06-09", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131071 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "context": 1000000, + "output": 128000 } }, - "umans-flash": { - "id": "umans-flash", - "name": "Umans Flash", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "nemotron-3-nano-omni": { + "id": "nemotron-3-nano-omni", + "name": "Nemotron Nano 3 Omni", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": true, + "release_date": "2026-04-28", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" @@ -46313,42 +43657,38 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 1, - "cache_read": 0.05 + "input": 0.5, + "output": 0.9 } }, - "umans-deepseek-v4-pro-0813": { - "id": "umans-deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "alibaba-qwen3-32b": { + "id": "alibaba-qwen3-32b", + "name": "Qwen3-32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "low", + "medium", "high", "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2025-04-30", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -46359,47 +43699,42 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393215 + "context": 131000, + "output": 40960 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 + "input": 0.25, + "output": 0.55 } }, - "umans-deepseek-v4-flash-0731": { - "id": "umans-deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "nemotron-nano-12b-v2-vl": { + "id": "nemotron-nano-12b-v2-vl", + "name": "Nemotron Nano 12B v2 VL", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", + "medium", "high", "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2024-10", + "release_date": "2025-12-01", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -46407,43 +43742,38 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393215 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0.2, + "output": 0.6 } }, - "umans-kimi-k3": { - "id": "umans-kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "openai-gpt-5.1-codex-max": { + "id": "openai-gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -46453,151 +43783,133 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "umans-coder": { - "id": "umans-coder", - "name": "Umans Coder", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "fal-ai/fast-sdxl": { + "id": "fal-ai/fast-sdxl", + "name": "Fast SDXL", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "stable-diffusion", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2023-07-26", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131071 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "context": 0, + "output": 0 } - } - } - }, - "siliconflow": { - "id": "siliconflow", - "env": [ - "SILICONFLOW_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.siliconflow.com/v1", - "name": "SiliconFlow", - "doc": "https://cloud.siliconflow.com/models", - "models": { - "ByteDance-Seed/Seed-OSS-36B-Instruct": { - "id": "ByteDance-Seed/Seed-OSS-36B-Instruct", - "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "seed", + }, + "fal-ai/elevenlabs/tts/multilingual-v2": { + "id": "fal-ai/elevenlabs/tts/multilingual-v2", + "name": "ElevenLabs Multilingual TTS v2", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "elevenlabs", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-04", - "last_updated": "2025-11-25", + "tool_call": false, + "temperature": false, + "release_date": "2023-08-22", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.21, - "output": 0.57 + "context": 0, + "output": 0 } }, - "nex-agi/Nex-N2-Pro": { - "id": "nex-agi/Nex-N2-Pro", - "name": "Nex-N2-Pro", - "description": "Open agentic MoE model (397B total, 17B active) for coding, tool use, and research workflows", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "fal-ai/flux/schnell": { + "id": "fal-ai/flux/schnell", + "name": "FLUX.1 [schnell]", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-08-01", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": true, "limit": { - "context": 262144, - "output": 256000 - }, - "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.25 + "context": 0, + "output": 0 } }, - "tencent/Hunyuan-A13B-Instruct": { - "id": "tencent/Hunyuan-A13B-Instruct", - "name": "tencent/Hunyuan-A13B-Instruct", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", + "fal-ai/stable-audio-25/text-to-audio": { + "id": "fal-ai/stable-audio-25/text-to-audio", + "name": "Stable Audio 2.5 (Text-to-Audio)", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-10-08", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "output": 0 + } + } + } + }, + "subconscious": { + "id": "subconscious", + "env": [ + "SUBCONSCIOUS_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://api.subconscious.dev/v1", + "name": "Subconscious", + "doc": "https://docs.subconscious.dev", + "models": { + "subconscious/glm-5.2": { + "id": "subconscious/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -46605,16 +43917,14 @@ "type": "toggle" }, { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-11-25", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -46623,34 +43933,36 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "tencent/Hy3": { - "id": "tencent/Hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + "subconscious/tim-qwen3.6-27b": { + "id": "subconscious/tim-qwen3.6-27b", + "name": "TIM-Qwen3.6 27B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ "text" @@ -46659,39 +43971,52 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 192000, - "output": 262144 + "context": 8192, + "input": 8192, + "output": 5000 }, "cost": { - "input": 0.132, - "output": 0.528, - "cache_read": 0.033 + "input": 0.3, + "output": 3, + "cache_read": 0.15 } - }, - "meituan-longcat/LongCat-2.0": { - "id": "meituan-longcat/LongCat-2.0", - "name": "LongCat-2.0", - "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", - "family": "longcat", + } + } + }, + "venice": { + "id": "venice", + "env": [ + "VENICE_API_KEY" + ], + "npm": "venice-ai-sdk-provider", + "name": "Venice AI", + "doc": "https://docs.venice.ai", + "models": { + "z-ai-glm-5-turbo": { + "id": "z-ai-glm-5-turbo", + "name": "GLM 5 Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "release_date": "2026-03-15", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -46700,32 +44025,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1049000, - "output": 131072 + "context": 200000, + "output": 32768 }, "cost": { - "input": 0.75, - "output": 2.95, - "cache_read": 0.015 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "google/gemma-4-12B-it": { - "id": "google/gemma-4-12B-it", - "name": "Gemma 4 12B IT", - "description": "Compact Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": false, - "reasoning": false, - "tool_call": true, + "grok-4-20-multi-agent": { + "id": "grok-4-20-multi-agent", + "name": "Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "structured_output": true, - "temperature": true, - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "release_date": "2026-03-12", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -46733,26 +44059,48 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 2000000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 1.42, + "output": 2.83, + "cache_read": 0.23, + "tiers": [ + { + "input": 2.83, + "output": 5.67, + "cache_read": 0.45, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.83, + "output": 5.67, + "cache_read": 0.45 + } } }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -46761,89 +44109,78 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 0.138, + "output": 0.275, + "cache_read": 0.028 } }, - "google/gemma-4-26B-A4B-it": { - "id": "google/gemma-4-26B-A4B-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "google-gemma-4-31b-it": { + "id": "google-gemma-4-31b-it", + "name": "Google Gemma 4 31B Instruct", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", "family": "gemma", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-04-03", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 - }, + "context": 256000, + "output": 8192 + }, "cost": { "input": 0.12, - "output": 0.4 + "output": 0.36, + "cache_read": 0.09 } }, - "Qwen/Qwen3-VL-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", - "name": "Qwen/Qwen3-VL-30B-A3B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "kimi-k2-6": { + "id": "kimi-k2-6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-05", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262000, - "output": 262000 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.29, - "output": 1 - } - }, - "Qwen/Qwen3-VL-8B-Instruct": { - "id": "Qwen/Qwen3-VL-8B-Instruct", - "name": "Qwen/Qwen3-VL-8B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-11-25", + "knowledge": "2025-01", + "release_date": "2026-04-20", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -46853,29 +44190,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.18, - "output": 0.68 + "input": 0.75, + "output": 3.5, + "cache_read": 0.16 } }, - "Qwen/Qwen3-VL-32B-Thinking": { - "id": "Qwen/Qwen3-VL-32B-Thinking", - "name": "Qwen/Qwen3-VL-32B-Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai-gpt-56-terra-pro": { + "id": "openai-gpt-56-terra-pro", + "name": "GPT-5.6 Terra Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-10-21", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -46887,56 +44237,26 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.5 + "input": 3.125, + "output": 18.75, + "cache_read": 0.3125 } }, - "Qwen/Qwen3-30B-A3B-Instruct-2507": { - "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "Qwen 3 235B A22B Instruct 2507", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-07-30", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.09, - "output": 0.3 - } - }, - "Qwen/Qwen3.5-27B": { - "id": "Qwen/Qwen3.5-27B", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2025-04-29", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -46945,31 +44265,45 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 2 + "input": 0.15, + "output": 0.75 } }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai-gpt-56-sol-pro": { + "id": "openai-gpt-56-sol-pro", + "name": "GPT-5.6 Sol Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -46977,36 +44311,38 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.6 + "input": 6.25, + "output": 37.5, + "cache_read": 0.625 } }, - "Qwen/Qwen3-14B": { - "id": "Qwen/Qwen3-14B", - "name": "Qwen/Qwen3-14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "nvidia-nemotron-cascade-2-30b-a3b": { + "id": "nvidia-nemotron-cascade-2-30b-a3b", + "name": "Nemotron Cascade 2 30B A3B", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", + "release_date": "2026-03-24", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -47015,29 +44351,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 256000, + "output": 32768 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 0.14, + "output": 0.8 } }, - "Qwen/Qwen3-VL-30B-A3B-Thinking": { - "id": "Qwen/Qwen3-VL-30B-A3B-Thinking", - "name": "Qwen/Qwen3-VL-30B-A3B-Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-opus-4-7-fast": { + "id": "claude-opus-4-7-fast", + "name": "Claude Opus 4.7 Fast", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-10-11", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-05-14", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -47049,39 +44386,45 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.29, - "output": 1 + "input": 36, + "output": 180, + "cache_read": 3.6, + "cache_write": 45 } }, - "Qwen/Qwen3-8B": { - "id": "Qwen/Qwen3-8B", - "name": "Qwen/Qwen3-8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "openai-gpt-55-pro": { + "id": "openai-gpt-55-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-24", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -47089,26 +44432,25 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.06, - "output": 0.06 + "input": 37.5, + "output": 225 } }, - "Qwen/Qwen2.5-72B-Instruct": { - "id": "Qwen/Qwen2.5-72B-Instruct", - "name": "Qwen/Qwen2.5-72B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "llama-3.3-70b": { + "id": "llama-3.3-70b", + "name": "Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2025-11-25", + "release_date": "2025-04-06", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -47117,71 +44459,77 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 33000, - "output": 4000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.59, - "output": 0.59 + "input": 0.7, + "output": 2.8 } }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen/Qwen3-32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen3-5-397b-a17b": { + "id": "qwen3-5-397b-a17b", + "name": "Qwen 3.5 397B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", + "release_date": "2026-02-16", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 0.75, + "output": 4.5 } }, - "Qwen/Qwen3-Coder-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "name": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-01", - "last_updated": "2025-11-25", + "knowledge": "2025-05", + "release_date": "2025-12-06", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -47189,56 +44537,42 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 198000, + "output": 32768 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 6, + "output": 30, + "cache_read": 0.6, + "cache_write": 7.5 } }, - "Qwen/Qwen2.5-7B-Instruct": { - "id": "Qwen/Qwen2.5-7B-Instruct", - "name": "Qwen/Qwen2.5-7B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "zai-org-glm-5": { + "id": "zai-org-glm-5", + "name": "GLM 5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 33000, - "output": 4000 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.05, - "output": 0.05 - } - }, - "Qwen/Qwen3.5-35B-A3B": { - "id": "Qwen/Qwen3.5-35B-A3B", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-02-11", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -47247,91 +44581,110 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 198000, + "output": 32000 }, "cost": { - "input": 0.24, - "output": 1.8 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "Qwen/Qwen3.5-122B-A10B": { - "id": "Qwen/Qwen3.5-122B-A10B", - "name": "Qwen3.5 122B-A10B", + "qwen3-5-35b-a3b": { + "id": "qwen3-5-35b-a3b", + "name": "Qwen 3.5 35B A3B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-02-25", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0.26, - "output": 2.08 + "input": 0.3125, + "output": 1.25, + "cache_read": 0.15625 } }, - "Qwen/Qwen3.5-9B": { - "id": "Qwen/Qwen3.5-9B", - "name": "Qwen/Qwen3.5-9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "venice-uncensored-role-play": { + "id": "venice-uncensored-role-play", + "name": "Venice Role Play Uncensored", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "venice", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-04-24", + "release_date": "2026-02-20", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.1, - "output": 0.15 + "input": 0.5, + "output": 2 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "qwen-3-6-plus": { + "id": "qwen-3-6-plus", + "name": "Qwen 3.6 Plus Uncensored", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", "family": "qwen", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "knowledge": "2025-04", + "release_date": "2026-04-06", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -47339,33 +44692,60 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.39, - "output": 2.34 + "input": 0.625, + "output": 3.75, + "cache_read": 0.0625, + "cache_write": 0.78, + "tiers": [ + { + "input": 2.5, + "output": 7.5, + "cache_read": 0.0625, + "cache_write": 0.78, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 7.5, + "cache_read": 0.0625, + "cache_write": 0.78 + } } }, - "Qwen/Qwen3.8-2.4T-A95B": { - "id": "Qwen/Qwen3.8-2.4T-A95B", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", + "zai-org-glm-4.6": { + "id": "zai-org-glm-4.6", + "name": "GLM 4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2025-04", + "release_date": "2024-04-01", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -47376,27 +44756,28 @@ }, "open_weights": true, "limit": { - "context": 1049000, - "output": 131000 + "context": 198000, + "output": 16384 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 0.43, + "output": 1.75, + "cache_read": 0.08 } }, - "Qwen/Qwen3-VL-32B-Instruct": { - "id": "Qwen/Qwen3-VL-32B-Instruct", - "name": "Qwen/Qwen3-VL-32B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai-gpt-4o-2024-11-20": { + "id": "openai-gpt-4o-2024-11-20", + "name": "GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-21", - "last_updated": "2025-11-25", + "knowledge": "2023-09", + "release_date": "2026-02-28", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -47408,29 +44789,41 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 3.125, + "output": 12.5 } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "name": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "grok-4-3": { + "id": "grok-4-3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-31", - "last_updated": "2025-11-25", + "release_date": "2026-04-18", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -47438,29 +44831,50 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 0.25, - "output": 1 + "input": 1.42, + "output": 2.83, + "cache_read": 0.23, + "tiers": [ + { + "input": 2.83, + "output": 5.67, + "cache_read": 0.45, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.83, + "output": 5.67, + "cache_read": 0.45 + } } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen-3-7-plus": { + "id": "qwen-3-7-plus", + "name": "Qwen 3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", "family": "qwen", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -47468,36 +44882,56 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 3.2 + "input": 0.5, + "output": 2, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 1.5, + "output": 6, + "cache_read": 0.15, + "cache_write": 1.875, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.5, + "output": 6, + "cache_read": 0.15, + "cache_write": 1.875 + } } }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "deepseek-ai/DeepSeek-V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "minimax-m27": { + "id": "minimax-m27", + "name": "MiniMax M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-25", - "last_updated": "2025-11-25", + "release_date": "2026-03-18", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -47506,42 +44940,44 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 164000, - "output": 164000 + "context": 198000, + "output": 32768 }, "cost": { - "input": 0.27, - "output": 1 + "input": 0.375, + "output": 1.5, + "cache_read": 0.06875 } }, - "deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "kimi-k2-7-code": { + "id": "kimi-k2-7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-13", + "last_updated": "2026-06-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -47549,34 +44985,32 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.014 + "input": 0.75, + "output": 3.5, + "cache_read": 0.16 } }, - "deepseek-ai/DeepSeek-R1": { - "id": "deepseek-ai/DeepSeek-R1", - "name": "deepseek-ai/DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-11-25", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -47585,28 +45019,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 164000, - "output": 164000 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 2.18 + "input": 1.65, + "output": 3.301, + "cache_read": 0.33 } }, - "deepseek-ai/DeepSeek-V4-Pro-0813": { - "id": "deepseek-ai/DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "qwen3-235b-a22b-thinking-2507": { + "id": "qwen3-235b-a22b-thinking-2507", + "name": "Qwen 3 235B A22B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, @@ -47614,9 +45052,8 @@ "field": "reasoning_content" }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2025-04-29", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -47627,124 +45064,100 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 + "input": 0.45, + "output": 3.5 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-07-31", + "release_date": "2025-01-15", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 198000, + "output": 64000 }, "cost": { - "input": 0.13, - "output": 0.28, - "cache_read": 0.028 + "input": 3.75, + "output": 18.75, + "cache_read": 0.375, + "cache_write": 4.69 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "openai-gpt-4o-mini-2024-07-18": { + "id": "openai-gpt-4o-mini-2024-07-18", + "name": "GPT-4o Mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-09", + "release_date": "2026-02-28", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.50162, - "output": 3.135, - "cache_read": 0.135 + "input": 0.1875, + "output": 0.75, + "cache_read": 0.09375 } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "deepseek-ai/DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -47752,30 +45165,34 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 0.42, - "cache_read": 0.135 + "input": 6, + "output": 30, + "cache_read": 0.6, + "cache_write": 7.5 } }, - "deepseek-ai/DeepSeek-V3": { - "id": "deepseek-ai/DeepSeek-V3", - "name": "deepseek-ai/DeepSeek-V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-29", + "last_updated": "2026-07-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -47783,39 +45200,45 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 1 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek-ai/DeepSeek-V3.1-Terminus": { - "id": "deepseek-ai/DeepSeek-V3.1-Terminus", - "name": "deepseek-ai/DeepSeek-V3.1-Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "openai-gpt-53-codex": { + "id": "openai-gpt-53-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-24", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -47823,39 +45246,45 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 1 + "input": 2.19, + "output": 17.5, + "cache_read": 0.219 } }, - "deepseek-ai/DeepSeek-V3.2-Exp": { - "id": "deepseek-ai/DeepSeek-V3.2-Exp", - "name": "deepseek-ai/DeepSeek-V3.2-Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "openai-gpt-54": { + "id": "openai-gpt-54", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-10-10", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -47863,36 +45292,41 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 1000000, + "input": 922000, + "output": 131072 }, "cost": { - "input": 0.27, - "output": 0.41 + "input": 3.13, + "output": 18.8, + "cache_read": 0.313 } }, - "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp": { - "id": "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", + "openai-gpt-56-terra": { + "id": "openai-gpt-56-terra", + "name": "GPT-5.6 Terra", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -47902,69 +45336,72 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 128000 }, "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.028 + "input": 3.125, + "output": 18.75, + "cache_read": 0.3125 } }, - "stepfun-ai/Step-3.5-Flash": { - "id": "stepfun-ai/Step-3.5-Flash", - "name": "stepfun-ai/Step-3.5-Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "family": "step", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "google-gemma-3-27b-it": { + "id": "google-gemma-3-27b-it", + "name": "Google Gemma 3 27B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "release_date": "2025-11-04", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 198000, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.12, + "output": 0.2 } }, - "MiniMaxAI/MiniMax-M3": { - "id": "MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "openai-gpt-56-luna": { + "id": "openai-gpt-56-luna", + "name": "GPT-5.6 Luna", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -47974,29 +45411,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 1.25, + "output": 7.5, + "cache_read": 0.125 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMaxAI/MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "aion-labs-aion-3-0-mini": { + "id": "aion-labs-aion-3-0-mini", + "name": "Aion 3.0 Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-06-15", + "structured_output": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text" @@ -48007,30 +45453,33 @@ }, "open_weights": false, "limit": { - "context": 197000, - "output": 131000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.875, + "output": 1.75, + "cache_read": 0.225 } }, - "inclusionAI/Ling-flash-2.0": { - "id": "inclusionAI/Ling-flash-2.0", - "name": "inclusionAI/Ling-flash-2.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", - "attachment": false, - "reasoning": false, + "claude-opus-4-8-fast": { + "id": "claude-opus-4-8-fast", + "name": "Claude Opus 4.8 Fast", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-09-18", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -48038,36 +45487,27 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 12, + "output": 60, + "cache_read": 1.2, + "cache_write": 15 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "moonshotai/Kimi-K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "gemma-4-uncensored": { + "id": "gemma-4-uncensored", + "name": "Gemma 4 Uncensored", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-06-15", + "release_date": "2026-04-13", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -48079,41 +45519,28 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 256000, + "output": 8192 }, "cost": { - "input": 0.77, - "output": 3.4, - "cache_read": 0.14 + "input": 0.1625, + "output": 0.5 } }, - "moonshotai/Kimi-K3": { - "id": "moonshotai/Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "qwen3-next-80b": { + "id": "qwen3-next-80b", + "name": "Qwen 3 Next 80b", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "release_date": "2025-04-29", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -48121,38 +45548,36 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 262000 + "context": 256000, + "output": 16384 }, "cost": { - "input": 2.7, - "output": 13.5, - "cache_read": 0.27 + "input": 0.35, + "output": 1.9 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "mistral-small-2603": { + "id": "mistral-small-2603", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -48164,37 +45589,28 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.85916, - "output": 3.8, - "cache_read": 0.17993 + "input": 0.1875, + "output": 0.75 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "moonshotai/Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", - "attachment": false, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -48204,39 +45620,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.45, - "output": 2.25, - "cache_read": 0.07 + "input": 6, + "output": 30, + "cache_read": 0.6, + "cache_write": 7.5 } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "zai-org/GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "zai-org-glm-4.7-flash": { + "id": "zai-org-glm-4.7-flash", + "name": "GLM 4.7 Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "knowledge": "2025-04", + "release_date": "2026-01-29", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -48247,43 +45666,42 @@ }, "open_weights": true, "limit": { - "context": 205000, - "output": 131000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.19, - "output": 3.74, - "cache_read": 0.6, - "cache_write": 0 + "input": 0.125, + "output": 0.5 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "google-gemma-4-26b-a4b-it": { + "id": "google-gemma-4-26b-a4b-it", + "name": "Google Gemma 4 26B A4B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", - "max" + "none", + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-04-02", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -48291,31 +45709,32 @@ }, "open_weights": true, "limit": { - "context": 1049000, - "output": 262000 + "context": 256000, + "output": 8192 }, "cost": { - "input": 1.302, - "output": 4.092, - "cache_read": 0.26, - "cache_write": 0 + "input": 0.13, + "output": 0.4, + "cache_read": 0.05 } }, - "zai-org/GLM-4.5-Air": { - "id": "zai-org/GLM-4.5-Air", - "name": "zai-org/GLM-4.5-Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", - "attachment": false, - "reasoning": false, + "grok-4-5": { + "id": "grok-4-5", + "name": "Grok 4.5", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-11-25", + "release_date": "2026-07-07", + "last_updated": "2026-07-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -48323,36 +45742,89 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 500000, + "output": 32000 }, "cost": { - "input": 0.14, - "output": 0.86 + "input": 2.27, + "output": 6.8, + "cache_read": 0.57, + "tiers": [ + { + "input": 4.53, + "output": 13.6, + "cache_read": 1.13, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4.53, + "output": 13.6, + "cache_read": 1.13 + } } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "zai-org/GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "grok-4-20": { + "id": "grok-4-20", + "name": "Grok 4.20", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-06-15", + "release_date": "2026-03-12", + "last_updated": "2026-06-11", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 128000 + }, + "cost": { + "input": 1.42, + "output": 2.83, + "cache_read": 0.23, + "tiers": [ + { + "input": 2.83, + "output": 5.67, + "cache_read": 0.45, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.83, + "output": 5.67, + "cache_read": 0.45 + } + } + }, + "hermes-3-llama-3.1-405b": { + "id": "hermes-3-llama-3.1-405b", + "name": "Hermes 3 Llama 3.1 405b", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "hermes", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2025-09-25", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -48363,85 +45835,62 @@ }, "open_weights": true, "limit": { - "context": 205000, - "output": 205000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.95, - "output": 2.55, - "cache_read": 0.2 + "input": 1.1, + "output": 3 } }, - "zai-org/GLM-5.3-Flash": { - "id": "zai-org/GLM-5.3-Flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + "aion-labs-aion-2-0": { + "id": "aion-labs-aion-2-0", + "name": "Aion 2.0", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "tool_call": false, + "release_date": "2026-03-24", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1049000, - "output": 262000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03, - "cache_write": 0 + "input": 1, + "output": 2, + "cache_read": 0.25 } }, - "zai-org/GLM-5V-Turbo": { - "id": "zai-org/GLM-5V-Turbo", - "name": "zai-org/GLM-5V-Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "qwen3-vl-235b-a22b": { + "id": "qwen3-vl-235b-a22b", + "name": "Qwen3 VL 235B", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "structured_output": true, + "release_date": "2026-01-16", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -48451,46 +45900,47 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24, - "cache_write": 0 + "input": 0.21, + "output": 1.9, + "cache_read": 0.1 } }, - "zai-org/GLM-5.3": { - "id": "zai-org/GLM-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "xiaomi-mimo-v2-5": { + "id": "xiaomi-mimo-v2-5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2024-12", + "release_date": "2026-06-11", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -48498,68 +45948,74 @@ }, "open_weights": true, "limit": { - "context": 1049000, - "output": 262000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "input": 0.14, + "output": 0.28, + "cache_read": 0.05 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "openai/gpt-oss-20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "venice-uncensored-1-2": { + "id": "venice-uncensored-1-2", + "name": "Venice Uncensored 1.2", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "venice", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-13", - "last_updated": "2025-11-25", + "release_date": "2026-04-01", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 8000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.04, - "output": 0.18 + "input": 0.2, + "output": 0.9 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "openai/gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "gemini-3-5-flash": { + "id": "gemini-3-5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-13", - "last_updated": "2025-11-25", + "knowledge": "2025-01", + "release_date": "2026-05-22", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -48567,76 +46023,74 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 8000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.05, - "output": 0.45 + "input": 1.55, + "output": 9.45, + "cache_read": 0.155, + "cache_write": 0.086 } - } - } - }, - "minimax-coding-plan": { - "id": "minimax-coding-plan", - "env": [ - "MINIMAX_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://api.minimax.io/anthropic/v1", - "name": "MiniMax Token Plan (minimax.io)", - "doc": "https://platform.minimax.io/docs/token-plan/intro", - "models": { - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-25", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-10", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 512000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 12, + "output": 60, + "cache_read": 1.2, + "cache_write": 15 } }, - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "openai-gpt-oss-120b": { + "id": "openai-gpt-oss-120b", + "name": "OpenAI GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2025-11-06", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -48647,28 +46101,27 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.07, + "output": 0.3 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "zai-org-glm-5-2": { + "id": "zai-org-glm-5-2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ "text" @@ -48679,28 +46132,27 @@ }, "open_weights": true, "limit": { - "context": 204800, + "context": 1000000, "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "nvidia-nemotron-3-nano-30b-a3b": { + "id": "nvidia-nemotron-3-nano-30b-a3b", + "name": "NVIDIA Nemotron 3 Nano 30B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-01-27", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -48711,60 +46163,60 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.075, + "output": 0.3 } }, - "MiniMax-M2.7-highspeed": { - "id": "MiniMax-M2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 6, + "output": 30, + "cache_read": 0.6, + "cache_write": 7.5 } }, - "MiniMax-M2": { - "id": "MiniMax-M2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "mistral-small-3-2-24b-instruct": { + "id": "mistral-small-3-2-24b-instruct", + "name": "Mistral Small 3.2 24B Instruct", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "structured_output": true, + "release_date": "2026-01-15", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -48775,29 +46227,89 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.09375, + "output": 0.25 } }, - "MiniMax-M2.5-highspeed": { - "id": "MiniMax-M2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, + "openai-gpt-54-pro": { + "id": "openai-gpt-54-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-06-11", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 37.5, + "output": 225, + "tiers": [ + { + "input": 75, + "output": 337.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 75, + "output": 337.5 + } + } + }, + "minimax-m3-preview": { + "id": "minimax-m3-preview", + "name": "MiniMax M3 Preview", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax-m3", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "release_date": "2026-06-12", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text", + "image", + "video" ], "output": [ "text" @@ -48805,47 +46317,44 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 524288, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } - } - } - }, - "alibaba-coding-plan": { - "id": "alibaba-coding-plan", - "env": [ - "ALIBABA_CODING_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://coding-intl.dashscope.aliyuncs.com/v1", - "name": "Alibaba Coding Plan", - "doc": "https://www.alibabacloud.com/help/en/model-studio/coding-plan", - "models": { - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + }, + "gemini-3-1-pro-preview": { + "id": "gemini-3-1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -48854,27 +46363,44 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 65536 + "output": 32768 }, "cost": { "input": 2.5, - "output": 7.5, + "output": 15, "cache_read": 0.5, - "cache_write": 3.125 + "cache_write": 0.5, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 0.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 0.5 + } } }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "Qwen3 Coder Next", + "qwen3-coder-480b-a35b-instruct-turbo": { + "id": "qwen3-coder-480b-a35b-instruct-turbo", + "name": "Qwen 3 Coder 480B Turbo", "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-03", - "last_updated": "2026-02-03", + "release_date": "2026-01-27", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -48885,33 +46411,38 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 256000, "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.35, + "output": 1.5, + "cache_read": 0.04 } }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", + "qwen3-6-27b": { + "id": "qwen3-6-27b", + "name": "Qwen 3.6 27B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "release_date": "2026-04-24", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -48924,38 +46455,43 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 256000, "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.325, + "output": 3.25 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "openai-gpt-52-codex": { + "id": "openai-gpt-52-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08", + "release_date": "2025-01-15", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -48963,36 +46499,28 @@ }, "open_weights": false, "limit": { - "context": 202752, - "output": 16384 + "context": 256000, + "input": 272000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2.19, + "output": 17.5, + "cache_read": 0.219 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "qwen-3-7-max": { + "id": "qwen-3-7-max", + "name": "Qwen 3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2026-05-22", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -49003,30 +46531,40 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2.7, + "output": 8.05, + "cache_read": 0.27, + "cache_write": 3.35 } }, - "qwen3-max-2026-01-23": { - "id": "qwen3-max-2026-01-23", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "olafangensan-glm-4.7-flash-heretic": { + "id": "olafangensan-glm-4.7-flash-heretic", + "name": "GLM 4.7 Flash Heretic", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-23", - "last_updated": "2026-01-23", + "structured_output": true, + "release_date": "2026-02-04", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -49035,40 +46573,46 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 24000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.07, + "output": 0.4, + "cache_read": 0.035 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "openai-gpt-54-mini": { + "id": "openai-gpt-54-mini", + "name": "GPT-5.4 Mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-27", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -49076,103 +46620,91 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.1875, - "output": 1.125, - "cache_write": 0.234375 - } - }, - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.9375, + "output": 5.625, + "cache_read": 0.09375 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "grok-build-0-1": { + "id": "grok-build-0-1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-05-21", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "input": 196601, - "output": 24576 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1, + "output": 2, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 4, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 4, + "cache_read": 0.4 + } } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "zai-org-glm-4.7": { + "id": "zai-org-glm-4.7", + "name": "GLM 4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2025-12-24", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -49183,77 +46715,33 @@ }, "open_weights": true, "limit": { - "context": 202752, + "context": 198000, "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.55, + "output": 2.65, + "cache_read": 0.11 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -49265,34 +46753,28 @@ "output": 64000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 3.6, + "output": 18, + "cache_read": 0.36, + "cache_write": 4.5 } - } - } - }, - "lilac": { - "id": "lilac", - "env": [ - "LILAC_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.getlilac.com/v1", - "name": "Lilac", - "doc": "https://docs.getlilac.com/inference/models", - "models": { - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + }, + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, @@ -49302,87 +46784,116 @@ "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-12-19", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262100, - "output": 262100 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.11, - "output": 0.35 + "input": 0.7, + "output": 3.75, + "cache_read": 0.07 } }, - "minimaxai/minimax-m3": { - "id": "minimaxai/minimax-m3", - "name": "MiniMax M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax-m3", + "openai-gpt-55": { + "id": "openai-gpt-55", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 1000000, + "input": 922000, + "output": 131072 }, "cost": { - "input": 0.28, - "output": 1.1, - "cache_read": 0.05 + "input": 6.25, + "output": 37.5, + "cache_read": 0.625, + "tiers": [ + { + "input": 12.5, + "output": 56.25, + "cache_read": 1.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 12.5, + "output": 56.25, + "cache_read": 1.25 + } } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "qwen3-5-9b": { + "id": "qwen3-5-9b", + "name": "Qwen 3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-03-05", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -49394,42 +46905,36 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 32768 }, "cost": { - "input": 0.7, - "output": 3.5, - "cache_read": 0.2 + "input": 0.1, + "output": 0.15 } }, - "zai-org/glm-5.2": { - "id": "zai-org/glm-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "mercury-2": { + "id": "mercury-2", + "name": "Mercury 2", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "max" + "none", + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-02-20", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -49438,77 +46943,33 @@ "text" ] }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 524288 - }, - "cost": { - "input": 0.9, - "output": 3, - "cache_read": 0.27 - } - } - } - }, - "moonshotai-cn": { - "id": "moonshotai-cn", - "env": [ - "MOONSHOT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.moonshot.cn/v1", - "name": "Moonshot AI (China)", - "doc": "https://platform.moonshot.cn/docs/api/chat", - "models": { - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 50000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.3125, + "output": 0.9375, + "cache_read": 0.03125 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "zai-org-glm-5-1": { + "id": "zai-org-glm-5-1", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, @@ -49517,14 +46978,11 @@ }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-04-07", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -49532,70 +46990,75 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 24000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 1.54, + "output": 4.84, + "cache_read": 0.286 } }, - "kimi-k2.7-code-highspeed": { - "id": "kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code HighSpeed", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "openai-gpt-56-luna-pro": { + "id": "openai-gpt-56-luna-pro", + "name": "GPT-5.6 Luna Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-nano", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 + "input": 1.25, + "output": 7.5, + "cache_read": 0.125 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "kimi-k2-5": { + "id": "kimi-k2-5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", - "high", - "max" + "medium", + "high" ] } ], @@ -49605,59 +47068,52 @@ }, "structured_output": true, "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2024-04", + "release_date": "2026-01-27", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 256000, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.56, + "output": 3.5, + "cache_read": 0.22 } - } - } - }, - "bothub": { - "id": "bothub", - "env": [ - "BOTHUB_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://openai.bothub.ru/v1", - "name": "Bothub", - "doc": "https://bothub.ru/models", - "models": { - "gemma-4-31b-it:free": { - "id": "gemma-4-31b-it:free", - "name": "Gemma 4 31B IT (free)", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + }, + "z-ai-glm-5v-turbo": { + "id": "z-ai-glm-5v-turbo", + "name": "GLM 5V Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-04-01", + "last_updated": "2026-06-11", "modalities": { "input": [ "text", @@ -49667,45 +47123,43 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 200000, "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 1.5, + "output": 5, + "cache_read": 0.3 } }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + "nvidia-nemotron-3-ultra-550b-a55b": { + "id": "nvidia-nemotron-3-ultra-550b-a55b", + "name": "NVIDIA Nemotron 3 Ultra", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-06-04", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -49713,19 +47167,19 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 256000, + "output": 32768 }, "cost": { - "input": 0.12, - "output": 0.44 + "input": 0.625, + "output": 3.125, + "cache_read": 0.1875 } }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + "aion-labs-aion-3-0": { + "id": "aion-labs-aion-3-0", + "name": "Aion 3.0", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -49734,18 +47188,15 @@ "values": [ "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text" @@ -49754,21 +47205,22 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, + "context": 128000, + "output": 32768 + }, "cost": { - "input": 0.1, - "output": 0.28 + "input": 3.75, + "output": 7.5, + "cache_read": 0.9375 } }, - "nemotron-3-ultra-550b-a55b:free": { - "id": "nemotron-3-ultra-550b-a55b:free", - "name": "Nemotron 3 Ultra (free)", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", + "openai-gpt-52": { + "id": "openai-gpt-52", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -49776,16 +47228,19 @@ "type": "effort", "values": [ "none", + "minimal", + "low", "medium", "high" ] } ], "tool_call": true, - "interleaved": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-13", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -49794,21 +47249,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "input": 272000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 2.19, + "output": 17.5, + "cache_read": 0.219 } }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -49816,17 +47273,16 @@ "type": "effort", "values": [ "none", - "high", - "max" + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2025-12-04", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -49837,19 +47293,20 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 160000, + "output": 32768 }, "cost": { - "input": 1.61, - "output": 4.84 + "input": 0.33, + "output": 0.48, + "cache_read": 0.16 } }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", + "openai-gpt-56-sol": { + "id": "openai-gpt-56-sol", + "name": "GPT-5.6 Sol", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -49857,16 +47314,14 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": false, "knowledge": "2026-02-16", @@ -49875,8 +47330,7 @@ "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -49884,85 +47338,68 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.06, - "output": 0.37 + "input": 6.25, + "output": 37.5, + "cache_read": 0.625 } }, - "muse-spark-1.3-contributor": { - "id": "muse-spark-1.3-contributor", - "name": "Muse Spark 1.3 Contributor", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", - "family": "muse", - "attachment": true, + "minimax-m25": { + "id": "minimax-m25", + "name": "MiniMax M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "interleaved": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "release_date": "2026-02-12", + "last_updated": "2026-06-11", "modalities": { "input": [ - "text", - "image", - "video", - "pdf", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 943718 + "context": 198000, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.2 + "input": 0.27, + "output": 0.95, + "cache_read": 0.03 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "llama-3.2-3b": { + "id": "llama-3.2-3b", + "name": "Llama 3.2 3B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2024-10-03", + "last_updated": "2026-06-11", "modalities": { "input": [ "text" @@ -49973,32 +47410,32 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.72, - "output": 5.41 + "input": 0.15, + "output": 0.6 } } } }, - "xiaomi-token-plan-cn": { - "id": "xiaomi-token-plan-cn", + "crossmodel": { + "id": "crossmodel", "env": [ - "XIAOMI_API_KEY" + "CROSSMODEL_API_KEY" ], "npm": "@ai-sdk/openai-compatible", - "api": "https://token-plan-cn.xiaomimimo.com/v1", - "name": "Xiaomi Token Plan (China)", - "doc": "https://platform.xiaomimimo.com/#/docs", + "api": "https://api.crossmodel.ai/v1", + "name": "CrossModel", + "doc": "https://www.crossmodel.ai/docs", "models": { - "mimo-v2.6-pro": { - "id": "mimo-v2.6-pro", - "name": "MiMo-V2.6-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": true, + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -50006,18 +47443,13 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -50025,21 +47457,34 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.47, + "output": 2.16, + "cache_read": 0.1, + "cache_write": 0.47, + "tiers": [ + { + "input": 0.62, + "output": 2.47, + "cache_read": 0.13, + "cache_write": 0.62, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -50047,96 +47492,47 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "mimo-v2.5-tts-voicedesign": { - "id": "mimo-v2.5-tts-voicedesign", - "name": "MiMo-V2.5-TTS-VoiceDesign", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" ], "output": [ - "audio" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mimo-v2.5-tts": { - "id": "mimo-v2.5-tts", - "name": "MiMo-V2.5-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ "text" - ], - "output": [ - "audio" ] }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 3.8, + "cache_read": 0.2, + "cache_write": 1, + "tiers": [ + { + "input": 1.2, + "output": 4.4, + "cache_read": 0.3, + "cache_write": 1.2, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -50145,13 +47541,10 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -50160,51 +47553,23 @@ "text" ] }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "mimo-v2-tts": { - "id": "mimo-v2-tts", - "name": "MiMo-V2-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.2, + "output": 4.4, + "cache_read": 0.3, + "cache_write": 1.2 } }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -50213,13 +47578,9 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -50230,21 +47591,34 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.6, + "output": 3, + "cache_read": 0.16, + "cache_write": 0.6, + "tiers": [ + { + "input": 0.8, + "output": 3.4, + "cache_read": 0.2, + "cache_write": 0.8, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "mimo-v2.6-flash": { - "id": "mimo-v2.6-flash", - "name": "MiMo-V2.6-Flash", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -50252,97 +47626,67 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "mimo-v2.5-tts-voiceclone": { - "id": "mimo-v2.5-tts-voiceclone", - "name": "MiMo-V2.5-TTS-VoiceClone", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" + "input": 0.9, + "output": 3.7, + "cache_read": 0.18, + "cache_write": 0.9, + "tiers": [ + { + "input": 1.1, + "output": 4.3, + "cache_read": 0.27, + "cache_write": 1.1, + "tier": { + "type": "context", + "size": 32000 + } + } ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 } - } - } - }, - "privatemode-ai": { - "id": "privatemode-ai", - "env": [ - "PRIVATEMODE_API_KEY", - "PRIVATEMODE_ENDPOINT" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "http://localhost:8080/v1", - "name": "Privatemode AI", - "doc": "https://docs.privatemode.ai/api/overview", - "models": { - "glm-flash-latest": { - "id": "glm-flash-latest", - "name": "GLM Flash (latest)", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + }, + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -50352,40 +47696,44 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, - "status": "beta", "cost": { - "input": 0.8897, - "output": 4.4718, - "cache_read": 0.0924 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02, + "cache_write": 0.2 } }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -50395,64 +47743,62 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "status": "beta", "cost": { - "input": 0.8897, - "output": 4.4718, - "cache_read": 0.0924 - } - }, - "whisper-large-v3": { - "id": "whisper-large-v3", - "name": "Whisper large-v3", - "description": "Open Whisper checkpoint for robust multilingual transcription and captioning", - "family": "whisper", - "attachment": true, - "reasoning": false, - "tool_call": false, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", - "modalities": { - "input": [ - "audio" + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 2.5, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 448, - "output": 4096 - }, - "cost": { - "input": 0.01618, - "output": 0 + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 5 + } } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -50462,27 +47808,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 1.791, - "output": 8.9436, - "cache_read": 0.1733 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075, + "cache_write": 0.75 } }, - "deepseek-ocr-2": { - "id": "deepseek-ocr-2", - "name": "DeepSeek OCR 2", - "description": "High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes", + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", "attachment": true, - "reasoning": false, - "tool_call": false, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -50492,84 +47853,135 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "status": "beta", "cost": { - "input": 0.8897, - "output": 1.4675, - "cache_read": 0.0924 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "qwen3-embedding-4b": { - "id": "qwen3-embedding-4b", - "name": "Qwen3-Embedding 4B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", - "attachment": false, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-06-06", - "last_updated": "2025-06-06", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 2560 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.1502, - "output": 0 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075, + "cache_write": 0.15 } }, - "voxtral-mini-3b": { - "id": "voxtral-mini-3b", - "name": "Voxtral Mini 3B", - "description": "Speech-to-text model for audio transcription, translation, and audio understanding", - "family": "voxtral", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-07", - "last_updated": "2025-07", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "audio" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 32000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.00462, - "output": 0 + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 10, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 10 + } } }, - "kimi-latest": { - "id": "kimi-latest", - "name": "Kimi (latest)", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -50578,11 +47990,10 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", @@ -50594,38 +48005,33 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 262144 + "context": 1000000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 1.791, - "output": 8.9436, - "cache_read": 0.1733 + "input": 0.16, + "output": 0.32, + "cache_read": 0.004, + "cache_write": 0.16 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -50636,123 +48042,122 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.4969, - "output": 1.9644, - "cache_read": 0.0462 + "input": 0.47, + "output": 0.94, + "cache_read": 0.005, + "cache_write": 0.47 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 131072 + "context": 1000000, + "output": 128000 }, - "status": "beta", "cost": { - "input": 1.791, - "output": 8.9436, - "cache_read": 0.1733 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "glm-latest": { - "id": "glm-latest", - "name": "GLM (latest)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 131072 + "context": 1000000, + "output": 128000 }, - "status": "beta", "cost": { - "input": 1.791, - "output": 8.9436, - "cache_read": 0.1733 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } - } - } - }, - "llmgateway-providers": { - "id": "llmgateway-providers", - "env": [ - "LLMGATEWAY_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.llmgateway.io/v1", - "name": "LLM Gateway", - "doc": "https://llmgateway.io/docs", - "models": { - "consensusprotocol/gemma-4-31b-it": { - "id": "consensusprotocol/gemma-4-31b-it", - "name": "Gemma 4 31B IT (Consensus Protocol)", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": false, + }, + "anthropic/claude-opus-4-8": { + "id": "anthropic/claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high", "xhigh", "max" @@ -50760,33 +48165,36 @@ } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.25, - "cache_read": 0.01 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "consensusprotocol/Qwen3.8-27B": { - "id": "consensusprotocol/Qwen3.8-27B", - "name": "Qwen3.8 27B (Consensus Protocol)", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -50795,15 +48203,17 @@ "values": [ "low", "medium", - "xhigh" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", @@ -50815,64 +48225,60 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.08, - "output": 0.35, - "cache_read": 0.05 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "consensusprotocol/glm-5.3-flash": { - "id": "consensusprotocol/glm-5.3-flash", - "name": "GLM-5.3 Flash (Consensus Protocol)", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.25, - "cache_read": 0.02 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "consensusprotocol/deepseek-v4.1-flash": { - "id": "consensusprotocol/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (Consensus Protocol)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -50880,18 +48286,21 @@ "type": "effort", "values": [ "low", + "medium", "high", - "xhigh", "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -50901,39 +48310,39 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.2, - "output": 0.6, - "cache_read": 0.005 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "consensusprotocol/gpt-oss-20b": { - "id": "consensusprotocol/gpt-oss-20b", - "name": "GPT OSS 20B (Consensus Protocol)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": true, + "tencent/hy3-preview": { + "id": "tencent/hy3-preview", + "name": "Hy3 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", - "medium", "high" ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text" @@ -50944,45 +48353,56 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 32768 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.04, - "output": 0.19, - "cache_read": 0.01 + "input": 0.19, + "output": 0.63, + "cache_read": 0.063, + "cache_write": 0.19, + "tiers": [ + { + "input": 0.25, + "output": 1, + "cache_read": 0.094, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 16000 + } + }, + { + "input": 0.32, + "output": 1.25, + "cache_read": 0.125, + "cache_write": 0.32, + "tier": { + "type": "context", + "size": 32000 + } + } + ] } }, - "consensusprotocol/deepseek-v4-flash": { - "id": "consensusprotocol/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (Consensus Protocol)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "moonshot/kimi-k2.7-code": { + "id": "moonshot/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -50990,44 +48410,38 @@ }, "open_weights": true, "limit": { - "context": 1050000, - "output": 393216 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.05, - "output": 0.1, - "cache_read": 0.01 + "input": 1, + "output": 4.16, + "cache_read": 0.18, + "cache_write": 1 } }, - "deepinfra/gemma-4-31b-it": { - "id": "deepinfra/gemma-4-31b-it", - "name": "Gemma 4 31B IT (DeepInfra)", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "moonshot/kimi-k2.5": { + "id": "moonshot/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -51035,27 +48449,34 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.13, - "output": 0.38 + "input": 0.62, + "output": 3.3, + "cache_read": 0.11, + "cache_write": 0.62 } }, - "deepinfra/qwen3-vl-235b-a22b-instruct": { - "id": "deepinfra/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct (DeepInfra)", - "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "moonshot/kimi-k2.6": { + "id": "moonshot/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -51067,45 +48488,39 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.2, - "output": 0.88, - "cache_read": 0.11 + "input": 1, + "output": 4.16, + "cache_read": 0.18, + "cache_write": 1 } }, - "deepinfra/mimo-v2.5": { - "id": "deepinfra/mimo-v2.5", - "name": "MiMo V2.5 (DeepInfra)", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "gemini/gemini-2.5-pro": { + "id": "gemini/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", @@ -51117,45 +48532,50 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08, + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "cache_write": 1.25, "tiers": [ { - "input": 0.8, - "output": 4, - "cache_read": 0.16, + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 2.5, "tier": { "type": "context", - "size": 256000 + "size": 200000 } } ], "context_over_200k": { - "input": 0.8, - "output": 4, - "cache_read": 0.16 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 2.5 } } }, - "deepinfra/qwen3.8-27b": { - "id": "deepinfra/qwen3.8-27b", - "name": "Qwen3.8 27B (DeepInfra)", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", + "gemini/gemini-2.5-flash": { + "id": "gemini/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -51165,41 +48585,47 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.2, + "input": 0.3, "output": 2.5, - "cache_read": 0.05 + "cache_read": 0.03, + "cache_write": 0.3 } }, - "deepinfra/nemotron-3.5-lightning": { - "id": "deepinfra/nemotron-3.5-lightning", - "name": "Nemotron 3.5 Lightning (DeepInfra)", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron", - "attachment": false, + "gemini/gemini-3.5-flash": { + "id": "gemini/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", "high" @@ -51209,214 +48635,220 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.08, - "output": 0.2 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 1.5 } }, - "deepinfra/qwen3.8-2.4t-a95b": { - "id": "deepinfra/qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B (DeepInfra)", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, + "gemini/gemini-2.5-flash-lite": { + "id": "gemini/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0.1 } }, - "deepinfra/muse-glimmer-30b": { - "id": "deepinfra/muse-glimmer-30b", - "name": "Muse Glimmer 30B (DeepInfra)", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", + "gemini/gemini-3.1-pro-preview": { + "id": "gemini/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.04 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4, + "cache_write": 4 + } } }, - "deepinfra/deepseek-v4.1-flash": { - "id": "deepinfra/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (DeepInfra)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "gemini/gemini-3-flash-preview": { + "id": "gemini/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 393216 + "output": 65536 }, "cost": { - "input": 0.2, - "output": 0.6, - "cache_read": 0.006 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.5 } }, - "deepinfra/ling-3.0-flash": { - "id": "deepinfra/ling-3.0-flash", - "name": "InclusionAI Ling 3.0 Flash (DeepInfra)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-02", - "last_updated": "2026-08-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.012 - } - }, - "deepinfra/qwen3-vl-30b-a3b-instruct": { - "id": "deepinfra/qwen3-vl-30b-a3b-instruct", - "name": "Qwen3 VL 30B A3B Instruct (DeepInfra)", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-10-05", - "last_updated": "2025-10-05", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -51424,27 +48856,53 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.32, + "output": 1.25, + "cache_read": 0.032, + "cache_write": 0.4, + "tiers": [ + { + "input": 0.96, + "output": 3.75, + "cache_read": 0.096, + "cache_write": 1.2, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.96, + "output": 3.75, + "cache_read": 0.096, + "cache_write": 1.2 + } } }, - "deepinfra/gemma-4-26b-a4b-it": { - "id": "deepinfra/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT (DeepInfra)", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -51453,147 +48911,124 @@ "text" ] }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.07, - "output": 0.34 - } - }, - "deepinfra/nemotron-3-ultra-550b": { - "id": "deepinfra/nemotron-3-ultra-550b", - "name": "Nemotron 3 Ultra 550B (DeepInfra)", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 2.2, - "cache_read": 0.1 + "input": 1.88, + "output": 5.63, + "cache_read": 0.375, + "cache_write": 2.35 } }, - "deepinfra/inkling-small": { - "id": "deepinfra/inkling-small", - "name": "Inkling Small (DeepInfra)", - "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", - "family": "ling", + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", "image", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 262144 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.45, - "output": 1.2, - "cache_read": 0.1 + "input": 0.19, + "output": 1.13, + "cache_read": 0.019, + "cache_write": 0.24, + "tiers": [ + { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075, + "cache_write": 0.94, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075, + "cache_write": 0.94 + } } }, - "deepinfra/mimo-v2.5-pro": { - "id": "deepinfra/mimo-v2.5-pro", - "name": "MiMo V2.5 Pro (DeepInfra)", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], - "tool_call": false, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2, + "input": 0.32, + "output": 1.88, + "cache_read": 0.032, + "cache_write": 0.4, "tiers": [ { - "input": 2, - "output": 6, - "cache_read": 0.4, + "input": 1.25, + "output": 7.5, + "cache_read": 0.124, + "cache_write": 1.57, "tier": { "type": "context", "size": 256000 @@ -51601,74 +49036,41 @@ } ], "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 + "input": 1.25, + "output": 7.5, + "cache_read": 0.124, + "cache_write": 1.57 } } }, - "deepinfra/qwen3.5-9b": { - "id": "deepinfra/qwen3.5-9b", - "name": "Qwen3.5 9B (DeepInfra)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.15 - } - }, - "deepinfra/inkling": { - "id": "deepinfra/inkling", - "name": "Inkling (DeepInfra)", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "minimal", - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -51676,37 +49078,41 @@ }, "open_weights": true, "limit": { - "context": 524288, - "output": 262144 + "context": 1000000, + "output": 65000 }, "cost": { - "input": 0.95, - "output": 4.05, - "cache_read": 0.16 + "input": 0.16, + "output": 0.32, + "cache_read": 0.004, + "cache_write": 0.16 } }, - "deepinfra/hy3": { - "id": "deepinfra/hy3", - "name": "Hy3 (DeepInfra)", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -51717,48 +49123,37 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 192000, - "output": 131072 + "context": 1000000, + "output": 65000 }, "cost": { - "input": 0.14, - "output": 0.58, - "cache_read": 0.035 + "input": 0.47, + "output": 0.94, + "cache_read": 0.005, + "cache_write": 0.47 } }, - "deepinfra/glm-5.1": { - "id": "deepinfra/glm-5.1", - "name": "GLM-5.1 (DeepInfra)", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -51766,41 +49161,46 @@ }, "open_weights": true, "limit": { - "context": 198000, - "output": 65536 + "context": 1024000, + "output": 512000 }, "cost": { - "input": 1.05, - "output": 3.5, - "cache_read": 0.205 + "input": 0.33, + "output": 1.32, + "cache_read": 0.066, + "cache_write": 0.33, + "tiers": [ + { + "input": 0.66, + "output": 2.63, + "cache_read": 0.132, + "cache_write": 0.66, + "tier": { + "type": "context", + "size": 512000 + } + } + ], + "context_over_200k": { + "input": 0.66, + "output": 2.63, + "cache_read": 0.132, + "cache_write": 0.66 + } } }, - "deepinfra/deepseek-v4-pro": { - "id": "deepinfra/deepseek-v4-pro", - "name": "DeepSeek V4 Pro (DeepInfra)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -51811,45 +49211,52 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 64000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.3, - "output": 2.6, - "cache_read": 0.1 + "input": 0.33, + "output": 1.32, + "cache_read": 0.066, + "cache_write": 0.42 } - }, - "deepinfra/step-3.7-flash": { - "id": "deepinfra/step-3.7-flash", - "name": "Step 3.7 Flash (DeepInfra)", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, + } + } + }, + "lmstudio": { + "id": "lmstudio", + "env": [ + "LMSTUDIO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "http://127.0.0.1:1234/v1", + "name": "LMStudio", + "doc": "https://lmstudio.ai/models", + "models": { + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -51857,28 +49264,26 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 131072, "output": 32768 }, "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "input": 0, + "output": 0 } }, - "deepinfra/deepseek-v3.2": { - "id": "deepinfra/deepseek-v3.2", - "name": "DeepSeek V3.2 (DeepInfra)", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", + "qwen/qwen3-30b-a3b-2507": { + "id": "qwen/qwen3-30b-a3b-2507", + "name": "Qwen3 30B A3B 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "knowledge": "2025-04", + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ "text" @@ -51889,38 +49294,26 @@ }, "open_weights": true, "limit": { - "context": 160000, - "output": 65536 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.26, - "output": 0.38, - "cache_read": 0.13 + "input": 0, + "output": 0 } }, - "deepinfra/glm-5.3": { - "id": "deepinfra/glm-5.3", - "name": "GLM-5.3 (DeepInfra)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "qwen/qwen3-coder-30b": { + "id": "qwen/qwen3-coder-30b", + "name": "Qwen3 Coder 30B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -51931,42 +49324,42 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.2 + "input": 0, + "output": 0 } - }, - "deepinfra/deepseek-v4-flash": { - "id": "deepinfra/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (DeepInfra)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + } + } + }, + "poolside": { + "id": "poolside", + "env": [ + "POOLSIDE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.poolside.ai/v1", + "name": "Poolside", + "doc": "https://platform.poolside.ai", + "models": { + "poolside/laguna-xs.2": { + "id": "poolside/laguna-xs.2", + "name": "Laguna XS.2", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-04-28", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -51977,45 +49370,35 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 393216 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.08, - "output": 0.18, - "cache_read": 0.016 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "cerebras/gemma-4-31b-it": { - "id": "cerebras/gemma-4-31b-it", - "name": "Gemma 4 31B IT (Cerebras)", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, + "poolside/laguna-m.1": { + "id": "poolside/laguna-m.1", + "name": "Laguna M.1", + "description": "Poolside's flagship agentic coding model for long-horizon work", + "family": "laguna", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-04-28", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -52023,26 +49406,32 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 262144, "output": 32768 }, "cost": { - "input": 0.99, - "output": 1.49 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "cerebras/qwen3-235b-a22b-instruct-2507": { - "id": "cerebras/qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507 (Cerebras)", - "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", - "family": "qwen", + "poolside/laguna-xs-2.1": { + "id": "poolside/laguna-xs-2.1", + "name": "Laguna XS 2.1", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": false, "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", + "release_date": "2026-07-02", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" @@ -52053,27 +49442,85 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 8192 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 1.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "cerebras/llama-3.3-70b-instruct": { - "id": "cerebras/llama-3.3-70b-instruct", - "name": "Llama 3.3 70B Instruct (Cerebras)", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + } + } + }, + "zenifra": { + "id": "zenifra", + "env": [ + "ZENIFRA_AI_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://ai.zenifra.com/v1", + "name": "Zenifra", + "doc": "https://docs.zenifra.com", + "models": { + "alibaba/qwen3.6-35b-a3b": { + "id": "alibaba/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "provider": { + "shape": "completions" + }, + "cost": { + "input": 0.19, + "output": 0.48 + } + } + } + }, + "zenmux": { + "id": "zenmux", + "env": [ + "ZENMUX_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://zenmux.ai/api/v1", + "name": "ZenMux", + "doc": "https://docs.zenmux.ai", + "models": { + "inclusionai/ling-1t": { + "id": "inclusionai/ling-1t", + "name": "Ling-1T", + "description": "Tool-capable chat model for instruction following and agentic application workflows", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-01-01", + "release_date": "2025-10-09", + "last_updated": "2025-10-09", "modalities": { "input": [ "text" @@ -52082,37 +49529,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 64000 }, "cost": { - "input": 0.85, - "output": 1.2 + "input": 0.56, + "output": 2.24, + "cache_read": 0.11 } }, - "cerebras/glm-4.7": { - "id": "cerebras/glm-4.7", - "name": "GLM-4.7 (Cerebras)", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, + "inclusionai/ring-2.6-1t": { + "id": "inclusionai/ring-2.6-1t", + "name": "inclusionAI: Ring-2.6-1T", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2025-12-31", + "release_date": "2026-05-07", + "last_updated": "2026-05-14", "modalities": { "input": [ "text" @@ -52123,36 +49562,27 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 262000, + "output": 65000 }, "cost": { - "input": 2.25, - "output": 2.75 + "input": 0.3, + "output": 2.5, + "cache_read": 0.06 } }, - "cerebras/gpt-oss-120b": { - "id": "cerebras/gpt-oss-120b", - "name": "GPT OSS 120B (Cerebras)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "inclusionai/ring-1t": { + "id": "inclusionai/ring-1t", + "name": "Ring-1T", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01-01", + "release_date": "2025-10-12", + "last_updated": "2025-10-12", "modalities": { "input": [ "text" @@ -52161,41 +49591,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 64000 }, "cost": { - "input": 0.35, - "output": 0.75 + "input": 0.56, + "output": 2.24, + "cache_read": 0.11 } }, - "groq/gpt-oss-20b": { - "id": "groq/gpt-oss-20b", - "name": "GPT OSS 20B (Groq)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "moonshotai/kimi-k2.7-code-free": { + "id": "moonshotai/kimi-k2.7-code-free", + "name": "Kimi K2.7 Code (Free)", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -52203,36 +49628,27 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32766 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.1, - "output": 0.5 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "groq/gpt-oss-120b": { - "id": "groq/gpt-oss-120b", - "name": "GPT OSS 120B (Groq)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "moonshotai/kimi-k2-thinking-turbo": { + "id": "moonshotai/kimi-k2-thinking-turbo", + "name": "Kimi K2 Thinking Turbo", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01-01", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ "text" @@ -52241,44 +49657,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32766 + "context": 262000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.75 + "input": 1.15, + "output": 8, + "cache_read": 0.15 } }, - "scx-ai-gp/glm-5.3-flash": { - "id": "scx-ai-gp/glm-5.3-flash", - "name": "GLM-5.3 Flash (SCX.ai)", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "video", - "pdf" + "video" ], "output": [ "text" @@ -52286,28 +49694,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.088, - "output": 0.25, - "cache_read": 0.025 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "scx-ai-gp/glm-5.2-fast": { - "id": "scx-ai-gp/glm-5.2-fast", - "name": "GLM-5.2 Turbo (SCX.ai)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-01-01", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ "text" @@ -52316,48 +49723,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 262000, + "output": 64000 }, "cost": { - "input": 2.2, - "output": 6.5, - "cache_read": 0.45 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "scx-ai-gp/qwen3.8-max": { - "id": "scx-ai-gp/qwen3.8-max", - "name": "Qwen3.8 Max (SCX.ai)", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2025-01-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", "image", - "video", - "pdf" + "video" ], "output": [ "text" @@ -52365,42 +49765,34 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 262000, + "output": 64000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 + "input": 0.58, + "output": 3.02, + "cache_read": 0.1 } }, - "scx-ai-gp/kimi-k3": { - "id": "scx-ai-gp/kimi-k3", - "name": "Kimi K3 (SCX.ai)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2025-01-01", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", @@ -52413,28 +49805,26 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 262140, + "output": 262140 }, "cost": { - "input": 3.5, - "output": 18, - "cache_read": 0.35 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "scx-ai-gp/deepseek-v4.1-flash": { - "id": "scx-ai-gp/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (SCX.ai)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-01-01", + "release_date": "2025-09-04", + "last_updated": "2025-09-04", "modalities": { "input": [ "text" @@ -52443,237 +49833,268 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 262000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.01 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "scx-ai-gp/glm-5.2": { - "id": "scx-ai-gp/glm-5.2", - "name": "GLM-5.2 (SCX.ai)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "baidu/ernie-5.0-thinking-preview": { + "id": "baidu/ernie-5.0-thinking-preview", + "name": "ERNIE 5.0", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2026-01-22", + "last_updated": "2026-01-22", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 64000 + }, + "cost": { + "input": 0.84, + "output": 3.37 + } + }, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.88, - "output": 2.55, - "cache_read": 0.16 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025 } }, - "scx-ai-gp/glm-5.3": { - "id": "scx-ai-gp/glm-5.3", - "name": "GLM-5.3 (SCX.ai)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-01-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "pdf", + "image", + "text", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048000, + "output": 64000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 1.25, + "output": 10, + "cache_read": 0.31, + "cache_write": 4.5 } }, - "scx-ai-gp/kimi-k2.7-code": { - "id": "scx-ai-gp/kimi-k2.7-code", - "name": "Kimi K2.7 Code (SCX.ai)", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text", + "pdf", "image", - "video" + "text", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048000, + "output": 64000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.3, + "output": 2.5, + "cache_read": 0.07, + "cache_write": 1 } }, - "zai/glm-4.6v": { - "id": "zai/glm-4.6v", - "name": "GLM-4.6V (Z AI)", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.05 + "input": 1.5, + "output": 9, + "cache_read": 0.15 } }, - "zai/glm-4.5": { - "id": "zai/glm-4.5", - "name": "GLM-4.5 (Z AI)", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-01-01", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ - "text" + "pdf", + "image", + "text", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 98304 + "context": 1048000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0.1, + "output": 0.4, + "cache_read": 0.03, + "cache_write": 1 } }, - "zai/glm-5.3-flash": { - "id": "zai/glm-5.3-flash", - "name": "GLM-5.3 Flash (Z AI)", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -52681,92 +50102,127 @@ "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2026-02-19", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", - "video", - "pdf" + "pdf", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1048000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 4.5 } }, - "zai/glm-4.6": { - "id": "zai/glm-4.6", - "name": "GLM-4.6 (Z AI)", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2025-01-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ + "text", + "image", + "pdf", + "audio" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048000, + "output": 64000 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 1 + } + }, + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-03-20", + "last_updated": "2025-03-20", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 1050000, + "output": 65530 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0.25, + "output": 1.5 } }, - "zai/glm-4.5-x": { - "id": "zai/glm-4.5-x", - "name": "GLM-4.5 X (Z AI)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "x-ai/grok-code-fast-1": { + "id": "x-ai/grok-code-fast-1", + "name": "Grok Code Fast 1", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-01-01", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text" @@ -52777,77 +50233,124 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 2.2, - "output": 8.9, - "cache_read": 0.45 + "input": 0.2, + "output": 1.5, + "cache_read": 0.02 } }, - "zai/glm-5": { - "id": "zai/glm-5", - "name": "GLM-5 (Z AI)", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202800, - "output": 131100 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "cache_write": 0, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "cache_write": 0, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "cache_write": 0 + } } }, - "zai/glm-4.5-airx": { - "id": "zai/glm-4.5-airx", - "name": "GLM-4.5 AirX (Z AI)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "x-ai/grok-4.1-fast-non-reasoning": { + "id": "x-ai/grok-4.1-fast-non-reasoning", + "name": "Grok 4.1 Fast Non Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-01-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 64000 + }, + "cost": { + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 + } + }, + "x-ai/grok-4": { + "id": "x-ai/grok-4", + "name": "Grok 4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ + "image", "text" ], "output": [ @@ -52856,20 +50359,19 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 1.1, - "output": 4.5, - "cache_read": 0.22 + "input": 3, + "output": 15, + "cache_read": 0.75 } }, - "zai/glm-4.5v": { - "id": "zai/glm-4.5v", - "name": "GLM-4.5V (Z AI)", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "x-ai/grok-4-fast": { + "id": "x-ai/grok-4-fast", + "name": "Grok 4 Fast", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -52878,47 +50380,46 @@ } ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "knowledge": "2025-01-01", + "release_date": "2025-09-19", + "last_updated": "2025-09-19", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16000 + "context": 2000000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "zai/glm-4-32b-0414-128k": { - "id": "zai/glm-4-32b-0414-128k", - "name": "GLM-4 32B (0414-128k) (Z AI)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "x-ai/grok-4.2-fast-non-reasoning": { + "id": "x-ai/grok-4.2-fast-non-reasoning", + "name": "Grok 4.2 Fast Non Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -52926,35 +50427,31 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 3, + "output": 9 } }, - "zai/glm-4.6v-flashx": { - "id": "zai/glm-4.6v-flashx", - "name": "GLM-4.6V FlashX (Z AI)", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", + "x-ai/grok-4.2-fast": { + "id": "x-ai/grok-4.2-fast", + "name": "Grok 4.2 Fast", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -52962,21 +50459,19 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16000 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.04, - "output": 0.4, - "cache_read": 0.004 + "input": 3, + "output": 9 } }, - "zai/glm-4.7": { - "id": "zai/glm-4.7", - "name": "GLM-4.7 (Z AI)", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, + "x-ai/grok-4.1-fast": { + "id": "x-ai/grok-4.1-fast", + "name": "Grok 4.1 Fast", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -52984,58 +50479,79 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2025-01-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 2000000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "zai/glm-5.2": { - "id": "zai/glm-5.2", - "name": "GLM-5.2 (Z AI)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "x-ai/grok-build-0.1": { + "id": "x-ai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 1, + "output": 2, + "cache_read": 0.2 + } + }, + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "GLM 4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-01-01", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -53044,46 +50560,67 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.28, + "output": 1.14, + "cache_read": 0.06 } }, - "zai/glm-5.1": { - "id": "zai/glm-5.1", - "name": "GLM-5.1 (Z AI)", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "z-ai/glm-4.5": { + "id": "z-ai/glm-4.5", + "name": "GLM 4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 64000 + }, + "cost": { + "input": 0.35, + "output": 1.54, + "cache_read": 0.07 + } + }, + "z-ai/glm-4.7-flashx": { + "id": "z-ai/glm-4.7-flashx", + "name": "GLM 4.7 FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-01-01", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -53092,22 +50629,21 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 200000, - "output": 128000 + "output": 64000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.07, + "output": 0.42, + "cache_read": 0.01 } }, - "zai/glm-4.7-flashx": { - "id": "zai/glm-4.7-flashx", - "name": "GLM-4.7 FlashX (Z AI)", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -53116,11 +50652,13 @@ } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2026-04-03", + "last_updated": "2026-04-03", "modalities": { "input": [ "text" @@ -53129,30 +50667,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 200000, - "output": 128000 + "output": 131072 }, "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.01 + "input": 0.8781, + "output": 3.5126, + "cache_read": 0.1903 } }, - "zai/glm-4.5-air": { - "id": "zai/glm-4.5-air", - "name": "GLM-4.5 Air (Z AI)", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "GLM 4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-01-01", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -53161,21 +50698,21 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 98304 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03 + "input": 0.35, + "output": 1.54, + "cache_read": 0.07 } }, - "zai/glm-5.3": { - "id": "zai/glm-5.3", - "name": "GLM-5.3 (Z AI)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, @@ -53183,17 +50720,16 @@ { "type": "effort", "values": [ - "low", "high", "max" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -53205,44 +50741,31 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 131072 }, "cost": { "input": 1.4, - "output": 4.4, + "output": 4.5, "cache_read": 0.26 } }, - "aws-bedrock/claude-haiku-4-5": { - "id": "aws-bedrock/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (AWS Bedrock)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "z-ai/glm-4.6v-flash-free": { + "id": "z-ai/glm-4.6v-flash-free", + "name": "GLM 4.6V Flash (Free)", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-01-01", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -53254,64 +50777,55 @@ "output": 64000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0, + "output": 0 } }, - "aws-bedrock/llama-4-maverick-17b-instruct": { - "id": "aws-bedrock/llama-4-maverick-17b-instruct", - "name": "Llama 4 Maverick 17B Instruct (AWS Bedrock)", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "z-ai/glm-4.7-flash-free": { + "id": "z-ai/glm-4.7-flash-free", + "name": "GLM 4.7 Flash (Free)", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-01-01", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 2048 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.24, - "output": 0.97 + "input": 0, + "output": 0 } }, - "aws-bedrock/claude-sonnet-4-5": { - "id": "aws-bedrock/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (AWS Bedrock)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "z-ai/glm-5.2-free": { + "id": "z-ai/glm-5.2-free", + "name": "GLM 5.2 (Free)", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - }, { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh", "max" ] } @@ -53319,14 +50833,44 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 + } + }, + "z-ai/glm-4.6v": { + "id": "z-ai/glm-4.6v", + "name": "GLM 4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -53335,44 +50879,33 @@ "open_weights": false, "limit": { "context": 200000, - "output": 8192 + "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.14, + "output": 0.42, + "cache_read": 0.03 } }, - "aws-bedrock/claude-fable-5-1": { - "id": "aws-bedrock/claude-fable-5-1", - "name": "Claude Fable 5.1 (AWS Bedrock)", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", + "z-ai/glm-5v-turbo": { + "id": "z-ai/glm-5v-turbo", + "name": "GLM 5V Turbo", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", "image", + "video", "pdf" ], "output": [ @@ -53381,135 +50914,132 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 200000, "output": 128000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 0.726, + "output": 3.1946, + "cache_read": 0.1743 } }, - "aws-bedrock/llama-4-scout-17b-instruct": { - "id": "aws-bedrock/llama-4-scout-17b-instruct", - "name": "Llama 4 Scout 17B Instruct (AWS Bedrock)", - "description": "Open Llama with long-context vision for efficient multimodal agents", - "family": "llama", + "z-ai/glm-4.6v-flash": { + "id": "z-ai/glm-4.6v-flash", + "name": "GLM 4.6V FlashX", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-01-01", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 2048 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.17, - "output": 0.66 + "input": 0.02, + "output": 0.21, + "cache_read": 0.0043 } }, - "aws-bedrock/claude-opus-4-1-20250805": { - "id": "aws-bedrock/claude-opus-4-1-20250805", - "name": "Claude Opus 4.1 (AWS Bedrock)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "z-ai/glm-4.5-air": { + "id": "z-ai/glm-4.5-air", + "name": "GLM 4.5 Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01-01", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, "open_weights": false, + "limit": { + "context": 128000, + "output": 64000 + }, + "cost": { + "input": 0.11, + "output": 0.56, + "cache_read": 0.02 + } + }, + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM 5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, "limit": { "context": 200000, - "output": 32000 + "output": 128000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.58, + "output": 2.6, + "cache_read": 0.14 } }, - "aws-bedrock/claude-opus-4-5-20251101": { - "id": "aws-bedrock/claude-opus-4-5-20251101", - "name": "Claude Opus 4.5 (AWS Bedrock)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "GLM 5 Turbo", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "knowledge": "2025-01-01", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -53518,20 +51048,17 @@ "open_weights": false, "limit": { "context": 200000, - "output": 32000 + "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.88, + "output": 3.48 } }, - "aws-bedrock/claude-opus-5": { - "id": "aws-bedrock/claude-opus-5", - "name": "Claude Opus 5 (AWS Bedrock)", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", + "openai/gpt-5.5-instant": { + "id": "openai/gpt-5.5-instant", + "name": "GPT-5.5 Instant", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -53540,18 +51067,16 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "structured_output": true, + "temperature": true, + "knowledge": "2025-12-01", + "release_date": "2026-05-05", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -53564,41 +51089,37 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 400000, "output": 128000 }, "cost": { "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "output": 30, + "cache_read": 0.5 } }, - "aws-bedrock/claude-fable-5": { - "id": "aws-bedrock/claude-fable-5", - "name": "Claude Fable 5 (AWS Bedrock)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2-Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": false, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -53611,21 +51132,22 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" + }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 21, + "output": 168 } }, - "aws-bedrock/claude-opus-4-8": { - "id": "aws-bedrock/claude-opus-4-8", - "name": "Claude Opus 4.8 (AWS Bedrock)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -53634,18 +51156,15 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -53658,51 +51177,35 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 400000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.25, + "output": 10, + "cache_read": 0.12 } }, - "aws-bedrock/claude-sonnet-4-5-20250929": { - "id": "aws-bedrock/claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5 (2025-09-29) (AWS Bedrock)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "openai/gpt-5.1-chat": { + "id": "openai/gpt-5.1-chat", + "name": "GPT-5.1 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-01-01", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", + "pdf", "image", - "pdf" + "text" ], "output": [ "text" @@ -53710,46 +51213,33 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 128000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 1.25, + "output": 10, + "cache_read": 0.12 } }, - "aws-bedrock/claude-sonnet-5": { - "id": "aws-bedrock/claude-sonnet-5", - "name": "Claude Sonnet 5 (AWS Bedrock)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -53757,46 +51247,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" + }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.2, + "output": 1.25 } }, - "aws-bedrock/claude-opus-4-6": { - "id": "aws-bedrock/claude-opus-4-6", - "name": "Claude Opus 4.6 (AWS Bedrock)", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "openai/gpt-5.3-chat": { + "id": "openai/gpt-5.3-chat", + "name": "GPT-5.3 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -53804,28 +51280,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 16380 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.75, + "output": 14 } }, - "aws-bedrock/grok-4-3": { - "id": "aws-bedrock/grok-4-3", - "name": "Grok 4.3 (AWS Bedrock)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -53833,10 +51309,10 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "knowledge": "2025-01-01", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -53848,36 +51324,23 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 400000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "output": 10, + "cache_read": 0.12 } }, - "aws-bedrock/claude-haiku-4-5-20251001": { - "id": "aws-bedrock/claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5 (2025-10-01) (AWS Bedrock)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -53886,22 +51349,19 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "temperature": false, + "knowledge": "2025-01-01", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ @@ -53910,21 +51370,23 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 400000, "output": 64000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" + }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 1.75, + "output": 14, + "cache_read": 0.17 } }, - "aws-bedrock/grok-4-6": { - "id": "aws-bedrock/grok-4-6", - "name": "Grok 4.6 (AWS Bedrock)", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -53933,21 +51395,18 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -53955,50 +51414,43 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 400000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 1.75, + "output": 14 } }, - "aws-bedrock/claude-sonnet-4-6": { - "id": "aws-bedrock/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6 (AWS Bedrock)", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1-Codex-Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-01-01", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" @@ -54006,21 +51458,23 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, "output": 64000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" + }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.25, + "output": 2, + "cache_read": 0.03 } }, - "aws-bedrock/claude-opus-4-7": { - "id": "aws-bedrock/claude-opus-4-7", - "name": "Claude Opus 4.7 (AWS Bedrock)", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -54029,23 +51483,19 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -54053,28 +51503,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" + }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 3.75, + "output": 18.75 } }, - "aws-bedrock/llama-3.1-70b-instruct": { - "id": "aws-bedrock/llama-3.1-70b-instruct", - "name": "Llama 3.1 70B Instruct (AWS Bedrock)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ "text" @@ -54085,45 +51536,43 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 2048 + "context": 400000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 0.75, + "output": 4.5 } }, - "google-ai-studio/gemini-3.6-flash": { - "id": "google-ai-studio/gemini-3.6-flash", - "name": "Gemini 3.6 Flash (Google AI Studio)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2025-08-31", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -54131,46 +51580,45 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.08333 + "input": 45, + "output": 225 } }, - "google-ai-studio/gemini-3.5-flash-lite": { - "id": "google-ai-studio/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite (Google AI Studio)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -54179,22 +51627,34 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.08333 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "google-ai-studio/gemini-3.1-pro-preview": { - "id": "google-ai-studio/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro (Preview) (Google AI Studio)", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "GPT-5 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -54207,18 +51667,14 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "knowledge": "2025-01-01", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -54226,43 +51682,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 1.25, + "output": 10, + "cache_read": 0.12 } }, - "google-ai-studio/gemini-3.5-flash": { - "id": "google-ai-studio/gemini-3.5-flash", - "name": "Gemini 3.5 Flash (Google AI Studio)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -54270,17 +51712,14 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "temperature": false, + "knowledge": "2025-01-01", + "release_date": "2026-01-15", + "last_updated": "2026-01-15", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -54289,51 +51728,44 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 0.08333 + "input": 1.75, + "output": 14, + "cache_read": 0.17 } }, - "google-ai-studio/gemini-2.5-pro": { - "id": "google-ai-studio/gemini-2.5-pro", - "name": "Gemini 2.5 Pro (Google AI Studio)", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" ] - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2025-01-01", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", "image", - "audio", - "video", + "text", "pdf" ], "output": [ @@ -54342,44 +51774,27 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "https://zenmux.ai/api/v1" }, "cost": { "input": 1.25, "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "cache_read": 0.12 } }, - "google-ai-studio/gemini-2.5-flash": { - "id": "google-ai-studio/gemini-2.5-flash", - "name": "Gemini 2.5 Flash (Google AI Studio)", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1, - "max": 24576 - }, { "type": "effort", "values": [ @@ -54391,16 +51806,14 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -54409,93 +51822,129 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "google-ai-studio/gemini-3.7-flash": { - "id": "google-ai-studio/gemini-3.7-flash", - "name": "Gemini 3.7 Flash (Google AI Studio)", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "video", "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.08333 + "input": 0.4, + "output": 2, + "cache_read": 0.08, + "tiers": [ + { + "input": 0.8, + "output": 4, + "cache_read": 0.16, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.8, + "output": 4, + "cache_read": 0.16 + } } }, - "google-ai-studio/gemini-3-flash-preview": { - "id": "google-ai-studio/gemini-3-flash-preview", - "name": "Gemini 3 Flash (Preview) (Google AI Studio)", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "xiaomi/mimo-v2-omni": { + "id": "xiaomi/mimo-v2-omni", + "name": "MiMo V2 Omni", + "description": "MiMo omni model for text, image, video, audio, and agents", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text", "image", - "video", "audio", + "video", "pdf" ], "output": [ @@ -54504,124 +51953,201 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 265000, + "output": 265000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 0.4, + "output": 2, + "cache_read": 0.08 } }, - "google-ai-studio/gemini-pro-latest": { - "id": "google-ai-studio/gemini-pro-latest", - "name": "Gemini Pro Latest (Google AI Studio)", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "xiaomi/mimo-v2-flash": { + "id": "xiaomi/mimo-v2-flash", + "name": "MiMo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-27", - "last_updated": "2026-02-27", + "knowledge": "2024-12-01", + "release_date": "2025-12-16", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 262144, "output": 65536 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 0.1, + "output": 0.3, + "cache_read": 0.01 } }, - "google-ai-studio/gemini-3.8-flash": { - "id": "google-ai-studio/gemini-3.8-flash", - "name": "Gemini 3.8 Flash (Google AI Studio)", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, + "xiaomi/mimo-v2-pro": { + "id": "xiaomi/mimo-v2-pro", + "name": "MiMo V2 Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, "open_weights": false, + "limit": { + "context": 1000000, + "output": 256000 + }, + "cost": { + "input": 1, + "output": 3, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } + } + }, + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 131072 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.08333 + "input": 1, + "output": 3, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "google-ai-studio/gemini-2.5-flash-lite": { - "id": "google-ai-studio/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash Lite (Google AI Studio)", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", + "sapiens-ai/agnes-1.5-pro": { + "id": "sapiens-ai/agnes-1.5-pro", + "name": "Agnes 1.5 Pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-21", + "last_updated": "2026-03-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.16, + "output": 0.8 + } + }, + "sapiens-ai/agnes-1.5-lite": { + "id": "sapiens-ai/agnes-1.5-lite", + "name": "Agnes 1.5 Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-03-26", + "last_updated": "2026-03-26", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -54629,27 +52155,24 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 + "input": 0.12, + "output": 0.6 } }, - "google-ai-studio/gemini-3.1-flash-lite": { - "id": "google-ai-studio/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite (Google AI Studio)", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "volcengine/doubao-seed-2.0-pro": { + "id": "volcengine/doubao-seed-2.0-pro", + "name": "Doubao-Seed-2.0-pro", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -54657,18 +52180,15 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "knowledge": "2026-02-14", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" @@ -54676,21 +52196,20 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 0.08333 + "input": 0.45, + "output": 2.24, + "cache_read": 0.09, + "cache_write": 0.0024 } }, - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (Anthropic)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "volcengine/doubao-seed-code": { + "id": "volcengine/doubao-seed-code", + "name": "Doubao-Seed-Code", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -54699,23 +52218,19 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-01-01", + "release_date": "2025-11-11", + "last_updated": "2025-11-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -54723,51 +52238,41 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 256000, "output": 64000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.17, + "output": 1.12, + "cache_read": 0.03 } }, - "anthropic/claude-sonnet-4-5": { - "id": "anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (Anthropic)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "volcengine/doubao-seed-1.8": { + "id": "volcengine/doubao-seed-1.8", + "name": "Doubao-Seed-1.8", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-01-01", + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -54775,21 +52280,20 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 256000, "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.11, + "output": 0.28, + "cache_read": 0.02, + "cache_write": 0.0024 } }, - "anthropic/claude-opus-5-5": { - "id": "anthropic/claude-opus-5-5", - "name": "Claude Opus 5.5 (Anthropic)", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", + "volcengine/doubao-seed-2.0-mini": { + "id": "volcengine/doubao-seed-2.0-mini", + "name": "Doubao-Seed-2.0-mini", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -54798,23 +52302,20 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "knowledge": "2026-02-14", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -54822,21 +52323,20 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 + "input": 0.03, + "output": 0.28, + "cache_read": 0.01, + "cache_write": 0.0024 } }, - "anthropic/claude-fable-5-1": { - "id": "anthropic/claude-fable-5-1", - "name": "Claude Fable 5.1 (Anthropic)", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", + "volcengine/doubao-seed-2.0-lite": { + "id": "volcengine/doubao-seed-2.0-lite", + "name": "Doubao-Seed-2.0-lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -54845,23 +52345,20 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "temperature": true, + "knowledge": "2026-02-14", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -54869,51 +52366,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 0.09, + "output": 0.51, + "cache_read": 0.02, + "cache_write": 0.0024 } }, - "anthropic/claude-opus-4-5-20251101": { - "id": "anthropic/claude-opus-4-5-20251101", - "name": "Claude Opus 4.5 (Anthropic)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "volcengine/doubao-seed-2.0-code": { + "id": "volcengine/doubao-seed-2.0-code", + "name": "Doubao Seed 2.0 Code", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "knowledge": "2025-01-01", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -54921,21 +52397,19 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 256000, "output": 32000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.9, + "output": 4.48 } }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5 (Anthropic)", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", + "anthropic/claude-sonnet-5-free": { + "id": "anthropic/claude-sonnet-5-free", + "name": "Claude Sonnet 5 (Free)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -54944,18 +52418,15 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", @@ -54971,43 +52442,32 @@ "context": 1000000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" + }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5 (Anthropic)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "anthropic/claude-3.5-haiku": { + "id": "anthropic/claude-3.5-haiku", + "name": "Claude 3.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2024-11-04", + "last_updated": "2024-11-04", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -55015,21 +52475,24 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 } }, - "anthropic/claude-opus-4-8": { - "id": "anthropic/claude-opus-4-8", - "name": "Claude Opus 4.8 (Anthropic)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -55038,18 +52501,15 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", @@ -55063,49 +52523,44 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "anthropic/claude-sonnet-4-5-20250929": { - "id": "anthropic/claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5 (2025-09-29) (Anthropic)", + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude Sonnet 4", "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-01-01", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ @@ -55114,9 +52569,13 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 1000000, "output": 64000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" + }, "cost": { "input": 3, "output": 15, @@ -55124,9 +52583,45 @@ "cache_write": 3.75 } }, + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", + "modalities": { + "input": [ + "image", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 + } + }, "anthropic/claude-sonnet-5": { "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5 (Anthropic)", + "name": "Claude Sonnet 5", "description": "Everyday Claude agent model for coding, planning, browsing, and general work", "family": "claude-sonnet", "attachment": true, @@ -55137,14 +52632,11 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, "knowledge": "2026-01-31", "release_date": "2026-06-30", @@ -55164,18 +52656,21 @@ "context": 1000000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" + }, "cost": { "input": 2, "output": 10, "cache_read": 0.2, - "cache_write": 2.5 + "cache_write": 4 } }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6 (Anthropic)", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -55184,18 +52679,15 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -55211,6 +52703,10 @@ "context": 1000000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" + }, "cost": { "input": 5, "output": 25, @@ -55218,11 +52714,10 @@ "cache_write": 6.25 } }, - "anthropic/claude-haiku-4-5-20251001": { - "id": "anthropic/claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5 (2025-10-01) (Anthropic)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "anthropic/claude-3.7-sonnet": { + "id": "anthropic/claude-3.7-sonnet", + "name": "Claude 3.7 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -55231,18 +52726,15 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-01-01", + "release_date": "2025-02-24", + "last_updated": "2025-02-24", "modalities": { "input": [ "text", @@ -55258,43 +52750,39 @@ "context": 200000, "output": 64000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" + }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6 (Anthropic)", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -55308,20 +52796,24 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7 (Anthropic)", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -55330,18 +52822,15 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", @@ -55357,246 +52846,256 @@ "context": 1000000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" + }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "deepseek/deepseek-v4.1-flash": { - "id": "deepseek/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (DeepSeek)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-01-01", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1050000, - "output": 393216 + "context": 200000, + "output": 64000 }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro (DeepSeek)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "high", - "max" + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-01-01", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1050000, - "output": 393216 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "runpod/kimi-k3": { - "id": "runpod/kimi-k3", - "name": "Kimi K3 (Runpod)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-18", + "last_updated": "2026-02-18", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { "input": 3, "output": 15, - "cache_read": 0.3 + "cache_read": 0.3, + "cache_write": 3.75 } }, - "runware/gemma-4-31b-it": { - "id": "runware/gemma-4-31b-it", - "name": "Gemma 4 31B IT (Runware)", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "high", - "xhigh", - "max" + "low", + "medium", + "high" ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-01-01", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 32000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.102, - "output": 0.297, - "cache_read": 0.012 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "runware/glm-5.3-flash": { - "id": "runware/glm-5.3-flash", - "name": "GLM-5.3 Flash (Runware)", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-05-31", + "release_date": "2026-02-06", + "last_updated": "2026-02-06", "modalities": { "input": [ - "text", "image", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "runware/kimi-k3": { - "id": "runware/kimi-k3", - "name": "Kimi K3 (Runware)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "tencent/hy3-preview": { + "id": "tencent/hy3-preview", + "name": "Hy3 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -55604,21 +53103,17 @@ "values": [ "none", "low", - "high", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -55626,97 +53121,72 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 256000, + "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.172, + "output": 0.572, + "cache_read": 0.058, + "cache_write": 0 } }, - "runware/deepseek-v4.1-flash": { - "id": "runware/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (Runware)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "stepfun/step-3": { + "id": "stepfun/step-3", + "name": "Step-3", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-01-01", + "release_date": "2025-07-31", + "last_updated": "2025-07-31", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 65536, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.01 + "input": 0.21, + "output": 0.57 } }, - "runware/kimi-k2.6": { - "id": "runware/kimi-k2.6", - "name": "Kimi K2.6 (Runware)", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "stepfun/step-3.7-flash": { + "id": "stepfun/step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2026-01-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -55724,47 +53194,40 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.6, - "output": 3.05, - "cache_read": 0.13 + "input": 0.2, + "output": 1.15 } }, - "runware/glm-5.2": { - "id": "runware/glm-5.2", - "name": "GLM-5.2 (Runware)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "stepfun/step-3.7-flash-free": { + "id": "stepfun/step-3.7-flash-free", + "name": "Step 3.7 Flash (Free)", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2026-01-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -55772,42 +53235,26 @@ }, "open_weights": true, "limit": { - "context": 1024000, - "output": 128000 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.8, - "output": 2.55, - "cache_read": 0.16 + "input": 0, + "output": 0 } }, - "runware/deepseek-v4-pro": { - "id": "runware/deepseek-v4-pro", - "name": "DeepSeek V4 Pro (Runware)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "stepfun/step-3.5-flash": { + "id": "stepfun/step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-01-01", + "release_date": "2026-02-02", + "last_updated": "2026-02-02", "modalities": { "input": [ "text" @@ -55816,39 +53263,26 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.961, - "output": 1.922, - "cache_read": 0.079 + "input": 0.1, + "output": 0.3 } }, - "runware/gpt-oss-120b": { - "id": "runware/gpt-oss-120b", - "name": "GPT OSS 120B (Runware)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "kuaishou/kat-coder-pro-v2": { + "id": "kuaishou/kat-coder-pro-v2", + "name": "KAT-Coder-Pro-V2", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-03-30", + "last_updated": "2026-03-30", "modalities": { "input": [ "text" @@ -55857,39 +53291,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 256000, + "output": 80000 }, "cost": { - "input": 0.032, - "output": 0.14, - "cache_read": 0.032 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "runware/glm-5.3": { - "id": "runware/glm-5.3", - "name": "GLM-5.3 (Runware)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen3-Coder-Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-01-01", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -55898,93 +53321,39 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 64000 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.2 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "runware/deepseek-v4-flash": { - "id": "runware/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (Runware)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.076, - "output": 0.153, - "cache_read": 0.014 - } - }, - "tencent/glm-5v-turbo": { - "id": "tencent/glm-5v-turbo", - "name": "GLM-5V Turbo (Tencent Cloud)", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" @@ -55992,91 +53361,50 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 - } - }, - "tencent/kimi-k3": { - "id": "tencent/kimi-k3", - "name": "Kimi K3 (Tencent Cloud)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "cache_write": 0.5, + "tiers": [ + { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24, + "cache_write": 1.5, + "tier": { + "type": "context", + "size": 256000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "context_over_200k": { + "input": 1.2, + "output": 4.8, + "cache_read": 0.24, + "cache_write": 1.5 + } } }, - "tencent/glm-5": { - "id": "tencent/glm-5", - "name": "GLM-5 (Tencent Cloud)", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -56085,124 +53413,83 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "tencent/kimi-k2.7-code-highspeed": { - "id": "tencent/kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed (Tencent Cloud)", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": true, + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3-Max-Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2026-01-23", + "last_updated": "2026-01-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 64000 }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 + "input": 1.2, + "output": 6 } }, - "tencent/kimi-k2.6": { - "id": "tencent/kimi-k2.6", - "name": "Kimi K2.6 (Tencent Cloud)", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "qwen/qwen3.5-flash": { + "id": "qwen/qwen3.5-flash", + "name": "Qwen3.5 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-01-01", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1020000, + "output": 1020000 }, "cost": { - "input": 0.858, - "output": 3.566, - "cache_read": 0.145 + "input": 0.1, + "output": 0.4 } }, - "tencent/minimax-m2.7": { - "id": "tencent/minimax-m2.7", - "name": "MiniMax M2.7 (Tencent Cloud)", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, + "qwen/qwen3.5-plus": { + "id": "qwen/qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -56210,63 +53497,49 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2025-01-01", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.8, + "output": 4.8 } }, - "tencent/mimo-v2.5-pro": { - "id": "tencent/mimo-v2.5-pro", - "name": "MiMo V2.5 Pro (Tencent Cloud)", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6-Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-03-30", + "last_updated": "2026-03-30", "modalities": { "input": [ "text" @@ -56275,20 +53548,22 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 64000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036, + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, "tiers": [ { "input": 2, "output": 6, - "cache_read": 0.4, + "cache_read": 0.2, + "cache_write": 2.5, "tier": { "type": "context", "size": 256000 @@ -56298,39 +53573,43 @@ "context_over_200k": { "input": 2, "output": 6, - "cache_read": 0.4 + "cache_read": 0.2, + "cache_write": 2.5 } } }, - "tencent/minimax-m3": { - "id": "tencent/minimax-m3", - "name": "MiniMax M3 (Tencent Cloud)", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -56339,32 +53618,31 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 131072 + "output": 384000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "tencent/glm-5.2": { - "id": "tencent/glm-5.2", - "name": "GLM-5.2 (Tencent Cloud)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], @@ -56372,10 +53650,11 @@ "interleaved": { "field": "reasoning_content" }, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -56387,39 +53666,30 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 384000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "tencent/hy3": { - "id": "tencent/hy3", - "name": "Hy3 (Tencent Cloud)", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + "deepseek/deepseek-v3.2-exp": { + "id": "deepseek/deepseek-v3.2-exp", + "name": "DeepSeek-V3.2-Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "knowledge": "2025-01-01", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text" @@ -56428,29 +53698,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 163000, + "output": 64000 }, "cost": { - "input": 0.132, - "output": 0.528, - "cache_read": 0.033 + "input": 0.22, + "output": 0.33 } }, - "tencent/hy-mt2-plus": { - "id": "tencent/hy-mt2-plus", - "name": "Hy-MT2 Plus (Tencent Cloud)", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + "deepseek/deepseek-chat": { + "id": "deepseek/deepseek-chat", + "name": "DeepSeek-V3.2 (Non-thinking Mode)", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2025-01-01", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -56461,41 +53729,31 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 128000, + "output": 64000 }, "cost": { - "input": 0.074, - "output": 0.295 + "input": 0.28, + "output": 0.42, + "cache_read": 0.03 } }, - "tencent/glm-5.1": { - "id": "tencent/glm-5.1", - "name": "GLM-5.1 (Tencent Cloud)", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-01-01", + "release_date": "2025-12-05", + "last_updated": "2025-12-05", "modalities": { "input": [ "text" @@ -56504,41 +53762,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 128000, + "output": 64000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.28, + "output": 0.43 } }, - "tencent/glm-5-turbo": { - "id": "tencent/glm-5-turbo", - "name": "GLM-5 Turbo (Tencent Cloud)", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax M2.7 highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "knowledge": "2025-01-01", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ "text" @@ -56549,45 +53794,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 204800, + "output": 131070 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 0.611, + "output": 2.4439 } }, - "tencent/deepseek-v4-pro": { - "id": "tencent/deepseek-v4-pro", - "name": "DeepSeek V4 Pro (Tencent Cloud)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-01-01", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -56596,43 +53826,34 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 393216 + "context": 204800, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.00363 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "tencent/hy4-preview": { - "id": "tencent/hy4-preview", - "name": "Hy4 Preview (Tencent Cloud)", - "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", - "family": "Hy", + "minimax/minimax-m2.5-lightning": { + "id": "minimax/minimax-m2.5-lightning", + "name": "MiniMax M2.5 highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", + "knowledge": "2025-01-01", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -56641,46 +53862,74 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 204800, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.834, - "output": 2.501, - "cache_read": 0.042 + "input": 0.6, + "output": 4.8, + "cache_read": 0.06, + "cache_write": 0.75 } }, - "tencent/kimi-k2.7-code": { - "id": "tencent/kimi-k2.7-code", - "name": "Kimi K2.7 Code (Tencent Cloud)", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01-01", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 204000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.38 + } + }, + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", @@ -56693,45 +53942,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 512000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.6, + "output": 2.4 } }, - "tencent/deepseek-v4-flash": { - "id": "tencent/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (Tencent Cloud)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-01-01", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ "text" @@ -56740,45 +53974,37 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 393216 + "context": 204000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.38 } }, - "meta-contributor/muse-spark-1.2-contributor": { - "id": "meta-contributor/muse-spark-1.2-contributor", - "name": "Muse Spark 1.2 Contributor (Meta Contributor)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "muse", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-06", - "last_updated": "2026-08-06", + "knowledge": "2025-01-01", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -56786,29 +54012,41 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 204800, + "output": 131070 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://zenmux.ai/api/anthropic/v1" }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 + "input": 0.3055, + "output": 1.2219 } - }, - "meta-contributor/muse-spark-1.3-contributor": { - "id": "meta-contributor/muse-spark-1.3-contributor", - "name": "Muse Spark 1.3 Contributor (Meta Contributor)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "muse", - "attachment": true, + } + } + }, + "kenari": { + "id": "kenari", + "env": [ + "KENARI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://kenari.id/v1", + "name": "Kenari", + "doc": "https://kenari.id/docs", + "models": { + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", "high", "xhigh" ] @@ -56817,46 +54055,41 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 + "input": 0, + "output": 0 } }, - "novita/glm-4.6v": { - "id": "novita/glm-4.6v", - "name": "GLM-4.6V (NovitaAI)", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "kimi-k2-6": { + "id": "kimi-k2-6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -56869,27 +54102,27 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.055 + "input": 0, + "output": 0 } }, - "novita/gemma-4-31b-it": { - "id": "novita/gemma-4-31b-it", - "name": "Gemma 4 31B IT (NovitaAI)", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "glm-5-1": { + "id": "glm-5-1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -56900,27 +54133,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.4 + "input": 0, + "output": 0 } }, - "novita/qwen3-vl-235b-a22b-instruct": { - "id": "novita/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct (NovitaAI)", - "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemma-4-31b-it": { + "id": "gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -56932,64 +54165,19 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 262144, "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.5 - } - }, - "novita/qwen3.7-max": { - "id": "novita/qwen3.7-max", - "name": "Qwen3.7 Max (NovitaAI)", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 3.75, - "cache_read": 0.25 + "input": 0, + "output": 0 } }, - "novita/mimo-v2.5": { - "id": "novita/mimo-v2.5", - "name": "MiMo V2.5 (NovitaAI)", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "grok-4-3": { + "id": "grok-4-3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -57004,102 +54192,81 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.168, - "output": 0.336, - "cache_read": 0.0034, - "tiers": [ - { - "input": 0.8, - "output": 4, - "cache_read": 0.16, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.8, - "output": 4, - "cache_read": 0.16 - } + "input": 0, + "output": 0 } }, - "novita/qwen3.8-27b": { - "id": "novita/qwen3.8-27b", - "name": "Qwen3.8 27B (NovitaAI)", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "qwen3-7-plus": { + "id": "qwen3-7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 64000 }, "cost": { - "input": 0.42, - "output": 3, - "cache_read": 0.085 + "input": 0, + "output": 0 } }, - "novita/qwen3-vl-235b-a22b-thinking": { - "id": "novita/qwen3-vl-235b-a22b-thinking", - "name": "Qwen3 VL 235B A22B Thinking (NovitaAI)", - "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "kimi-k2-7-code": { + "id": "kimi-k2-7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -57107,40 +54274,36 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.98, - "output": 3.95 + "input": 0, + "output": 0 } }, - "novita/qwen3.8-2.4t-a95b": { - "id": "novita/qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B (NovitaAI)", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -57152,42 +54315,38 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 131072 + "output": 384000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 0, + "output": 0 } }, - "novita/glm-5.3-flash": { - "id": "novita/glm-5.3-flash", - "name": "GLM-5.3 Flash (NovitaAI)", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + "deepseek-v4-flash:free": { + "id": "deepseek-v4-flash:free", + "name": "DeepSeek V4 Flash (Free)", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -57195,67 +54354,75 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 0, + "output": 0 } }, - "novita/glm-4.6": { - "id": "novita/glm-4.6", - "name": "GLM-4.6 (NovitaAI)", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.55, - "output": 2.2, - "cache_read": 0.11 + "input": 0, + "output": 0 } }, - "novita/qwen3-235b-a22b-instruct-2507": { - "id": "novita/qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507 (NovitaAI)", - "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -57263,19 +54430,19 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 512000, + "output": 128000 }, "cost": { - "input": 0.09, - "output": 0.58 + "input": 0, + "output": 0 } }, - "novita/qwen3.8-max": { - "id": "novita/qwen3.8-max", - "name": "Qwen3.8 Max (NovitaAI)", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -57284,23 +54451,21 @@ "values": [ "low", "medium", - "xhigh" + "high", + "xhigh", + "max" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ @@ -57310,19 +54475,18 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 128000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 0, + "output": 0 } }, - "novita/kimi-k3": { - "id": "novita/kimi-k3", - "name": "Kimi K3 (NovitaAI)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "gpt-5-4-mini": { + "id": "gpt-5-4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -57330,58 +54494,58 @@ "type": "effort", "values": [ "none", - "minimal", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0, + "output": 0 } }, - "novita/llama-4-maverick-17b-instruct": { - "id": "novita/llama-4-maverick-17b-instruct", - "name": "Llama 4 Maverick 17B Instruct (NovitaAI)", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", + "mimo-v2-5": { + "id": "mimo-v2-5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -57390,42 +54554,35 @@ "open_weights": true, "limit": { "context": 1048576, - "output": 8192 + "output": 131072 }, "cost": { - "input": 0.27, - "output": 0.85 + "input": 0, + "output": 0 } }, - "novita/glm-5": { - "id": "novita/glm-5", - "name": "GLM-5 (NovitaAI)", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -57436,20 +54593,19 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0, + "output": 0 } }, - "novita/deepseek-v4.1-flash": { - "id": "novita/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (NovitaAI)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "gpt-5-5": { + "id": "gpt-5-5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -57457,65 +54613,61 @@ "type": "effort", "values": [ "none", - "minimal", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 393216 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.006 + "input": 0, + "output": 0 } }, - "novita/minimax-m2.5": { - "id": "novita/minimax-m2.5", - "name": "MiniMax M2.5 (NovitaAI)", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "glm-5-2": { + "id": "glm-5-2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -57526,82 +54678,32 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131100 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0, + "output": 0 } }, - "novita/kimi-k2.6": { - "id": "novita/kimi-k2.6", - "name": "Kimi K2.6 (NovitaAI)", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "grok-build-0-1": { + "id": "grok-build-0-1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.8, - "output": 3.4, - "cache_read": 0.16 - } - }, - "novita/qwen3-max": { - "id": "novita/qwen3-max", - "name": "Qwen3 Max (NovitaAI)", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text" + "pdf" ], "output": [ "text" @@ -57609,70 +54711,80 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.845, - "output": 3.38 + "input": 0, + "output": 0 } }, - "novita/glm-4.5v": { - "id": "novita/glm-4.5v", - "name": "GLM-4.5V (NovitaAI)", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 16000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11 + "input": 0, + "output": 0 } }, - "novita/llama-4-scout-17b-instruct": { - "id": "novita/llama-4-scout-17b-instruct", - "name": "Llama 4 Scout 17B Instruct (NovitaAI)", - "description": "Open Llama with long-context vision for efficient multimodal agents", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -57681,24 +54793,26 @@ "open_weights": true, "limit": { "context": 131072, - "output": 131072 + "output": 32768 }, "cost": { - "input": 0.18, - "output": 0.59 + "input": 0, + "output": 0 } }, - "novita/qwen3-235b-a22b-fp8": { - "id": "novita/qwen3-235b-a22b-fp8", - "name": "Qwen3 235B A22B FP8 (NovitaAI)", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "mimo-v2-5-pro": { + "id": "mimo-v2-5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -57707,29 +54821,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 40960, - "output": 20000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0, + "output": 0 } }, - "novita/qwen3-coder-480b-a35b-instruct": { - "id": "novita/qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct (NovitaAI)", - "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", - "family": "qwen", + "deepseek-v4-pro:free": { + "id": "deepseek-v4-pro:free", + "name": "DeepSeek V4 Pro (Free)", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -57740,99 +54863,86 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.38, - "output": 1.55 + "input": 0, + "output": 0 } }, - "novita/qwen3-coder-30b-a3b-instruct": { - "id": "novita/qwen3-coder-30b-a3b-instruct", - "name": "Qwen3 Coder 30B A3B Instruct (NovitaAI)", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", - "attachment": false, + "gpt-image-2": { + "id": "gpt-image-2", + "name": "GPT-Image-2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "tool_call": false, + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 160000, - "output": 32768 + "context": 272000, + "output": 16384 }, "cost": { - "input": 0.07, - "output": 0.27 + "input": 0, + "output": 0 } - }, - "novita/ling-3.0-flash": { - "id": "novita/ling-3.0-flash", - "name": "InclusionAI Ling 3.0 Flash (NovitaAI)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", - "attachment": false, + } + } + }, + "openai": { + "id": "openai", + "env": [ + "OPENAI_API_KEY" + ], + "npm": "@ai-sdk/openai", + "name": "OpenAI", + "doc": "https://platform.openai.com/docs/models", + "models": { + "gpt-6-sol": { + "id": "gpt-6-sol", + "name": "GPT-6 Sol", + "description": "OpenAI model for complex coding and agentic workflows", + "family": "gpt-sol", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none" + "none", + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-02", - "last_updated": "2026-08-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.012 - } - }, - "novita/qwen3-vl-30b-a3b-instruct": { - "id": "novita/qwen3-vl-30b-a3b-instruct", - "name": "Qwen3 VL 30B A3B Instruct (NovitaAI)", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-10-05", - "last_updated": "2025-10-05", + "structured_output": true, + "temperature": false, + "knowledge": "2026-04-20", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -57840,59 +54950,182 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } }, "cost": { - "input": 0.2, - "output": 0.7 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5, + "tiers": [ + { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 15, + "cache_read": 0.4, + "cache_write": 5 + } } }, - "novita/gemma-4-26b-a4b-it": { - "id": "novita/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT (NovitaAI)", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, - "reasoning": false, + "gpt-6-luna": { + "id": "gpt-6-luna", + "name": "GPT-6 Luna", + "description": "OpenAI's most efficient model for focused, high-volume tasks", + "family": "gpt-luna", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "structured_output": true, + "temperature": false, + "knowledge": "2026-05-18", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 0.2, + "output": 1, + "cache_read": 0.02, + "cache_write": 0.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 0.1, + "output": 0.5, + "cache_read": 0.01, + "cache_write": 0.125, + "tiers": [ + { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 0.2, + "output": 0.75, + "cache_read": 0.02, + "cache_write": 0.25 + } } }, - "novita/llama-3-70b-instruct": { - "id": "novita/llama-3-70b-instruct", - "name": "Llama 3 70B Instruct (NovitaAI)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2024-04-18", - "last_updated": "2024-04-18", + "o3": { + "id": "o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -57900,25 +55133,27 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 8000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.51, - "output": 0.74 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "novita/qwen3-235b-a22b-thinking-2507": { - "id": "novita/qwen3-235b-a22b-thinking-2507", - "name": "Qwen3 235B A22B Thinking 2507 (NovitaAI)", - "description": "Tool-capable chat model for instruction following and agentic application workflows", + "text-embedding-3-large": { + "id": "text-embedding-3-large", + "name": "text-embedding-3-large", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "tool_call": false, + "temperature": false, + "knowledge": "2024-01", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -57929,50 +55164,62 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 8191, + "output": 3072 }, "cost": { - "input": 0.3, - "output": 3 + "input": 0.13, + "output": 0 } }, - "novita/llama-3.3-70b-instruct": { - "id": "novita/llama-3.3-70b-instruct", - "name": "Llama 3.3 70B Instruct (NovitaAI)", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, + "gpt-5.2-pro": { + "id": "gpt-5.2-pro", + "name": "GPT-5.2 Pro", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 120000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.135, - "output": 0.4 + "input": 21, + "output": 168 } }, - "novita/qwen3.6-35b-a3b": { - "id": "novita/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B A3B (NovitaAI)", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "gpt-5.6": { + "id": "gpt-5.6", + "name": "GPT-5.6", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -57980,7 +55227,6 @@ "type": "effort", "values": [ "none", - "minimal", "low", "medium", "high", @@ -57990,86 +55236,136 @@ } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 64000 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 10, + "output": 60, + "cache_read": 1, + "cache_write": 12.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } }, "cost": { - "input": 0.248, - "output": 1.485 + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "novita/qwen3-next-80b-a3b-instruct": { - "id": "novita/qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct (NovitaAI)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "novita/minimax-m2.7": { - "id": "novita/minimax-m2.7", - "name": "MiniMax M2.7 (NovitaAI)", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "gpt-3.5-turbo": { + "id": "gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" @@ -58078,101 +55374,105 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131100 + "context": 16385, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.5, + "output": 1.5, + "cache_read": 0 } }, - "novita/mimo-v2.5-pro": { - "id": "novita/mimo-v2.5-pro", - "name": "MiMo V2.5 Pro (NovitaAI)", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", - "low", - "medium", "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] }, - "structured_output": false, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 272000 + }, + "cost": { + "input": 15, + "output": 120 + } + }, + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.522, - "output": 1.044, - "cache_read": 0.0043, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "novita/glm-4.7": { - "id": "novita/glm-4.7", - "name": "GLM-4.7 (NovitaAI)", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "gpt-4": { + "id": "gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ "text" @@ -58181,45 +55481,43 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 128000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 30, + "output": 60 } }, - "novita/qwen3.8-flash": { - "id": "novita/qwen3.8-flash", - "name": "Qwen3.8 Flash (NovitaAI)", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -58227,41 +55525,98 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "novita/qwen35-397b-a17b": { - "id": "novita/qwen35-397b-a17b", - "name": "Qwen3.5 397B A17B (NovitaAI)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.5", + "o3-pro": { + "id": "o3-pro", + "name": "o3-pro", + "description": "High-effort o3 tier for difficult technical reasoning and careful answers", + "family": "o-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 20, + "output": 80 + } + }, + "chatgpt-image-latest": { + "id": "chatgpt-image-latest", + "name": "chatgpt-image-latest", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 0, + "input": 0, + "output": 0 + } + }, + "gpt-4o-2024-05-13": { + "id": "gpt-4o-2024-05-13", + "name": "GPT-4o (2024-05-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ "text", @@ -58273,75 +55628,74 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 64000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 5, + "output": 15 } }, - "novita/glm-5.2": { - "id": "novita/glm-5.2", - "name": "GLM-5.2 (NovitaAI)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "minimal", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "novita/ernie-4.5-vl-424b-a47b": { - "id": "novita/ernie-4.5-vl-424b-a47b", - "name": "ERNIE 4.5 VL 424B A47B (NovitaAI)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ernie", + "gpt-5-chat-latest": { + "id": "gpt-5-chat-latest", + "name": "GPT-5 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": false, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -58353,71 +55707,90 @@ }, "open_weights": false, "limit": { - "context": 123000, - "output": 16000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.42, - "output": 1.25 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "novita/hy3": { - "id": "novita/hy3", - "name": "Hy3 (NovitaAI)", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", + "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 192000, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.58, - "cache_read": 0.035 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "novita/kimi-k2": { - "id": "novita/kimi-k2", - "name": "Kimi K2 (NovitaAI)", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "gpt-5.3-codex-spark": { + "id": "gpt-5.3-codex-spark", + "name": "GPT-5.3 Codex Spark", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex-spark", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-07-11", - "last_updated": "2025-07-11", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -58425,123 +55798,111 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "input": 100000, + "output": 32000 }, "cost": { - "input": 0.57, - "output": 2.3 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "novita/glm-5.1": { - "id": "novita/glm-5.1", - "name": "GLM-5.1 (NovitaAI)", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "gpt-5.1-codex-max": { + "id": "gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.38, - "output": 4.4, - "cache_read": 0.26 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "novita/step-3.7-flash": { - "id": "novita/step-3.7-flash", - "name": "Step 3.7 Flash (NovitaAI)", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "gpt-5.3-chat-latest": { + "id": "gpt-5.3-chat-latest", + "name": "GPT-5.3 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "knowledge": "2025-08-31", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 256000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "novita/llama-3.2-3b-instruct": { - "id": "novita/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct (NovitaAI)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "gpt-4o-2024-08-06": { + "id": "gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2024-09-18", + "knowledge": "2023-09", + "release_date": "2024-08-06", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -58549,32 +55910,27 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 32000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.03, - "output": 0.05 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "novita/deepseek-v3.2": { - "id": "novita/deepseek-v3.2", - "name": "DeepSeek V3.2 (NovitaAI)", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", + "text-embedding-ada-002": { + "id": "text-embedding-ada-002", + "name": "text-embedding-ada-002", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2022-12", + "release_date": "2022-12-15", + "last_updated": "2022-12-15", "modalities": { "input": [ "text" @@ -58583,22 +55939,21 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 65536 + "context": 8192, + "output": 1536 }, "cost": { - "input": 0.269, - "output": 0.4, - "cache_read": 0.1345 + "input": 0.1, + "output": 0 } }, - "novita/glm-5.3": { - "id": "novita/glm-5.3", - "name": "GLM-5.3 (NovitaAI)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -58606,16 +55961,17 @@ "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ "text" @@ -58624,203 +55980,122 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.4, + "input": 1.1, "output": 4.4, - "cache_read": 0.26 + "cache_read": 0.55 } }, - "novita/kimi-k2.7-code": { - "id": "novita/kimi-k2.7-code", - "name": "Kimi K2.7 Code (NovitaAI)", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - }, - "novita/minimax-m2.1": { - "id": "novita/minimax-m2.1", - "name": "MiniMax M2.1 (NovitaAI)", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "novita/deepseek-v4-flash": { - "id": "novita/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (NovitaAI)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "minimal", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1050000, - "output": 393216 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 - } - }, - "embercloud/glm-4.5": { - "id": "embercloud/glm-4.5", - "name": "GLM-4.5 (EmberCloud)", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131000, - "output": 96000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "embercloud/qwen3-coder-next": { - "id": "embercloud/qwen3-coder-next", - "name": "Qwen3 Coder Next (EmberCloud)", - "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", - "family": "qwen", + "text-embedding-3-small": { + "id": "text-embedding-3-small", + "name": "text-embedding-3-small", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", + "tool_call": false, + "temperature": false, + "knowledge": "2024-01", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -58829,30 +56104,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 8191, + "output": 1536 }, "cost": { - "input": 0.108, - "output": 0.675, - "cache_read": 0.06 + "input": 0.02, + "output": 0 } }, - "embercloud/glm-5": { - "id": "embercloud/glm-5", - "name": "GLM-5 (EmberCloud)", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "minimal", "low", "medium", "high", @@ -58862,149 +56135,134 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 203000, - "output": 131000 - }, - "cost": { - "input": 0.72, - "output": 2.3, - "cache_read": 0.144 - } - }, - "embercloud/kimi-k2.5": { - "id": "embercloud/kimi-k2.5", - "name": "Kimi K2.5 (EmberCloud)", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.405, - "output": 1.98, - "cache_read": 0.225 - } - }, - "embercloud/glm-4.7-flash": { - "id": "embercloud/glm-4.7-flash", - "name": "GLM-4.7 Flash (EmberCloud)", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1050000, + "input": 922000, + "output": 128000 }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131000 + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 2.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } }, "cost": { - "input": 0.06, - "output": 0.4, - "cache_read": 0.01 + "input": 1, + "output": 6, + "cache_read": 0.1, + "cache_write": 1.25, + "tiers": [ + { + "input": 2, + "output": 9, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 9, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "embercloud/glm-4.7": { - "id": "embercloud/glm-4.7", - "name": "GLM-4.7 (EmberCloud)", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.38, - "output": 1.98, - "cache_read": 0.19 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "embercloud/glm-5.2": { - "id": "embercloud/glm-5.2", - "name": "GLM-5.2 (EmberCloud)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "minimal", "low", "medium", "high", @@ -59014,150 +56272,103 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 203000, - "output": 131000 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } }, "cost": { - "input": 1.26, - "output": 3.96, - "cache_read": 0.234 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "cache_write": 6.25 + } } }, - "embercloud/glm-5.1": { - "id": "embercloud/glm-5.1", - "name": "GLM-5.1 (EmberCloud)", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "gpt-5.1-chat-latest": { + "id": "gpt-5.1-chat-latest", + "name": "GPT-5.1 Chat", + "description": "Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "medium" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 203000, - "output": 131000 - }, - "cost": { - "input": 0.931, - "output": 2.93, - "cache_read": 0.173 - } - }, - "embercloud/glm-4.5-air": { - "id": "embercloud/glm-4.5-air", - "name": "GLM-4.5 Air (EmberCloud)", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131000, - "output": 96000 - }, - "cost": { - "input": 0.13, - "output": 0.85, - "cache_read": 0.025 - } - }, - "perplexity/sonar-pro": { - "id": "perplexity/sonar-pro", - "name": "Sonar Pro (Perplexity)", - "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", - "family": "sonar-pro", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8192 - }, - "cost": { - "input": 3, - "output": 15 - } - }, - "perplexity/sonar": { - "id": "perplexity/sonar", - "name": "Sonar (Perplexity)", - "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", - "family": "sonar", - "attachment": false, - "reasoning": false, - "tool_call": false, "structured_output": true, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -59165,30 +56376,40 @@ }, "open_weights": false, "limit": { - "context": 130000, - "output": 4096 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1, - "output": 1 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "perplexity/sonar-reasoning-pro": { - "id": "perplexity/sonar-reasoning-pro", - "name": "Sonar Reasoning Pro (Perplexity)", - "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", - "family": "sonar-reasoning", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gpt-5.2-chat-latest": { + "id": "gpt-5.2-chat-latest", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium" + ] + } + ], + "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -59197,35 +56418,34 @@ "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 16384 }, "cost": { - "input": 2, - "output": 8 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "azure-ai-foundry/grok-4-1-fast-reasoning": { - "id": "azure-ai-foundry/grok-4-1-fast-reasoning", - "name": "Grok 4.1 Fast Reasoning (Azure AI Foundry)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "o4-mini-deep-research": { + "id": "o4-mini-deep-research", + "name": "o4-mini-deep-research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "medium" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-06-26", + "last_updated": "2024-06-26", "modalities": { "input": [ "text", @@ -59237,71 +56457,60 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.2, - "output": 0.5 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "azure-ai-foundry/grok-4-1-fast-non-reasoning": { - "id": "azure-ai-foundry/grok-4-1-fast-non-reasoning", - "name": "Grok 4.1 Fast Non-Reasoning (Azure AI Foundry)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "gpt-image-1.5": { + "id": "gpt-image-1.5", + "name": "gpt-image-1.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "tool_call": false, + "temperature": false, + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 - }, - "cost": { - "input": 0.2, - "output": 0.5 + "context": 0, + "input": 0, + "output": 0 } }, - "azure-ai-foundry/grok-4-3": { - "id": "azure-ai-foundry/grok-4-3", - "name": "Grok 4.3 (Azure AI Foundry)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -59309,62 +56518,32 @@ }, "open_weights": false, "limit": { - "context": 20000, - "output": 8192 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "meta/muse-spark-1.3": { - "id": "meta/muse-spark-1.3", - "name": "Muse Spark 1.3 (Meta)", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", - "family": "muse", + "gpt-4o-2024-11-20": { + "id": "gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2023-09", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", "modalities": { "input": [ "text", - "image", - "video", - "pdf", - "audio" + "image" ], "output": [ "text" @@ -59372,45 +56551,43 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "meta/muse-spark-1.1": { - "id": "meta/muse-spark-1.1", - "name": "Muse Spark 1.1 (Meta)", - "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", - "family": "muse", + "o1": { + "id": "o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text", "image", - "pdf", - "video" + "pdf" ], "output": [ "text" @@ -59418,46 +56595,42 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "meta/muse-spark-1.2": { - "id": "meta/muse-spark-1.2", - "name": "Muse Spark 1.2 (Meta)", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", + "o1-pro": { + "id": "o1-pro", + "name": "o1-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2025-03-19", + "last_updated": "2025-03-19", "modalities": { "input": [ "text", - "image", - "video", - "pdf", - "audio" + "image" ], "output": [ "text" @@ -59465,18 +56638,17 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 150, + "output": 600 } }, - "azure/gpt-5.4": { - "id": "azure/gpt-5.4", - "name": "GPT-5.4 (Azure)", + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", "family": "gpt", "attachment": true, @@ -59495,7 +56667,7 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-08-31", "release_date": "2026-03-05", "last_updated": "2026-03-05", @@ -59515,23 +56687,57 @@ "input": 922000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, "cost": { "input": 2.5, "output": 15, - "cache_read": 0.25 + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "azure/gpt-5.4-pro": { - "id": "azure/gpt-5.4-pro", - "name": "GPT-5.4 Pro (Azure)", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", "medium", "high", "xhigh" @@ -59539,11 +56745,11 @@ } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": false, "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -59555,31 +56761,50 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, + "input": 272000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, "cost": { - "input": 30, - "output": 180 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "azure/gpt-3.5-turbo": { - "id": "azure/gpt-3.5-turbo", - "name": "GPT-3.5 Turbo (Azure)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", "family": "gpt", - "attachment": false, + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -59587,39 +56812,35 @@ }, "open_weights": false, "limit": { - "context": 16385, - "output": 4096 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "azure/gpt-5.4-nano": { - "id": "azure/gpt-5.4-nano", - "name": "GPT-5.4 Nano (Azure)", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "o3-deep-research": { + "id": "o3-deep-research", + "name": "o3-deep-research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", - "medium", - "high", - "xhigh" + "medium" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-06-26", + "last_updated": "2024-06-26", "modalities": { "input": [ "text", @@ -59631,45 +56852,43 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 10, + "output": 40, + "cache_read": 2.5 } }, - "azure/gpt-5.2-codex": { - "id": "azure/gpt-5.2-codex", - "name": "GPT-5.2 Codex (Azure)", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -59682,67 +56901,51 @@ "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "azure/gpt-5.1-codex": { - "id": "azure/gpt-5.1-codex", - "name": "GPT-5.1 Codex (Azure)", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", + "gpt-image-1": { + "id": "gpt-image-1", + "name": "gpt-image-1", + "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", + "family": "gpt-image", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2025-04-24", + "last_updated": "2025-04-24", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 272000 - }, - "cost": { - "input": 1.25, - "output": 10 + "context": 0, + "input": 0, + "output": 0 } }, - "azure/gpt-4o": { - "id": "azure/gpt-4o", - "name": "GPT-4o (Azure)", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -59755,39 +56958,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "azure/gpt-5-mini": { - "id": "azure/gpt-5-mini", - "name": "GPT-5 Mini (Azure)", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", + "gpt-4-turbo": { + "id": "gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ "text", @@ -59799,70 +56991,54 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 10, + "output": 30 } }, - "azure/gpt-5.2-pro": { - "id": "azure/gpt-5.2-pro", - "name": "GPT-5.2 Pro (Azure)", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", + "gpt-image-1-mini": { + "id": "gpt-image-1-mini", + "name": "gpt-image-1-mini", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2025-09-26", + "last_updated": "2025-09-26", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 272000 - }, - "cost": { - "input": 21, - "output": 168 + "context": 0, + "input": 0, + "output": 0 } }, - "azure/o4-mini": { - "id": "azure/o4-mini", - "name": "o4 Mini (Azure)", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -59872,9 +57048,9 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -59886,90 +57062,27 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 - } - }, - "azure/o3-mini": { - "id": "azure/o3-mini", - "name": "o3 Mini (Azure)", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 - } - }, - "azure/gpt-4": { - "id": "azure/gpt-4", - "name": "GPT-4 (Azure)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 30, - "output": 60 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "azure/gpt-5.3-codex": { - "id": "azure/gpt-5.3-codex", - "name": "GPT-5.3 Codex (Azure)", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", "high", "xhigh" @@ -59978,15 +57091,14 @@ ], "tool_call": true, "structured_output": false, - "temperature": true, + "temperature": false, "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -59994,76 +57106,57 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "azure/gpt-4.1-nano": { - "id": "azure/gpt-4.1-nano", - "name": "GPT-4.1 Nano (Azure)", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "azure/gpt-5-nano": { - "id": "azure/gpt-5-nano", - "name": "GPT-5 Nano (Azure)", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": false, + "gpt-5.5-pro": { + "id": "gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -60071,39 +57164,42 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "azure/o1": { - "id": "azure/o1", - "name": "o1 (Azure)", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, + "reasoning": false, + "tool_call": true, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", @@ -60116,26 +57212,27 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "azure/gpt-6-astra": { - "id": "azure/gpt-6-astra", - "name": "GPT-6 Astra (Azure)", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -60147,9 +57244,9 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", @@ -60166,39 +57263,80 @@ "input": 922000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 10, + "output": 60, + "cache_read": 1, + "cache_write": 12.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } + }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "azure/gpt-5.1": { - "id": "azure/gpt-5.1", - "name": "GPT-5.1 (Azure)", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", @@ -60220,18 +57358,17 @@ "cache_read": 0.125 } }, - "azure/gpt-5.4-mini": { - "id": "azure/gpt-5.4-mini", - "name": "GPT-5.4 Mini (Azure)", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -60241,14 +57378,15 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -60261,64 +57399,47 @@ "output": 128000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "azure/gpt-5.6-luna": { - "id": "azure/gpt-5.6-luna", - "name": "GPT-5.6 Luna (Azure)", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", + "gpt-image-2": { + "id": "gpt-image-2", + "name": "gpt-image-2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 0, + "input": 0, + "output": 0 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 + "input": 5, + "output": 30, + "cache_read": 1.25 } }, - "azure/gpt-5.2": { - "id": "azure/gpt-5.2", - "name": "GPT-5.2 (Azure)", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", "family": "gpt", "attachment": true, "reasoning": true, @@ -60329,17 +57450,16 @@ "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -60356,14 +57476,14 @@ "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "azure/gpt-5.5": { - "id": "azure/gpt-5.5", - "name": "GPT-5.5 (Azure)", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", "description": "Default frontier GPT for coding, computer use, research, and knowledge work", "family": "gpt", "attachment": true, @@ -60402,6 +57522,22 @@ "input": 922000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, "cost": { "input": 5, "output": 30, @@ -60424,19 +57560,30 @@ } } }, - "azure/gpt-4.1": { - "id": "azure/gpt-4.1", - "name": "GPT-4.1 (Azure)", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", + "gpt-6-astra": { + "id": "gpt-6-astra", + "name": "GPT-6 Astra", + "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", + "family": "gpt-astra", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "temperature": false, + "release_date": "2026-09-04", + "last_updated": "2026-09-04", "modalities": { "input": [ "text", @@ -60449,133 +57596,185 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 20, + "output": 100, + "cache_read": 2, + "cache_write": 25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + }, + "pro": { + "provider": { + "body": { + "reasoning": { + "mode": "pro" + } + } + } + } + } }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5, + "tiers": [ + { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 20, + "output": 75, + "cache_read": 2, + "cache_write": 25 + } } - }, - "azure/gpt-4.1-mini": { - "id": "azure/gpt-4.1-mini", - "name": "GPT-4.1 Mini (Azure)", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + } + } + }, + "berget": { + "id": "berget", + "env": [ + "BERGET_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.berget.ai/v1", + "name": "Berget.AI", + "doc": "https://api.berget.ai", + "models": { + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2023-12", + "release_date": "2025-04-27", + "last_updated": "2025-04-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.99, + "output": 0.99 } }, - "azure/gpt-5.6-terra": { - "id": "azure/gpt-5.6-terra", - "name": "GPT-5.6 Terra (Azure)", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.83, + "output": 3.85, + "cache_read": 0.16 } }, - "azure/gpt-4-turbo": { - "id": "azure/gpt-4-turbo", - "name": "GPT-4 Turbo (Azure)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "knowledge": "2025-12", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ + "audio", + "image", "text", - "image" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 4096 + "output": 8192 }, "cost": { - "input": 10, - "output": 30 + "input": 0.275, + "output": 0.55 } }, - "azure/gpt-oss-120b": { - "id": "azure/gpt-oss-120b", - "name": "GPT OSS 120B (Azure)", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT-OSS-120B", "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", "family": "gpt-oss", "attachment": false, @@ -60591,8 +57790,9 @@ } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, + "knowledge": "2025-08", "release_date": "2025-08-05", "last_updated": "2025-08-05", "modalities": { @@ -60605,142 +57805,176 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.22, + "output": 0.83 } }, - "azure/gpt-5.1-codex-mini": { - "id": "azure/gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini (Azure)", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "mistralai/Mistral-Medium-3.5-128B": { + "id": "mistralai/Mistral-Medium-3.5-128B", + "name": "Mistral Medium 3.5 128B", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "structured_output": true, + "temperature": true, + "knowledge": "2026-04", + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 1.65, + "output": 5.5 } }, - "azure/o3": { - "id": "azure/o3", - "name": "o3 (Azure)", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", - "attachment": true, - "reasoning": false, - "tool_call": false, + "mistralai/Mistral-Small-3.2-24B-Instruct-2506": { + "id": "mistralai/Mistral-Small-3.2-24B-Instruct-2506", + "name": "Mistral Small 3.2 24B Instruct 2506", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-10-01", + "last_updated": "2025-10-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 32000, + "output": 8192 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.33, + "output": 0.33 } }, - "azure/gpt-5": { - "id": "azure/gpt-5", - "name": "GPT-5 (Azure)", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "zai-org/GLM-4.7": { + "id": "zai-org/GLM-4.7", + "name": "GLM 4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.77, + "output": 2.75 + } + }, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", + "none", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 524288, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.54, + "output": 4.84 } - }, - "azure/gpt-5.6-sol": { - "id": "azure/gpt-5.6-sol", - "name": "GPT-5.6 Sol (Azure)", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + } + } + }, + "snowflake-cortex": { + "id": "snowflake-cortex", + "env": [ + "SNOWFLAKE_ACCOUNT", + "SNOWFLAKE_CORTEX_PAT" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://${SNOWFLAKE_ACCOUNT}.snowflakecomputing.com/api/v2/cortex/v1", + "name": "Snowflake Cortex", + "doc": "https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-rest-api", + "models": { + "openai-gpt-5.1": { + "id": "openai-gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -60748,25 +57982,23 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -60774,30 +58006,23 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, + "input": 272000, "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5 } }, - "alibaba/qwen-flash": { - "id": "alibaba/qwen-flash", - "name": "Qwen Flash (Alibaba Cloud)", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, + "snowflake-llama3.3-70b": { + "id": "snowflake-llama3.3-70b", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -60806,24 +58031,18 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0.0625 + "context": 128000, + "output": 4096 } }, - "alibaba/qwen3.7-max": { - "id": "alibaba/qwen3.7-max", - "name": "Qwen3.7 Max (Alibaba Cloud)", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "openai-gpt-5.2": { + "id": "openai-gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -60833,20 +58052,20 @@ "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -60854,31 +58073,33 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "context": 400000, + "input": 272000, + "output": 128000 } }, - "alibaba/qwen-coder-plus": { - "id": "alibaba/qwen-coder-plus", - "name": "Qwen Coder Plus (Alibaba Cloud)", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2024-09-18", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -60886,31 +58107,28 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.502, - "output": 1.004 + "context": 200000, + "output": 16384 } }, - "alibaba/qwen3-vl-plus": { - "id": "alibaba/qwen3-vl-plus", - "name": "Qwen3 VL Plus (Alibaba Cloud)", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -60918,29 +58136,44 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, - "cost": { - "input": 0.2, - "output": 1.6, - "cache_read": 0.04, - "cache_write": 0.25 + "status": "beta", + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } } }, - "alibaba/qwen-max": { - "id": "alibaba/qwen-max", - "name": "Qwen Max (Alibaba Cloud)", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": false, + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek-R1", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" @@ -60949,46 +58182,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 1.6, - "output": 6.4 + "context": 128000, + "output": 32768 } }, - "alibaba/qwen3.8-max": { - "id": "alibaba/qwen3.8-max", - "name": "Qwen3.8 Max (Alibaba Cloud)", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ @@ -60998,20 +58214,14 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 + "output": 128000 } }, - "alibaba/kimi-k3": { - "id": "alibaba/kimi-k3", - "name": "Kimi K3 (Alibaba Cloud)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "openai-gpt-5": { + "id": "openai-gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -61022,44 +58232,39 @@ "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 400000, + "input": 272000, + "output": 128000 }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } + "status": "beta" }, - "alibaba/glm-5": { - "id": "alibaba/glm-5", - "name": "GLM-5 (Alibaba Cloud)", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "openai-gpt-5.5": { + "id": "openai-gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -61069,42 +58274,68 @@ "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "cost": { - "input": 0.573, - "output": 2.58 + "status": "beta" + }, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 128000 } }, - "alibaba/deepseek-v4.1-flash": { - "id": "alibaba/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (Alibaba Cloud)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "openai-gpt-5-nano": { + "id": "openai-gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -61115,18 +58346,16 @@ "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -61136,33 +58365,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 393216 + "context": 400000, + "input": 272000, + "output": 128000 }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 - } + "status": "beta" }, - "alibaba/qwen-plus": { - "id": "alibaba/qwen-plus", - "name": "Qwen Plus (Alibaba Cloud)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -61170,82 +58402,55 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32000 - }, - "cost": { - "input": 0.4, - "output": 1.2, - "cache_read": 0.08, - "cache_write": 0.5 + "context": 200000, + "output": 16384 } }, - "alibaba/qwen3.7-flash": { - "id": "alibaba/qwen3.7-flash", - "name": "Qwen3.7 Flash (Alibaba Cloud)", - "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", - "family": "qwen", + "mistral-large2": { + "id": "mistral-large2", + "name": "Mistral Large (latest)", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 991000, - "output": 65536 - }, - "cost": { - "input": 0.03, - "output": 0.13, - "cache_read": 0.006, - "cache_write": 0.0375 + "context": 262144, + "output": 262144 } }, - "alibaba/qwen3-max": { - "id": "alibaba/qwen3-max", - "name": "Qwen3 Max (Alibaba Cloud)", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", + "openai-gpt-4.1": { + "id": "openai-gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -61253,31 +58458,32 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32800 - }, - "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24, - "cache_write": 1.5 + "context": 1047576, + "output": 32768 } }, - "alibaba/qwen3-vl-flash": { - "id": "alibaba/qwen3-vl-flash", - "name": "Qwen3 VL Flash (Alibaba Cloud)", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -61285,20 +58491,15 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01 + "context": 1000000, + "output": 16384 } }, - "alibaba/kimi-k2.5": { - "id": "alibaba/kimi-k2.5", - "name": "Kimi K2.5 (Alibaba Cloud)", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "openai-gpt-5.4": { + "id": "openai-gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -61309,46 +58510,87 @@ "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 98304 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "cost": { - "input": 0.574, - "output": 3.011 + "status": "beta", + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } } }, - "alibaba/qwen3.6-35b-a3b": { - "id": "alibaba/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B A3B (Alibaba Cloud)", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "gemini-3.1-pro": { + "id": "gemini-3.1-pro", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + } + }, + "openai-gpt-5-mini": { + "id": "openai-gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -61359,99 +58601,114 @@ "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 272000, + "input": 272000, + "output": 8192 }, - "cost": { - "input": 0.375, - "output": 2.25 - } - }, - "alibaba/qwen3.6-flash": { - "id": "alibaba/qwen3.6-flash", - "name": "Qwen3.6 Flash (Alibaba Cloud)", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, + "status": "beta" + } + } + }, + "tencent-token-plan": { + "id": "tencent-token-plan", + "env": [ + "TENCENT_TOKEN_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.lkeap.cloud.tencent.com/plan/v3", + "name": "Tencent Token Plan", + "doc": "https://cloud.tencent.com/document/product/1823/130060", + "models": { + "hy3": { + "id": "hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.05, - "cache_write": 0.3125 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "alibaba/qwen3-coder-flash": { - "id": "alibaba/qwen3-coder-flash", - "name": "Qwen3 Coder Flash (Alibaba Cloud)", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + } + } + }, + "github-models": { + "id": "github-models", + "env": [ + "GITHUB_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://models.github.ai/inference", + "name": "GitHub Models", + "doc": "https://docs.github.com/en/github-models", + "models": { + "ai21-labs/ai21-jamba-1.5-mini": { + "id": "ai21-labs/ai21-jamba-1.5-mini", + "name": "AI21 Jamba 1.5 Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "jamba", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2024-03", + "release_date": "2024-08-29", + "last_updated": "2024-08-29", "modalities": { "input": [ "text" @@ -61462,29 +58719,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 256000, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 1.5, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0, + "output": 0 } }, - "alibaba/qwen3-coder-plus": { - "id": "alibaba/qwen3-coder-plus", - "name": "Qwen3 Coder Plus (Alibaba Cloud)", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", + "ai21-labs/ai21-jamba-1.5-large": { + "id": "ai21-labs/ai21-jamba-1.5-large", + "name": "AI21 Jamba 1.5 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "jamba", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "knowledge": "2024-03", + "release_date": "2024-08-29", + "last_updated": "2024-08-29", "modalities": { "input": [ "text" @@ -61495,93 +58750,61 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 66000 + "context": 256000, + "output": 4096 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.2, - "cache_write": 1.25 + "input": 0, + "output": 0 } }, - "alibaba/qwen3.8-flash": { - "id": "alibaba/qwen3.8-flash", - "name": "Qwen3.8 Flash (Alibaba Cloud)", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "core42/jais-30b-chat": { + "id": "core42/jais-30b-chat", + "name": "JAIS 30b Chat", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "jais", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": true, + "knowledge": "2023-03", + "release_date": "2023-08-30", + "last_updated": "2023-08-30", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 8192, + "output": 2048 }, "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016, - "cache_write": 0.2 + "input": 0, + "output": 0 } }, - "alibaba/qwen35-397b-a17b": { - "id": "alibaba/qwen35-397b-a17b", - "name": "Qwen3.5 397B A17B (Alibaba Cloud)", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.5", - "attachment": true, + "xai/grok-3-mini": { + "id": "xai/grok-3-mini", + "name": "Grok 3 Mini", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "knowledge": "2024-10", + "release_date": "2024-12-09", + "last_updated": "2024-12-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -61589,28 +58812,27 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0, + "output": 0 } }, - "alibaba/qwen3.6-max-preview": { - "id": "alibaba/qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview (Alibaba Cloud)", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "xai/grok-3": { + "id": "xai/grok-3", + "name": "Grok 3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "knowledge": "2024-10", + "release_date": "2024-12-09", + "last_updated": "2024-12-09", "modalities": { "input": [ "text" @@ -61621,44 +58843,27 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.3, - "output": 7.8, - "cache_read": 0.13 + "input": 0, + "output": 0 } }, - "alibaba/glm-5.2": { - "id": "alibaba/glm-5.2", - "name": "GLM-5.2 (Alibaba Cloud)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "microsoft/phi-3.5-moe-instruct": { + "id": "microsoft/phi-3.5-moe-instruct", + "name": "Phi-3.5-MoE instruct (128k)", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2023-10", + "release_date": "2024-08-20", + "last_updated": "2024-08-20", "modalities": { "input": [ "text" @@ -61669,80 +58874,58 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.28 + "input": 0, + "output": 0 } }, - "alibaba/qwen-omni-turbo": { - "id": "alibaba/qwen-omni-turbo", - "name": "Qwen Omni Turbo (Alibaba Cloud)", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "microsoft/phi-3-small-128k-instruct": { + "id": "microsoft/phi-3-small-128k-instruct", + "name": "Phi-3-small instruct (128k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01-19", - "last_updated": "2025-03-26", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 8192 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0, + "output": 0 } }, - "alibaba/deepseek-v4-pro": { - "id": "alibaba/deepseek-v4-pro", - "name": "DeepSeek V4 Pro (Alibaba Cloud)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "microsoft/phi-3.5-mini-instruct": { + "id": "microsoft/phi-3.5-mini-instruct", + "name": "Phi-3.5-mini instruct (128k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-10", + "release_date": "2024-08-20", + "last_updated": "2024-08-20", "modalities": { "input": [ "text" @@ -61753,94 +58936,58 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 393216 + "context": 128000, + "output": 4096 }, "cost": { - "input": 2.4, - "output": 4.8, - "cache_read": 0.2 + "input": 0, + "output": 0 } }, - "alibaba/qwen3.6-plus": { - "id": "alibaba/qwen3.6-plus", - "name": "Qwen3.6 Plus (Alibaba Cloud)", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": true, + "microsoft/phi-3-medium-128k-instruct": { + "id": "microsoft/phi-3-medium-128k-instruct", + "name": "Phi-3-medium instruct (128k)", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 0, + "output": 0 } }, - "alibaba/qwen-plus-latest": { - "id": "alibaba/qwen-plus-latest", - "name": "Qwen Plus Latest (Alibaba Cloud)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "microsoft/phi-3-small-8k-instruct": { + "id": "microsoft/phi-3-small-8k-instruct", + "name": "Phi-3-small instruct (8k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2024-09-09", - "last_updated": "2024-09-09", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -61849,40 +58996,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32000 + "context": 8192, + "output": 2048 }, "cost": { - "input": 0.4, - "output": 1.2, - "cache_read": 0.08, - "cache_write": 0.5 + "input": 0, + "output": 0 } }, - "alibaba/glm-5.3": { - "id": "alibaba/glm-5.3", - "name": "GLM-5.3 (Alibaba Cloud)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "microsoft/phi-4-reasoning": { + "id": "microsoft/phi-4-reasoning", + "name": "Phi-4-Reasoning", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -61893,47 +59029,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.28 + "input": 0, + "output": 0 } }, - "alibaba/qwen3.7-plus": { - "id": "alibaba/qwen3.7-plus", - "name": "Qwen3.7 Plus (Alibaba Cloud)", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, + "microsoft/mai-ds-r1": { + "id": "microsoft/mai-ds-r1", + "name": "MAI-DS-R1", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mai", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2024-06", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -61941,46 +59060,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 65536, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "cache_write": 0.5 + "input": 0, + "output": 0 } }, - "alibaba/deepseek-v4-flash": { - "id": "alibaba/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (Alibaba Cloud)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "microsoft/phi-4-mini-instruct": { + "id": "microsoft/phi-4-mini-instruct", + "name": "Phi-4-mini-instruct", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -61991,44 +59091,27 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 393216 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 0.4, - "cache_read": 0.04 + "input": 0, + "output": 0 } }, - "baidu/glm-5": { - "id": "baidu/glm-5", - "name": "GLM-5 (Baidu)", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "microsoft/phi-4": { + "id": "microsoft/phi-4", + "name": "Phi-4", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -62039,42 +59122,27 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 16000, + "output": 4096 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 - } + "input": 0, + "output": 0 + } }, - "baidu/deepseek-v4.1-flash": { - "id": "baidu/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (Baidu)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "microsoft/phi-3.5-vision-instruct": { + "id": "microsoft/phi-3.5-vision-instruct", + "name": "Phi-3.5-vision instruct (128k)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "phi", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2023-10", + "release_date": "2024-08-20", + "last_updated": "2024-08-20", "modalities": { "input": [ "text", @@ -62086,50 +59154,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.006 + "input": 0, + "output": 0 } }, - "baidu/kimi-k2.6": { - "id": "baidu/kimi-k2.6", - "name": "Kimi K2.6 (Baidu)", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "microsoft/phi-3-medium-4k-instruct": { + "id": "microsoft/phi-3-medium-4k-instruct", + "name": "Phi-3-medium instruct (4k)", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -62137,47 +59185,32 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 4096, + "output": 1024 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0, + "output": 0 } }, - "baidu/glm-5.2": { - "id": "baidu/glm-5.2", - "name": "GLM-5.2 (Baidu)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "microsoft/phi-4-multimodal-instruct": { + "id": "microsoft/phi-4-multimodal-instruct", + "name": "Phi-4-multimodal-instruct", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "phi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -62185,44 +59218,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0 } }, - "baidu/glm-5.1": { - "id": "baidu/glm-5.1", - "name": "GLM-5.1 (Baidu)", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "microsoft/phi-4-mini-reasoning": { + "id": "microsoft/phi-4-mini-reasoning", + "name": "Phi-4-mini-reasoning", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -62233,45 +59249,27 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0 } }, - "baidu/deepseek-v4-pro": { - "id": "baidu/deepseek-v4-pro", - "name": "DeepSeek V4 Pro (Baidu)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "microsoft/phi-3-mini-128k-instruct": { + "id": "microsoft/phi-3-mini-128k-instruct", + "name": "Phi-3-mini instruct (128k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -62282,37 +59280,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.042 + "input": 0, + "output": 0 } }, - "baidu/glm-5.3": { - "id": "baidu/glm-5.3", - "name": "GLM-5.3 (Baidu)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "microsoft/phi-3-mini-4k-instruct": { + "id": "microsoft/phi-3-mini-4k-instruct", + "name": "Phi-3-mini instruct (4k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -62323,133 +59311,154 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 4096, + "output": 1024 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0 } }, - "baidu/deepseek-v4-flash": { - "id": "baidu/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (Baidu)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "openai/o3": { + "id": "openai/o3", + "name": "OpenAI o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning_options": [], + "tool_call": false, + "temperature": false, + "knowledge": "2024-04", + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.014 + "input": 0, + "output": 0 } }, - "gonka24/glm-5.3-flash": { - "id": "gonka24/glm-5.3-flash", - "name": "GLM-5.3 Flash (Gonka24)", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "openai/o1-mini": { + "id": "openai/o1-mini", + "name": "OpenAI o1-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], + "tool_call": false, + "temperature": false, + "knowledge": "2023-10", + "release_date": "2024-09-12", + "last_updated": "2024-12-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2023-10", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ + "text", + "image", + "audio" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "OpenAI o4-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": false, + "knowledge": "2024-04", + "release_date": "2025-01-31", + "last_updated": "2025-01-31", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 200000, - "output": 16384 + "output": 100000 }, "cost": { - "input": 0.15, - "output": 0.3, - "cache_read": 0.035 + "input": 0, + "output": 0 } }, - "gonka24/minimax-m2.7": { - "id": "gonka24/minimax-m2.7", - "name": "MiniMax M2.7 (Gonka24)", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "openai/o1-preview": { + "id": "openai/o1-preview", + "name": "OpenAI o1-preview", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "reasoning_options": [], + "tool_call": false, + "temperature": false, + "knowledge": "2023-10", + "release_date": "2024-09-12", + "last_updated": "2024-09-12", "modalities": { "input": [ "text" @@ -62458,44 +59467,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131100 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.08, - "output": 0.32, - "cache_read": 0.017 + "input": 0, + "output": 0 } }, - "gonka24/deepseek-v4-flash": { - "id": "gonka24/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (Gonka24)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "OpenAI o3-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning_options": [], + "tool_call": false, + "temperature": false, + "knowledge": "2024-04", + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ "text" @@ -62504,40 +59498,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 390000, - "output": 16384 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.065, - "output": 0.116, - "cache_read": 0.012 + "input": 0, + "output": 0 } }, - "bytedance/seed-1-6-flash-250715": { - "id": "bytedance/seed-1-6-flash-250715", - "name": "Seed 1.6 Flash (250715) (ByteDance)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -62549,38 +59531,27 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.07, - "output": 0.3, - "cache_read": 0.015 + "input": 0, + "output": 0 } }, - "bytedance/seed-1-6-250915": { - "id": "bytedance/seed-1-6-250915", - "name": "Seed 1.6 (250915) (ByteDance)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, + "openai/o1": { + "id": "openai/o1", + "name": "OpenAI o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "reasoning_options": [], + "tool_call": false, + "temperature": false, + "knowledge": "2023-10", + "release_date": "2024-09-12", + "last_updated": "2024-12-17", "modalities": { "input": [ "text", @@ -62592,38 +59563,26 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05 + "input": 0, + "output": 0 } }, - "bytedance/seed-1-8-251228": { - "id": "bytedance/seed-1-8-251228", - "name": "Seed 1.8 (251228) (ByteDance)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -62635,82 +59594,62 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05 + "input": 0, + "output": 0 } }, - "bytedance/glm-4.7": { - "id": "bytedance/glm-4.7", - "name": "GLM-4.7 (ByteDance)", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0, + "output": 0 } }, - "bytedance/seed-1-6-250615": { - "id": "bytedance/seed-1-6-250615", - "name": "Seed 1.6 (250615) (ByteDance)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-06-25", - "last_updated": "2025-06-25", + "knowledge": "2023-10", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ "text" @@ -62718,93 +59657,59 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05 + "input": 0, + "output": 0 } }, - "bytedance/glm-5.2": { - "id": "bytedance/glm-5.2", - "name": "GLM-5.2 (ByteDance)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "mistral-ai/mistral-small-2503": { + "id": "mistral-ai/mistral-small-2503", + "name": "Mistral Small 3.1", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-09", + "release_date": "2025-03-01", + "last_updated": "2025-03-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1024000, - "output": 128000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0 } }, - "bytedance/deepseek-v4-pro": { - "id": "bytedance/deepseek-v4-pro", - "name": "DeepSeek V4 Pro (ByteDance)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "mistral-ai/mistral-nemo": { + "id": "mistral-ai/mistral-nemo", + "name": "Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-03", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text" @@ -62815,74 +59720,59 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 + "input": 0, + "output": 0 } }, - "bytedance/deepseek-v3.2": { - "id": "bytedance/deepseek-v3.2", - "name": "DeepSeek V3.2 (ByteDance)", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", + "mistral-ai/mistral-medium-2505": { + "id": "mistral-ai/mistral-medium-2505", + "name": "Mistral Medium 3 (25.05)", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "knowledge": "2024-09", + "release_date": "2025-05-01", + "last_updated": "2025-05-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 128000, "output": 32768 }, "cost": { - "input": 0.28, - "output": 0.42, - "cache_read": 0.056 + "input": 0, + "output": 0 } }, - "bytedance/gpt-oss-120b": { - "id": "bytedance/gpt-oss-120b", - "name": "GPT OSS 120B (ByteDance)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "mistral-ai/mistral-large-2411": { + "id": "mistral-ai/mistral-large-2411", + "name": "Mistral Large 24.11", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-09", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", "modalities": { "input": [ "text" @@ -62891,47 +59781,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 32000 + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.02 + "input": 0, + "output": 0 } }, - "bytedance/deepseek-v4-flash": { - "id": "bytedance/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (ByteDance)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "mistral-ai/ministral-3b": { + "id": "mistral-ai/ministral-3b", + "name": "Ministral 3B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-03", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ "text" @@ -62942,41 +59814,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.014 + "input": 0, + "output": 0 } }, - "xai/grok-4-5": { - "id": "xai/grok-4-5", - "name": "Grok 4.5 (xAI)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "mistral-ai/codestral-2501": { + "id": "mistral-ai/codestral-2501", + "name": "Codestral 25.01", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "codestral", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "knowledge": "2024-03", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -62984,41 +59845,30 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 32000, + "output": 8192 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3 + "input": 0, + "output": 0 } }, - "xai/grok-build-0-1": { - "id": "xai/grok-build-0-1", - "name": "Grok Build 0.1 (xAI)", - "description": "Grok coding model for agentic engineering, edits, and codebase workflows", - "family": "grok-build", - "attachment": true, + "cohere/cohere-command-a": { + "id": "cohere/cohere-command-a", + "name": "Cohere Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-05-20", - "last_updated": "2026-05-20", + "knowledge": "2024-03", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -63026,48 +59876,29 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 4, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 4, - "cache_read": 0.4 - } + "input": 0, + "output": 0 } }, - "xai/grok-4": { - "id": "xai/grok-4", - "name": "Grok 4 (xAI)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "cohere/cohere-command-r-08-2024": { + "id": "cohere/cohere-command-r-08-2024", + "name": "Cohere Command R 08-2024", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "knowledge": "2024-03", + "release_date": "2024-08-01", + "last_updated": "2024-08-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -63075,31 +59906,29 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 + "input": 0, + "output": 0 } }, - "xai/grok-4-20-beta-0309-non-reasoning": { - "id": "xai/grok-4-20-beta-0309-non-reasoning", - "name": "Grok 4.20 Beta Non-Reasoning (0309) (xAI)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "cohere/cohere-command-r-plus": { + "id": "cohere/cohere-command-r-plus", + "name": "Cohere Command R+", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "knowledge": "2024-03", + "release_date": "2024-04-04", + "last_updated": "2024-08-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -63107,58 +59936,29 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0, + "output": 0 } }, - "xai/grok-4-3": { - "id": "xai/grok-4-3", - "name": "Grok 4.3 (xAI)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "cohere/cohere-command-r-plus-08-2024": { + "id": "cohere/cohere-command-r-plus-08-2024", + "name": "Cohere Command R+ 08-2024", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "knowledge": "2024-03", + "release_date": "2024-08-01", + "last_updated": "2024-08-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -63166,57 +59966,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0, + "output": 0 } }, - "xai/grok-4-20-beta-0309-reasoning": { - "id": "xai/grok-4-20-beta-0309-reasoning", - "name": "Grok 4.20 Beta Reasoning (0309) (xAI)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "cohere/cohere-command-r": { + "id": "cohere/cohere-command-r", + "name": "Cohere Command R", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "knowledge": "2024-03", + "release_date": "2024-03-11", + "last_updated": "2024-08-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -63224,148 +59997,91 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0, + "output": 0 } }, - "xai/grok-4-6": { - "id": "xai/grok-4-6", - "name": "Grok 4.6 (xAI)", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, + "meta/meta-llama-3-8b-instruct": { + "id": "meta/meta-llama-3-8b-instruct", + "name": "Meta-Llama-3-8B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2023-12", + "release_date": "2024-04-18", + "last_updated": "2024-04-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 8192, + "output": 2048 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 0, + "output": 0 } }, - "xai/grok-4-7": { - "id": "xai/grok-4-7", - "name": "Grok 4.7 (xAI)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "meta/llama-3.2-11b-vision-instruct": { + "id": "meta/llama-3.2-11b-vision-instruct", + "name": "Llama-3.2-11B-Vision-Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 0, + "output": 0 } }, - "ranoai/deepseek-v4-flash": { - "id": "ranoai/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (RanoAI)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "meta/meta-llama-3.1-405b-instruct": { + "id": "meta/meta-llama-3.1-405b-instruct", + "name": "Meta-Llama-3.1-405B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -63376,46 +60092,31 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 393216 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0, + "output": 0 } }, - "canopywave/kimi-k3": { - "id": "canopywave/kimi-k3", - "name": "Kimi K3 (CanopyWave)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "meta/llama-4-scout-17b-16e-instruct": { + "id": "meta/llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout 17B 16E Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -63423,50 +60124,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 128000, + "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0, + "output": 0 } }, - "canopywave/kimi-k2.6": { - "id": "canopywave/kimi-k2.6", - "name": "Kimi K2.6 (CanopyWave)", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "meta/llama-3.3-70b-instruct": { + "id": "meta/llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -63474,44 +60155,27 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 128000, "output": 32768 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0, + "output": 0 } }, - "canopywave/glm-5.2": { - "id": "canopywave/glm-5.2", - "name": "GLM-5.2 (CanopyWave)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "meta/meta-llama-3.1-70b-instruct": { + "id": "meta/meta-llama-3.1-70b-instruct", + "name": "Meta-Llama-3.1-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -63522,45 +60186,27 @@ }, "open_weights": true, "limit": { - "context": 200000, + "context": 128000, "output": 32768 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0 } }, - "canopywave/deepseek-v4-pro": { - "id": "canopywave/deepseek-v4-pro", - "name": "DeepSeek V4 Pro (CanopyWave)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "meta/meta-llama-3.1-8b-instruct": { + "id": "meta/meta-llama-3.1-8b-instruct", + "name": "Meta-Llama-3.1-8B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -63571,48 +60217,32 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 393216 + "context": 128000, + "output": 32768 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.01 + "input": 0, + "output": 0 } }, - "canopywave/deepseek-v4-flash": { - "id": "canopywave/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (CanopyWave)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "meta/llama-3.2-90b-vision-instruct": { + "id": "meta/llama-3.2-90b-vision-instruct", + "name": "Llama-3.2-90B-Vision-Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -63620,45 +60250,31 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 393216 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.03 + "input": 0, + "output": 0 } }, - "fireworks/kimi-k3": { - "id": "fireworks/kimi-k3", - "name": "Kimi K3 (Fireworks AI)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "meta/llama-4-maverick-17b-128e-instruct-fp8": { + "id": "meta/llama-4-maverick-17b-128e-instruct-fp8", + "name": "Llama 4 Maverick 17B 128E Instruct FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -63666,45 +60282,30 @@ }, "open_weights": true, "limit": { - "context": 1040384, - "output": 1040384 + "context": 128000, + "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0, + "output": 0 } }, - "fireworks/deepseek-v4.1-flash": { - "id": "fireworks/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (Fireworks AI)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "meta/meta-llama-3-70b-instruct": { + "id": "meta/meta-llama-3-70b-instruct", + "name": "Meta-Llama-3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2023-12", + "release_date": "2024-04-18", + "last_updated": "2024-04-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -63712,45 +60313,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 8192, + "output": 2048 }, "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 + "input": 0, + "output": 0 } }, - "fireworks/kimi-k3-fast": { - "id": "fireworks/kimi-k3-fast", - "name": "Kimi K3 Fast (Fireworks AI)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "deepseek/deepseek-r1-0528": { + "id": "deepseek/deepseek-r1-0528", + "name": "DeepSeek-R1-0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2024-06", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -63758,44 +60344,27 @@ }, "open_weights": true, "limit": { - "context": 1040384, - "output": 1040384 + "context": 65536, + "output": 8192 }, "cost": { - "input": 4.5, - "output": 22.5, - "cache_read": 0.45 + "input": 0, + "output": 0 } }, - "fireworks/deepseek-v4-pro": { - "id": "fireworks/deepseek-v4-pro", - "name": "DeepSeek V4 Pro (Fireworks AI)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "deepseek/deepseek-v3-0324": { + "id": "deepseek/deepseek-v3-0324", + "name": "DeepSeek-V3-0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-06", + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ "text" @@ -63806,44 +60375,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 + "input": 0, + "output": 0 } }, - "fireworks/deepseek-v4-flash": { - "id": "fireworks/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (Fireworks AI)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "deepseek/deepseek-r1": { + "id": "deepseek/deepseek-r1", + "name": "DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-06", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -63854,37 +60406,76 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 65536, + "output": 8192 }, "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 + "input": 0, + "output": 0 + } + } + } + }, + "neuralwatt": { + "id": "neuralwatt", + "env": [ + "NEURALWATT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.neuralwatt.com/v1", + "name": "Neuralwatt", + "doc": "https://portal.neuralwatt.com/docs", + "models": { + "kimi-k2.5-fast": { + "id": "kimi-k2.5-fast", + "name": "Kimi K2.5 Fast", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-01-27", + "last_updated": "2026-01-27", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262128, + "output": 262128 + }, + "cost": { + "input": 0.52, + "output": 2.59, + "cache_read": 0.13 } }, - "sakana/fugu-max": { - "id": "sakana/fugu-max", - "name": "Fugu Max (Sakana AI)", - "description": "Multi-agent model for routing expert agents across complex analytical tasks", - "family": "fugu", + "kimi-k2.6-flex": { + "id": "kimi-k2.6-flex", + "name": "Kimi K2.6 Flex", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -63894,113 +60485,158 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 262128, + "output": 262128 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 0.345, + "output": 1.61, + "cache_read": 0.08625 } }, - "sakana/fugu-ultra-v2.0": { - "id": "sakana/fugu-ultra-v2.0", - "name": "Fugu Ultra v2.0 (Sakana AI)", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, + "glm-5.2-short-fast-flex": { + "id": "glm-5.2-short-fast-flex", + "name": "GLM 5.2 Short Fast Flex", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-06-17", + "last_updated": "2026-06-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 199984, + "output": 199984 + }, + "cost": { + "input": 0.725, + "output": 2.25, + "cache_read": 0.18125 + } + }, + "glm-5.2-flex": { + "id": "glm-5.2-flex", + "name": "GLM 5.2 Flex", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "none", + "minimal", + "low", + "medium", "high", "xhigh", "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 1048560, + "output": 1048560 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.725, + "output": 2.25, + "cache_read": 0.18125 } }, - "sakana/fugu-ultra": { - "id": "sakana/fugu-ultra", - "name": "Fugu Ultra (Sakana AI)", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM 5.2", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "none", + "minimal", + "low", + "medium", "high", "xhigh", "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "temperature": true, + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 1048560, + "output": 1048560 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 1.45, + "output": 4.5, + "cache_read": 0.3625 } }, - "scx-ai/gemma-4-31b-it": { - "id": "scx-ai/gemma-4-31b-it", - "name": "Gemma 4 31B IT (SCX.ai (Turbo))", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "glm-5.2-short-fast": { + "id": "glm-5.2-short-fast", + "name": "GLM 5.2 Short Fast", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ "text" @@ -64011,27 +60647,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 199984, + "output": 199984 }, "cost": { - "input": 0.3, - "output": 0.91 + "input": 1.45, + "output": 4.5, + "cache_read": 0.3625 } }, - "scx-ai/llama-4-maverick-17b-instruct": { - "id": "scx-ai/llama-4-maverick-17b-instruct", - "name": "Llama 4 Maverick 17B Instruct (SCX.ai (Turbo))", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", + "qwen3.5-397b-fast": { + "id": "qwen3.5-397b-fast", + "name": "Qwen3.5 397B Fast", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ "text" @@ -64042,35 +60677,30 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 262128, + "output": 262128 }, "cost": { - "input": 0.53, - "output": 1.62 + "input": 0.69, + "output": 4.14, + "cache_read": 0.1725 } }, - "scx-ai/qwen3-32b": { - "id": "scx-ai/qwen3-32b", - "name": "Qwen3 32B (SCX.ai (Turbo))", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "kimi-k2.6-fast": { + "id": "kimi-k2.6-fast", + "name": "Kimi K2.6 Fast", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -64078,41 +60708,30 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 8192 + "context": 262128, + "output": 262128 }, "cost": { - "input": 0.36, - "output": 0.87 + "input": 0.69, + "output": 3.22, + "cache_read": 0.1725 } }, - "scx-ai/minimax-m2.7": { - "id": "scx-ai/minimax-m2.7", - "name": "MiniMax M2.7 (SCX.ai (Turbo))", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "qwen3.6-35b-fast": { + "id": "qwen3.6-35b-fast", + "name": "Qwen3.6 35B Fast", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "qwen3.6", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -64120,37 +60739,46 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 196608 + "context": 131056, + "output": 131056 }, "cost": { - "input": 0.48, - "output": 1.79, - "cache_read": 0.05 + "input": 0.29, + "output": 1.15, + "cache_read": 0.0725 } }, - "scx-ai/gpt-oss-120b": { - "id": "scx-ai/gpt-oss-120b", - "name": "GPT OSS 120B (SCX.ai (Turbo))", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "glm-5.2-short-flex": { + "id": "glm-5.2-short-flex", + "name": "GLM 5.2 Short Flex", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ "text" @@ -64161,37 +60789,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 199984, + "output": 199984 }, "cost": { - "input": 0.17, - "output": 0.55 + "input": 0.725, + "output": 2.25, + "cache_read": 0.18125 } }, - "together-ai/qwen3.8-2.4t-a95b": { - "id": "together-ai/qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B (Together AI)", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", + "glm-5.2-fast": { + "id": "glm-5.2-fast", + "name": "GLM 5.2 Fast", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ "text" @@ -64202,23 +60819,26 @@ }, "open_weights": true, "limit": { - "context": 1010000, - "output": 909000 + "context": 1048560, + "output": 1048560 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 1.45, + "output": 4.5, + "cache_read": 0.3625 } }, - "together-ai/glm-5.3-flash": { - "id": "together-ai/glm-5.3-flash", - "name": "GLM-5.3 Flash (Together AI)", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + "glm-5.2-short": { + "id": "glm-5.2-short", + "name": "GLM 5.2 Short", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -64230,19 +60850,18 @@ "xhigh", "max" ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -64250,41 +60869,29 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 943717 + "context": 199984, + "output": 199984 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 1.45, + "output": 4.5, + "cache_read": 0.3625 } }, - "together-ai/muse-glimmer-30b": { - "id": "together-ai/muse-glimmer-30b", - "name": "Muse Glimmer 30B (Together AI)", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", + "kimi-k2.7-code-flex": { + "id": "kimi-k2.7-code-flex", + "name": "Kimi K2.7 Code Flex", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", @@ -64296,20 +60903,20 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 117964 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.35, - "output": 1.5, - "cache_read": 0.04 + "input": 0.475, + "output": 2, + "cache_read": 0.11875 } }, - "together-ai/kimi-k3": { - "id": "together-ai/kimi-k3", - "name": "Kimi K3 (Together AI)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -64318,15 +60925,16 @@ } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -64334,40 +60942,34 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 262128, + "output": 262128 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.69, + "output": 3.22, + "cache_read": 0.1725 } }, - "together-ai/deepseek-v4.1-flash": { - "id": "together-ai/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (Together AI)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", @@ -64379,46 +60981,33 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 262128, + "output": 262128 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.006 + "input": 0.52, + "output": 2.59, + "cache_read": 0.13 } }, - "together-ai/inkling": { - "id": "together-ai/inkling", - "name": "Inkling (Together AI)", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" @@ -64426,39 +61015,38 @@ }, "open_weights": true, "limit": { - "context": 524288, - "output": 471859 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.17 + "input": 0.95, + "output": 4, + "cache_read": 0.2375 } }, - "together-ai/glm-4.7": { - "id": "together-ai/glm-4.7", - "name": "GLM-4.7 (Together AI)", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" } ], - "tool_call": false, + "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -64466,19 +61054,20 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 128000 + "context": 131056, + "output": 131056 }, "cost": { - "input": 0.45, - "output": 2 + "input": 0.29, + "output": 1.15, + "cache_read": 0.0725 } }, - "together-ai/minimax-m3": { - "id": "together-ai/minimax-m3", - "name": "MiniMax M3 (Together AI)", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "Qwen/Qwen3.5-397B-A17B-FP8": { + "id": "Qwen/Qwen3.5-397B-A17B-FP8", + "name": "Qwen3.5 397B A17B FP8", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -64486,11 +61075,13 @@ "type": "toggle" } ], - "tool_call": false, - "structured_output": true, + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ "text" @@ -64501,41 +61092,39 @@ }, "open_weights": true, "limit": { - "context": 524288, - "output": 131072 + "context": 262128, + "output": 262128 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.69, + "output": 4.14, + "cache_read": 0.1725 } - }, - "together-ai/deepseek-v4-pro": { - "id": "together-ai/deepseek-v4-pro", - "name": "DeepSeek V4 Pro (Together AI)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + } + } + }, + "siliconflow-cn": { + "id": "siliconflow-cn", + "env": [ + "SILICONFLOW_CN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.siliconflow.cn/v1", + "name": "SiliconFlow (China)", + "doc": "https://cloud.siliconflow.com/models", + "models": { + "baidu/ERNIE-4.5-300B-A47B": { + "id": "baidu/ERNIE-4.5-300B-A47B", + "name": "baidu/ERNIE-4.5-300B-A47B", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "ernie", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2025-07-02", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -64544,39 +61133,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 163840 + "context": 131000, + "output": 131000 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.13 + "input": 0.28, + "output": 1.1 } }, - "together-ai/gpt-oss-120b": { - "id": "together-ai/gpt-oss-120b", - "name": "GPT OSS 120B (Together AI)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "ByteDance-Seed/Seed-OSS-36B-Instruct": { + "id": "ByteDance-Seed/Seed-OSS-36B-Instruct", + "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "seed", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-09-04", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -64585,39 +61163,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.21, + "output": 0.57 } }, - "together-ai/glm-5.3": { - "id": "together-ai/glm-5.3", - "name": "GLM-5.3 (Together AI)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "stepfun-ai/Step-3.5-Flash": { + "id": "stepfun-ai/Step-3.5-Flash", + "name": "stepfun-ai/Step-3.5-Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "family": "step", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -64626,43 +61194,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 943717 + "context": 262000, + "output": 262000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.1, + "output": 0.3 } }, - "together-ai/deepseek-v4-flash": { - "id": "together-ai/deepseek-v4-flash", - "name": "DeepSeek V4 Flash (Together AI)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "inclusionAI/Ling-flash-2.0": { + "id": "inclusionAI/Ling-flash-2.0", + "name": "inclusionAI/Ling-flash-2.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2025-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -64671,46 +61224,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 131000, + "output": 131000 }, "cost": { "input": 0.14, - "output": 0.28, - "cache_read": 0.03 + "output": 0.57 } }, - "xiaomi/mimo-v2.6-pro": { - "id": "xiaomi/mimo-v2.6-pro", - "name": "MiMo V2.6 Pro (Xiaomi)", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": true, + "Pro/moonshotai/Kimi-K2.6": { + "id": "Pro/moonshotai/Kimi-K2.6", + "name": "Pro/moonshotai/Kimi-K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" @@ -64718,48 +61261,35 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo V2.5 (Xiaomi)", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, + "Pro/moonshotai/Kimi-K2.5": { + "id": "Pro/moonshotai/Kimi-K2.5", + "name": "Pro/moonshotai/Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" @@ -64767,58 +61297,34 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028, - "tiers": [ - { - "input": 0.8, - "output": 4, - "cache_read": 0.16, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.8, - "output": 4, - "cache_read": 0.16 - } + "input": 0.45, + "output": 2.25 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo V2.5 Pro (Xiaomi)", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "Pro/zai-org/GLM-5": { + "id": "Pro/zai-org/GLM-5", + "name": "Pro/zai-org/GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -64829,60 +61335,33 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 205000, + "output": 205000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 1, + "output": 3.2 } }, - "xiaomi/mimo-v2.6-flash": { - "id": "xiaomi/mimo-v2.6-flash", - "name": "MiMo V2.6 Flash (Xiaomi)", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, + "Pro/zai-org/GLM-5.1": { + "id": "Pro/zai-org/GLM-5.1", + "name": "Pro/zai-org/GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -64890,45 +61369,37 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 205000, + "output": 205000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 1.4, + "output": 4.4, + "cache_write": 0 } }, - "quartz/gemini-3.1-pro-preview": { - "id": "quartz/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro (Preview) (Quartz)", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, + "Pro/deepseek-ai/DeepSeek-R1": { + "id": "Pro/deepseek-ai/DeepSeek-R1", + "name": "Pro/deepseek-ai/DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2025-05-28", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -64936,55 +61407,31 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 164000, + "output": 164000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.5, + "output": 2.18 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax M2.5 (MiniMax)", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "Pro/deepseek-ai/DeepSeek-V3.1-Terminus": { + "id": "Pro/deepseek-ai/DeepSeek-V3.1-Terminus", + "name": "Pro/deepseek-ai/DeepSeek-V3.1-Terminus", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2025-09-29", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -64993,41 +61440,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131100 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.27, + "output": 1 } }, - "minimax/minimax-m2.1-lightning": { - "id": "minimax/minimax-m2.1-lightning", - "name": "MiniMax M2.1 Lightning (MiniMax)", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "Pro/deepseek-ai/DeepSeek-V3.2": { + "id": "Pro/deepseek-ai/DeepSeek-V3.2", + "name": "Pro/deepseek-ai/DeepSeek-V3.2", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2025-12-03", + "last_updated": "2025-12-03", "modalities": { "input": [ "text" @@ -65038,27 +61477,26 @@ }, "open_weights": false, "limit": { - "context": 196608, - "output": 131072 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.12, - "output": 0.48 + "input": 0.27, + "output": 0.42 } }, - "minimax/minimax-text-01": { - "id": "minimax/minimax-text-01", - "name": "MiniMax Text 01 (MiniMax)", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "Pro/deepseek-ai/DeepSeek-V3": { + "id": "Pro/deepseek-ai/DeepSeek-V3", + "name": "Pro/deepseek-ai/DeepSeek-V3", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-01-15", - "last_updated": "2025-01-15", + "release_date": "2024-12-26", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -65069,38 +61507,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.2, - "output": 1.1 + "input": 0.25, + "output": 1 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax M2.7 (MiniMax)", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "Pro/MiniMaxAI/MiniMax-M2.5": { + "id": "Pro/MiniMaxAI/MiniMax-M2.5", + "name": "Pro/MiniMaxAI/MiniMax-M2.5", + "description": "Frontier MiniMax model for engineering, office tasks, and agentic reasoning", "family": "minimax", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -65109,43 +61538,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131100 + "context": 192000, + "output": 131000 }, "cost": { "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "output": 1.22 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax M3 (MiniMax)", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen/Qwen3.6-35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "knowledge": "2025-04", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", @@ -65158,42 +61572,36 @@ }, "open_weights": true, "limit": { - "context": 512000, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 + "input": 0.23, + "output": 1.86 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax M2.7 Highspeed (MiniMax)", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen/Qwen3.5-397B-A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -65201,39 +61609,33 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131100 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06 + "input": 0.29, + "output": 1.74 } }, - "minimax/minimax-m2.5-highspeed": { - "id": "minimax/minimax-m2.5-highspeed", - "name": "MiniMax M2.5 Highspeed (MiniMax)", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "release_date": "2025-07-28", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -65242,44 +61644,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131100 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.03 + "input": 0.13, + "output": 0.6 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax M2.1 (MiniMax)", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", + "Qwen/Qwen3.5-122B-A10B": { + "id": "Qwen/Qwen3.5-122B-A10B", + "name": "Qwen/Qwen3.5-122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2025-04", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -65287,41 +61683,36 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.27, - "output": 1.1 + "input": 0.29, + "output": 2.32 } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax M2 (MiniMax)", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", + "Qwen/Qwen3.5-27B": { + "id": "Qwen/Qwen3.5-27B", + "name": "Qwen/Qwen3.5-27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "knowledge": "2025-04", + "release_date": "2026-02-25", + "last_updated": "2026-02-25", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -65329,27 +61720,36 @@ }, "open_weights": true, "limit": { - "context": 196608, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 1, - "cache_read": 0.03 + "input": 0.26, + "output": 2.09 } }, - "inference.net/llama-3.2-11b-instruct": { - "id": "inference.net/llama-3.2-11b-instruct", - "name": "Llama 3.2 11B Instruct (Inference.net)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "Qwen/Qwen3-8B": { + "id": "Qwen/Qwen3-8B", + "name": "Qwen/Qwen3-8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -65360,138 +61760,113 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.07, - "output": 0.33 + "input": 0.06, + "output": 0.06 } }, - "azure-anthropic/claude-opus-5": { - "id": "azure-anthropic/claude-opus-5", - "name": "Claude Opus 5 (Azure Anthropic)", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, + "Qwen/Qwen3.5-4B": { + "id": "Qwen/Qwen3.5-4B", + "name": "Qwen/Qwen3.5-4B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0 } }, - "azure-anthropic/claude-fable-5": { - "id": "azure-anthropic/claude-fable-5", - "name": "Claude Fable 5 (Azure Anthropic)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, + "Qwen/Qwen3.5-9B": { + "id": "Qwen/Qwen3.5-9B", + "name": "Qwen/Qwen3.5-9B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.22, + "output": 1.74 } }, - "azure-anthropic/claude-opus-4-8": { - "id": "azure-anthropic/claude-opus-4-8", - "name": "Claude Opus 4.8 (Azure Anthropic)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen/Qwen3-32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -65499,46 +61874,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.14, + "output": 0.57 } }, - "azure-anthropic/claude-sonnet-5": { - "id": "azure-anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5 (Azure Anthropic)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "Qwen/Qwen3-14B": { + "id": "Qwen/Qwen3-14B", + "name": "Qwen/Qwen3-14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -65546,92 +61914,68 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.07, + "output": 0.28 } }, - "azure-anthropic/claude-opus-4-6": { - "id": "azure-anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6 (Azure Anthropic)", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "Qwen/Qwen3.5-35B-A3B": { + "id": "Qwen/Qwen3.5-35B-A3B", + "name": "Qwen/Qwen3.5-35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2025-04", + "release_date": "2026-02-25", + "last_updated": "2026-02-25", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.23, + "output": 1.86 } }, - "azure-anthropic/claude-opus-4-7": { - "id": "azure-anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7 (Azure Anthropic)", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "Qwen/Qwen3-VL-32B-Thinking": { + "id": "Qwen/Qwen3-VL-32B-Thinking", + "name": "Qwen/Qwen3-VL-32B-Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "release_date": "2025-10-21", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -65639,44 +61983,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.2, + "output": 1.5 } }, - "atria/atria-dawn-preview": { - "id": "atria/atria-dawn-preview", - "name": "Atria Dawn Preview (Atria)", - "description": "Preview model for early access evaluation, prototyping, and compatibility testing", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "Qwen/Qwen3-VL-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", + "name": "Qwen/Qwen3-VL-30B-A3B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-09-12", - "last_updated": "2026-09-12", + "release_date": "2025-10-05", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -65684,27 +62014,26 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.29, + "output": 1 } }, - "mistral/mistral-large-latest": { - "id": "mistral/mistral-large-latest", - "name": "Mistral Large Latest (Mistral AI)", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", + "Qwen/Qwen2.5-72B-Instruct": { + "id": "Qwen/Qwen2.5-72B-Instruct", + "name": "Qwen/Qwen2.5-72B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", + "release_date": "2024-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -65713,28 +62042,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 262144 + "context": 33000, + "output": 4000 }, "cost": { - "input": 4, - "output": 12 + "input": 0.59, + "output": 0.59 } }, - "mistral/ministral-8b-2512": { - "id": "mistral/ministral-8b-2512", - "name": "Ministral 8B (Mistral AI)", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "Qwen/Qwen3-VL-32B-Instruct": { + "id": "Qwen/Qwen3-VL-32B-Instruct", + "name": "Qwen/Qwen3-VL-32B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "release_date": "2025-10-21", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -65746,58 +62075,58 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 0.2, + "output": 0.6 } }, - "mistral/devstral-2512": { - "id": "mistral/devstral-2512", - "name": "Devstral 2 (Mistral AI)", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "Qwen/Qwen3-VL-235B-A22B-Thinking": { + "id": "Qwen/Qwen3-VL-235B-A22B-Thinking", + "name": "Qwen/Qwen3-VL-235B-A22B-Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "release_date": "2025-10-04", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.4, - "output": 2 + "input": 0.45, + "output": 3.5 } }, - "mistral/mistral-large-2512": { - "id": "mistral/mistral-large-2512", - "name": "Mistral Large 3 (Mistral AI)", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", + "Qwen/Qwen3-VL-8B-Instruct": { + "id": "Qwen/Qwen3-VL-8B-Instruct", + "name": "Qwen/Qwen3-VL-8B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "release_date": "2025-10-15", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -65807,28 +62136,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 0.18, + "output": 0.68 } }, - "mistral/ministral-14b-2512": { - "id": "mistral/ministral-14b-2512", - "name": "Ministral 14B (Mistral AI)", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "Qwen/Qwen3-VL-30B-A3B-Thinking": { + "id": "Qwen/Qwen3-VL-30B-A3B-Thinking", + "name": "Qwen/Qwen3-VL-30B-A3B-Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "release_date": "2025-10-11", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -65840,62 +62170,59 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 0.29, + "output": 1 } }, - "mistral/mistral-small-2506": { - "id": "mistral/mistral-small-2506", - "name": "Mistral Small 3.2 (Mistral AI)", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, + "Qwen/Qwen3-30B-A3B-Instruct-2507": { + "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", + "release_date": "2025-07-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.1, + "input": 0.09, "output": 0.3 } }, - "mistral/ministral-3b-2512": { - "id": "mistral/ministral-3b-2512", - "name": "Ministral 3B (Mistral AI)", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, + "Qwen/Qwen3-Coder-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "name": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "release_date": "2025-08-01", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -65903,26 +62230,26 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 0.07, + "output": 0.28 } }, - "mistral/codestral-2508": { - "id": "mistral/codestral-2508", - "name": "Codestral (Mistral AI)", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "codestral", + "Qwen/Qwen3-Coder-480B-A35B-Instruct": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "name": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-07-30", - "last_updated": "2025-07-30", + "release_date": "2025-07-31", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -65933,44 +62260,30 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0.25, + "output": 1 } }, - "vertex-anthropic/claude-haiku-4-5": { - "id": "vertex-anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (Vertex AI (Anthropic))", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "Qwen/Qwen3-VL-235B-A22B-Instruct": { + "id": "Qwen/Qwen3-VL-235B-A22B-Instruct", + "name": "Qwen/Qwen3-VL-235B-A22B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2025-10-04", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -65978,50 +62291,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1 + "input": 0.3, + "output": 1.5 } }, - "vertex-anthropic/claude-sonnet-4-5": { - "id": "vertex-anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (Vertex AI (Anthropic))", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "Qwen/Qwen2.5-7B-Instruct": { + "id": "Qwen/Qwen2.5-7B-Instruct", + "name": "Qwen/Qwen2.5-7B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2024-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -66029,96 +62321,63 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 33000, + "output": 4000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.05, + "output": 0.05 } }, - "vertex-anthropic/claude-opus-4-5-20251101": { - "id": "vertex-anthropic/claude-opus-4-5-20251101", - "name": "Claude Opus 4.5 (Vertex AI (Anthropic))", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "PaddlePaddle/PaddleOCR-VL-1.5": { + "id": "PaddlePaddle/PaddleOCR-VL-1.5", + "name": "PaddlePaddle/PaddleOCR-VL-1.5", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "release_date": "2026-01-29", + "last_updated": "2026-01-29", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 16384, + "output": 16384 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0, + "output": 0 } }, - "vertex-anthropic/claude-sonnet-5": { - "id": "vertex-anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5 (Vertex AI (Anthropic))", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "tencent/Hunyuan-A13B-Instruct": { + "id": "tencent/Hunyuan-A13B-Instruct", + "name": "tencent/Hunyuan-A13B-Instruct", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "release_date": "2025-06-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -66126,97 +62385,72 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.14, + "output": 0.57 } }, - "vertex-anthropic/claude-opus-4-6": { - "id": "vertex-anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6 (Vertex AI (Anthropic))", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1049000, + "output": 262000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 1.4, + "output": 4.4, + "cache_write": 0 } }, - "vertex-anthropic/claude-sonnet-4-6": { - "id": "vertex-anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6 (Vertex AI (Anthropic))", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "zai-org/GLM-4.5-Air": { + "id": "zai-org/GLM-4.5-Air", + "name": "zai-org/GLM-4.5-Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2025-07-28", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -66224,73 +62458,33 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.14, + "output": 0.86 } }, - "vertex-anthropic/claude-opus-4-7": { - "id": "vertex-anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7 (Vertex AI (Anthropic))", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "deepseek-ai/DeepSeek-R1": { + "id": "deepseek-ai/DeepSeek-R1", + "name": "deepseek-ai/DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 - } - }, - "vertex-openai/qwen3-235b-a22b-instruct-2507": { - "id": "vertex-openai/qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507 (Vertex AI (OpenAI-compatible))", - "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", + "release_date": "2025-05-28", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -66299,42 +62493,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.22, - "output": 0.88 + "input": 0.5, + "output": 2.18 } }, - "vertex-openai/qwen3-next-80b-a3b-thinking": { - "id": "vertex-openai/qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B Thinking (Vertex AI (OpenAI-compatible))", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", + "deepseek-ai/DeepSeek-V3.1-Terminus": { + "id": "deepseek-ai/DeepSeek-V3.1-Terminus", + "name": "deepseek-ai/DeepSeek-V3.1-Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2025-09-29", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -66343,45 +62528,39 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.15, - "output": 1.2 + "input": 0.27, + "output": 1 } }, - "vertex-openai/glm-5": { - "id": "vertex-openai/glm-5", - "name": "GLM-5 (Vertex AI (OpenAI-compatible))", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -66392,28 +62571,28 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 32768 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.1 + "input": 0.14, + "output": 0.28, + "cache_read": 0.003 } }, - "vertex-openai/qwen3-coder-480b-a35b-instruct": { - "id": "vertex-openai/qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct (Vertex AI (OpenAI-compatible))", - "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", - "family": "qwen", + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "deepseek-ai/DeepSeek-V4-Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -66424,32 +62603,32 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 1049000, + "output": 393000 }, "cost": { - "input": 0.22, - "output": 1.8, - "cache_read": 0.022 + "input": 1.74, + "output": 3.48, + "cache_read": 0.145 } }, - "vertex-openai/kimi-k2-thinking": { - "id": "vertex-openai/kimi-k2-thinking", - "name": "Kimi K2 Thinking (Vertex AI (OpenAI-compatible))", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "deepseek-ai/DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "release_date": "2025-12-03", + "last_updated": "2025-12-03", "modalities": { "input": [ "text" @@ -66458,33 +62637,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.06 + "input": 0.27, + "output": 0.42 } }, - "vertex-openai/qwen3-next-80b-a3b-instruct": { - "id": "vertex-openai/qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct (Vertex AI (OpenAI-compatible))", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "deepseek-ai/DeepSeek-OCR": { + "id": "deepseek-ai/DeepSeek-OCR", + "name": "deepseek-ai/DeepSeek-OCR", + "description": "OCR model for extracting structured text from documents and screenshots", + "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2025-10-20", + "last_updated": "2025-10-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -66492,35 +62668,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 1.2 + "input": 0, + "output": 0 } }, - "vertex-openai/glm-4.7": { - "id": "vertex-openai/glm-4.7", - "name": "GLM-4.7 (Vertex AI (OpenAI-compatible))", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", + "deepseek-ai/DeepSeek-V3": { + "id": "deepseek-ai/DeepSeek-V3", + "name": "deepseek-ai/DeepSeek-V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2024-12-26", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -66529,32 +62696,45 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 128000 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 0.25, + "output": 1 } - }, - "vertex-openai/grok-4-20-non-reasoning": { - "id": "vertex-openai/grok-4-20-non-reasoning", - "name": "Grok 4.20 Non-Reasoning (Vertex AI (OpenAI-compatible))", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", + } + } + }, + "merge-gateway": { + "id": "merge-gateway", + "env": [ + "MERGE_GATEWAY_API_KEY" + ], + "npm": "merge-gateway-ai-sdk-provider", + "name": "Merge Gateway", + "doc": "https://docs.merge.dev/merge-gateway", + "models": { + "xai/grok-4.3": { + "id": "xai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", "family": "grok", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -66562,7 +62742,7 @@ }, "open_weights": false, "limit": { - "context": 2000000, + "context": 1000000, "output": 30000 }, "cost": { @@ -66587,76 +62767,24 @@ } } }, - "vertex-openai/grok-4-6": { - "id": "vertex-openai/grok-4-6", - "name": "Grok 4.6 (Vertex AI (OpenAI-compatible))", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", + "xai/grok-4.20-0309-reasoning": { + "id": "xai/grok-4.20-0309-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "vertex-openai/grok-4-20-reasoning": { - "id": "vertex-openai/grok-4-20-reasoning", - "name": "Grok 4.20 Reasoning (Vertex AI (OpenAI-compatible))", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, "release_date": "2026-03-09", "last_updated": "2026-03-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -66664,7 +62792,7 @@ }, "open_weights": false, "limit": { - "context": 2000000, + "context": 1000000, "output": 30000 }, "cost": { @@ -66689,201 +62817,641 @@ } } }, - "vertex-openai/deepseek-v3.2": { - "id": "vertex-openai/deepseek-v3.2", - "name": "DeepSeek V3.2 (Vertex AI (OpenAI-compatible))", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 65536 - }, - "cost": { - "input": 0.56, - "output": 1.68, - "cache_read": 0.056 - } - }, - "google-vertex/gemini-3.6-flash": { - "id": "google-vertex/gemini-3.6-flash", - "name": "Gemini 3.6 Flash (Google Vertex AI)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.08333 + "input": 0.95, + "output": 4 } }, - "google-vertex/gemini-3.5-flash-lite": { - "id": "google-vertex/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite (Google Vertex AI)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 32768 } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.08333 + "input": 0.6, + "output": 2.5 } }, - "google-vertex/gemini-3.1-pro-preview": { - "id": "google-vertex/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro (Preview) (Google Vertex AI)", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } + "input": 0.6, + "output": 3 + } + }, + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image", + "video" ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.95, + "output": 4 + } + }, + "moonshotai/kimi-k2.7-code-highspeed": { + "id": "moonshotai/kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 32768 } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.9, + "output": 8 } }, - "google-vertex/gemini-3.5-flash": { - "id": "google-vertex/gemini-3.5-flash", - "name": "Gemini 3.5 Flash (Google Vertex AI)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "mistral/codestral-latest": { + "id": "mistral/codestral-latest", + "name": "Codestral (latest)", + "description": "Mistral code model for completions, refactors, and developer IDE workflows", + "family": "codestral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-05-29", + "last_updated": "2025-01-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 4096 + }, + "cost": { + "input": 0.3, + "output": 0.9 + } + }, + "mistral/mistral-large-latest": { + "id": "mistral/mistral-large-latest", + "name": "Mistral Large (latest)", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "mistral/devstral-small-2507": { + "id": "mistral/devstral-small-2507", + "name": "Devstral Small", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-07-10", + "last_updated": "2025-07-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0.1, + "output": 0.3 + } + }, + "mistral/pixtral-large-latest": { + "id": "mistral/pixtral-large-latest", + "name": "Pixtral Large (latest)", + "description": "Mistral's larger vision model for document-heavy image understanding and chat", + "family": "pixtral", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2024-11-04", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "mistral/mistral-medium-latest": { + "id": "mistral/mistral-medium-latest", + "name": "Mistral Medium (latest)", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.4, + "output": 2 + } + }, + "mistral/mistral-small-latest": { + "id": "mistral/mistral-small-latest", + "name": "Mistral Small (latest)", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.15, + "output": 0.6 + } + }, + "mistral/mistral-medium-2505": { + "id": "mistral/mistral-medium-2505", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.4, + "output": 2 + } + }, + "mistral/mistral-large-2411": { + "id": "mistral/mistral-large-2411", + "name": "Mistral Large 2.1", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-18", + "last_updated": "2024-11-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 2, + "output": 6 + } + }, + "mistral/magistral-medium-latest": { + "id": "mistral/magistral-medium-latest", + "name": "Magistral Medium (latest)", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-medium", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2025-03-17", + "last_updated": "2025-03-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2, + "output": 5 + } + }, + "mistral/devstral-medium-latest": { + "id": "mistral/devstral-medium-latest", + "name": "Devstral 2 (latest)", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2 + } + }, + "mistral/devstral-2512": { + "id": "mistral/devstral-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2 + } + }, + "mistral/mistral-large-2512": { + "id": "mistral/mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "mistral/devstral-medium-2507": { + "id": "mistral/devstral-medium-2507", + "name": "Devstral Medium", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-07-10", + "last_updated": "2025-07-10", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 128000 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2 + } + }, + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -66901,8 +63469,8 @@ "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", @@ -66921,29 +63489,20 @@ "output": 65536 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 0.08333 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "google-vertex/gemini-2.5-pro": { - "id": "google-vertex/gemini-2.5-pro", - "name": "Gemini 2.5 Pro (Google Vertex AI)", + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", "description": "Google's proven reasoning model for coding, math, and multimodal analysis", "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - }, { "type": "budget_tokens", "min": 128, @@ -66995,26 +63554,21 @@ } } }, - "google-vertex/gemini-2.5-flash": { - "id": "google-vertex/gemini-2.5-flash", - "name": "Gemini 2.5 Flash (Google Vertex AI)", + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1, - "max": 24576 + "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 0, + "max": 24576 } ], "tool_call": true, @@ -67038,18 +63592,19 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65535 + "output": 65536 }, "cost": { "input": 0.3, "output": 2.5, - "cache_read": 0.03 + "cache_read": 0.03, + "input_audio": 1 } }, - "google-vertex/gemini-3.7-flash": { - "id": "google-vertex/gemini-3.7-flash", - "name": "Gemini 3.7 Flash (Google Vertex AI)", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "family": "gemini-flash", "attachment": true, "reasoning": true, @@ -67067,9 +63622,9 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", @@ -67088,24 +63643,55 @@ "output": 65536 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.08333 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 } }, - "google-vertex/gemini-3-flash-preview": { - "id": "google-vertex/gemini-3-flash-preview", - "name": "Gemini 3 Flash (Preview) (Google Vertex AI)", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + } + }, + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -67116,8 +63702,8 @@ "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", @@ -67133,25 +63719,42 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65535 + "output": 65536 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "google-vertex/gemini-3.8-flash": { - "id": "google-vertex/gemini-3.8-flash", - "name": "Gemini 3.8 Flash (Google Vertex AI)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini", + "google/gemini-flash-lite-latest": { + "id": "google/gemini-flash-lite-latest", + "name": "Gemini Flash-Lite Latest", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -67161,13 +63764,16 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2025-01", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", "image", - "audio" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -67179,19 +63785,28 @@ "output": 65536 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.08333 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "google-vertex/gemini-2.5-flash-lite": { - "id": "google-vertex/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash Lite (Google Vertex AI)", + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", "family": "gemini-flash-lite", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 512, + "max": 24576 + } + ], "tool_call": true, "structured_output": true, "temperature": true, @@ -67213,26 +63828,26 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65535 + "output": 65536 }, "cost": { "input": 0.1, "output": 0.4, - "cache_read": 0.01 + "cache_read": 0.01, + "input_audio": 0.3 } }, - "google-vertex/gemini-3.1-flash-lite": { - "id": "google-vertex/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite (Google Vertex AI)", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -67243,8 +63858,8 @@ "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", @@ -67263,90 +63878,87 @@ "output": 65536 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 0.08333 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "aws-mantle/gpt-6-astra": { - "id": "aws-mantle/gpt-6-astra", - "name": "GPT-6 Astra (AWS Mantle)", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "context": 262144, + "output": 32768 } }, - "aws-mantle/gpt-5.6-luna": { - "id": "aws-mantle/gpt-5.6-luna", - "name": "GPT-5.6 Luna (AWS Mantle)", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", + "google/gemini-3-pro-preview": { + "id": "google/gemini-3-pro-preview", + "name": "Gemini 3 Pro Preview", + "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -67355,46 +63967,61 @@ }, "open_weights": false, "limit": { - "context": 921600, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.22, - "output": 1.32, - "cache_read": 0.022, - "cache_write": 0.275 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "aws-mantle/gpt-5.6-terra": { - "id": "aws-mantle/gpt-5.6-terra", - "name": "GPT-5.6 Terra (AWS Mantle)", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -67403,46 +64030,46 @@ }, "open_weights": false, "limit": { - "context": 921600, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.2, - "output": 13.2, - "cache_read": 0.22, - "cache_write": 2.75 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 } }, - "aws-mantle/gpt-5.6-sol": { - "id": "aws-mantle/gpt-5.6-sol", - "name": "GPT-5.6 Sol (AWS Mantle)", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + "google/gemini-flash-latest": { + "id": "google/gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -67451,45 +64078,46 @@ }, "open_weights": false, "limit": { - "context": 921600, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 4.4, - "output": 22, - "cache_read": 0.44, - "cache_write": 5.5 + "input": 0.3, + "output": 2.5, + "cache_read": 0.075, + "input_audio": 1 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4 (OpenAI)", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -67498,43 +64126,44 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro (OpenAI)", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -67542,31 +64171,43 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 30, - "output": 180 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5 Turbo (OpenAI)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", "family": "gpt", - "attachment": false, - "reasoning": false, + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -67574,30 +64215,34 @@ }, "open_weights": false, "limit": { - "context": 16385, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openai/gpt-4o-transcribe": { - "id": "openai/gpt-4o-transcribe", - "name": "GPT-4o Transcribe (OpenAI)", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", "family": "gpt", - "attachment": false, + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-03-20", - "last_updated": "2025-03-20", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", - "audio" + "image", + "pdf" ], "output": [ "text" @@ -67605,42 +64250,33 @@ }, "open_weights": false, "limit": { - "context": 16000, - "output": 2000 + "context": 128000, + "output": 16384 }, "cost": { "input": 2.5, - "output": 10 + "output": 10, + "cache_read": 1.25 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro (OpenAI)", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -67648,32 +64284,50 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 + } + }, + "openai/gpt-4o-2024-05-13": { + "id": "openai/gpt-4o-2024-05-13", + "name": "GPT-4o (2024-05-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", + "modalities": { + "input": [ + "text", + "image" ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 5, + "output": 15 } }, "openai/gpt-5.4-nano": { "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 Nano (OpenAI)", + "name": "GPT-5.4 nano", "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", "family": "gpt-nano", "attachment": true, @@ -67692,7 +64346,7 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-08-31", "release_date": "2026-03-17", "last_updated": "2026-03-17", @@ -67717,24 +64371,24 @@ "cache_read": 0.02 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o (OpenAI)", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "openai/gpt-5-chat-latest": { + "id": "openai/gpt-5-chat-latest", + "name": "GPT-5 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -67742,39 +64396,29 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2.5, + "input": 1.25, "output": 10, - "cache_read": 1.25 + "cache_read": 0.125 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini (OpenAI)", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", + "openai/gpt-5.3-chat-latest": { + "id": "openai/gpt-5.3-chat-latest", + "name": "GPT-5.3 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", @@ -67786,39 +64430,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2 Pro (OpenAI)", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", + "openai/gpt-4o-2024-08-06": { + "id": "openai/gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-08-06", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", @@ -67830,21 +64463,21 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 272000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 21, - "output": 168 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4 Mini (OpenAI)", - "description": "Fast o-series model for compact reasoning, coding, and tool use", + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", "family": "o-mini", - "attachment": true, + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -67860,12 +64493,11 @@ "structured_output": true, "temperature": false, "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -67879,25 +64511,38 @@ "cost": { "input": 1.1, "output": 4.4, - "cache_read": 0.275 + "cache_read": 0.55 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3 Mini (OpenAI)", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": false, - "reasoning": false, - "tool_call": false, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -67905,31 +64550,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai/gpt-4": { - "id": "openai/gpt-4", - "name": "GPT-4 (OpenAI)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, + "openai/gpt-5.1-chat-latest": { + "id": "openai/gpt-5.1-chat-latest", + "name": "GPT-5.1 Chat", + "description": "Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -67937,43 +64585,33 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 128000, + "output": 16384 }, "cost": { - "input": 30, - "output": 60 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex (OpenAI)", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "openai/gpt-5.2-chat-latest": { + "id": "openai/gpt-5.2-chat-latest", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, + "structured_output": true, + "temperature": false, "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -67981,9 +64619,8 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { "input": 1.75, @@ -67993,7 +64630,7 @@ }, "openai/gpt-4.1-nano": { "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 Nano (OpenAI)", + "name": "GPT-4.1 nano", "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", "family": "gpt-nano", "attachment": true, @@ -68015,7 +64652,7 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1047576, "output": 32768 }, "cost": { @@ -68024,33 +64661,23 @@ "cache_read": 0.025 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano (OpenAI)", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "openai/gpt-4o-2024-11-20": { + "id": "openai/gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -68058,19 +64685,18 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, "openai/o1": { "id": "openai/o1", - "name": "o1 (OpenAI)", + "name": "o1", "description": "O-series reasoning model for hard analysis, math, coding, and planning", "family": "o", "attachment": true, @@ -68085,7 +64711,7 @@ ] } ], - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": false, "knowledge": "2023-09", @@ -68112,31 +64738,36 @@ "cache_read": 7.5 } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5 Pro (OpenAI)", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high" + "none", + "low", + "medium", + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -68144,39 +64775,73 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 272000 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 15, - "output": 120 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "openai/gpt-6-astra": { - "id": "openai/gpt-6-astra", - "name": "GPT-6 Astra (OpenAI)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "release_date": "2026-09-03", - "last_updated": "2026-09-03", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -68188,42 +64853,90 @@ }, "open_weights": false, "limit": { - "context": 1050000, + "context": 400000, + "input": 272000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1 (OpenAI)", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", "family": "gpt", "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -68240,24 +64953,24 @@ "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o Mini (OpenAI)", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", "family": "gpt-mini", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -68270,40 +64983,39 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 Mini (OpenAI)", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -68320,37 +65032,24 @@ "output": 128000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna (OpenAI)", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", @@ -68363,21 +65062,19 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2 (OpenAI)", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", "family": "gpt", "attachment": true, "reasoning": true, @@ -68388,17 +65085,16 @@ "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -68415,14 +65111,14 @@ "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, "openai/gpt-5.5": { "id": "openai/gpt-5.5", - "name": "GPT-5.5 (OpenAI)", + "name": "GPT-5.5", "description": "Default frontier GPT for coding, computer use, research, and knowledge work", "family": "gpt", "attachment": true, @@ -68461,6 +65157,22 @@ "input": 922000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, "cost": { "input": 5, "output": 30, @@ -68483,345 +65195,339 @@ } } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1 (OpenAI)", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, + "zai/glm-4.7": { + "id": "zai/glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 Mini (OpenAI)", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + "zai/glm-4.5": { + "id": "zai/glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "openai/gpt-6-luna": { - "id": "openai/gpt-6-luna", - "name": "GPT-6 Luna (OpenAI)", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, + "zai/glm-4.7-flashx": { + "id": "zai/glm-4.7-flashx", + "name": "GLM-4.7-FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.5, + "input": 0.07, + "output": 0.4, "cache_read": 0.01, - "cache_write": 0.125 + "cache_write": 0 } }, - "openai/gpt-4o-mini-transcribe": { - "id": "openai/gpt-4o-mini-transcribe", - "name": "GPT-4o Mini Transcribe (OpenAI)", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "gpt", + "zai/glm-5.1": { + "id": "zai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-03-20", - "last_updated": "2025-03-20", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16000, - "output": 2000 + "context": 200000, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 5 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT-5.6 Terra (OpenAI)", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, + "zai/glm-4.6": { + "id": "zai/glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo (OpenAI)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, + "zai/glm-5.2": { + "id": "zai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 50000 + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 10, - "output": 30 + "input": 1.4, + "output": 4.4 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3 (OpenAI)", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "zai/glm-4.5-air": { + "id": "zai/glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 98304 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.2, + "output": 1.1, + "cache_read": 0.03, + "cache_write": 0 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5 (OpenAI)", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "zai/glm-5": { + "id": "zai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0 } }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol (OpenAI)", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, + "zai/glm-5-turbo": { + "id": "zai/glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -68829,43 +65535,38 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 131072 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5 + "input": 1.2, + "output": 4, + "cache_read": 0.24, + "cache_write": 0 } }, - "openai/gpt-6-sol": { - "id": "openai/gpt-6-sol", - "name": "GPT-6 Sol (OpenAI)", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", + "anthropic/claude-haiku-4-5-20251001": { + "id": "anthropic/claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", @@ -68878,68 +65579,71 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "moonshot/kimi-k3": { - "id": "moonshot/kimi-k3", - "name": "Kimi K3 (Moonshot AI)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "anthropic/claude-opus-4-1-20250805": { + "id": "anthropic/claude-opus-4-1-20250805", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 200000, + "output": 32000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "moonshot/kimi-k2.7-code-highspeed": { - "id": "moonshot/kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed (Moonshot AI)", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -68952,309 +65656,334 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "moonshot/kimi-k2.6": { - "id": "moonshot/kimi-k2.6", - "name": "Kimi K2.6 (Moonshot AI)", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "anthropic/claude-opus-4-5-20251101": { + "id": "anthropic/claude-opus-4-5-20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "moonshot/kimi-k2.5": { - "id": "moonshot/kimi-k2.5", - "name": "Kimi K2.5 (Moonshot AI)", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "anthropic/claude-opus-4-8": { + "id": "anthropic/claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 128000 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 5, + "output": 25 } }, - "moonshot/kimi-k2.7-code": { - "id": "moonshot/kimi-k2.7-code", - "name": "Kimi K2.7 Code (Moonshot AI)", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "anthropic/claude-opus-4-20250514": { + "id": "anthropic/claude-opus-4-20250514", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } - } - } - }, - "above": { - "id": "above", - "env": [ - "ABOVE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.above.dev/v1", - "name": "above.dev", - "doc": "https://above.dev/docs", - "models": { - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM 5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + }, + "anthropic/claude-sonnet-4-20250514": { + "id": "anthropic/claude-sonnet-4-20250514", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.165, - "output": 0.55, - "cache_read": 0.0319 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "glm-5.2-fast": { - "id": "glm-5.2-fast", - "name": "GLM 5.2 Fast", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", + "low", + "medium", "high", "max" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 2.31, - "output": 7.26, - "cache_read": 0.231 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen 3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": false, + "anthropic/claude-sonnet-4-5-20250929": { + "id": "anthropic/claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -69262,20 +65991,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 2.2, - "output": 6.6, - "cache_read": 0.275 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek-v4-flash-vision-exp": { - "id": "deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision (Exp)", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -69286,60 +66016,86 @@ "type": "effort", "values": [ "low", + "medium", "high", "max" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 64000 }, "cost": { - "input": 0.242, - "output": 0.726, - "reasoning": 0.726, - "cache_read": 0.0077 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo V2.5 Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "cohere/command-a-03-2025": { + "id": "cohere/command-a-03-2025", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8000 }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "cohere/command-r-08-2024": { + "id": "cohere/command-r-08-2024", + "name": "Command R", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -69350,40 +66106,56 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.5077, - "output": 1.0154, - "cache_read": 0.0042 + "input": 0.15, + "output": 0.6 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "cohere/command-r7b-12-2024": { + "id": "cohere/command-r7b-12-2024", + "name": "Command R7B", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2024-12-02", + "last_updated": "2024-12-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "structured_output": true, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.0375, + "output": 0.15 + } + }, + "cohere/command-r-plus-08-2024": { + "id": "cohere/command-r-plus-08-2024", + "name": "Command R+", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", + "attachment": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -69394,20 +66166,19 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 4000 }, "cost": { - "input": 1.54, - "output": 4.84, - "cache_read": 0.154 + "input": 2.5, + "output": 10 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "alibaba/qwen3.7-max": { + "id": "alibaba/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -69415,21 +66186,15 @@ "type": "toggle" }, { - "type": "effort", - "values": [ - "high", - "max" - ] + "type": "budget_tokens", + "min": 1, + "max": 250000 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -69438,92 +66203,75 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 65536 }, "cost": { - "input": 0.726, - "output": 2.178, - "reasoning": 2.178, - "cache_read": 0.0242 + "input": 1.65, + "output": 4.95 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "alibaba/qwen3.6-plus": { + "id": "alibaba/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "budget_tokens", + "min": 1, + "max": 250000 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 65536 }, "cost": { - "input": 0.165, - "output": 0.66, - "reasoning": 0.66, - "cache_read": 0.0033 + "input": 0.5, + "output": 3 } - } - } - }, - "kilo": { - "id": "kilo", - "env": [ - "KILO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.kilo.ai/api/gateway", - "name": "Kilo Gateway", - "doc": "https://kilo.ai", - "models": { - "sao10k/l3-lunaris-8b": { - "id": "sao10k/l3-lunaris-8b", - "name": "Sao10K: Llama 3 8B Lunaris", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + }, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2024-08-13", - "last_updated": "2024-08-13", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -69532,28 +66280,34 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 7372 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.04, - "output": 0.05 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "sao10k/l3.3-euryale-70b": { - "id": "sao10k/l3.3-euryale-70b", - "name": "Sao10K: Llama 3.3 Euryale 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2024-12-18", - "last_updated": "2024-12-18", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -69562,28 +66316,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.65, - "output": 0.75 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "sao10k/l3.1-euryale-70b": { - "id": "sao10k/l3.1-euryale-70b", - "name": "Sao10K: Llama 3.1 Euryale 70B v2.2", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-08-28", - "last_updated": "2024-08-28", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -69592,28 +66347,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.85, - "output": 0.85 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "kwaipilot/kat-coder-pro-v2.5": { - "id": "kwaipilot/kat-coder-pro-v2.5", - "name": "Kwaipilot: KAT-Coder-Pro V2.5", - "description": "KAT-Coder-Pro V2.5 is a flagship-level Agentic Coding model that can directly hand over an entire issue or an entire business workflow to it, allowing it to autonomously locate and make...", - "family": "kat-coder", + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-10", - "last_updated": "2026-07-10", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -69622,237 +66379,213 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 235929 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.74, - "output": 2.96, - "cache_read": 0.15 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "stealth/claude-opus-4.6": { - "id": "stealth/claude-opus-4.6", - "name": "Stealth: Claude Opus 4.6 (20% off)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 4, - "output": 20, - "reasoning": 0, - "cache_read": 0.4, - "cache_write": 5 + "input": 0.3, + "output": 1.2 } }, - "stealth/claude-opus-4.7": { - "id": "stealth/claude-opus-4.7", - "name": "Stealth: Claude Opus 4.7 (20% off)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 128000 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 512000, "output": 128000 }, "cost": { - "input": 4, - "output": 20, - "reasoning": 0, - "cache_read": 0.4, - "cache_write": 5 + "input": 0.6, + "output": 2.4 } }, - "stealth/claude-opus-4.8": { - "id": "stealth/claude-opus-4.8", - "name": "Stealth: Claude Opus 4.8 (20% off)", - "description": "Your prompts and completions may be retained and used to train or improve the provider's services. This third-party-served variant of Claude Opus 4.8 is offered at 20% lower cost than standard Claude Opus 4.8 pricing and is not served by Anthropic or Kilo Code.", - "family": "claude-opus", - "attachment": true, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 196608, "output": 128000 }, "cost": { - "input": 4, - "output": 20, - "reasoning": 0, - "cache_read": 0.4, - "cache_write": 5 + "input": 0.3, + "output": 1.2 } }, - "stealth/claude-sonnet-4.6": { - "id": "stealth/claude-sonnet-4.6", - "name": "Stealth: Claude Sonnet 4.6 (20% off)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2.4, - "output": 12, - "reasoning": 0, - "cache_read": 0.24, - "cache_write": 3 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "stealth/qwen3.6-plus": { - "id": "stealth/qwen3.6-plus", - "name": "Stealth: Qwen3.6 Plus (50% off)", - "description": "Your prompts and completions may be retained and used to train or improve the provider's services. This third-party-served variant of Qwen3.6 Plus is offered at 50% lower cost than standard Qwen3.6 Plus pricing and is not served by Alibaba or Kilo Code. Note: a surcharge applies to long-context workloads exceeding 256K input tokens.", - "family": "qwen3.6", + "minimax/minimax-m2.5-highspeed": { + "id": "minimax/minimax-m2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-13", + "last_updated": "2026-02-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 + } + } + } + }, + "qihang-ai": { + "id": "qihang-ai", + "env": [ + "QIHANG_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.qhaigc.net/v1", + "name": "QiHang", + "doc": "https://www.qhaigc.net/docs", + "models": { + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "knowledge": "2025-07-31", + "release_date": "2025-10-01", + "last_updated": "2025-10-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -69860,43 +66593,35 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 0, - "cache_read": 0.025, - "cache_write": 0.3125 + "input": 0.14, + "output": 0.71 } }, - "bytedance-seed/seed-1.6-flash": { - "id": "bytedance-seed/seed-1.6-flash", - "name": "ByteDance Seed: Seed 1.6 Flash", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "image", "text", - "video" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -69904,43 +66629,45 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 0.09, + "output": 0.71, + "tiers": [ + { + "input": 0.09, + "output": 0.71, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 0.09, + "output": 0.71 + } } }, - "bytedance-seed/seed-2.0-lite": { - "id": "bytedance-seed/seed-2.0-lite", - "name": "Seed 2.0 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-03", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -69948,40 +66675,32 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.25, - "output": 2 + "input": 0.71, + "output": 3.57 } }, - "bytedance-seed/seed-2-1-turbo": { - "id": "bytedance-seed/seed-2-1-turbo", - "name": "ByteDance Seed: Seed 2.1 Turbo", - "description": "Seed 2.1 Turbo is a multimodal model from ByteDance Seed for coding and long-horizon agent workflows. It is suited for end-to-end software delivery, multi-step task execution, and understanding visual and...", - "family": "seed", + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -69989,43 +66708,33 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 2.5 + "input": 0.25, + "output": 2 } }, - "bytedance-seed/seed-2.0-mini": { - "id": "bytedance-seed/seed-2.0-mini", - "name": "Seed 2.0 Mini", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -70033,41 +66742,32 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 0.43, + "output": 2.14 } }, - "bytedance-seed/seed-2.0-code": { - "id": "bytedance-seed/seed-2.0-code", - "name": "Seed 2.0 Code", - "description": "Seed 2.0 Code is a model from ByteDance Seed optimized for agentic coding. It is suited for frontend development, multilingual programming tasks, and coding-agent workflows in tools such as Claude...", - "family": "seed", + "gemini-3-pro-preview": { + "id": "gemini-3-pro-preview", + "name": "Gemini 3 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-11", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ @@ -70076,40 +66776,31 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1000000, + "output": 65000 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.57, + "output": 3.43 } }, - "bytedance-seed/seed-1.6": { - "id": "bytedance-seed/seed-1.6", - "name": "ByteDance Seed: Seed 1.6", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5-Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ - "image", "text", - "video" + "image" ], "output": [ "text" @@ -70117,42 +66808,35 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 2 + "input": 0.04, + "output": 0.29 } }, - "~moonshotai/kimi-latest": { - "id": "~moonshotai/kimi-latest", - "name": "MoonshotAI: Kimi Latest", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -70161,38 +66845,45 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 943718 + "output": 65536 }, "cost": { - "input": 1.4989, - "output": 10.758, - "cache_read": 0.3 + "input": 0.07, + "output": 0.43, + "tiers": [ + { + "input": 0.07, + "output": 0.43, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 0.07, + "output": 0.43 + } } }, - "poolside/laguna-s-2.1:free": { - "id": "poolside/laguna-s-2.1:free", - "name": "Poolside: Laguna S 2.1 (free)", - "description": "Laguna S 2.1 is the latest coding agent model from [Poolside](). Laguna S 2.1 is a 118B total parameter model with 8B active parameters, scoring 70.2% on Terminal-Bench 2.1 and...", - "family": "laguna-s", - "attachment": false, + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -70200,75 +66891,75 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.14, + "output": 1.14 } - }, - "poolside/laguna-s-2.1": { - "id": "poolside/laguna-s-2.1", - "name": "Poolside: Laguna S 2.1", - "description": "Laguna S 2.1 is the latest coding agent model from [Poolside](). Laguna S 2.1 is a 118B total parameter model with 8B active parameters, scoring 70.2% on Terminal-Bench 2.1 and...", - "family": "laguna-s", + } + } + }, + "xiaomi-token-plan-ams": { + "id": "xiaomi-token-plan-ams", + "env": [ + "XIAOMI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://token-plan-ams.xiaomimimo.com/v1", + "name": "Xiaomi Token Plan (Europe)", + "doc": "https://platform.xiaomimimo.com/#/docs", + "models": { + "mimo-v2.5-tts": { + "id": "mimo-v2.5-tts", + "name": "MiMo-V2.5-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.01 + "input": 0, + "output": 0 } }, - "poolside/laguna-xs-2.1:free": { - "id": "poolside/laguna-xs-2.1:free", - "name": "Poolside: Laguna XS 2.1 (free)", - "description": "Laguna XS 2.1 is the latest coding agent model in the 33B-A3B category from [Poolside](https://poolside.ai/) and a step forward from their Laguna XS.2 model (released in April 2026). It combines...", - "family": "laguna", + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-07-02", - "last_updated": "2026-07-02", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -70277,37 +66968,37 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 1048576, + "output": 131072 }, "cost": { "input": 0, - "output": 0 + "output": 0, + "cache_read": 0 } }, - "poolside/laguna-xs-2.1": { - "id": "poolside/laguna-xs-2.1", - "name": "Poolside: Laguna XS 2.1", - "description": "Laguna XS 2.1 is the latest coding agent model in the 33B-A3B category from [Poolside](https://poolside.ai/) and a step forward from their Laguna XS.2 model (released in April 2026). It combines...", - "family": "laguna", + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-07-02", - "last_updated": "2026-07-02", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -70318,799 +67009,570 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1048576, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.05 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "prism-ml/ternary-bonsai-2-27b": { - "id": "prism-ml/ternary-bonsai-2-27b", - "name": "PrismML: Ternary Bonsai 2 27B", - "description": "Bonsai 2 27B is a 27B-parameter reasoning model from PrismML derived from Qwen3.8-27B. It supports coding, mathematics, tool calling, and image understanding with a 262K-token context window. Ternary compression shrinks...", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-18", - "last_updated": "2026-09-18", + "mimo-v2-tts": { + "id": "mimo-v2-tts", + "name": "MiMo-V2-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.075, - "output": 0.5 + "input": 0, + "output": 0 } }, - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "mimo-v2.5-tts-voicedesign": { + "id": "mimo-v2.5-tts-voicedesign", + "name": "MiMo-V2.5-TTS-VoiceDesign", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "mimo-v2.5-tts-voiceclone": { + "id": "mimo-v2.5-tts-voiceclone", + "name": "MiMo-V2.5-TTS-VoiceClone", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0 } - }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + } + } + }, + "modelscope": { + "id": "modelscope", + "env": [ + "MODELSCOPE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api-inference.modelscope.cn/v1", + "name": "ModelScope", + "doc": "https://modelscope.cn/docs/model-service/API-Inference/intro", + "models": { + "Qwen/Qwen3-30B-A3B-Thinking-2507": { + "id": "Qwen/Qwen3-30B-A3B-Thinking-2507", + "name": "Qwen3 30B A3B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0 } }, - "anthropic/claude-3-haiku": { - "id": "anthropic/claude-3-haiku", - "name": "Anthropic: Claude 3 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude", - "attachment": true, - "reasoning": false, + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen3-235B-A22B-Thinking-2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2024-03-13", - "last_updated": "2024-03-13", + "knowledge": "2025-04", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 4096 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.03, - "cache_write": 0.3 + "input": 0, + "output": 0 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Claude Opus 4.8 is Anthropic's most capable generally available model in the Opus family. It supports text, image, and file inputs with text output, with reasoning support and a 1M-token...", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "Qwen/Qwen3-Coder-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "name": "Qwen3 Coder 30B A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2025-04", + "release_date": "2025-07-31", + "last_updated": "2025-07-31", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "Qwen/Qwen3-30B-A3B-Instruct-2507": { + "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen3 30B A3B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-04", + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0, + "output": 0 } }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5", - "description": "Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work. It is particularly strong at end-to-end software tasks, code review and bug finding, visual analysis...", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "knowledge": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-07-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0 } }, - "anthropic/claude-fable-5.1": { - "id": "anthropic/claude-fable-5.1", - "name": "Claude Fable 5.1", - "description": "Claude Fable 5.1 improves on Claude Fable 5 across the board, with the biggest gains in agentic coding, long-running agentic workflows, and knowledge work: long code refactors, front-end and visual...", - "family": "claude-fable", - "attachment": true, + "ZhipuAI/GLM-4.6": { + "id": "ZhipuAI/GLM-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 202752, + "output": 98304 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 0, + "output": 0 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "ZhipuAI/GLM-4.5": { + "id": "ZhipuAI/GLM-4.5", + "name": "GLM-4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 131072, + "output": 98304 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0 } - }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude Fable 5 is a Mythos-class model from Anthropic, built for autonomous knowledge work and coding. It supports text, image, and file inputs with text output, with reasoning support and...", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + } + } + }, + "groq": { + "id": "groq", + "env": [ + "GROQ_API_KEY" + ], + "npm": "@ai-sdk/groq", + "name": "Groq", + "doc": "https://console.groq.com/docs/models", + "models": { + "llama-3.3-70b-versatile": { + "id": "llama-3.3-70b-versatile", + "name": "Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.59, + "output": 0.79 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Anthropic: Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "llama-3.1-8b-instant": { + "id": "llama-3.1-8b-instant", + "name": "Llama 3.1 8B", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "image", - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.05, + "output": 0.08 } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Sonnet 5 is Anthropic's most capable Sonnet-class model, with frontier performance across coding, agents, and professional work. It supports adaptive thinking with selectable reasoning effort levels (low, medium, high, max,...", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "whisper-large-v3-turbo": { + "id": "whisper-large-v3-turbo", + "name": "Whisper Large V3 Turbo", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "context": 0, + "output": 0 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "whisper-large-v3": { + "id": "whisper-large-v3", + "name": "Whisper", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2023-09-01", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "context": 0, + "output": 0 } }, - "anthropic/claude-opus-5.5": { - "id": "anthropic/claude-opus-5.5", - "name": "Claude Opus 5.5", - "description": "Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code...", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "meta-llama/llama-prompt-guard-2-86m": { + "id": "meta-llama/llama-prompt-guard-2-86m", + "name": "Prompt Guard 2 86M", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-05-29", + "last_updated": "2025-05-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 512, + "output": 512 }, + "status": "beta", "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 + "input": 0.04, + "output": 0.04 } }, - "nex-agi/nex-n2.5-mini": { - "id": "nex-agi/nex-n2.5-mini", - "name": "Nex AGI: Nex-N2.5-Mini", - "description": "Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...", - "family": "agi", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "medium", - "high" - ] - } - ], + "meta-llama/llama-prompt-guard-2-22m": { + "id": "meta-llama/llama-prompt-guard-2-22m", + "name": "Llama Prompt Guard 2 22M", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "temperature": false, + "release_date": "2025-05-29", + "last_updated": "2025-05-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 235929 + "context": 512, + "output": 512 }, + "status": "beta", "cost": { - "input": 0.025, - "output": 0.1, - "cache_read": 0.0025 + "input": 0.03, + "output": 0.03 } }, - "nex-agi/nex-n2.5-mini:free": { - "id": "nex-agi/nex-n2.5-mini:free", - "name": "Nex AGI: Nex-N2.5-Mini (free)", - "description": "Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...", - "family": "agi", + "meta-llama/llama-4-scout-17b-16e-instruct": { + "id": "meta-llama/llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout 17B 16E", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -71120,28 +67582,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 235929 + "context": 131072, + "output": 8192 }, + "status": "beta", "cost": { - "input": 0, - "output": 0 + "input": 0.11, + "output": 0.34 } }, - "nex-agi/nex-n2.5-pro:free": { - "id": "nex-agi/nex-n2.5-pro:free", - "name": "Nex AGI: Nex-N2.5-Pro (free)", - "description": "Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...", - "family": "agi", - "attachment": true, + "openai/gpt-oss-safeguard-20b": { + "id": "openai/gpt-oss-safeguard-20b", + "name": "Safety GPT OSS 20B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", "medium", "high" ] @@ -71150,39 +67613,39 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "release_date": "2025-10-29", + "last_updated": "2026-06-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 235929 + "context": 131072, + "output": 65536 }, + "status": "beta", "cost": { - "input": 0, - "output": 0 + "input": 0.075, + "output": 0.3 } }, - "nex-agi/nex-n2.5-pro": { - "id": "nex-agi/nex-n2.5-pro", - "name": "Nex AGI: Nex-N2.5-Pro", - "description": "Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...", - "family": "agi", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", "medium", "high" ] @@ -71191,41 +67654,8 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 235929 - }, - "cost": { - "input": 0.075, - "output": 0.25, - "cache_read": 0.015 - } - }, - "cohere/command-r-08-2024": { - "id": "cohere/command-r-08-2024", - "name": "Command R", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "release_date": "2025-08-05", + "last_updated": "2025-10-21", "modalities": { "input": [ "text" @@ -71236,26 +67666,28 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 131072, + "output": 65536 }, "cost": { "input": 0.15, - "output": 0.6 + "output": 0.6, + "cache_read": 0.075 } }, - "cohere/command-a-plus": { - "id": "cohere/command-a-plus", - "name": "Cohere: Command A+", - "description": "Command A+ is Cohere's flagship model for enterprise agentic workflows. It accepts text and image inputs with a 192K context window, supports native tool calling with strict tool schemas, structured...", - "family": "command-a", - "attachment": true, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } @@ -71263,110 +67695,90 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "release_date": "2025-08-05", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 192000, - "output": 64000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.5, - "cache_read": 0.15 + "input": 0.075, + "output": 0.3, + "cache_read": 0.0375 } }, - "cohere/command-a": { - "id": "cohere/command-a", - "name": "Cohere: Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", + "canopylabs/orpheus-v1-english": { + "id": "canopylabs/orpheus-v1-english", + "name": "Canopy Labs Orpheus V1 English", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "canopylabs", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "temperature": false, + "release_date": "2025-12-19", + "last_updated": "2025-12-19", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 4000, + "output": 50000 }, - "cost": { - "input": 2.5, - "output": 10 - } + "status": "beta" }, - "cohere/north-mini-code:free": { - "id": "cohere/north-mini-code:free", - "name": "Cohere: North Mini Code (free)", - "description": "North Mini Code is Cohere's first agentic coding model and the debut of its North family. A sparse mixture-of-experts model with 30B total parameters and 3B active, it is optimized...", - "family": "north", + "canopylabs/orpheus-arabic-saudi": { + "id": "canopylabs/orpheus-arabic-saudi", + "name": "Canopy Labs Orpheus Arabic Saudi", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "canopylabs", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 4000, + "output": 50000 }, - "cost": { - "input": 0, - "output": 0 - } + "status": "beta" }, - "cohere/command-r7b-12-2024": { - "id": "cohere/command-r7b-12-2024", - "name": "Command R7B", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-r", + "groq/compound": { + "id": "groq/compound", + "name": "Compound", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "groq", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-12-02", - "last_updated": "2024-12-02", + "release_date": "2025-09-04", + "last_updated": "2025-09-04", "modalities": { "input": [ "text" @@ -71375,29 +67787,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 0.0375, - "output": 0.15 + "context": 131072, + "output": 8192 } }, - "cohere/command-r-plus-08-2024": { - "id": "cohere/command-r-plus-08-2024", - "name": "Command R+", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", + "groq/compound-mini": { + "id": "groq/compound-mini", + "name": "Compound Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "groq", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "release_date": "2025-09-04", + "last_updated": "2025-09-04", "modalities": { "input": [ "text" @@ -71406,21 +67812,17 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 2.5, - "output": 10 + "context": 131072, + "output": 8192 } }, - "deepseek/deepseek-chat-v3.1": { - "id": "deepseek/deepseek-chat-v3.1", - "name": "DeepSeek: DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "qwen/qwen3-32b": { + "id": "qwen/qwen3-32b", + "name": "Qwen3-32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -71428,15 +67830,14 @@ "type": "effort", "values": [ "none", - "high" + "default" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "release_date": "2025-06-11", + "last_updated": "2025-06-12", "modalities": { "input": [ "text" @@ -71445,41 +67846,45 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163840, - "output": 32768 + "context": 131072, + "output": 40960 }, + "status": "beta", "cost": { - "input": 0.27, - "output": 1, - "cache_read": 0.135 + "input": 0.29, + "output": 0.59 } - }, - "deepseek/deepseek-v4-flash-0731": { - "id": "deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.", - "family": "deepseek-flash", + } + } + }, + "mixlayer": { + "id": "mixlayer", + "env": [ + "MIXLAYER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://models.mixlayer.ai/v1", + "name": "Mixlayer", + "doc": "https://docs.mixlayer.com", + "models": { + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -71490,43 +67895,33 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 943718 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.028 + "input": 0.3, + "output": 2.4 } }, - "deepseek/deepseek-v4.1-flash": { - "id": "deepseek/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the cost-efficient tier of the V4.1 family. DeepSeek reports that it exceeds V4 Pro on performance, speed, and task...", - "family": "deepseek-flash", - "attachment": true, + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5 35B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -71534,42 +67929,33 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 943718 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.006 + "input": 0.25, + "output": 1.3 } }, - "deepseek/deepseek-v4-flash-vision-exp": { - "id": "deepseek/deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "DeepSeek V4 Flash Vision Exp is an experimental vision-enabled version of [DeepSeek V4 Flash 0731](https://openrouter.ai/deepseek/deepseek-v4-flash-0731) from DeepSeek, adding image understanding while matching the base model on text capabilities including agents,...", - "family": "deepseek-flash", - "attachment": true, + "qwen/qwen3.5-9b": { + "id": "qwen/qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -71577,36 +67963,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 943718 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.028 + "input": 0.1, + "output": 0.4 } }, - "deepseek/deepseek-r1-distill-llama-70b": { - "id": "deepseek/deepseek-r1-distill-llama-70b", - "name": "DeepSeek: R1 Distill Llama 70B (retires Sep 28)", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek", + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2025-01-23", - "last_updated": "2025-01-23", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -71615,69 +67995,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 7372 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.8, - "output": 0.8 + "input": 0.6, + "output": 3.6 } }, - "deepseek/deepseek-v3.1-terminus": { - "id": "deepseek/deepseek-v3.1-terminus", - "name": "DeepSeek: DeepSeek V3.1 Terminus (retires Sep 28)", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B A10B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-22", - "last_updated": "2025-09-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.27, - "output": 1, - "cache_read": 0.135 - } - }, - "deepseek/deepseek-chat": { - "id": "deepseek/deepseek-chat", - "name": "DeepSeek Chat", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -71688,68 +68031,101 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.32, - "output": 0.89 + "input": 0.4, + "output": 3.2 } - }, - "deepseek/deepseek-v4-pro-0813": { - "id": "deepseek/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.", - "family": "deepseek-thinking", - "attachment": false, + } + } + }, + "orcarouter": { + "id": "orcarouter", + "env": [ + "ORCAROUTER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.orcarouter.ai/v1", + "name": "OrcaRouter", + "doc": "https://docs.orcarouter.ai", + "models": { + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1024000, - "output": 384000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 + "input": 2.5, + "output": 15, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "deepseek/deepseek-v3.2-exp": { - "id": "deepseek/deepseek-v3.2-exp", - "name": "DeepSeek: DeepSeek V3.2 Exp (retires Sep 28)", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } @@ -71757,11 +68133,16 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -71769,136 +68150,160 @@ }, "open_weights": false, "limit": { - "context": 163840, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.27, - "output": 0.41 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "deepseek/deepseek-chat-v3-0324": { - "id": "deepseek/deepseek-chat-v3-0324", - "name": "DeepSeek: DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163840, - "output": 147456 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.13, + "output": 0.38 } }, - "deepseek/deepseek-r1": { - "id": "deepseek/deepseek-r1", - "name": "DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", + "medium", "high" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 64000, - "output": 16000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.7, - "output": 2.5 + "input": 4, + "output": 18, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", - "attachment": false, + "google/gemini-flash-lite-latest": { + "id": "google/gemini-flash-lite-latest", + "name": "Gemini Flash-Lite Latest", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-01", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1024000, - "output": 384000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.6, - "output": 3.2, - "cache_read": 0.135 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025 } }, - "deepseek/deepseek-r1-0528": { - "id": "deepseek/deepseek-r1-0528", - "name": "DeepSeek: R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek", - "attachment": false, + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", + "medium", "high" ] } @@ -71906,11 +68311,16 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -71918,27 +68328,29 @@ }, "open_weights": false, "limit": { - "context": 163840, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.7, - "output": 2.5, - "cache_read": 0.35 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "input_audio": 0.3 } }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } @@ -71946,54 +68358,65 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.269, - "output": 0.4, - "cache_read": 0.1345 + "input": 4, + "output": 18, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", - "attachment": false, + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -72001,30 +68424,44 @@ }, "open_weights": true, "limit": { - "context": 1024000, - "output": 384000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0.06, + "output": 0.33 } }, - "unbiased/pareto": { - "id": "unbiased/pareto", - "name": "Pareto", - "description": "Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer", + "google/gemini-3-pro-preview": { + "id": "google/gemini-3-pro-preview", + "name": "Gemini 3 Pro Preview", + "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", + "family": "gemini-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -72032,80 +68469,146 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.25 + "input": 4, + "output": 18, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "tencent/hy3-preview": { - "id": "tencent/hy3-preview", - "name": "Hy3 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", + "medium", "high" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ + "text", + "image", + "video", + "audio", + "pdf" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 + } + }, + "google/gemini-flash-latest": { + "id": "google/gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", + "modalities": { + "input": [ + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.18, - "output": 0.6, - "cache_read": 0.06 + "input": 0.5, + "output": 3, + "cache_read": 0.075, + "input_audio": 1 } }, - "tencent/hunyuan-a13b-instruct": { - "id": "tencent/hunyuan-a13b-instruct", - "name": "Tencent: Hunyuan A13B Instruct", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", - "attachment": false, + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-08", - "last_updated": "2025-07-08", + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -72113,26 +68616,32 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "tencent/hy-mt2-30b-a3b": { - "id": "tencent/hy-mt2-30b-a3b", - "name": "Tencent: Hy-MT2-30B-A3B", - "description": "Hy-MT2-30B-A3B is Tencent's flagship translation model in the Hy-MT2 family. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and...", - "family": "Hy", + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-20", - "last_updated": "2026-08-20", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -72141,28 +68650,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.074, - "output": 0.295 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "tencent/hy-mt2-1.8b": { - "id": "tencent/hy-mt2-1.8b", - "name": "Tencent: Hy-MT2-1.8B", - "description": "Hy-MT2-1.8B is a compact 1.8B-parameter translation model from Tencent. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and style-guided...", - "family": "Hy", + "z-ai/glm-4.5": { + "id": "z-ai/glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2026-08-20", - "last_updated": "2026-08-20", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -72171,38 +68683,34 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 4096 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.044, - "output": 0.177 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "tencent/hy3": { - "id": "tencent/hy3", - "name": "Hy3", - "description": "Hy3 is a 295B-parameter Mixture-of-Experts model from Tencent (21B active, 192 experts with top-8 routing) built for reasoning, agentic workflows, and real-world production use. It supports a configurable reasoning effort:...", - "family": "Hy", + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -72213,28 +68721,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 192000, - "output": 128000 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.0825, - "output": 0.33, - "cache_read": 0.020625 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "tencent/hy-mt2-7b": { - "id": "tencent/hy-mt2-7b", - "name": "Tencent: Hy-MT2-7B", - "description": "Hy-MT2-7B is a 7B-parameter translation model from Tencent. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and style-guided translation.", - "family": "Hy", + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2026-08-19", - "last_updated": "2026-08-19", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -72243,38 +68752,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.074, - "output": 0.295 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "tencent/hy4-preview": { - "id": "tencent/hy4-preview", - "name": "Hy4 preview", - "description": "Tencent: Hy4 preview is a mixture-of-experts model from Tencent, with 49B active parameters out of 770B total. It is designed for coding agents, complex tool-use workflows, and productivity tasks that...", - "family": "Hy", + "z-ai/glm-4.5-air": { + "id": "z-ai/glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -72285,26 +68787,31 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 64000 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.834, - "output": 2.501, - "cache_read": 0.042 + "input": 0.2, + "output": 1.1, + "cache_read": 0.03, + "cache_write": 0 } }, - "anthracite-org/magnum-v4-72b": { - "id": "anthracite-org/magnum-v4-72b", - "name": "Magnum v4 72B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -72313,28 +68820,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 5 + "input": 1, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0 } }, - "meta-llama/llama-4-scout": { - "id": "meta-llama/llama-4-scout", - "name": "Meta: Llama 4 Scout", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2 Pro", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -72346,30 +68866,33 @@ }, "open_weights": false, "limit": { - "context": 327680, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 21, + "output": 168 } }, - "meta-llama/llama-guard-4-12b": { - "id": "meta-llama/llama-guard-4-12b", - "name": "Meta: Llama Guard 4 12B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-04-30", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -72377,30 +68900,32 @@ }, "open_weights": false, "limit": { - "context": 163840, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.18, - "output": 0.18 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "meta-llama/llama-4-maverick": { - "id": "meta-llama/llama-4-maverick", - "name": "Meta: Llama 4 Maverick", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": true, + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -72408,88 +68933,105 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 16385, + "output": 4096 }, "cost": { - "input": 0.1875, - "output": 0.6525 + "input": 0.5, + "output": 1.5, + "cache_read": 0 } }, - "meta-llama/llama-3.3-70b-instruct": { - "id": "meta-llama/llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 272000 }, "cost": { - "input": 0.1, - "output": 0.32 + "input": 15, + "output": 120 } }, - "meta-llama/llama-3.1-8b-instruct": { - "id": "meta-llama/llama-3.1-8b-instruct", - "name": "Llama-3.1-8B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.02, - "output": 0.04 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "meta-llama/llama-3.2-1b-instruct": { - "id": "meta-llama/llama-3.2-1b-instruct", - "name": "Meta: Llama 3.2 1B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "openai/gpt-4": { + "id": "openai/gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ "text" @@ -72500,29 +69042,31 @@ }, "open_weights": false, "limit": { - "context": 60000, - "output": 54000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.027, - "output": 0.201 + "input": 30, + "output": 60 } }, - "meta-llama/llama-3.2-3b-instruct": { - "id": "meta-llama/llama-3.2-3b-instruct", - "name": "Meta: Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "openai/gpt-4o-2024-05-13": { + "id": "openai/gpt-4o-2024-05-13", + "name": "GPT-4o (2024-05-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -72530,74 +69074,67 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.05, - "output": 0.33 + "input": 5, + "output": 15 } }, - "meta-llama/llama-3.1-70b-instruct": { - "id": "meta-llama/llama-3.1-70b-instruct", - "name": "Llama-3.1-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 0.4 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "~google/gemini-flash-latest": { - "id": "~google/gemini-flash-latest", - "name": "Google: Gemini Flash Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "openai/gpt-5-chat-latest": { + "id": "openai/gpt-5-chat-latest", + "name": "GPT-5 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "video", - "pdf", - "audio" + "image" ], "output": [ "text" @@ -72605,46 +69142,34 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "~google/gemini-pro-latest": { - "id": "~google/gemini-pro-latest", - "name": "Google: Gemini Pro Latest", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "audio", - "pdf", - "image", "text", - "video" + "image" ], "output": [ "text" @@ -72652,42 +69177,34 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "perceptron/perceptron-mk1": { - "id": "perceptron/perceptron-mk1", - "name": "Perceptron: Perceptron Mk1", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": false, + "reasoning_options": [], + "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -72695,124 +69212,100 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "z-ai/glm-4.6v": { - "id": "z-ai/glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "openai/gpt-5.3-chat-latest": { + "id": "openai/gpt-5.3-chat-latest", + "name": "GPT-5.3 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "knowledge": "2025-08-31", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.055 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "z-ai/glm-4.5": { - "id": "z-ai/glm-4.5", - "name": "GLM-4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "openai/gpt-4o-2024-08-06": { + "id": "openai/gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2023-09", + "release_date": "2024-08-06", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "z-ai/glm-5v-turbo": { - "id": "z-ai/glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "image", "text", - "video" + "image" ], "output": [ "text" @@ -72820,161 +69313,135 @@ }, "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "z-ai/glm-5.3-flash": { - "id": "z-ai/glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks. Its hybrid sparse and linear attention architecture maintains accurate long-context behavior while...", - "family": "glm-flash", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 198000, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.43, - "output": 1.75, - "cache_read": 0.08 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "openai/gpt-5.1-chat-latest": { + "id": "openai/gpt-5.1-chat-latest", + "name": "GPT-5.1 Chat", + "description": "Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations", + "family": "gpt-codex", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 198000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 1.92, - "cache_read": 0.12 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "z-ai/glm-4.5v": { - "id": "z-ai/glm-4.5v", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "openai/gpt-5.2-chat-latest": { + "id": "openai/gpt-5.2-chat-latest", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -72984,44 +69451,34 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, + "context": 128000, "output": 16384 }, "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "z-ai/glm-5.3-flashx": { - "id": "z-ai/glm-5.3-flashx", - "name": "Z.ai: GLM 5.3 FlashX", - "description": "GLM-5.3-FlashX is the high-speed variant of Z.ai's GLM-5.3-Flash, a native multimodal model delivering inference speeds of up to 200 tokens/s. Built on the same hybrid sparse and linear attention architecture...", - "family": "glm", + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-09-18", - "last_updated": "2026-09-18", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -73029,201 +69486,254 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.37, - "output": 1.25, - "cache_read": 0.075 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "z-ai/glm-4.7-flash": { - "id": "z-ai/glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "openai/gpt-4o-2024-11-20": { + "id": "openai/gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2023-09", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.0605, - "output": 0.4 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.4, - "output": 1.75, - "cache_read": 0.08 + "input": 5, + "output": 22.5, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "GLM 5.2 is a large-scale reasoning model from Z.ai. It supports text input and output with a 1M-token context window, and is suited for long-horizon agent workflows, project-level software engineering,...", - "family": "glm", - "attachment": false, + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "xhigh" - ] + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } } - ], + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -73231,151 +69741,215 @@ }, "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "z-ai/glm-5.2:free": { - "id": "z-ai/glm-5.2:free", - "name": "Z.ai: GLM 5.2 (free)", - "description": "GLM 5.2 is a large-scale reasoning model from Z.ai. It supports text input and output with a 1M-token context window, and is suited for long-horizon agent workflows, project-level software engineering,...", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "xhigh" - ] - } - ], - "tool_call": false, + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 10, + "output": 30 + } + }, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 29491 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "z-ai/glm-4.5-air": { - "id": "z-ai/glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", - "attachment": false, + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], "tool_call": true, "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.13, - "output": 0.85, - "cache_read": 0.025 + "input": 60, + "output": 270, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "z-ai/glm-5.3": { - "id": "z-ai/glm-5.3", - "name": "GLM-5.3", - "description": "GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks. It supports text input and output with a 1M-token context window, and improves...", - "family": "glm", - "attachment": false, + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "inference-net/schematron-v2-small": { - "id": "inference-net/schematron-v2-small", - "name": "Inference.net: Schematron V2 Small", - "description": "Schematron V2 Small is a 3B-parameter HTML-to-JSON extraction model from Inference.net. It prioritizes extraction quality for complex schemas and long pages. Extraction instructions must be supplied through a JSON schema...", - "attachment": false, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-12", - "last_updated": "2026-09-12", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -73384,28 +69958,32 @@ "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.23, - "cache_read": 0.05 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "inference-net/schematron-v2-turbo": { - "id": "inference-net/schematron-v2-turbo", - "name": "Inference.net: Schematron V2 Turbo", - "description": "Schematron V2 Turbo is a 3B-parameter HTML-to-JSON extraction model from Inference.net. It prioritizes throughput for high-volume extraction workloads. Extraction instructions must be supplied through a JSON schema in response_format rather...", + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-09-12", - "last_updated": "2026-09-12", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -73413,232 +69991,243 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.03, - "output": 0.15, - "cache_read": 0.03 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "thinkingmachines/inkling-small:free": { - "id": "thinkingmachines/inkling-small:free", - "name": "Thinking Machines: Inkling Small (free)", - "description": "Inkling Small is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 12B active parameters out of 276B total. It is positioned as the smaller, more efficient member of...", - "family": "ling", + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "thinkingmachines/inkling-small": { - "id": "thinkingmachines/inkling-small", - "name": "Inkling Small", - "description": "Inkling Small is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 12B active parameters out of 276B total. It is positioned as the smaller, more efficient member of...", - "family": "ling", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.45, - "output": 1.2, - "cache_read": 0.1 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "thinkingmachines/inkling": { - "id": "thinkingmachines/inkling", - "name": "Inkling", - "description": "Inkling is an open-weight multimodal mixture-of-experts model from Thinking Machines Lab, with 41B active parameters out of 975B total. It is designed for general-purpose reasoning, coding, agentic and tool-use systems,...", - "family": "ling", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 471859 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.95, - "output": 4.05, - "cache_read": 0.16 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "meituan/longcat-2.0": { - "id": "meituan/longcat-2.0", - "name": "Meituan: LongCat 2.0", - "description": "LongCat 2.0 is a sparse mixture-of-experts language model from Meituan, with 48B active parameters out of 1.6T total. It is suited for coding, repository-level changes, long-horizon problem solving, and agentic...", - "family": "longcat", + "kimi/kimi-k2.5": { + "id": "kimi/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-20", - "last_updated": "2026-07-20", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048756, + "context": 262144, "output": 262144 }, "cost": { - "input": 0.75, + "input": 0.6, "output": 3, - "cache_read": 0.015 + "cache_read": 0.1 } }, - "openrouter/bodybuilder": { - "id": "openrouter/bodybuilder", - "name": "Body Builder (beta)", - "description": "Preview model for early access evaluation, prototyping, and compatibility testing", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-03-15", - "last_updated": "2026-03-15", + "kimi/kimi-k2.6": { + "id": "kimi/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 262144, + "output": 262144 }, - "status": "beta", "cost": { - "input": 0, - "output": 0 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "openrouter/free": { - "id": "openrouter/free", - "name": "OpenRouter Free Models Router", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "orcarouter/auto": { + "id": "orcarouter/auto", + "name": "OrcaRouter Auto", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "release_date": "2025-01-01", + "last_updated": "2026-05-14", "modalities": { "input": [ "text", @@ -73650,27 +70239,32 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { "input": 0, "output": 0 } }, - "openrouter/pareto-code": { - "id": "openrouter/pareto-code", - "name": "Pareto Code Router", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-05-01", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -73679,63 +70273,68 @@ "open_weights": false, "limit": { "context": 200000, - "output": 65536 + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openrouter/auto": { - "id": "openrouter/auto", - "name": "Auto Router", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude Sonnet 4 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-03-15", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "audio", - "image", - "pdf", "text", - "video" + "image", + "pdf" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { - "context": 2000000, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "perplexity/sonar-pro": { - "id": "perplexity/sonar-pro", - "name": "Perplexity: Sonar Pro", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-03-07", - "last_updated": "2025-03-07", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -73744,37 +70343,42 @@ "open_weights": false, "limit": { "context": 200000, - "output": 8000 + "output": 64000 }, "cost": { - "input": 3, - "output": 15 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "perplexity/sonar-deep-research": { - "id": "perplexity/sonar-deep-research", - "name": "Perplexity: Sonar Deep Research", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar-deep-research", - "attachment": false, + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-03-07", - "last_updated": "2025-03-07", + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -73782,31 +70386,54 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 115200 + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 2, - "output": 8, - "reasoning": 3 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "perplexity/sonar": { - "id": "perplexity/sonar", - "name": "Perplexity: Sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar", + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-01-27", - "last_updated": "2025-01-27", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -73814,39 +70441,34 @@ }, "open_weights": false, "limit": { - "context": 127072, - "output": 114364 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1, - "output": 1 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "perplexity/sonar-reasoning-pro": { - "id": "perplexity/sonar-reasoning-pro", - "name": "Perplexity: Sonar Reasoning Pro", - "description": "Web-grounded reasoning model for multi-step research and cited answers", - "family": "sonar-reasoning", + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-03-07", - "last_updated": "2025-03-07", + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -73854,38 +70476,43 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 115200 + "context": 200000, + "output": 64000 }, "cost": { - "input": 2, - "output": 8 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "perplexity/sonar-pro-search": { - "id": "perplexity/sonar-pro-search", - "name": "Perplexity: Sonar Pro Search", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", + "medium", "high" ] } ], - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "release_date": "2025-10-30", - "last_updated": "2025-10-30", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -73893,46 +70520,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8000 + "context": 1000000, + "output": 64000 }, "cost": { "input": 3, - "output": 15 + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "meta/muse-spark-1.3": { - "id": "meta/muse-spark-1.3", - "name": "Muse Spark 1.3", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It is designed to keep track of information across extended tasks, work through...", - "family": "muse", + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Claude Opus 4 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", "image", - "video", - "pdf", - "audio" + "pdf" ], "output": [ "text" @@ -73940,20 +70555,21 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "meta/muse-glimmer-30b": { - "id": "meta/muse-glimmer-30b", - "name": "Muse Glimmer 30B", - "description": "Muse Glimmer 30B is a dense, open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark and optimized for autonomous agents on consumer hardware. It is suited for long-horizon...", - "family": "muse", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -73962,17 +70578,15 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -73983,48 +70597,56 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 0.3, - "output": 1.1, - "cache_read": 0.04 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "meta/muse-spark-1.1": { - "id": "meta/muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Muse Spark 1.1 is a multimodal reasoning model from Meta, built for agentic tasks. It accepts text, images, video, audio, and PDF documents and returns text, with a 1M-token context...", - "family": "muse", + "grok/grok-4.3": { + "id": "grok/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", - "video", - "pdf", - "audio" + "pdf" ], "output": [ "text" @@ -74032,93 +70654,80 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 1000000, + "output": 30000 }, "cost": { "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "meta/muse-spark-1.2": { - "id": "meta/muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a reasoning model from Meta, designed for complex agentic tasks. It accepts text, images, video, audio, and PDF documents, returns text, and offers a 1M-token context...", - "family": "muse", + "qwen/qwen3.6-35b-a3b": { + "id": "qwen/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", "video", - "pdf", "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 943718 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 0.248, + "output": 1.485 } }, - "meta/muse-spark-1.2-contributor": { - "id": "meta/muse-spark-1.2-contributor", - "name": "Meta: Muse Spark 1.2 Contributor", - "description": "Muse Spark 1.2 contributor tier is a reasoning model from Meta designed for developers who want to start building at an even lower cost. It’s meaningfully cheaper than Muse Spark...", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-08-21", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text", - "image", - "video", - "pdf", - "audio" + "text" ], "output": [ "text" @@ -74126,47 +70735,32 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 + "input": 0.359, + "output": 1.434 } }, - "meta/muse-spark-1.3-contributor": { - "id": "meta/muse-spark-1.3-contributor", - "name": "Meta: Muse Spark 1.3 Contributor", - "description": "Muse Spark 1.3 Contributor is the cost-efficient contributor tier of Meta’s multimodal reasoning model for experimentation, learning, and early-stage agentic, multi-agent, and coding workflows. It is designed to track information...", - "family": "muse", - "attachment": true, + "qwen/qwen3.5-plus": { + "id": "qwen/qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", "image", - "video", - "pdf", - "audio" + "video" ], "output": [ "text" @@ -74174,182 +70768,169 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 + "input": 0.115, + "output": 0.688, + "reasoning": 2.4 } }, - "nousresearch/hermes-3-llama-3.1-70b": { - "id": "nousresearch/hermes-3-llama-3.1-70b", - "name": "Nous: Hermes 3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "nousresearch", - "attachment": false, - "reasoning": false, - "tool_call": false, + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-08-18", - "last_updated": "2024-08-18", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.7, - "output": 0.7 + "input": 0.086, + "output": 0.688 } }, - "nousresearch/hermes-3-llama-3.1-405b": { - "id": "nousresearch/hermes-3-llama-3.1-405b", - "name": "Nous: Hermes 3 405B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "nousresearch", - "attachment": false, - "reasoning": false, - "tool_call": false, + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-08-16", - "last_updated": "2024-08-16", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1, - "output": 1 + "input": 0.057, + "output": 0.459 } }, - "nousresearch/hermes-4-405b": { - "id": "nousresearch/hermes-4-405b", - "name": "Nous: Hermes 4 405B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "nousresearch", - "attachment": false, + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 117964 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1, - "output": 3 + "input": 0.172, + "output": 1.032 } }, - "~z-ai/glm-flash-latest": { - "id": "~z-ai/glm-flash-latest", - "name": "Z.ai: GLM Flash Latest", - "description": "This model always redirects to the latest model in the GLM Flash family.", - "family": "glm-flash", + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.075, - "output": 0.25, - "cache_read": 0.015 + "input": 0.115, + "output": 0.917 } }, - "~z-ai/glm-latest": { - "id": "~z-ai/glm-latest", - "name": "Z.ai: GLM Latest", - "description": "This model always redirects to the latest GLM model from Z.ai.", - "family": "glm", + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-19", - "last_updated": "2026-08-19", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -74357,35 +70938,51 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.5614, - "output": 1.7644, - "cache_read": 0.10426 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "arcee-ai/trinity-large-thinking": { - "id": "arcee-ai/trinity-large-thinking", - "name": "Trinity Large Thinking", - "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", - "family": "trinity", + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-05-28", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -74396,259 +70993,193 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 80000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.25, - "output": 0.8, - "cache_read": 0.06 + "input": 0.19, + "output": 0.37, + "cache_read": 0.0028 } }, - "~openai/gpt-terra-latest": { - "id": "~openai/gpt-terra-latest", - "name": "OpenAI: GPT Terra Latest", - "description": "This model always redirects to the latest model in the OpenAI GPT Terra family.", - "family": "gpt-terra", - "attachment": true, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "pdf", - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.56, + "output": 1.12, + "cache_read": 0.003625 } }, - "~openai/gpt-luna-latest": { - "id": "~openai/gpt-luna-latest", - "name": "OpenAI: GPT Luna Latest", - "description": "This model always redirects to the latest model in the OpenAI GPT Luna family.", - "family": "gpt-luna", + "deepseek/deepseek-reasoner": { + "id": "deepseek/deepseek-reasoner", + "name": "DeepSeek Reasoner", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ - "pdf", - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125 + "input": 0.435, + "output": 0.87, + "cache_read": 0.028 } }, - "~openai/gpt-sol-latest": { - "id": "~openai/gpt-sol-latest", - "name": "OpenAI: GPT Sol Latest", - "description": "This model always redirects to the latest model in the OpenAI GPT Sol family.", - "family": "gpt-sol", + "deepseek/deepseek-chat": { + "id": "deepseek/deepseek-chat", + "name": "DeepSeek Chat", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ - "pdf", - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "~openai/gpt-astra-latest": { - "id": "~openai/gpt-astra-latest", - "name": "OpenAI: GPT Astra Latest ($$$$)", - "description": "This model always redirects to the latest model in the OpenAI GPT Astra family.", - "family": "gpt-astra", - "attachment": true, + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "pdf", - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "~openai/gpt-mini-latest": { - "id": "~openai/gpt-mini-latest", - "name": "OpenAI: GPT Mini Latest", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "pdf", - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "cognitivecomputations/dolphin-mistral-24b-venice-edition": { - "id": "cognitivecomputations/dolphin-mistral-24b-venice-edition", - "name": "Venice: Uncensored", - "description": "Venice Uncensored Dolphin Mistral 24B Venice Edition is a fine-tuned variant of Mistral-Small-24B-Instruct-2501, developed by dphn.ai in collaboration with Venice.ai. This model is designed as an “uncensored” instruct-tuned LLM, preserving...", - "family": "mistral", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -74657,27 +71188,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.9 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "thedrummer/skyfall-36b-v2": { - "id": "thedrummer/skyfall-36b-v2", - "name": "TheDrummer: Skyfall 36B V2", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "minimax/minimax-m2.5-highspeed": { + "id": "minimax/minimax-m2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-03-10", - "last_updated": "2025-03-10", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -74686,31 +71220,46 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 29491 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.55, - "output": 0.8, - "cache_read": 0.25 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } - }, - "thedrummer/unslopnemo-12b": { - "id": "thedrummer/unslopnemo-12b", - "name": "TheDrummer: UnslopNemo 12B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + } + } + }, + "helicone": { + "id": "helicone", + "env": [ + "HELICONE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://ai-gateway.helicone.ai/v1", + "name": "Helicone", + "doc": "https://helicone.ai/models", + "models": { + "chatgpt-4o-latest": { + "id": "chatgpt-4o-latest", + "name": "OpenAI ChatGPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "release_date": "2024-11-08", - "last_updated": "2024-11-08", + "knowledge": "2024-08", + "release_date": "2024-08-14", + "last_updated": "2024-08-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -74718,28 +71267,31 @@ }, "open_weights": false, "limit": { - "context": 1024000, - "output": 819200 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.4, - "output": 0.4 + "input": 5, + "output": 20, + "cache_read": 2.5 } }, - "thedrummer/cydonia-24b-v4.1": { - "id": "thedrummer/cydonia-24b-v4.1", - "name": "TheDrummer: Cydonia 24B V4.1", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "gpt-4.1-mini-2025-04-14": { + "id": "gpt-4.1-mini-2025-04-14", + "name": "OpenAI GPT-4.1 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "release_date": "2025-09-27", - "last_updated": "2025-09-27", + "knowledge": "2025-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -74747,39 +71299,30 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 0.5, - "cache_read": 0.15 + "input": 0.39999999999999997, + "output": 1.5999999999999999, + "cache_read": 0.09999999999999999 } }, - "baidu/ernie-4.5-vl-424b-a47b": { - "id": "baidu/ernie-4.5-vl-424b-a47b", - "name": "Baidu: ERNIE 4.5 VL 424B A47B (retires Oct 8)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ernie", - "attachment": true, + "deepseek-v3.1-terminus": { + "id": "deepseek-v3.1-terminus", + "name": "DeepSeek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "knowledge": "2025-09", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -74788,42 +71331,31 @@ }, "open_weights": false, "limit": { - "context": 123000, - "output": 16000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.42, - "output": 1.25 + "input": 0.27, + "output": 1, + "cache_read": 0.21600000000000003 } }, - "x-ai/grok-4.20-multi-agent": { - "id": "x-ai/grok-4.20-multi-agent", - "name": "SpaceXAI: Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": false, - "structured_output": true, + "claude-3.5-haiku": { + "id": "claude-3.5-haiku", + "name": "Anthropic: Claude 3.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "knowledge": "2024-10", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -74831,44 +71363,31 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 1800000 + "context": 200000, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.7999999999999999, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 } }, - "x-ai/grok-4.7": { - "id": "x-ai/grok-4.7", - "name": "Grok 4.7", - "description": "Grok 4.7 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "llama-3.1-8b-instruct": { + "id": "llama-3.1-8b-instruct", + "name": "Meta Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "knowledge": "2024-07", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -74876,43 +71395,30 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 450000 + "context": 16384, + "output": 16384 }, "cost": { - "input": 1.6, - "output": 4.8, - "cache_read": 0.4 + "input": 0.02, + "output": 0.049999999999999996 } }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "Grok 4.3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "o3": { + "id": "o3", + "name": "OpenAI o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": false, + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -74920,42 +71426,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 900000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "x-ai/grok-4.5": { - "id": "x-ai/grok-4.5", - "name": "Grok 4.5", - "description": "Grok 4.5 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "llama-prompt-guard-2-86m": { + "id": "llama-prompt-guard-2-86m", + "name": "Meta Llama Prompt Guard 2 86M", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -74963,40 +71457,29 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 450000 + "context": 512, + "output": 2 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3 + "input": 0.01, + "output": 0.01 } }, - "x-ai/grok-build-0.1": { - "id": "x-ai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Grok coding model for agentic engineering, edits, and codebase workflows", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3 Coder 30B A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2025-07", + "release_date": "2025-07-31", + "last_updated": "2025-07-31", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75004,41 +71487,29 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 230400 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 + "input": 0.09999999999999999, + "output": 0.3 } }, - "x-ai/grok-4.20": { - "id": "x-ai/grok-4.20", - "name": "SpaceXAI: Grok 4.20", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "hermes-2-pro-llama-3-8b": { + "id": "hermes-2-pro-llama-3-8b", + "name": "Hermes 2 Pro Llama 3 8B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "knowledge": "2024-05", + "release_date": "2024-05-27", + "last_updated": "2024-05-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75046,44 +71517,29 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 1800000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.14, + "output": 0.14 } }, - "x-ai/grok-4.6": { - "id": "x-ai/grok-4.6", - "name": "Grok 4.6", - "description": "Grok 4.6 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "deepseek-v3": { + "id": "deepseek-v3", + "name": "DeepSeek V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2024-12", + "release_date": "2024-12-26", + "last_updated": "2024-12-26", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75091,44 +71547,30 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 450000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 0.56, + "output": 1.68, + "cache_read": 0.07 } }, - "~anthropic/claude-opus-latest": { - "id": "~anthropic/claude-opus-latest", - "name": "Anthropic: Claude Opus Latest", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "grok-code-fast-1": { + "id": "grok-code-fast-1", + "name": "xAI Grok Code Fast 1", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2024-08", + "release_date": "2024-08-25", + "last_updated": "2024-08-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75136,42 +71578,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "output": 10000 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 + "input": 0.19999999999999998, + "output": 1.5, + "cache_read": 0.02 } }, - "~anthropic/claude-haiku-latest": { - "id": "~anthropic/claude-haiku-latest", - "name": "Anthropic: Claude Haiku Latest", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "o1-mini": { + "id": "o1-mini", + "name": "OpenAI: o1-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75179,47 +71609,31 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 65536 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "~anthropic/claude-sonnet-latest": { - "id": "~anthropic/claude-sonnet-latest", - "name": "Anthropic: Claude Sonnet Latest", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "deepseek-r1-distill-llama-70b": { + "id": "deepseek-r1-distill-llama-70b", + "name": "DeepSeek R1 Distill Llama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75227,45 +71641,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.03, + "output": 0.13 } }, - "~anthropic/claude-fable-latest": { - "id": "~anthropic/claude-fable-latest", - "name": "Anthropic: Claude Fable Latest ($$$$)", - "description": "This model always redirects to the latest model in the Claude Fable family.", - "family": "claude-fable", - "attachment": true, + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -75273,43 +71672,40 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 40960 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 0.29, + "output": 0.59 } }, - "upstage/solar-pro-3": { - "id": "upstage/solar-pro-3", - "name": "Upstage: Solar Pro 3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "solar-pro", + "claude-sonnet-4": { + "id": "claude-sonnet-4", + "name": "Anthropic: Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "knowledge": "2025-05", + "release_date": "2025-05-14", + "last_updated": "2025-05-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -75317,44 +71713,33 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 3, + "output": 15, + "cache_read": 0.30000000000000004, + "cache_write": 3.75 } }, - "upstage/solar-pro4": { - "id": "upstage/solar-pro4", - "name": "Upstage: Solar Pro 4", - "description": "Solar Pro 4 is a large language model from Upstage. It is suited for agentic workflows, office productivity, document-intensive work, and coding.", - "family": "solar", + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -75362,44 +71747,30 @@ }, "open_weights": false, "limit": { - "context": 524288, - "output": 131072 + "context": 262000, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.14, + "output": 1.4 } }, - "upstage/solar-mini4": { - "id": "upstage/solar-mini4", - "name": "Upstage: Solar Mini 4", - "description": "Solar Mini 4 is Upstage's compact, cost-efficient language model, a 35B-parameter mixture-of-experts with 3B active parameters and a 524K context window. It is built for agentic use cases where response...", - "family": "solar", + "llama-4-maverick": { + "id": "llama-4-maverick", + "name": "Meta Llama 4 Maverick 17B 128E", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -75407,38 +71778,26 @@ }, "open_weights": false, "limit": { - "context": 524288, - "output": 131072 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 + "input": 0.15, + "output": 0.6 } }, - "~deepseek/deepseek-flash-latest": { - "id": "~deepseek/deepseek-flash-latest", - "name": "DeepSeek: DeepSeek Flash Latest", - "description": "This model always redirects to the latest model in the DeepSeek Flash family.", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "mistral-nemo": { + "id": "mistral-nemo", + "name": "Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-09-14", - "last_updated": "2026-09-14", + "knowledge": "2024-07", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", @@ -75450,38 +71809,26 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 128000, + "output": 16400 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01 + "input": 20, + "output": 40 } }, - "~deepseek/deepseek-pro-latest": { - "id": "~deepseek/deepseek-pro-latest", - "name": "DeepSeek: DeepSeek Pro Latest", - "description": "This model always redirects to the latest model in the DeepSeek Pro family.", - "family": "deepseek", + "llama-3.3-70b-versatile": { + "id": "llama-3.3-70b-versatile", + "name": "Meta Llama 3.3 70B Versatile", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-14", - "last_updated": "2026-09-14", + "knowledge": "2024-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -75492,38 +71839,26 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 393216 + "context": 131072, + "output": 32678 }, "cost": { - "input": 0.39996, - "output": 1.19988, - "cache_read": 0.012726 + "input": 0.59, + "output": 0.7899999999999999 } }, - "~deepseek/deepseek-v4-flash-latest": { - "id": "~deepseek/deepseek-v4-flash-latest", - "name": "DeepSeek: DeepSeek V4 Flash Latest", - "description": "This model always redirects to the latest model in the DeepSeek V4 Flash family.", - "family": "deepseek", + "gemma2-9b-it": { + "id": "gemma2-9b-it", + "name": "Google Gemma 2", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-01", - "last_updated": "2026-08-01", + "knowledge": "2024-06", + "release_date": "2024-06-25", + "last_updated": "2024-06-25", "modalities": { "input": [ "text" @@ -75534,43 +71869,37 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.038, - "output": 0.55, - "cache_read": 0.0228 + "input": 0.01, + "output": 0.03 } }, - "~x-ai/grok-latest": { - "id": "~x-ai/grok-latest", - "name": "xAI: Grok Latest", - "description": "This model always redirects to the latest Grok model from xAI.", - "family": "grok", - "attachment": true, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Google Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "knowledge": "2025-06", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -75578,30 +71907,32 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 450000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.6, - "output": 4.8, - "cache_read": 0.4 + "input": 1.25, + "output": 10, + "cache_read": 0.3125, + "cache_write": 1.25 } }, - "bytedance/ui-tars-1.5-7b": { - "id": "bytedance/ui-tars-1.5-7b", - "name": "ByteDance: UI-TARS 7B ", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "gpt-5": { + "id": "gpt-5", + "name": "OpenAI GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "tool_call": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -75609,38 +71940,31 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 2048 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.1 + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 } }, - "liquid/lfm-2.5-2.6b:free": { - "id": "liquid/lfm-2.5-2.6b:free", - "name": "LiquidAI: LFM2.5-2.6B (free)", - "description": "LFM2.5-2.6B is a compact reasoning model from Liquid AI. It is suited for agent workflows, data extraction, RAG, and long-context processing. Liquid advises against using it for agentic coding or...", - "family": "liquid", + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Anthropic: Claude 4.5 Haiku (20251001)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "knowledge": "2025-10", + "release_date": "2025-10-01", + "last_updated": "2025-10-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -75648,145 +71972,134 @@ }, "open_weights": false, "limit": { - "context": 65536, + "context": 200000, "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 5, + "cache_read": 0.09999999999999999, + "cache_write": 1.25 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "claude-4.5-haiku": { + "id": "claude-4.5-haiku", + "name": "Anthropic: Claude 4.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-10", + "release_date": "2025-10-01", + "last_updated": "2025-10-01", "modalities": { "input": [ - "image", "text", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.09, - "output": 0.34, - "cache_read": 0.05 + "input": 1, + "output": 5, + "cache_read": 0.09999999999999999, + "cache_write": 1.25 } }, - "google/gemini-3.1-flash-image-preview": { - "id": "google/gemini-3.1-flash-image-preview", - "name": "Nano Banana 2 Preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "high" - ] - } - ], + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "OpenAI: GPT-5 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-01", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 58982 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 3 + "input": 15, + "output": 120 } }, - "google/lyria-3-pro-preview": { - "id": "google/lyria-3-pro-preview", - "name": "Lyria 3 Pro Preview", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "lyria", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Google Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], + "tool_call": true, "temperature": true, - "release_date": "2026-03-25", - "last_updated": "2026-03-25", + "knowledge": "2025-06", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { "context": 1048576, - "output": 65536 + "output": 65535 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 2.5, + "cache_read": 0.075, + "cache_write": 0.3 } }, - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma 3 27B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "gpt-4o": { + "id": "gpt-4o", + "name": "OpenAI GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "knowledge": "2024-05", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ "text", @@ -75796,64 +72109,61 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.08, - "output": 0.16 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "google/gemini-2.5-flash-image": { - "id": "google/gemini-2.5-flash-image", - "name": "Nano Banana", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, + "o4-mini": { + "id": "o4-mini", + "name": "OpenAI o4 Mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, + "tool_call": true, + "temperature": false, "knowledge": "2024-06", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "google/gemma-2-27b-it": { - "id": "google/gemma-2-27b-it", - "name": "Google: Gemma 2 27B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "llama-3.1-8b-instant": { + "id": "llama-3.1-8b-instant", + "name": "Meta Llama 3.1 8B Instant", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "release_date": "2024-07-13", - "last_updated": "2024-07-13", + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ "text" @@ -75864,45 +72174,29 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 2048 + "context": 131072, + "output": 32678 }, "cost": { - "input": 0.65, - "output": 0.65 + "input": 0.049999999999999996, + "output": 0.08 } }, - "google/gemini-3.6-flash": { - "id": "google/gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Gemini 3.6 Flash is a high-efficiency model from Google for coding, agentic workflows, and web and app development. It is designed to produce polished outputs with fewer unnecessary edits and...", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "llama-prompt-guard-2-22m": { + "id": "llama-prompt-guard-2-22m", + "name": "Meta Llama Prompt Guard 2 22M", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2024-10", + "release_date": "2024-10-01", + "last_updated": "2024-10-01", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -75910,134 +72204,102 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 512, + "output": 2 }, "cost": { - "input": 0.375, - "output": 1.875, - "reasoning": 1.875, - "cache_read": 0.0375, - "cache_write": 0.020833 + "input": 0.01, + "output": 0.01 } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "o3-pro": { + "id": "o3-pro", + "name": "OpenAI o3 Pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ - "image", "text", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.042, - "output": 0.22 + "input": 20, + "output": 80 } }, - "google/gemini-3-pro-image-preview": { - "id": "google/gemini-3-pro-image-preview", - "name": "Nano Banana Pro Preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, + "grok-3-mini": { + "id": "grok-3-mini", + "name": "xAI Grok 3 Mini", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ - "image", - "text", - "pdf" + "text" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 32768 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1, - "output": 6, - "reasoning": 6, - "cache_read": 0.1, - "cache_write": 0.1875 + "input": 0.3, + "output": 0.5, + "cache_read": 0.075 } }, - "google/gemini-3.5-flash-lite": { - "id": "google/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Gemini 3.5 Flash Lite is a high-efficiency model from Google with upgraded agentic capabilities. It is suited for subagents that execute focused tasks within complex, multi-agent workflows.", - "family": "gemini-flash-lite", - "attachment": true, + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "Anthropic: Claude Opus 4.1 (20250805)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2025-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -76045,123 +72307,102 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.15, - "output": 1.25, - "reasoning": 1.25, - "cache_read": 0.015, - "cache_write": 0.041667 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "google/gemini-3.1-flash-image": { - "id": "google/gemini-3.1-flash-image", - "name": "Nano Banana 2", - "description": "Gemini 3.1 Flash Image, a.k.a. \"Nano Banana 2,\" is Google’s latest state of the art image generation and editing model, delivering Pro-level visual quality at Flash speed. It combines advanced...", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, + "deepseek-tng-r1t2-chimera": { + "id": "deepseek-tng-r1t2-chimera", + "name": "DeepSeek TNG R1T2 Chimera", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2025-07", + "release_date": "2025-07-02", + "last_updated": "2025-07-02", "modalities": { "input": [ - "image", - "text", - "pdf" + "text" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 130000, + "output": 163840 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.3, + "output": 1.2 } }, - "google/lyria-3-clip-preview": { - "id": "google/lyria-3-clip-preview", - "name": "Lyria 3 Clip Preview", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "lyria", - "attachment": true, + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Meta Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2026-03-25", - "last_updated": "2026-03-25", + "knowledge": "2024-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 16400 }, "cost": { - "input": 0, - "output": 0 + "input": 0.13, + "output": 0.39 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "sonar-reasoning-pro": { + "id": "sonar-reasoning-pro", + "name": "Perplexity Sonar Reasoning Pro", + "description": "Web-grounded reasoning model for multi-step research and cited answers", + "family": "sonar-reasoning", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" ] } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -76169,80 +72410,61 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 127000, + "output": 4096 }, "cost": { - "input": 1, - "output": 6, - "reasoning": 6, - "cache_read": 0.1, - "cache_write": 0.1875 + "input": 2, + "output": 8 } }, - "google/gemma-3-12b-it": { - "id": "google/gemma-3-12b-it", - "name": "Gemma 3 12B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "grok-3": { + "id": "grok-3", + "name": "xAI Grok 3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 131072, - "output": 16384 + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.15 + "input": 3, + "output": 15, + "cache_read": 0.75 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "glm-4.6": { + "id": "glm-4.6", + "name": "Zai GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "knowledge": "2024-07", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -76250,45 +72472,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.75, - "output": 4.5, - "reasoning": 4.5, - "cache_read": 0.075, - "cache_write": 0.041667 + "input": 0.44999999999999996, + "output": 1.5 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2025-11", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -76296,43 +72502,30 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 10, - "reasoning": 10, - "cache_read": 0.125, - "cache_write": 0.375 + "input": 0.48, + "output": 2 } }, - "google/gemini-3-pro-image": { - "id": "google/gemini-3-pro-image", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro is Google’s most advanced image-generation and editing model, built on Gemini 3 Pro. It extends the original Nano Banana with significantly improved multimodal reasoning, real-world grounding, and...", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "xAI Grok 4.1 Fast Non-Reasoning", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "grok", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2025-11", + "release_date": "2025-11-17", + "last_updated": "2025-11-17", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text", @@ -76341,46 +72534,33 @@ }, "open_weights": false, "limit": { - "context": 65536, - "output": 32768 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "input": 0.19999999999999998, + "output": 0.5, + "cache_read": 0.049999999999999996 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "qwen3-coder": { + "id": "qwen3-coder", + "name": "Qwen3 Coder 480B A35B Instruct Turbo", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2025-07", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text", "image", "audio", - "video", - "pdf" + "video" ], "output": [ "text" @@ -76388,139 +72568,94 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333 + "input": 0.22, + "output": 0.95 } }, - "google/gemini-3.1-flash-lite-image": { - "id": "google/gemini-3.1-flash-lite-image", - "name": "Nano Banana 2 Lite", - "description": "Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) is Google's fastest, most cost-efficient Gemini image model, built for high-velocity developer pipelines and rapid-fire visual exploration. It delivers text-to-image generation...", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "gpt-5-chat-latest": { + "id": "gpt-5-chat-latest", + "name": "OpenAI GPT-5 Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2024-09", + "release_date": "2024-09-30", + "last_updated": "2024-09-30", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 58982 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 1.5 + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 } }, - "google/gemini-3.7-flash": { - "id": "google/gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "Gemini 3.7 Flash is a multimodal model from Google for fast agentic workflows, coding, and complex multi-step reasoning. It is designed for tasks that require responsive performance and reliable multi-step...", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "OpenAI: GPT-5.1 Codex", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-codex", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667 + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 } }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "o3-mini": { + "id": "o3-mini", + "name": "OpenAI o3 Mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "temperature": false, + "knowledge": "2023-10", + "release_date": "2023-10-01", + "last_updated": "2023-10-01", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -76528,49 +72663,30 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "mistral-large-2411": { + "id": "mistral-large-2411", + "name": "Mistral-Large", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2024-07", + "release_date": "2024-07-24", + "last_updated": "2024-07-24", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -76578,46 +72694,40 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.041667 + "input": 2, + "output": 6 } }, - "google/gemini-3.8-flash": { - "id": "google/gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Gemini 3.8 Flash is Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows.", - "family": "gemini-flash", - "attachment": true, + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Google Gemini 2.5 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 512, + "max": 24576 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2025-07", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -76626,29 +72736,27 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65536 + "output": 65535 }, "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667 + "input": 0.09999999999999999, + "output": 0.39999999999999997, + "cache_read": 0.024999999999999998, + "cache_write": 0.09999999999999999 } }, - "google/gemma-3-4b-it": { - "id": "google/gemma-3-4b-it", - "name": "Gemma 3 4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "llama-guard-4": { + "id": "llama-guard-4", + "name": "Meta Llama Guard 4 12B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", @@ -76658,45 +72766,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 131072, - "output": 16384 + "output": 1024 }, "cost": { - "input": 0.05, - "output": 0.1 + "input": 0.21, + "output": 0.21 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Anthropic: Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2025-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image", - "audio", - "video", - "pdf" + "image" ], "output": [ "text" @@ -76704,43 +72809,32 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.1, - "output": 0.4, - "reasoning": 0.4, - "cache_read": 0.01, - "cache_write": 0.083333 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "google/gemini-2.5-pro-preview": { - "id": "google/gemini-2.5-pro-preview", - "name": "Google: Gemini 2.5 Pro Preview 06-05", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "claude-3.7-sonnet": { + "id": "claude-3.7-sonnet", + "name": "Anthropic: Claude 3.7 Sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "knowledge": "2025-02", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ - "pdf", - "image", "text", - "audio" + "image" ], "output": [ "text" @@ -76748,132 +72842,98 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.25, - "output": 10, - "reasoning": 10, - "cache_read": 0.125, - "cache_write": 0.375 + "input": 3, + "output": 15, + "cache_read": 0.30000000000000004, + "cache_write": 3.75 } }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "OpenAI: GPT-5.1 Codex Mini", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-codex", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.083333 + "output": 2, + "cache_read": 0.024999999999999998 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "gpt-5.1-chat-latest": { + "id": "gpt-5.1-chat-latest", + "name": "OpenAI GPT-5.1 Chat", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-codex", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.125, - "output": 0.75, - "reasoning": 0.75, - "cache_read": 0.0125, - "cache_write": 0.041667 + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 } }, - "writer/palmyra-x5": { - "id": "writer/palmyra-x5", - "name": "Writer: Palmyra X5", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "palmyra", + "claude-3-haiku-20240307": { + "id": "claude-3-haiku-20240307", + "name": "Anthropic: Claude 3 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2026-01-21", - "last_updated": "2026-01-21", + "knowledge": "2024-03", + "release_date": "2024-03-07", + "last_updated": "2024-03-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -76881,29 +72941,33 @@ }, "open_weights": false, "limit": { - "context": 1040000, - "output": 8192 + "context": 200000, + "output": 4096 }, "cost": { - "input": 0.6, - "output": 6 + "input": 0.25, + "output": 1.25, + "cache_read": 0.03, + "cache_write": 0.3 } }, - "ibm-granite/granite-4.0-h-micro": { - "id": "ibm-granite/granite-4.0-h-micro", - "name": "IBM: Granite 4.0 Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "granite", + "grok-4-fast-reasoning": { + "id": "grok-4-fast-reasoning", + "name": "xAI: Grok 4 Fast Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-10-20", - "last_updated": "2025-10-20", + "knowledge": "2025-09", + "release_date": "2025-09-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -76911,39 +72975,31 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 117900 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.017, - "output": 0.112 + "input": 0.19999999999999998, + "output": 0.5, + "cache_read": 0.049999999999999996 } }, - "ibm-granite/granite-4.2-8b": { - "id": "ibm-granite/granite-4.2-8b", - "name": "IBM: Granite 4.2 8B", - "description": "Granite 4.2 8B is a dense reasoning model from IBM. It is suited for mathematics, code generation, multilingual dialogue, and agentic workflows that need multi-step reasoning. It supports full, low-effort,...", - "family": "granite", + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "OpenAI GPT-4.1 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-31", - "last_updated": "2026-08-31", + "knowledge": "2025-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -76951,28 +73007,37 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.06, - "output": 0.25, - "cache_read": 0.015 + "input": 0.09999999999999999, + "output": 0.39999999999999997, + "cache_read": 0.024999999999999998 } }, - "mistralai/mistral-nemo": { - "id": "mistralai/mistral-nemo", - "name": "Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "OpenAI GPT-OSS 120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ "text" @@ -76981,32 +73046,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 131072, - "output": 16384 + "output": 131072 }, "cost": { - "input": 0.019, - "output": 0.03 + "input": 0.04, + "output": 0.16 } }, - "mistralai/ministral-8b-2512": { - "id": "mistralai/ministral-8b-2512", - "name": "Mistral: Ministral 3 8B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, + "sonar": { + "id": "sonar", + "name": "Perplexity Sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "knowledge": "2025-01", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -77014,27 +73078,26 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 209715 + "context": 127000, + "output": 4096 }, "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.015 + "input": 1, + "output": 1 } }, - "mistralai/mistral-small-24b-instruct-2501": { - "id": "mistralai/mistral-small-24b-instruct-2501", - "name": "Mistral: Mistral Small 3", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "qwen2.5-coder-7b-fast": { + "id": "qwen2.5-coder-7b-fast", + "name": "Qwen2.5 Coder 7B fast", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, "temperature": true, - "release_date": "2025-01-30", - "last_updated": "2025-01-30", + "knowledge": "2024-09", + "release_date": "2024-09-15", + "last_updated": "2024-09-15", "modalities": { "input": [ "text" @@ -77045,30 +73108,29 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 32000, + "output": 8192 }, "cost": { - "input": 0.05, - "output": 0.08 + "input": 0.03, + "output": 0.09 } }, - "mistralai/mistral-saba": { - "id": "mistralai/mistral-saba", - "name": "Mistral: Saba", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": true, + "o1": { + "id": "o1", + "name": "OpenAI: o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-02-17", - "last_updated": "2025-02-17", + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -77076,41 +73138,31 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 26214 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.2, - "output": 0.6, - "cache_read": 0.02 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "mistralai/mistral-medium-3-5": { - "id": "mistralai/mistral-medium-3-5", - "name": "Mistral: Mistral Medium 3.5", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, + "ernie-4.5-21b-a3b-thinking": { + "id": "ernie-4.5-21b-a3b-thinking", + "name": "Baidu Ernie 4.5 21B A3B Thinking", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "ernie", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "knowledge": "2025-03", + "release_date": "2025-03-16", + "last_updated": "2025-03-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -77118,31 +73170,30 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 209715 + "context": 128000, + "output": 8000 }, "cost": { - "input": 1.5, - "output": 7.5 + "input": 0.07, + "output": 0.28 } }, - "mistralai/mistral-medium-3.1": { - "id": "mistralai/mistral-medium-3.1", - "name": "Mistral: Mistral Medium 3.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, + "llama-4-scout": { + "id": "llama-4-scout", + "name": "Meta Llama 4 Scout 17B 16E", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-13", - "last_updated": "2025-08-13", + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -77151,29 +73202,27 @@ "open_weights": false, "limit": { "context": 131072, - "output": 104857 + "output": 8192 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.04 + "input": 0.08, + "output": 0.3 } }, - "mistralai/mistral-small-3.2-24b-instruct": { - "id": "mistralai/mistral-small-3.2-24b-instruct", - "name": "Mistral: Mistral Small 3.2 24B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, + "sonar-pro": { + "id": "sonar-pro", + "name": "Perplexity Sonar Pro", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2025-06-20", - "last_updated": "2025-06-20", + "knowledge": "2025-01", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -77182,30 +73231,30 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 16384 + "context": 200000, + "output": 4096 }, "cost": { - "input": 0.09375, - "output": 0.25 + "input": 3, + "output": 15 } }, - "mistralai/mistral-large": { - "id": "mistralai/mistral-large", - "name": "Mistral Large", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, + "gpt-4.1": { + "id": "gpt-4.1", + "name": "OpenAI GPT-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-02-26", - "last_updated": "2024-02-26", + "knowledge": "2025-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" @@ -77213,37 +73262,37 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 102400 + "context": 1047576, + "output": 32768 }, "cost": { "input": 2, - "output": 6, - "cache_read": 0.2 + "output": 8, + "cache_read": 0.5 } }, - "mistralai/mistral-small-2603": { - "id": "mistralai/mistral-small-2603", - "name": "Mistral Small 4", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "Anthropic: Claude Sonnet 4.5 (20250929)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "knowledge": "2025-09", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", @@ -77253,34 +73302,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 209715 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 3, + "output": 15, + "cache_read": 0.30000000000000004, + "cache_write": 3.75 } }, - "mistralai/mistral-medium-3": { - "id": "mistralai/mistral-medium-3", - "name": "Mistral: Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, + "deepseek-reasoner": { + "id": "deepseek-reasoner", + "name": "DeepSeek Reasoner", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "knowledge": "2025-01", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -77288,64 +73336,75 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 104857 + "context": 128000, + "output": 64000 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.04 + "input": 0.56, + "output": 1.68, + "cache_read": 0.07 } }, - "mistralai/voxtral-small-24b-2507": { - "id": "mistralai/voxtral-small-24b-2507", - "name": "Voxtral Small 24B 2507", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "voxtral", - "attachment": true, - "reasoning": false, + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "xAI Grok 4.1 Fast Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", + "knowledge": "2025-11", + "release_date": "2025-11-17", + "last_updated": "2025-11-17", "modalities": { "input": [ "text", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 26214 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01 + "input": 0.19999999999999998, + "output": 0.5, + "cache_read": 0.049999999999999996 } }, - "mistralai/mistral-large-2407": { - "id": "mistralai/mistral-large-2407", - "name": "Mistral Large 2407", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, - "reasoning": false, + "gemini-3-pro-preview": { + "id": "gemini-3-pro-preview", + "name": "Google Gemini 3 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-11-19", - "last_updated": "2024-11-19", + "knowledge": "2025-11", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", - "pdf" + "image", + "audio", + "video" ], "output": [ "text" @@ -77353,27 +73412,27 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 104857 + "context": 1048576, + "output": 65536 }, "cost": { "input": 2, - "output": 6, - "cache_read": 0.2 + "output": 12, + "cache_read": 0.19999999999999998 } }, - "mistralai/ministral-14b-2512": { - "id": "mistralai/ministral-14b-2512", - "name": "Mistral: Ministral 3 14B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "OpenAI GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", @@ -77385,27 +73444,27 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 209715 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.02 + "input": 0.25, + "output": 2, + "cache_read": 0.024999999999999998 } }, - "mistralai/mistral-small-3.1-24b-instruct": { - "id": "mistralai/mistral-small-3.1-24b-instruct", - "name": "Mistral: Mistral Small 3.1 24B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "OpenAI GPT-4.1 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-03-17", - "last_updated": "2025-03-17", + "knowledge": "2025-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -77417,30 +73476,31 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 102400 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.351, - "output": 0.555 + "input": 0.39999999999999997, + "output": 1.5999999999999999, + "cache_read": 0.09999999999999999 } }, - "mistralai/ministral-3b-2512": { - "id": "mistralai/ministral-3b-2512", - "name": "Mistral: Ministral 3 3B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "sonar-reasoning": { + "id": "sonar-reasoning", + "name": "Perplexity Sonar Reasoning", + "description": "Web-grounded reasoning model for multi-step research and cited answers", + "family": "sonar-reasoning", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "knowledge": "2025-01", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -77448,31 +73508,40 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 104857 + "context": 127000, + "output": 4096 }, "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.01 + "input": 1, + "output": 5 } }, - "mistralai/mixtral-8x22b-instruct": { - "id": "mistralai/mixtral-8x22b-instruct", - "name": "Mistral: Mixtral 8x22B Instruct", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "sonar-deep-research": { + "id": "sonar-deep-research", + "name": "Perplexity Sonar Deep Research", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar-deep-research", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, "temperature": true, - "release_date": "2024-04-17", - "last_updated": "2024-04-17", + "knowledge": "2025-01", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -77480,74 +73549,29 @@ }, "open_weights": false, "limit": { - "context": 65536, - "output": 52428 + "context": 127000, + "output": 4096 }, "cost": { "input": 2, - "output": 6, - "cache_read": 0.2 + "output": 8 } }, - "mistralai/codestral-2508": { - "id": "mistralai/codestral-2508", - "name": "Mistral: Codestral 2508", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "codestral", - "attachment": true, + "kimi-k2-0905": { + "id": "kimi-k2-0905", + "name": "Kimi K2 (09/05)", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-01", - "last_updated": "2025-08-01", + "knowledge": "2025-09", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "pdf" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 204800 - }, - "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.03 - } - }, - "sakana/fugu-max": { - "id": "sakana/fugu-max", - "name": "Sakana: Fugu Max", - "description": "Fugu Max is the cost-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to route...", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", - "modalities": { - "input": [ - "text", - "image", - "pdf" ], "output": [ "text" @@ -77555,42 +73579,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 0.5, + "output": 2, + "cache_read": 0.39999999999999997 } }, - "sakana/fugu-ultra-v2": { - "id": "sakana/fugu-ultra-v2", - "name": "Sakana: Fugu Ultra v2", - "description": "Fugu Ultra v2 is the higher-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to...", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "OpenAI GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": false, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -77598,41 +73611,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.049999999999999996, + "output": 0.39999999999999997, + "cache_read": 0.005 } }, - "sakana/fugu-ultra": { - "id": "sakana/fugu-ultra", - "name": "Fugu Ultra", - "description": "Fugu Ultra is the higher-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to route...", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], + "grok-4": { + "id": "grok-4", + "name": "xAI Grok 4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-09", + "last_updated": "2024-07-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -77640,41 +73642,33 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 3, + "output": 15, + "cache_read": 0.75 } }, - "sakana/sakana-namazu": { - "id": "sakana/sakana-namazu", - "name": "Sakana Namazu", - "description": "Sakana Namazu is a Japanese-specialized reasoning model from Sakana AI, based on Kimi K2.6 with additional training for Japanese language and business contexts. It is suited for Japanese instruction following,...", - "family": "sakana-namazu", - "attachment": true, + "qwen3-235b-a22b-thinking": { + "id": "qwen3-235b-a22b-thinking", + "name": "Qwen3 235B A22B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -77683,38 +73677,39 @@ "open_weights": false, "limit": { "context": 262144, - "output": 65536 + "output": 81920 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.15 + "input": 0.3, + "output": 2.9000000000000004 } }, - "inclusionai/ling-3.0-flash-fin:free": { - "id": "inclusionai/ling-3.0-flash-fin:free", - "name": "inclusionAI: Ling 3.0 Flash Fin (free)", - "description": "Ling 3.0 Flash Fin is a finance-focused mixture-of-experts model from InclusionAI, built on Ling 3.0 Flash with 5.1B active parameters out of 124B total. It is designed for real-world investment...", - "family": "ling", + "claude-opus-4": { + "id": "claude-opus-4", + "name": "Anthropic: Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "knowledge": "2025-05", + "release_date": "2025-05-14", + "last_updated": "2025-05-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -77722,38 +73717,33 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "inclusionai/ling-3.0-flash-sante:free": { - "id": "inclusionai/ling-3.0-flash-sante:free", - "name": "inclusionAI: Ling 3.0 Flash Sante (free)", - "description": "Ling 3.0 Flash Sante is a health and medicine-focused mixture-of-experts model from InclusionAI, built on Ling 3.0 Flash with 5.1B active parameters out of 124B total. It is designed for...", - "family": "ling", + "grok-4-fast-non-reasoning": { + "id": "grok-4-fast-non-reasoning", + "name": "xAI Grok 4 Fast Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "knowledge": "2025-09", + "release_date": "2025-09-19", + "last_updated": "2025-09-19", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -77761,38 +73751,41 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.19999999999999998, + "output": 0.5, + "cache_read": 0.049999999999999996 } }, - "inclusionai/ling-3.0-flash": { - "id": "inclusionai/ling-3.0-flash", - "name": "inclusionAI: Ling 3.0 Flash", - "description": "*Ling-3.0-flash* is a *124B-parameter Mixture-of-Experts (MoE) model*, with approximately *5.1B parameters activated per token*. The model is designed with *token efficiency and production-scale agentic inference* as key priorities, enabling developers...", - "family": "ling", + "claude-4.5-opus": { + "id": "claude-4.5-opus", + "name": "Anthropic: Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-07-23", - "last_updated": "2026-07-23", + "knowledge": "2025-11", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -77800,36 +73793,28 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.012 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "inclusionai/ling-3.0-flash-vl": { - "id": "inclusionai/ling-3.0-flash-vl", - "name": "inclusionAI: Ling 3.0 Flash VL", - "description": "Ling 3.0 Flash VL builds on Ling 3.0 Flash (124B total / 5.5B active MoE from InclusionAI), further strengthening its language capabilities while adding native visual perception and advanced visual...", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "qwen3-vl-235b-a22b-instruct": { + "id": "qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-09", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", @@ -77842,36 +73827,26 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.012 + "input": 0.3, + "output": 1.5 } }, - "inclusionai/ling-3.0-flash-fin": { - "id": "inclusionai/ling-3.0-flash-fin", - "name": "inclusionAI: Ling 3.0 Flash Fin", - "description": "Ling 3.0 Flash Fin is a finance-focused mixture-of-experts model from InclusionAI, built on Ling 3.0 Flash with 5.1B active parameters out of 124B total. It is designed for real-world investment...", - "family": "ling", + "kimi-k2-0711": { + "id": "kimi-k2-0711", + "name": "Kimi K2 (07/11)", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -77882,80 +73857,57 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.012 + "input": 0.5700000000000001, + "output": 2.3 } }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "gemma-3-12b-it": { + "id": "gemma-3-12b-it", + "name": "Google Gemma 3 12B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2024-12", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 131072, + "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.049999999999999996, + "output": 0.09999999999999999 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "OpenAI GPT-4o-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2024-07", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", @@ -77965,38 +73917,39 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "OpenAI GPT-OSS 20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", + "medium", "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "knowledge": "2024-06", + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ "text" @@ -78005,39 +73958,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 98304 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 0.049999999999999996, + "output": 0.19999999999999998 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "claude-3.5-sonnet-v2": { + "id": "claude-3.5-sonnet-v2", + "name": "Anthropic: Claude 3.5 Sonnet v2", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2024-10", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ "text", @@ -78047,28 +73989,61 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 3 + "input": 3, + "output": 15, + "cache_read": 0.30000000000000004, + "cache_write": 3.75 } }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "MoonshotAI: Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "qwen3-30b-a3b": { + "id": "qwen3-30b-a3b", + "name": "Qwen3 30B A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-09-04", - "last_updated": "2025-09-04", + "knowledge": "2025-06", + "release_date": "2025-06-01", + "last_updated": "2025-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 41000, + "output": 41000 + }, + "cost": { + "input": 0.08, + "output": 0.29 + } + }, + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "OpenAI: GPT-5 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -78079,26 +74054,27 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 98304 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.5 + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 } }, - "moonshotai/kimi-k2": { - "id": "moonshotai/kimi-k2", - "name": "MoonshotAI: Kimi K2 0711", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-07-11", - "last_updated": "2025-07-11", + "knowledge": "2025-09", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ "text" @@ -78109,35 +74085,36 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 163840, + "output": 65536 }, "cost": { - "input": 0.57, - "output": 2.3 + "input": 0.27, + "output": 0.41 } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "MoonshotAI: Kimi K2.7 Code is a coding-focused model in Moonshot AI's Kimi K2 family, built to complete end-to-end programming tasks reliably over long contexts. It uses a native multimodal mixture-of-experts...", - "family": "kimi-k2", - "attachment": true, + "claude-4.5-sonnet": { + "id": "claude-4.5-sonnet", + "name": "Anthropic: Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2025-09", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", @@ -78147,30 +74124,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.7062, - "output": 3.3, - "cache_read": 0.18 + "input": 3, + "output": 15, + "cache_read": 0.30000000000000004, + "cache_write": 3.75 } }, - "kilo-auto/small": { - "id": "kilo-auto/small", - "name": "Auto Small", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistral-small": { + "id": "mistral-small", + "name": "Mistral Small 3.2", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", "modalities": { "input": [ "text", @@ -78180,35 +74157,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.4, - "reasoning": 0, - "cache_read": 0.005 + "input": 0.075, + "output": 0.2 } }, - "kilo-auto/efficient": { - "id": "kilo-auto/efficient", - "name": "Auto Efficient", - "description": "Routes each request to the cheapest model that gets the job done, based on continuously benchmarked accuracy and cost.", - "family": "auto", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "llama-3.1-8b-instruct-turbo": { + "id": "llama-3.1-8b-instruct-turbo", + "name": "Meta Llama 3.1 8B Instruct Turbo", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "knowledge": "2024-07", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -78216,142 +74189,153 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.325, - "output": 1.95, - "reasoning": 0, - "cache_read": 0.0325, - "cache_write": 0.40625 + "input": 0.02, + "output": 0.03 } }, - "kilo-auto/free": { - "id": "kilo-auto/free", - "name": "Auto Free", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", + "gpt-5.1": { + "id": "gpt-5.1", + "name": "OpenAI GPT-5.1", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 256000, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "reasoning": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.12500000000000003 } - }, - "kilo-auto/frontier": { - "id": "kilo-auto/frontier", - "name": "Auto Frontier", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": true, + } + } + }, + "zai": { + "id": "zai", + "env": [ + "ZHIPU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.z.ai/api/paas/v4", + "name": "Z.AI", + "doc": "https://docs.z.ai/guides/overview/pricing", + "models": { + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "reasoning": 0, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "kilo-auto/balanced": { - "id": "kilo-auto/balanced", - "name": "Auto Balanced", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", + "glm-4.5v": { + "id": "glm-4.5v", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 64000, + "output": 16384 }, "cost": { - "input": 0.325, - "output": 1.95, - "reasoning": 0, - "cache_read": 0.0325, - "cache_write": 0.40625 + "input": 0.6, + "output": 1.8 } }, - "rekaai/reka-flash-3": { - "id": "rekaai/reka-flash-3", - "name": "Reka Flash 3", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "reka", + "glm-4.5": { + "id": "glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -78360,69 +74344,75 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 58982 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.1, - "output": 0.2 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "rekaai/reka-edge": { - "id": "rekaai/reka-edge", - "name": "Reka Edge", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "reka", - "attachment": true, - "reasoning": false, + "glm-4.7-flashx": { + "id": "glm-4.7-flashx", + "name": "GLM-4.7-FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "output": 14745 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 0.07, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0 } }, - "nvidia/nemotron-3-nano-30b-a3b": { - "id": "nvidia/nemotron-3-nano-30b-a3b", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -78433,40 +74423,36 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 235929 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.03 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "nvidia/nemotron-3.5-content-safety:free": { - "id": "nvidia/nemotron-3.5-content-safety:free", - "name": "NVIDIA: Nemotron 3.5 Content Safety (free)", - "description": "NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).", - "family": "nemotron", - "attachment": true, + "glm-4.6": { + "id": "glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -78474,35 +74460,40 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "nvidia/nemotron-3.5-lightning": { - "id": "nvidia/nemotron-3.5-lightning", - "name": "Nemotron 3.5 Lightning 30B A3B", - "description": "NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that...", - "family": "nemotron", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -78513,41 +74504,38 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 1000000, "output": 131072 }, "cost": { - "input": 0.065, - "output": 0.18 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free": { - "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free", - "name": "NVIDIA: Nemotron 3 Nano Omni (free)", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", + "glm-4.6v": { + "id": "glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", "image", - "video", - "audio" + "video" ], "output": [ "text" @@ -78555,76 +74543,73 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 0.9 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.08, - "output": 0.45 + "input": 1.2, + "output": 4, + "cache_read": 0.24, + "cache_write": 0 } }, - "nvidia/nemotron-3-ultra-550b-a55b:free": { - "id": "nvidia/nemotron-3-ultra-550b-a55b:free", - "name": "NVIDIA: Nemotron 3 Ultra (free)", - "description": "NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).", - "family": "nemotron", + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -78635,36 +74620,33 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 1.1, + "cache_read": 0.03, + "cache_write": 0 } }, - "nvidia/nemotron-3-ultra-550b-a55b": { - "id": "nvidia/nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...", - "family": "nemotron", + "glm-4.7-flash": { + "id": "glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -78675,37 +74657,33 @@ }, "open_weights": true, "limit": { - "context": 202800, - "output": 182520 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 2.2, - "cache_read": 0.1 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "nvidia/nemotron-3-super-120b-a12b:free": { - "id": "nvidia/nemotron-3-super-120b-a12b:free", - "name": "NVIDIA: Nemotron 3 Super (free)", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "glm-4.5-flash": { + "id": "glm-4.5-flash", + "name": "GLM-4.5-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -78716,39 +74694,38 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 235929 + "context": 131072, + "output": 98304 }, "cost": { "input": 0, - "output": 0 + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "nvidia/nemotron-3.5-content-safety": { - "id": "nvidia/nemotron-3.5-content-safety", - "name": "Nemotron 3.5 Content Safety", - "description": "NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting...", - "family": "nemotron", - "attachment": true, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], - "tool_call": false, - "structured_output": false, + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -78756,35 +74733,36 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 117964 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 1, + "output": 3.2, + "cache_read": 0.2, + "cache_write": 0 } }, - "nvidia/nemotron-3.5-lightning:free": { - "id": "nvidia/nemotron-3.5-lightning:free", - "name": "NVIDIA: Nemotron 3.5 Lightning (free)", - "description": "NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).", - "family": "nemotron", + "glm-5-turbo": { + "id": "glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text" @@ -78793,282 +74771,291 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 200000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.2, + "output": 4, + "cache_read": 0.24, + "cache_write": 0 } - }, - "xiaomi/mimo-v2.6-pro": { - "id": "xiaomi/mimo-v2.6-pro", - "name": "MiMo-V2.6-Pro", - "description": "MiMo-V2.6-Pro is the flagship foundation model developed by Xiaomi. Built at a scale of over 1T parameters, it is designed to push the ceiling of capability for the most demanding...", - "family": "mimo", + } + } + }, + "nearai": { + "id": "nearai", + "env": [ + "NEARAI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://cloud-api.near.ai/v1", + "name": "NEAR AI Cloud", + "doc": "https://docs.near.ai/", + "models": { + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", "image", + "video", "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 + "output": 65536 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 }, + "cost": { + "input": 0.13, + "output": 0.4, + "cache_read": 0.026 + } + }, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", "audio", - "video" + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.003, + "input": 1.25, + "output": 10, + "cache_read": 0.125, "tiers": [ { - "input": 0.8, - "output": 4, - "cache_read": 0.16, + "input": 2.5, + "output": 15, + "cache_read": 0.25, "tier": { "type": "context", - "size": 256000 + "size": 200000 } } ], "context_over_200k": { - "input": 0.8, - "output": 4, - "cache_read": 0.16 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } } }, - "xiaomi/mimo-v2.6-pro-ultraspeed": { - "id": "xiaomi/mimo-v2.6-pro-ultraspeed", - "name": "MiMo-V2.6-Pro-UltraSpeed", - "description": "MiMo-V2.6-Pro-UltraSpeed is the fast speed edition of Xiaomi's flagship foundation model, MiMo-V2.6-Pro. Built from the same 1T MiMo-V2.6-Pro checkpoint, it matches the original model in quality while delivering roughly 10x...", - "family": "mimo", + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", "audio", - "video" + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 + "output": 65536 }, "cost": { - "input": 4.35, - "output": 8.7, - "cache_read": 0.036 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 + "output": 65536 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.004, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 } }, - "xiaomi/mimo-v2.6-flash": { - "id": "xiaomi/mimo-v2.6-flash", - "name": "MiMo-V2.6-Flash", - "description": "MiMo-V2.6-Flash is an open-source foundation model developed by Xiaomi. Built on a Mixture-of-Experts architecture with 309B total parameters and 15B activated per token, it employs a hybrid attention mechanism for...", - "family": "mimo", + "google/gemini-3-pro": { + "id": "google/gemini-3-pro", + "name": "Gemini 3 Pro Preview", + "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", "image", + "video", "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 1.25, + "output": 15, + "cache_read": 0 } }, - "undi95/remm-slerp-l2-13b": { - "id": "undi95/remm-slerp-l2-13b", - "name": "ReMM SLERP 13B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2023-07-22", - "last_updated": "2023-07-22", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -79076,25 +75063,27 @@ }, "open_weights": false, "limit": { - "context": 6144, - "output": 5529 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.35, - "output": 0.65 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "input_audio": 0.3 } }, - "gryphe/mythomax-l2-13b": { - "id": "gryphe/mythomax-l2-13b", - "name": "MythoMax 13B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "Qwen/Qwen3-Embedding-0.6B": { + "id": "Qwen/Qwen3-Embedding-0.6B", + "name": "Qwen3 Embedding 0.6B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2023-07-02", - "last_updated": "2023-07-02", + "temperature": false, + "release_date": "2025-06-03", + "last_updated": "2025-06-03", "modalities": { "input": [ "text" @@ -79103,36 +75092,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4096, - "output": 3686 + "context": 40960, + "output": 1024 }, "cost": { - "input": 0.08, - "output": 0.11 + "input": 0.01, + "output": 0 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "Qwen/Qwen3-Reranker-0.6B": { + "id": "Qwen/Qwen3-Reranker-0.6B", + "name": "Qwen3 Reranker 0.6B", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-03", + "last_updated": "2025-06-03", "modalities": { "input": [ "text" @@ -79143,58 +75123,67 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 40960, + "output": 1024 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.01, + "output": 0.01 } }, - "minimax/minimax-01": { - "id": "minimax/minimax-01", - "name": "MiniMax: MiniMax-01", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + "Qwen/Qwen3.6-35B-A3B-FP8": { + "id": "Qwen/Qwen3.6-35B-A3B-FP8", + "name": "Qwen 3.6 35B A3B FP8", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-01-15", - "last_updated": "2025-01-15", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000192, - "output": 900172 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 1.1 + "input": 0.17, + "output": 1.1, + "cache_read": 0.056 } }, - "minimax/minimax-m2-her": { - "id": "minimax/minimax-m2-her", - "name": "MiniMax-M2 Her", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "Qwen/Qwen3.5-122B-A10B": { + "id": "Qwen/Qwen3.5-122B-A10B", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-01-23", - "last_updated": "2026-01-23", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text" @@ -79203,37 +75192,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 2048 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.4, + "output": 3.2 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "Qwen/Qwen3-30B-A3B-Instruct-2507": { + "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen3 30B-A3B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ "text" @@ -79244,188 +75224,161 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.15, + "output": 0.55 } }, - "minimax/minimax-m1": { - "id": "minimax/minimax-m1", - "name": "MiniMax: MiniMax M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "Qwen/Qwen3-VL-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", + "name": "Qwen3-VL 30B-A3B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 40000 + "context": 256000, + "output": 32768 }, "cost": { - "input": 0.4, - "output": 2.2 + "input": 0.15, + "output": 0.55 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 512000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", + "low", + "medium", "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", + "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "mancer/weaver": { - "id": "mancer/weaver", - "name": "Mancer: Weaver (alpha)", - "description": "An attempt to recreate Claude-style verbosity, but don't expect the same level of coherence or memory. Meant for use in roleplay/narrative situations.", - "family": "alpha", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2023-08-02", - "last_updated": "2023-08-02", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -79433,36 +75386,40 @@ }, "open_weights": false, "limit": { - "context": 8000, - "output": 6000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.4, - "output": 0.75 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "stepfun/step-3.7-flash:free": { - "id": "stepfun/step-3.7-flash:free", - "name": "StepFun: Step 3.7 Flash (free)", - "description": "Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters per token. The model supports a 256K context window and exposes selectable reasoning levels (high/medium/low), letting callers trade off speed, cost, and depth of reasoning. Designed for coding, agentic workflows, structured outputs, and long-context productivity tasks.", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -79472,38 +75429,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "reasoning": 0, - "cache_read": 0 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "stepfun/step-3.5-flash": { - "id": "stepfun/step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", + "medium", "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ "text" @@ -79512,80 +75472,76 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "stepfun/step-3.7-flash": { - "id": "stepfun/step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters...", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 230400 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "input": 1.8, + "output": 15.5, + "cache_read": 0.18 } }, - "dots-studio/dots-3-note-preview:free": { - "id": "dots-studio/dots-3-note-preview:free", - "name": "Dots Studio: Dots3-Note Preview (free)", - "description": "Dots3-Note Preview is an open-weight mixture-of-experts model from Dots Studio, with 16B active parameters out of 280B total. It is the lightest model in the Dots 3 family and is...", + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -79597,26 +75553,26 @@ }, "open_weights": false, "limit": { - "context": 512000, - "output": 460800 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "inception/mercury-2.5": { - "id": "inception/mercury-2.5", - "name": "Inception: Mercury 2.5", - "description": "Mercury 2.5 is the fastest reasoning LLM, and the latest diffusion LLM (dLLM) from Inception. Instead of generating tokens sequentially, Mercury 2.5 produces and refines multiple tokens in parallel, achieving...", - "family": "mercury", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT-OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -79626,8 +75582,8 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -79636,23 +75592,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 260000, - "output": 65536 + "context": 131000, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02 + "input": 0.15, + "output": 0.55 } }, - "inception/mercury-2": { - "id": "inception/mercury-2", - "name": "Inception: Mercury 2", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", - "attachment": false, + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -79661,50 +75616,22 @@ "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-03-04", - "last_updated": "2026-03-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 50000 - }, - "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.025 - } - }, - "amazon/nova-lite-v1": { - "id": "amazon/nova-lite-v1", - "name": "Amazon: Nova Lite 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -79712,19 +75639,37 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 5120 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.06, - "output": 0.24 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "amazon/nova-2-lite-v1": { - "id": "amazon/nova-2-lite-v1", - "name": "Amazon: Nova 2 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -79732,21 +75677,23 @@ "type": "effort", "values": [ "none", - "high" + "low", + "medium", + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" @@ -79754,30 +75701,34 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65535 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "amazon/nova-pro-v1": { - "id": "amazon/nova-pro-v1", - "name": "Amazon: Nova Pro 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -79785,26 +75736,39 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 5120 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.8, - "output": 3.2 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "amazon/nova-premier-v1": { - "id": "amazon/nova-premier-v1", - "name": "Amazon: Nova Premier 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-10-31", - "last_updated": "2025-10-31", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -79816,30 +75780,34 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 12.5, - "cache_read": 0.625 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "amazon/nova-micro-v1": { - "id": "amazon/nova-micro-v1", - "name": "Amazon: Nova Micro 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -79847,28 +75815,43 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 5120 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.035, - "output": 0.14 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "relace/relace-search": { - "id": "relace/relace-search", - "name": "Relace: Relace Search", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, - "reasoning": false, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -79876,65 +75859,73 @@ }, "open_weights": false, "limit": { - "context": 256000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 1, - "output": 3 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "relace/relace-apply-3": { - "id": "relace/relace-apply-3", - "name": "Relace: Relace Apply 3", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "openai/whisper-large-v3": { + "id": "openai/whisper-large-v3", + "name": "Whisper Large v3", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": false, - "release_date": "2025-09-26", - "last_updated": "2025-09-26", + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 128000 + "context": 448, + "output": 448 }, "cost": { - "input": 0.85, - "output": 1.25 + "input": 0.01, + "output": 0 } }, - "aion-labs/aion-2.0": { - "id": "aion-labs/aion-2.0", - "name": "AionLabs: Aion-2.0", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", + "medium", "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -79942,30 +75933,46 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.8, - "output": 1.6, - "cache_read": 0.2 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "aion-labs/aion-rp-llama-3.1-8b": { - "id": "aion-labs/aion-rp-llama-3.1-8b", - "name": "AionLabs: Aion-RP 1.0 (8B)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-02-04", - "last_updated": "2025-02-04", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -79973,36 +75980,58 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 29491 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.8, - "output": 1.6 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "aion-labs/aion-3.0": { - "id": "aion-labs/aion-3.0", - "name": "AionLabs: Aion-3.0", - "description": "Aion-3.0 is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It uses a collaborative generation process in which multiple specialized models each contribute...", - "attachment": false, + "anthropic/claude-sonnet-4-5": { + "id": "anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-07-07", - "last_updated": "2026-07-07", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -80010,39 +76039,38 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { "input": 3, - "output": 6, - "cache_read": 0.75 + "output": 15.5, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "aion-labs/aion-3.5": { - "id": "aion-labs/aion-3.5", - "name": "AionLabs: Aion 3.5", - "description": "Aion 3.5 is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It uses a collaborative generation process in which multiple specialized models each...", - "attachment": false, + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -80050,39 +76078,42 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 3, - "output": 6, - "cache_read": 0.75 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "aion-labs/aion-3.5-mini": { - "id": "aion-labs/aion-3.5-mini", - "name": "AionLabs: Aion 3.5 Mini", - "description": "Aion 3.5 Mini is a multi-model roleplaying and storytelling system from AionLabs, built on the GLM family of models. It is the smaller, lower-cost sibling of Aion 3.5 and uses...", - "attachment": false, + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -80090,37 +76121,85 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 - }, + "context": 200000, + "output": 64000 + }, "cost": { - "input": 0.7, - "output": 1.4, - "cache_read": 0.18 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "aion-labs/aion-3.0-mini": { - "id": "aion-labs/aion-3.0-mini", - "name": "AionLabs: Aion-3.0-Mini", - "description": "Aion-3.0 Mini is a multi-model roleplaying and storytelling system from AionLabs, built on the DeepSeek family of models. It uses a collaborative generation process in which multiple specialized models each...", - "attachment": false, + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-07-07", - "last_updated": "2026-07-07", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ + "text", + "image", + "pdf" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", + "modalities": { + "input": [ + "text", + "image", + "pdf" ], "output": [ "text" @@ -80128,27 +76207,36 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.7, - "output": 1.4, - "cache_read": 0.18 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "qwen/qwen3-235b-a22b-2507": { - "id": "qwen/qwen3-235b-a22b-2507", - "name": "Qwen: Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "zai-org/GLM-5.1-FP8": { + "id": "zai-org/GLM-5.1-FP8", + "name": "GLM-5.1 FP8", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ "text" @@ -80157,69 +76245,70 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 235929 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0.1495, - "output": 0.598 + "input": 0.85, + "output": 3.3 } }, - "qwen/qwen3-vl-235b-a22b-instruct": { - "id": "qwen/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "black-forest-labs/FLUX.2-klein-4B": { + "id": "black-forest-labs/FLUX.2-klein-4B", + "name": "FLUX.2 Klein 4B", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "image" ] }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.26, - "output": 1.04 + "input": 1, + "output": 1 } - }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + } + } + }, + "llmgateway": { + "id": "llmgateway", + "env": [ + "LLMGATEWAY_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.llmgateway.io/v1", + "name": "LLM Gateway", + "doc": "https://llmgateway.io/docs", + "models": { + "qwen-coder-plus": { + "id": "qwen-coder-plus", + "name": "Qwen Coder Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2024-09-18", + "last_updated": "2024-09-18", "modalities": { "input": [ "text" @@ -80230,44 +76319,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "output": 8192 }, "cost": { - "input": 1.475, - "output": 4.425, - "cache_read": 0.295, - "cache_write": 1.84375 + "input": 0.502, + "output": 1.004 } }, - "qwen/qwen3.8-27b": { - "id": "qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Qwen3.8 27B is an open-weight dense vision-language model from Qwen. It is suited for coding, professional workflows, research, multimodal interaction, and long-running agent tasks, with flexible thinking that can be...", - "family": "qwen", + "mistral-large-latest": { + "id": "mistral-large-latest", + "name": "Mistral Large (latest)", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -80275,37 +76350,27 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 262144 }, "cost": { - "input": 0.425, - "output": 2.55, - "cache_read": 0.085, - "cache_write": 0.53125 + "input": 4, + "output": 12 } }, - "qwen/qwen3-vl-235b-a22b-thinking": { - "id": "qwen/qwen3-vl-235b-a22b-thinking", + "qwen3-vl-235b-a22b-thinking": { + "id": "qwen3-vl-235b-a22b-thinking", "name": "Qwen3 VL 235B A22B Thinking", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", @@ -80318,35 +76383,25 @@ "open_weights": true, "limit": { "context": 131072, - "output": 32768 + "output": 8192 }, "cost": { - "input": 0.4, - "output": 4 + "input": 0.98, + "output": 3.95 } }, - "qwen/qwen3.8-2.4t-a95b": { - "id": "qwen/qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", + "devstral-small-2507": { + "id": "devstral-small-2507", + "name": "Devstral Small", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2025-05", + "release_date": "2025-07-10", + "last_updated": "2025-07-10", "modalities": { "input": [ "text" @@ -80357,42 +76412,32 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "output": 128000 }, + "status": "deprecated", "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 + "input": 0.1, + "output": 0.3 } }, - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5 27B", + "qwen3-vl-30b-a3b-thinking": { + "id": "qwen3-vl-30b-a3b-thinking", + "name": "Qwen3 VL 30B A3B Thinking", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2025-10-02", + "last_updated": "2025-10-02", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -80400,66 +76445,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.195, - "output": 1.56 + "input": 0.2, + "output": 1 } }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 235929 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.15, - "output": 1.2 - } - }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -80470,24 +76480,25 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 235929 + "context": 1050000, + "output": 384000 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "qwen/qwen3-coder": { - "id": "qwen/qwen3-coder", - "name": "Qwen: Qwen3 Coder 480B A35B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, + "knowledge": "2025-04", "release_date": "2025-07-23", "last_updated": "2025-07-23", "modalities": { @@ -80500,36 +76511,28 @@ }, "open_weights": false, "limit": { - "context": 262144, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.975, - "output": 4.875 + "input": 6, + "output": 60, + "cache_read": 1.2, + "cache_write": 7.5 } }, - "qwen/qwen3-32b": { - "id": "qwen/qwen3-32b", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "minimax-m2.7-highspeed": { + "id": "minimax-m2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -80540,23 +76543,25 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.08, - "output": 0.28 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "qwen/qwen-plus": { - "id": "qwen/qwen-plus", + "qwen-plus": { + "id": "qwen-plus", "name": "Qwen Plus", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2024-04", "release_date": "2024-01-25", @@ -80571,42 +76576,36 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 131072, "output": 32768 }, "cost": { - "input": 0.26, - "output": 0.78, - "cache_read": 0.052, - "cache_write": 0.325 + "input": 0.4, + "output": 1.2, + "reasoning": 4, + "cache_read": 0.08, + "cache_write": 0.5 } }, - "qwen/qwen3.7-flash": { - "id": "qwen/qwen3.7-flash", - "name": "Qwen3.7 Flash", - "description": "Qwen3.7 Flash is a vision-language reasoning model from Alibaba. It is suited for multimodal agents, visual coding, search, and computer interaction, with strengths in object recognition, spatial understanding, and real-world...", - "family": "qwen", + "o3": { + "id": "o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -80614,30 +76613,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 991000, - "output": 65536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.03, - "output": 0.13, - "cache_read": 0.006, - "cache_write": 0.038 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "nemotron-3-ultra-550b": { + "id": "nemotron-3-ultra-550b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -80646,31 +76646,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "output": 128000 }, "cost": { - "input": 0.78, - "output": 3.9, - "cache_read": 0.156, - "cache_write": 0.975 + "input": 0.5, + "output": 2.5, + "cache_read": 0.15 } }, - "qwen/qwen3-coder-30b-a3b-instruct": { - "id": "qwen/qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -80681,41 +76679,33 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 235929 + "context": 228700, + "output": 131072 }, "cost": { - "input": 0.2925, - "output": 1.4625 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "qwen/qwen3.8-omni-flash": { - "id": "qwen/qwen3.8-omni-flash", - "name": "Qwen3.8 Omni Flash", - "description": "Qwen3.8 Omni Flash is an omni-modal reasoning model from Alibaba, the first Qwen model built around agentic capabilities with native audio-video understanding. It is suited for audio-video analysis and summarization,...", - "family": "qwen", + "grok-4-20-beta-0309-non-reasoning": { + "id": "grok-4-20-beta-0309-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -80723,69 +76713,102 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016 + "input": 2, + "output": 6, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "qwen/qwen3-vl-30b-a3b-instruct": { - "id": "qwen/qwen3-vl-30b-a3b-instruct", - "name": "Qwen: Qwen3 VL 30B A3B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.13, - "output": 0.52 + "input": 0.38, + "output": 1.98, + "cache_read": 0.19, + "cache_write": 0 } }, - "qwen/qwen3-235b-a22b-thinking-2507": { - "id": "qwen/qwen3-235b-a22b-thinking-2507", - "name": "Qwen: Qwen3 235B A22B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", + "low", + "medium", "high" ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -80793,36 +76816,28 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.23, - "output": 2.3 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 0.08333 } }, - "qwen/qwen3-235b-a22b": { - "id": "qwen/qwen3-235b-a22b", - "name": "Qwen3 235B-A22B", + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct (2507)", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-07-08", + "last_updated": "2025-07-08", "modalities": { "input": [ "text" @@ -80833,40 +76848,29 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 262144, "output": 8192 }, "cost": { - "input": 0.455, - "output": 1.82 + "input": 0.09, + "output": 0.58 } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, + "llama-3-70b-instruct": { + "id": "llama-3-70b-instruct", + "name": "Llama 3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2024-04-18", + "last_updated": "2024-04-18", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -80874,26 +76878,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 8192, + "output": 8000 }, "cost": { - "input": 0.26, - "output": 2.08 + "input": 0.51, + "output": 0.74 } }, - "qwen/qwen-plus-2025-07-28": { - "id": "qwen/qwen-plus-2025-07-28", - "name": "Qwen: Qwen Plus 0728", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-09-08", - "last_updated": "2025-09-08", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -80902,28 +76906,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 262000, + "output": 65536 }, "cost": { - "input": 0.26, - "output": 0.78 + "input": 0.07, + "output": 0.27 } }, - "qwen/qwen3.6-35b-a3b": { - "id": "qwen/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "seed-1-8-251228": { + "id": "seed-1-8-251228", + "name": "Seed 1.8 (251228)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", + "low", + "medium", "high" ] } @@ -80931,13 +76937,12 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2025-12-18", + "last_updated": "2025-12-18", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -80945,28 +76950,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 235929 + "context": 256000, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 1, + "input": 0.25, + "output": 2, "cache_read": 0.05 } }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "kimi-k2": { + "id": "kimi-k2", + "name": "Kimi K2", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2025-07-11", + "last_updated": "2025-07-11", "modalities": { "input": [ "text" @@ -80977,79 +76981,73 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 256000, "output": 16384 }, "cost": { - "input": 0.0975, - "output": 0.78 + "input": 0.57, + "output": 2.3, + "cache_read": 0.5 } }, - "qwen/qwen3-vl-30b-a3b-thinking": { - "id": "qwen/qwen3-vl-30b-a3b-thinking", - "name": "Qwen: Qwen3 VL 30B A3B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "llama-3.1-70b-instruct": { + "id": "llama-3.1-70b-instruct", + "name": "Llama 3.1 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 2048 }, + "status": "beta", "cost": { - "input": 0.2, - "output": 2.4 + "input": 0.72, + "output": 0.72 } }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "gpt-5.2-pro": { + "id": "gpt-5.2-pro", + "name": "GPT-5.2 Pro", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "high" + "medium", + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -81057,28 +77055,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.1875, - "output": 1.125, - "cache_write": 0.234375 + "input": 21, + "output": 168 } }, - "qwen/qwen3-coder-flash": { - "id": "qwen/qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "minimax-m2.1": { + "id": "minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -81087,30 +77084,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.195, - "output": 0.975, - "cache_read": 0.039, - "cache_write": 0.24375 + "input": 0.27, + "output": 1.1 } }, - "qwen/qwen-2.5-coder-32b-instruct": { - "id": "qwen/qwen-2.5-coder-32b-instruct", - "name": "Qwen2.5 Coder 32B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2024-11-11", - "last_updated": "2024-11-11", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -81119,78 +77115,60 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 32768, - "output": 29491 + "output": 16384 }, "cost": { - "input": 0.66, - "output": 1 + "input": 0.1, + "output": 0.3, + "reasoning": 8.4 } }, - "qwen/qwen3.5-flash-02-23": { - "id": "qwen/qwen3.5-flash-02-23", - "name": "Qwen: Qwen3.5-Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.5", + "pixtral-large-latest": { + "id": "pixtral-large-latest", + "name": "Pixtral Large (latest)", + "description": "Mistral's larger vision model for document-heavy image understanding and chat", + "family": "pixtral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-25", - "last_updated": "2026-02-25", + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2024-11-04", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.065, - "output": 0.26 + "input": 4, + "output": 12 } }, - "qwen/qwen3-max-thinking": { - "id": "qwen/qwen3-max-thinking", - "name": "Qwen: Qwen3 Max Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "glm-4-32b-0414-128k": { + "id": "glm-4-32b-0414-128k", + "name": "GLM-4 32B (0414-128k)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-09", - "last_updated": "2026-02-09", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text" @@ -81201,59 +77179,28 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.78, - "output": 3.9 - } - }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.65, - "output": 3.25, - "cache_read": 0.13, - "cache_write": 0.8125 + "input": 0.1, + "output": 0.1 } }, - "qwen/qwen3.5-9b": { - "id": "qwen/qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "seed-1-6-flash-250715": { + "id": "seed-1-6-flash-250715", + "name": "Seed 1.6 Flash (250715)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", + "low", + "medium", "high" ] } @@ -81261,13 +77208,12 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -81276,311 +77222,270 @@ "open_weights": true, "limit": { "context": 256000, - "output": 32768 + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.15 + "input": 0.07, + "output": 0.3, + "cache_read": 0.015 } }, - "qwen/qwen3.5-plus-20260420": { - "id": "qwen/qwen3.5-plus-20260420", - "name": "Qwen: Qwen3.5 Plus 2026-04-20", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.5", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.8, - "cache_write": 0.375 + "input": 0.15, + "output": 1.2 } }, - "qwen/qwen3.8-flash": { - "id": "qwen/qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen3.8 Flash is a multimodal reasoning model from Alibaba. It is suited for coding assistance, agentic workflows, visual understanding, document and codebase analysis, desktop interaction, chart analysis, and long-video analysis.", + "qwen3-vl-8b-instruct": { + "id": "qwen3-vl-8b-instruct", + "name": "Qwen3 VL 8B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2025-08-19", + "last_updated": "2025-08-19", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016, - "cache_write": 0.2 + "input": 0.08, + "output": 0.5 } }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-4o-mini-search-preview": { + "id": "gpt-4o-mini-search-preview", + "name": "GPT-4o Mini Search Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2024-10-01", + "last_updated": "2024-10-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, + "context": 128000, "output": 16384 }, "cost": { - "input": 0.1625, - "output": 1.3 + "input": 0.15, + "output": 0.6 } }, - "qwen/qwen3.6-max-preview": { - "id": "qwen/qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "glm-4.5v": { + "id": "glm-4.5v", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.027, - "output": 6.162, - "cache_write": 1.28375 + "input": 0.6, + "output": 1.8, + "cache_read": 0.11 } }, - "qwen/qwen3-30b-a3b": { - "id": "qwen/qwen3-30b-a3b", - "name": "Qwen3 30B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 16384 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.13, - "output": 0.52 + "input": 0.4, + "output": 1.6, + "cache_read": 0.08, + "cache_write": 0.5 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", + "low", + "medium", "high" ] + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video" + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.39, - "output": 2.34 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "qwen/qwen3-8b": { - "id": "qwen/qwen3-8b", - "name": "Qwen: Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", + "low", + "medium", "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.117, - "output": 0.455 - } - }, - "qwen/qwen3-vl-32b-instruct": { - "id": "qwen/qwen3-vl-32b-instruct", - "name": "Qwen: Qwen3 VL 32B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-10-23", - "last_updated": "2025-10-23", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -81592,60 +77497,34 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.104, - "output": 0.416 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "qwen/qwen3-vl-8b-instruct": { - "id": "qwen/qwen3-vl-8b-instruct", - "name": "Qwen: Qwen3 VL 8B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-10-14", - "last_updated": "2025-10-14", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ + "text", "image", - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.117, - "output": 0.455 - } - }, - "qwen/qwen3-30b-a3b-instruct-2507": { - "id": "qwen/qwen3-30b-a3b-instruct-2507", - "name": "Qwen: Qwen3 30B A3B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-29", - "last_updated": "2025-07-29", - "modalities": { - "input": [ - "text" + "pdf" ], "output": [ "text" @@ -81653,58 +77532,63 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.13, - "output": 0.52 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "qwen/qwen2.5-vl-72b-instruct": { - "id": "qwen/qwen2.5-vl-72b-instruct", - "name": "Qwen: Qwen2.5 VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-02-01", - "last_updated": "2025-02-01", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 115200 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.8, - "output": 1, - "cache_read": 0.4 + "input": 0.248, + "output": 1.485 } }, - "qwen/qwen-2.5-7b-instruct": { - "id": "qwen/qwen-2.5-7b-instruct", - "name": "Qwen: Qwen2.5 7B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "gpt-3.5-turbo": { + "id": "gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2024-10-16", - "last_updated": "2024-10-16", + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" @@ -81715,26 +77599,29 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 29491 + "context": 16385, + "output": 4096 }, "cost": { - "input": 0.1, - "output": 0.2 + "input": 0.5, + "output": 1.5, + "cache_read": 0 } }, - "qwen/qwen3.5-plus-02-15": { - "id": "qwen/qwen3.5-plus-02-15", - "name": "Qwen: Qwen3.5 Plus 2026-02-15", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.5", + "seed-1-6-250615": { + "id": "seed-1-6-250615", + "name": "Seed 1.6 (250615)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", + "low", + "medium", "high" ] } @@ -81742,40 +77629,40 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "release_date": "2025-06-25", + "last_updated": "2025-06-25", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 256000, + "output": 8192 }, "cost": { - "input": 0.26, - "output": 1.56 + "input": 0.25, + "output": 2, + "cache_read": 0.05 } }, - "qwen/qwen-2.5-72b-instruct": { - "id": "qwen/qwen-2.5-72b-instruct", - "name": "Qwen2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" @@ -81786,34 +77673,33 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.36, - "output": 0.4 + "input": 1.25, + "output": 3.75, + "cache_read": 0.125, + "cache_write": 3.125 } }, - "qwen/qwen3-30b-a3b-thinking-2507": { - "id": "qwen/qwen3-30b-a3b-thinking-2507", - "name": "Qwen: Qwen3 30B A3B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "glm-4.5": { + "id": "glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -81822,84 +77708,95 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 81920, - "output": 32768 + "context": 131000, + "output": 98304 }, "cost": { - "input": 0.2, - "output": 2.4 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "qwen/qwen3.6-27b": { - "id": "qwen/qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262140 + "context": 400000, + "input": 272000, + "output": 272000 }, "cost": { - "input": 0.45, - "output": 2.7 + "input": 15, + "output": 120 } }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", + "low", + "medium", "high" ] + }, + { + "type": "budget_tokens", + "min": 1, + "max": 24576 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -81907,78 +77804,33 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.325, - "output": 1.95, - "cache_write": 0.40625 - } - }, - "qwen/qwen3-14b": { - "id": "qwen/qwen3-14b", - "name": "Qwen: Qwen3 14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 40960, - "output": 16384 - }, - "cost": { - "input": 0.2275, - "output": 0.91 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "qwen/qwen3-vl-8b-thinking": { - "id": "qwen/qwen3-vl-8b-thinking", - "name": "Qwen: Qwen3 VL 8B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-14", - "last_updated": "2025-10-14", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -81986,36 +77838,27 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.18, - "output": 2.1 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Qwen3.7-Plus is a cost-effective model in Alibaba's Qwen3.7 series. It supports text and image input with text output, building on the series' text capabilities with a comprehensive upgrade to its...", - "family": "qwen", + "ministral-8b-2512": { + "id": "ministral-8b-2512", + "name": "Ministral 8B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "mistral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", @@ -82025,90 +77868,74 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 8192 }, "cost": { - "input": 0.32, - "output": 1.28, - "cache_read": 0.064, - "cache_write": 0.4 + "input": 0.15, + "output": 0.15 } }, - "qwen/qwen3.8-27b:free": { - "id": "qwen/qwen3.8-27b:free", - "name": "Qwen: Qwen3.8 27B (free)", - "description": "Qwen3.8 27B is an open-weight dense vision-language model from Qwen. It is suited for coding, professional workflows, research, multimodal interaction, and long-running agent tasks, with flexible thinking that can be...", - "family": "qwen", + "gpt-4": { + "id": "gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 30, + "output": 60 } }, - "qwen/qwen3.8-max-0902": { - "id": "qwen/qwen3.8-max-0902", - "name": "Qwen3.8 Max 0902", - "description": "Qwen3.8 Max 0902 is an updated snapshot of Qwen3.8 Max from Alibaba's Qwen team. It is a 2.4-trillion-parameter mixture-of-experts model that accepts text, image, and video input and returns text,...", - "family": "qwen", + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -82116,28 +77943,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 200000, + "output": 100000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "morph/morph-v3-large": { - "id": "morph/morph-v3-large", - "name": "Morph: Morph V3 Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "morph", + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "release_date": "2025-07-07", - "last_updated": "2025-07-07", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" @@ -82149,72 +77975,76 @@ "open_weights": false, "limit": { "context": 262144, - "output": 131072 + "output": 65536 }, "cost": { - "input": 0.9, - "output": 1.9 + "input": 0.845, + "output": 3.38, + "cache_read": 0.6, + "cache_write": 3.75 } }, - "morph/morph-v3-fast": { - "id": "morph/morph-v3-fast", - "name": "Morph: Morph V3 Fast", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "morph", + "mistral-small-2506": { + "id": "mistral-small-2506", + "name": "Mistral Small 3.2", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2025-07-07", - "last_updated": "2025-07-07", + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 81920, - "output": 38000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.8, - "output": 1.2 + "input": 0.1, + "output": 0.3 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -82223,85 +78053,77 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "cache_write": 0.08333 } }, - "openai/gpt-5.4-image-2": { - "id": "openai/gpt-5.4-image-2", - "name": "OpenAI: GPT-5.4 Image 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "qwen-plus-latest": { + "id": "qwen-plus-latest", + "name": "Qwen Plus Latest", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": false, + "reasoning": false, + "tool_call": true, "structured_output": true, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "temperature": true, + "release_date": "2025-01-25", + "last_updated": "2025-01-25", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { - "context": 272000, - "output": 128000 + "context": 1000000, + "output": 8192 }, "cost": { - "input": 8, - "output": 15, - "cache_read": 2 + "input": 0.4, + "output": 1.2, + "cache_read": 0.08, + "cache_write": 0.5 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", "high", "xhigh" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -82314,64 +78136,74 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 30, - "output": 180 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "openai/gpt-audio": { - "id": "openai/gpt-audio", - "name": "OpenAI: GPT Audio", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", + "gemma-4-31b-it": { + "id": "gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "audio", - "pdf" + "image" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.13, + "output": 0.38 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "grok-4-3": { + "id": "grok-4-3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ @@ -82380,134 +78212,118 @@ }, "open_weights": false, "limit": { - "context": 16385, - "output": 4096 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 1.25, + "output": 2.5, + "cache_read": 0.3125, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "openai/gpt-6-astra-pro": { - "id": "openai/gpt-6-astra-pro", - "name": "OpenAI: GPT-6 Astra Pro ($$$$)", - "description": "GPT-6 Astra Pro is the same underlying model as [GPT-6 Astra](https://openrouter.ai/openai/gpt-6-astra), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", - "family": "gpt", + "llama-4-scout-17b-instruct": { + "id": "llama-4-scout-17b-instruct", + "name": "Llama 4 Scout 17B Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "pdf", - "image", - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "output": 128000 + "context": 131072, + "output": 2048 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.18, + "output": 0.59 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 4096 }, "cost": { - "input": 30, - "output": 180 + "input": 0.13, + "output": 0.4 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "sonar-reasoning-pro": { + "id": "sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", + "family": "sonar-reasoning", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "tool_call": false, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "pdf", - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -82515,29 +78331,26 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 2, + "output": 8 } }, - "openai/gpt-4o-2024-05-13": { - "id": "openai/gpt-4o-2024-05-13", - "name": "GPT-4o (2024-05-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "grok-4-20-non-reasoning": { + "id": "grok-4-20-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", @@ -82550,52 +78363,103 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 5, - "output": 15 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "openai/gpt-4o-mini-2024-07-18": { - "id": "openai/gpt-4o-mini-2024-07-18", - "name": "OpenAI: GPT-4o-mini (2024-07-18)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "glm-4.7-flashx": { + "id": "glm-4.7-flashx", + "name": "GLM-4.7-FlashX", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0.07, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0 + } + }, + "qwen3-30b-a3b-instruct-2507": { + "id": "qwen3-30b-a3b-instruct-2507", + "name": "Qwen3 30B A3B Instruct (2507)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2025-07-08", + "last_updated": "2025-07-08", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262000, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.1, + "output": 0.3 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -82604,306 +78468,317 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.931, + "output": 2.93, + "cache_read": 0.173, + "cache_write": 0 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 1050000, + "output": 384000 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "openai/gpt-chat-latest": { - "id": "openai/gpt-chat-latest", - "name": "OpenAI: GPT Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": false, + "qwen3-next-80b-a3b-thinking": { + "id": "qwen3-next-80b-a3b-thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-05-05", - "last_updated": "2026-05-05", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.15, + "output": 1.2 } }, - "openai/gpt-audio-mini": { - "id": "openai/gpt-audio-mini", - "name": "OpenAI: GPT Audio Mini", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", - "attachment": true, - "reasoning": false, + "glm-4.6": { + "id": "glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "audio", - "pdf" + "text" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 2.4 + "input": 0.55, + "output": 2.2, + "cache_read": 0.11, + "cache_write": 0 } }, - "openai/gpt-5.6-sol-pro": { - "id": "openai/gpt-5.6-sol-pro", - "name": "GPT-5.6 Sol", - "description": "GPT-5.6 Sol Pro is the same underlying model as [GPT-5.6 Sol](https://openrouter.ai/openai/gpt-5.6-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", - "family": "gpt-sol", + "qwen35-397b-a17b": { + "id": "qwen35-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5 + "input": 0.6, + "output": 3.6 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "qwen3-235b-a22b-thinking-2507": { + "id": "qwen3-235b-a22b-thinking-2507", + "name": "Qwen3 235B A22B Thinking (2507)", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2025-07-08", + "last_updated": "2025-07-08", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262000, + "output": 8192 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.2, + "output": 0.6 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.6, + "output": 2.5, + "cache_read": 0.06 + } + }, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "image", "text", + "image", "pdf" ], "output": [ @@ -82912,42 +78787,42 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 21, - "output": 168 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", + "gemini-pro-latest": { + "id": "gemini-pro-latest", + "name": "Gemini Pro Latest", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-02-27", + "last_updated": "2026-02-27", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text" @@ -82955,41 +78830,35 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": true, + "glm-4.5-x": { + "id": "glm-4.5-x", + "name": "GLM-4.5 X", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -82997,31 +78866,44 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 128000, + "output": 16384 }, + "status": "beta", "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 2.2, + "output": 8.9, + "cache_read": 0.45 } }, - "openai/gpt-4": { - "id": "openai/gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ @@ -83030,44 +78912,45 @@ }, "open_weights": false, "limit": { - "context": 8191, - "output": 4096 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 30, - "output": 60 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], - "tool_call": true, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -83075,34 +78958,32 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "Grok 4.1 Fast Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text" @@ -83110,19 +78991,19 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", "family": "gpt-nano", "attachment": true, "reasoning": true, @@ -83130,24 +79011,24 @@ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -83160,33 +79041,33 @@ "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "openai/o1": { - "id": "openai/o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "grok-4-20-beta-0309-reasoning": { + "id": "grok-4-20-beta-0309-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", @@ -83199,73 +79080,97 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 2, + "output": 6, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high" + "none", + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "image", - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 15, - "output": 120 + "input": 1.26, + "output": 3.96, + "cache_read": 0.234, + "cache_write": 0 } }, - "openai/gpt-3.5-turbo-instruct": { - "id": "openai/gpt-3.5-turbo-instruct", - "name": "OpenAI: GPT-3.5 Turbo Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2023-09-28", - "last_updated": "2023-09-28", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", - "pdf" + "image", + "video" ], "output": [ "text" @@ -83273,32 +79178,30 @@ }, "open_weights": false, "limit": { - "context": 4095, - "output": 3685 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.5, - "output": 2 + "input": 0.15, + "output": 0.9, + "cache_read": 0.015 } }, - "openai/gpt-4o-2024-08-06": { - "id": "openai/gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "glm-4.5-airx": { + "id": "glm-4.5-airx", + "name": "GLM-4.5 AirX", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-08-06", - "last_updated": "2024-08-06", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -83310,41 +79213,29 @@ "output": 16384 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 1.1, + "output": 4.5, + "cache_read": 0.22 } }, - "openai/gpt-6-astra": { - "id": "openai/gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's flagship model for demanding end-to-end work. It is suited for advanced analysis, software engineering, deep research, scientific work, and document creation, with particular strengths in long-horizon...", - "family": "gpt-astra", + "gpt-5-chat-latest": { + "id": "gpt-5-chat-latest", + "name": "GPT-5 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "temperature": true, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -83352,95 +79243,99 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openai/gpt-5.6-luna-pro": { - "id": "openai/gpt-5.6-luna-pro", - "name": "GPT-5.6 Luna", - "description": "GPT-5.6 Luna Pro is the same underlying model as [GPT-5.6 Luna](https://openrouter.ai/openai/gpt-5.6-luna), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", - "family": "gpt-luna", + "ministral-3b-2512": { + "id": "ministral-3b-2512", + "name": "Ministral 3B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "mistral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 + "input": 0.1, + "output": 0.1 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "qwen2-5-vl-72b-instruct": { + "id": "qwen2-5-vl-72b-instruct", + "name": "Qwen2.5-VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 8192 + }, + "cost": { + "input": 0.13, + "output": 0.4 + } + }, + "glm-4.6v-flashx": { + "id": "glm-4.6v-flashx", + "name": "GLM-4.6V FlashX", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text" @@ -83448,76 +79343,65 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.04, + "output": 0.4, + "cache_read": 0.004 } }, - "openai/gpt-5-image-mini": { - "id": "openai/gpt-5-image-mini", - "name": "OpenAI: GPT-5 Image Mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-10-16", - "last_updated": "2025-10-16", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "video" ], "output": [ - "image", "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, + "context": 512000, "output": 128000 }, "cost": { - "input": 2.5, - "output": 2, - "cache_read": 0.25 + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + "qwen3-vl-plus": { + "id": "qwen3-vl-plus", + "name": "Qwen3-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -83525,42 +79409,50 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.2, + "output": 1.6, + "reasoning": 4.8, + "cache_read": 0.04, + "cache_write": 0.25 } }, - "openai/o3-pro": { - "id": "openai/o3-pro", - "name": "o3-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "high" + "low", + "medium", + "high", + "xhigh" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", - "pdf", - "image" + "image", + "pdf" ], "output": [ "text" @@ -83569,19 +79461,21 @@ "open_weights": false, "limit": { "context": 200000, - "output": 100000 + "output": 64000 }, "cost": { - "input": 20, - "output": 80 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -83595,47 +79489,43 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.018, - "output": 0.09 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openai/gpt-oss-safeguard-20b": { - "id": "openai/gpt-oss-safeguard-20b", - "name": "GPT OSS Safeguard 20B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", + "minimax-m2": { + "id": "minimax-m2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ "text" @@ -83646,38 +79536,43 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 196608, + "output": 128000 }, "cost": { - "input": 0.075, - "output": 0.3, - "cache_read": 0.0375 + "input": 0.2, + "output": 1, + "cache_read": 0.03 } }, - "openai/o3-mini-high": { - "id": "openai/o3-mini-high", - "name": "OpenAI: o3 Mini High", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high" + "low", + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "release_date": "2025-02-12", - "last_updated": "2025-02-12", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", + "image", "pdf" ], "output": [ @@ -83686,63 +79581,33 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/gpt-3.5-turbo-16k": { - "id": "openai/gpt-3.5-turbo-16k", - "name": "OpenAI: GPT-3.5 Turbo 16k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "gpt-5.3-chat-latest": { + "id": "gpt-5.3-chat-latest", + "name": "GPT-5.3 Chat (latest)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2023-08-28", - "last_updated": "2023-08-28", + "knowledge": "2025-08-31", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16385, - "output": 4096 - }, - "cost": { - "input": 3, - "output": 4 - } - }, - "openai/gpt-5.2-chat": { - "id": "openai/gpt-5.2-chat", - "name": "OpenAI: GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-12-10", - "last_updated": "2025-12-10", - "modalities": { - "input": [ - "pdf", - "image", - "text" + "image" ], "output": [ "text" @@ -83751,7 +79616,7 @@ "open_weights": false, "limit": { "context": 128000, - "output": 32000 + "output": 16384 }, "cost": { "input": 1.75, @@ -83759,36 +79624,32 @@ "cache_read": 0.175 } }, - "openai/gpt-6-sol-pro": { - "id": "openai/gpt-6-sol-pro", - "name": "OpenAI: GPT-6 Sol Pro", - "description": "GPT-6 Sol Pro is the same underlying model as [GPT-6 Sol](https://openrouter.ai/openai/gpt-6-sol), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", - "family": "gpt", + "grok-4-5": { + "id": "grok-4-5", + "name": "Grok 4.5", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ - "pdf", - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -83796,136 +79657,99 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "output": 128000 + "context": 500000, + "output": 500000 }, "cost": { "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "output": 6, + "cache_read": 0.5 } }, - "openai/gpt-5-image": { - "id": "openai/gpt-5-image", - "name": "OpenAI: GPT-5 Image ($$$$)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "claude-3-opus": { + "id": "claude-3-opus", + "name": "Claude 3 Opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, + "reasoning": false, + "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-10-14", - "last_updated": "2025-10-14", + "release_date": "2024-03-04", + "last_updated": "2024-03-04", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 200000, + "output": 4096 }, "cost": { - "input": 10, - "output": 10, - "cache_read": 1.25 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen-omni-turbo": { + "id": "qwen-omni-turbo", + "name": "Qwen-Omni Turbo", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-01-19", + "last_updated": "2025-03-26", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 32768, + "output": 2048 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.2, + "output": 0.8 } }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "GPT-5.6 Luna is a fast, cost-efficient model in OpenAI's GPT-5.6 series. It is suited for high-volume, latency-sensitive tasks such as chat, classification, and lightweight agentic workflows, providing capable reasoning for...", - "family": "gpt-luna", - "attachment": true, + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -83933,28 +79757,27 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -83965,14 +79788,13 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -83985,23 +79807,22 @@ "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -84013,13 +79834,13 @@ "structured_output": true, "temperature": false, "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -84037,71 +79858,54 @@ "cache_read": 0.175 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "qwen-flash": { + "id": "qwen-flash", + "name": "Qwen Flash", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -84109,64 +79913,67 @@ }, "open_weights": false, "limit": { - "context": 1047576, + "context": 1000000, "output": 32768 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.05, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0.0625 } }, - "openai/gpt-4o-2024-11-20": { - "id": "openai/gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "qwen3-4b-fp8": { + "id": "qwen3-4b-fp8", + "name": "Qwen3 4B FP8", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 16384 + "output": 8192 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.03, + "output": 0.03 } }, - "openai/gpt-3.5-turbo-0613": { - "id": "openai/gpt-3.5-turbo-0613", - "name": "OpenAI: GPT-3.5 Turbo (older v0613)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", + "image", + "audio", + "video", "pdf" ], "output": [ @@ -84175,53 +79982,51 @@ }, "open_weights": false, "limit": { - "context": 4095, - "output": 3685 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1, - "output": 2 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "qwen2-5-vl-32b-instruct": { + "id": "qwen2-5-vl-32b-instruct", + "name": "Qwen2.5 VL 32B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-03-15", + "last_updated": "2025-03-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 1.4, + "output": 4.2 } }, - "openai/gpt-5.6-terra-pro": { - "id": "openai/gpt-5.6-terra-pro", - "name": "GPT-5.6 Terra", - "description": "GPT-5.6 Terra Pro is the same underlying model as [GPT-5.6 Terra](https://openrouter.ai/openai/gpt-5.6-terra), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", - "family": "gpt-terra", + "gpt-5.6-luna": { + "id": "gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -84260,43 +80065,38 @@ "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 1, + "output": 6, + "cache_read": 0.1 } }, - "openai/gpt-6-luna": { - "id": "openai/gpt-6-luna", - "name": "GPT-6 Luna", - "description": "GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol. It is suited for high-volume and latency-sensitive workloads such as chat, classification, and lightweight agentic...", - "family": "gpt-luna", + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -84304,110 +80104,102 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "openai/o4-mini-high": { - "id": "openai/o4-mini-high", - "name": "OpenAI: o4 Mini High", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "image", "text", - "pdf" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 0.405, + "output": 1.98, + "cache_read": 0.225 } }, - "openai/gpt-6-luna-pro": { - "id": "openai/gpt-6-luna-pro", - "name": "OpenAI: GPT-6 Luna Pro", - "description": "GPT-6 Luna Pro is the same underlying model as [GPT-6 Luna](https://openrouter.ai/openai/gpt-6-luna), served with `reasoning.mode` set to `pro` for higher-quality responses on complex tasks. Learn more in OpenAI's docs: https://developers.openai.com/api/docs/guides/reasoning#reasoning-mode", - "family": "gpt", + "seed-1-6-250915": { + "id": "seed-1-6-250915", + "name": "Seed 1.6 (250915)", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ - "pdf", - "image", - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "output": 128000 + "context": 256000, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125 + "input": 0.25, + "output": 2, + "cache_read": 0.05 } }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", + "gpt-5.6-terra": { + "id": "gpt-5.6-terra", "name": "GPT-5.6 Terra", - "description": "GPT-5.6 Terra is a balanced model in OpenAI's GPT-5.6 series, positioned between the flagship Sol tier and the cost-efficient Luna tier. It is suited for everyday coding, reasoning, and agentic...", - "family": "gpt-terra", + "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -84446,30 +80238,36 @@ "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "gpt-5.2-chat-latest": { + "id": "gpt-5.2-chat-latest", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -84478,35 +80276,27 @@ "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 16384 }, "cost": { - "input": 10, - "output": 30 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "minimax-text-01": { + "id": "minimax-text-01", + "name": "MiniMax Text 01", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-01-15", + "last_updated": "2025-01-15", "modalities": { "input": [ "text" @@ -84517,44 +80307,31 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.03, - "output": 0.17, - "cache_read": 0.03 + "input": 0.2, + "output": 1.1 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text" @@ -84562,43 +80339,40 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.03 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -84606,44 +80380,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 32766 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.05, + "output": 0.25 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "qwen-vl-max": { + "id": "qwen-vl-max", + "name": "Qwen-VL Max", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-08", + "last_updated": "2025-08-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -84651,77 +80411,55 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.8, + "output": 3.2 } }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "GPT-5.6 Sol is the flagship model in OpenAI's GPT-5.6 series. It is suited for complex reasoning, coding, and agentic workflows, and is particularly strong at command-line and multi-step coding tasks...", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, + "llama-3-8b-instruct": { + "id": "llama-3-8b-instruct", + "name": "Llama 3 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5 + "input": 0.04, + "output": 0.04 } }, - "openai/gpt-6-sol": { - "id": "openai/gpt-6-sol", - "name": "GPT-6 Sol", - "description": "GPT-6 Sol is the cost-efficient high-end model in OpenAI's GPT-6 series, positioned below the flagship GPT-6 Astra and above the fast GPT-6 Luna tier. It is suited for demanding professional...", - "family": "gpt-sol", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -84731,11 +80469,10 @@ } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", @@ -84748,44 +80485,31 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "openai/o1-pro": { - "id": "openai/o1-pro", - "name": "o1-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "sonar": { + "id": "sonar", + "name": "Sonar", + "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", + "family": "sonar", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2025-03-19", - "last_updated": "2025-03-19", + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -84793,26 +80517,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 130000, + "output": 4096 }, "cost": { - "input": 150, - "output": 600 + "input": 1, + "output": 1 } }, - "microsoft/phi-4": { - "id": "microsoft/phi-4", - "name": "Microsoft: Phi 4", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", + "qwen-max": { + "id": "qwen-max", + "name": "Qwen Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "release_date": "2025-01-10", - "last_updated": "2025-01-10", + "knowledge": "2024-04", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", "modalities": { "input": [ "text" @@ -84823,28 +80547,42 @@ }, "open_weights": false, "limit": { - "context": 16384, - "output": 14745 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.07, - "output": 0.14 + "input": 1.6, + "output": 6.4 } }, - "microsoft/wizardlm-2-8x22b": { - "id": "microsoft/wizardlm-2-8x22b", - "name": "WizardLM-2 8x22B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2024-04-16", - "last_updated": "2024-04-16", + "o1": { + "id": "o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -84852,38 +80590,27 @@ }, "open_weights": false, "limit": { - "context": 65535, - "output": 8000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.62, - "output": 0.62 + "input": 15, + "output": 60, + "cache_read": 7.5 } - } - } - }, - "venice": { - "id": "venice", - "env": [ - "VENICE_API_KEY" - ], - "npm": "venice-ai-sdk-provider", - "name": "Venice AI", - "doc": "https://docs.venice.ai", - "models": { - "qwen-3-8-2-4t-a95b": { - "id": "qwen-3-8-2-4t-a95b", - "name": "Qwen 3.8 2.4T", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + }, + "minimax-m2.5-highspeed": { + "id": "minimax-m2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-13", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -84894,41 +80621,39 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.3125 + "input": 0.6, + "output": 2.4, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "openai-gpt-56-luna": { - "id": "openai-gpt-56-luna", - "name": "GPT-5.6 Luna", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-luna", + "deepseek-v3.1": { + "id": "deepseek-v3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ "text", @@ -84938,58 +80663,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 0.3125, - "tiers": [ - { - "input": 0.5, - "output": 2.25, - "cache_read": 0.05, - "cache_write": 0.625, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.5, - "output": 2.25, - "cache_read": 0.05, - "cache_write": 0.625 - } + "input": 0.56, + "output": 1.68, + "cache_read": 0.112 } }, - "grok-4-5": { - "id": "grok-4-5", - "name": "Grok 4.5", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "ministral-14b-2512": { + "id": "ministral-14b-2512", + "name": "Ministral 14B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "mistral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2026-07-07", - "last_updated": "2026-07-08", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", @@ -84999,47 +80695,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 32000 + "context": 262144, + "output": 8192 }, "cost": { - "input": 2.27, - "output": 6.8, - "cache_read": 0.34, - "tiers": [ - { - "input": 4.53, - "output": 13.6, - "cache_read": 0.68, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4.53, - "output": 13.6, - "cache_read": 0.68 - } + "input": 0.2, + "output": 0.2 } }, - "aion-labs-aion-3-5": { - "id": "aion-labs-aion-3-5", - "name": "Aion 3.5", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "sonar-pro": { + "id": "sonar-pro", + "name": "Sonar Pro", + "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", + "family": "sonar-pro", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -85047,42 +80728,36 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 8192 }, "cost": { - "input": 3.75, - "output": 7.5, - "cache_read": 0.9375 + "input": 3, + "output": 15 } }, - "kimi-k3-fast-api": { - "id": "kimi-k3-fast-api", - "name": "Kimi K3 Fast", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k3", + "glm-4.6v": { + "id": "glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -85090,113 +80765,98 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 4.5, - "output": 22.5, - "cache_read": 0.45 + "input": 0.3, + "output": 0.9, + "cache_read": 0.05 } }, - "qwen3-6-27b": { - "id": "qwen3-6-27b", - "name": "Qwen 3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-24", - "last_updated": "2026-06-11", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.325, - "output": 3.25 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "hermes-3-llama-3.1-405b": { - "id": "hermes-3-llama-3.1-405b", - "name": "Hermes 3 Llama 3.1 405b", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "hermes", - "attachment": false, + "claude-haiku-4-5-free": { + "id": "claude-haiku-4-5-free", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, "reasoning": false, - "tool_call": false, - "release_date": "2025-09-25", - "last_updated": "2026-06-11", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 200000, + "output": 200000 }, "cost": { - "input": 1.1, - "output": 3 + "input": 0, + "output": 0 } }, - "kimi-k2-7-code": { - "id": "kimi-k2-7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "minimax-m2.1-lightning": { + "id": "minimax-m2.1-lightning", + "name": "MiniMax M2.1 Lightning", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-13", - "last_updated": "2026-06-16", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -85204,98 +80864,141 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 196608, + "output": 131072 }, "cost": { - "input": 0.75, - "output": 3.5, - "cache_read": 0.16 + "input": 0.12, + "output": 0.48 } }, - "google-gemma-3-27b-it": { - "id": "google-gemma-3-27b-it", - "name": "Google Gemma 3 27B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-04", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 198000, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.12, - "output": 0.2 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "qwen-3-7-max": { - "id": "qwen-3-7-max", - "name": "Qwen 3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-05-22", - "last_updated": "2026-06-11", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 65536 + "output": 131072 }, "cost": { - "input": 2.7, - "output": 8.05, - "cache_read": 0.27, - "cache_write": 3.35 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028, + "tiers": [ + { + "input": 0.8, + "output": 4, + "cache_read": 0.16, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.8, + "output": 4, + "cache_read": 0.16 + } } }, - "qwen-3-8-max": { - "id": "qwen-3-8-max", - "name": "Qwen 3.8 Max", - "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", - "family": "qwen", + "mimo-v2-omni": { + "id": "mimo-v2-omni", + "name": "MiMo-V2-Omni", + "description": "MiMo omni model for text, image, video, audio, and agents", + "family": "mimo", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-07-22", - "last_updated": "2026-07-19", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text", "image", - "video" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -85303,22 +81006,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 256000, "output": 131072 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.3125, - "cache_write": 3.125 + "input": 0.4, + "output": 2, + "cache_read": 0.08 } }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -85326,47 +81028,18 @@ "values": [ "none", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-08-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 0.175, - "output": 0.35, - "cache_read": 0.035 - } - }, - "venice-uncensored-role-play": { - "id": "venice-uncensored-role-play", - "name": "Venice Role Play Uncensored", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "venice", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2026-02-20", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -85376,125 +81049,120 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 2 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "grok-build-0-1": { - "id": "grok-build-0-1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-06-11", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 4, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 4, - "cache_read": 0.4 - } + "input": 0.4, + "output": 2.2, + "cache_read": 0.08 } }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "Qwen 3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2026-06-11", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.75 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "openai-gpt-6-luna": { - "id": "openai-gpt-6-luna", - "name": "GPT-6 Luna", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-luna", + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -85502,107 +81170,92 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.125, - "output": 0.625, - "cache_read": 0.0125, - "cache_write": 0.15625, + "input": 2, + "output": 12, + "cache_read": 0.2, "tiers": [ { - "input": 0.25, - "output": 0.9375, - "cache_read": 0.025, - "cache_write": 0.3125, + "input": 4, + "output": 18, + "cache_read": 0.4, "tier": { "type": "context", - "size": 272000 + "size": 200000 } } ], "context_over_200k": { - "input": 0.25, - "output": 0.9375, - "cache_read": 0.025, - "cache_write": 0.3125 + "input": 4, + "output": 18, + "cache_read": 0.4 } } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k3", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", + "medium", "high", "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-17", + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 128000 }, "cost": { - "input": 3.75, - "output": 18.75, - "cache_read": 0.375 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai-gpt-54-mini": { - "id": "openai-gpt-54-mini", - "name": "GPT-5.4 Mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt", - "attachment": true, + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-27", - "last_updated": "2026-06-11", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -85610,71 +81263,58 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.9375, - "output": 5.625, - "cache_read": 0.09375 + "input": 0.108, + "output": 0.675, + "cache_read": 0.06 } }, - "qwen-3-8-flash": { - "id": "qwen-3-8-flash", - "name": "Qwen 3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "llama-4-maverick-17b-instruct": { + "id": "llama-4-maverick-17b-instruct", + "name": "Llama 4 Maverick 17B Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048576, + "output": 2048 }, "cost": { - "input": 0.14, - "output": 0.49, - "cache_read": 0.014 + "input": 0.27, + "output": 0.85 } }, - "olafangensan-glm-4.7-flash-heretic": { - "id": "olafangensan-glm-4.7-flash-heretic", - "name": "GLM 4.7 Flash Heretic", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", + "qwen3-coder-480b-a35b-instruct": { + "id": "qwen3-coder-480b-a35b-instruct", + "name": "Qwen3-Coder 480B-A35B Instruct", + "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "release_date": "2026-02-04", - "last_updated": "2026-06-11", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -85685,72 +81325,109 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 24000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.035 + "input": 0.3, + "output": 1.3 } }, - "deepseek-v4-flash-0731-fast": { - "id": "deepseek-v4-flash-0731-fast", - "name": "DeepSeek V4 Flash 0731 Fast", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", + "devstral-2512": { + "id": "devstral-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2 + } + }, + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", + "medium", "high", + "xhigh", "max" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-08-09", - "last_updated": "2026-08-11", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.35, - "output": 0.7, - "cache_read": 0.0875 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gemma-4-26b-a4b-it": { + "id": "gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-12-06", - "last_updated": "2026-06-11", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -85760,40 +81437,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 198000, + "context": 262144, "output": 32768 }, "cost": { - "input": 6, - "output": 30, - "cache_read": 0.6, - "cache_write": 7.5 + "input": 0.07, + "output": 0.34 } }, - "zai-org-glm-5-2": { - "id": "zai-org-glm-5-2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text" @@ -85804,44 +81470,30 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.1, + "output": 0.15 } }, - "claude-opus-4-8-fast": { - "id": "claude-opus-4-8-fast", - "name": "Claude Opus 4.8 Fast", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "qwq-plus": { + "id": "qwq-plus", + "name": "QwQ Plus", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-06-11", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-03-05", + "last_updated": "2025-03-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -85849,28 +81501,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 12, - "output": 60, - "cache_read": 1.2, - "cache_write": 15 + "input": 0.8, + "output": 2.4 } }, - "z-ai-glm-5-turbo": { - "id": "z-ai-glm-5-turbo", - "name": "GLM 5 Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "Grok 4.1 Fast Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -85880,31 +81529,32 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-06-11", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 32768 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "zai-org-glm-4.6": { - "id": "zai-org-glm-4.6", - "name": "GLM 4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "glm-5-2": { + "id": "glm-5-2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, @@ -85912,21 +81562,20 @@ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2024-04-01", - "last_updated": "2026-06-11", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -85937,39 +81586,27 @@ }, "open_weights": true, "limit": { - "context": 198000, - "output": 16384 + "context": 1024000, + "output": 131072 }, "cost": { - "input": 0.43, - "output": 1.75, - "cache_read": 0.08 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "openai-gpt-54-pro": { - "id": "openai-gpt-54-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt", + "qwen3-vl-30b-a3b-instruct": { + "id": "qwen3-vl-30b-a3b-instruct", + "name": "Qwen3 VL 30B A3B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-10-02", + "last_updated": "2025-10-02", "modalities": { "input": [ "text", @@ -85979,75 +81616,28 @@ "text" ] }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 37.5, - "output": 225, - "tiers": [ - { - "input": 75, - "output": 337.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 75, - "output": 337.5 - } - } - }, - "aion-labs-aion-3-0": { - "id": "aion-labs-aion-3-0", - "name": "Aion 3.0", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32768 + "context": 131072, + "output": 8192 }, "cost": { - "input": 3.75, - "output": 7.5, - "cache_read": 0.9375 + "input": 0.2, + "output": 0.7 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "qwen3-vl-flash": { + "id": "qwen3-vl-flash", + "name": "Qwen3 VL Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-01-15", - "last_updated": "2026-06-11", + "release_date": "2025-10-09", + "last_updated": "2025-10-09", "modalities": { "input": [ "text", @@ -86059,41 +81649,39 @@ }, "open_weights": false, "limit": { - "context": 198000, - "output": 64000 + "context": 262144, + "output": 32000 }, "cost": { - "input": 3.75, - "output": 18.75, - "cache_read": 0.375, - "cache_write": 4.69 + "input": 0.05, + "output": 0.4, + "cache_read": 0.01 } }, - "claude-opus-5-5": { - "id": "claude-opus-5-5", - "name": "Claude Opus 5.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-18", - "last_updated": "2026-09-22", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -86105,37 +81693,65 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 4.8, - "output": 24, - "cache_read": 0.24, - "cache_write": 6 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "kimi-k2-5": { - "id": "kimi-k2-5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2024-04", - "release_date": "2026-01-27", - "last_updated": "2026-06-11", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 + } + }, + "llama-3.1-nemotron-ultra-253b": { + "id": "llama-3.1-nemotron-ultra-253b", + "name": "Llama 3.1 Nemotron Ultra 253B", + "description": "Flagship Nemotron model for high-throughput reasoning and complex agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "release_date": "2025-04-07", + "last_updated": "2025-04-07", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -86143,40 +81759,27 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.56, - "output": 3.5, - "cache_read": 0.22 + "input": 0.6, + "output": 1.8 } }, - "claude-fable-5-1": { - "id": "claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "gpt-4-turbo": { + "id": "gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-08-29", - "last_updated": "2026-09-01", + "structured_output": false, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ "text", @@ -86188,36 +81791,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 12, - "output": 60, - "cache_read": 0.3, - "cache_write": 15 + "input": 10, + "output": 30 } }, - "gemini-3-7-flash": { - "id": "gemini-3-7-flash", - "name": "Gemini 3.7 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen3-coder-flash": { + "id": "qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -86229,27 +81825,39 @@ "output": 65536 }, "cost": { - "input": 0.9375, - "output": 4.6875, - "cache_read": 0.09375 + "input": 0.3, + "output": 1.5, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "grok-4-20": { - "id": "grok-4-20", - "name": "Grok 4.20", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "grok-build-0-1": { + "id": "grok-build-0-1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "release_date": "2026-03-12", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -86257,18 +81865,18 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 128000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 1.42, - "output": 2.83, - "cache_read": 0.23, + "input": 1, + "output": 2, + "cache_read": 0.2, "tiers": [ { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45, + "input": 2, + "output": 4, + "cache_read": 0.4, "tier": { "type": "context", "size": 200000 @@ -86276,38 +81884,36 @@ } ], "context_over_200k": { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45 + "input": 2, + "output": 4, + "cache_read": 0.4 } } }, - "openai-gpt-56-sol-pro": { - "id": "openai-gpt-56-sol-pro", - "name": "GPT-5.6 Sol Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-sol", + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -86319,52 +81925,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "seed-2-1-turbo": { - "id": "seed-2-1-turbo", - "name": "Seed 2.1 Turbo", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "grok-4": { + "id": "grok-4", + "name": "Grok 4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-28", - "last_updated": "2026-07-24", + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -86373,62 +81959,49 @@ "open_weights": false, "limit": { "context": 256000, - "output": 65536 + "output": 256000 }, "cost": { - "input": 0.625, - "output": 3.125, - "cache_read": 0.125 + "input": 3, + "output": 15, + "cache_read": 0.75 } }, - "z-ai-glm-5v-turbo": { - "id": "z-ai-glm-5v-turbo", - "name": "GLM 5V Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "llama-3.2-11b-instruct": { + "id": "llama-3.2-11b-instruct", + "name": "Llama 3.2 11B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-06-11", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32768 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.5, - "output": 5, - "cache_read": 0.3 + "input": 0.07, + "output": 0.33 } }, - "openai-gpt-55-pro": { - "id": "openai-gpt-55-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt", + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -86442,11 +82015,11 @@ } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-24", - "last_updated": "2026-06-11", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -86458,75 +82031,54 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, "input": 922000, "output": 128000 }, "cost": { - "input": 37.5, - "output": 225 + "input": 30, + "output": 180 } }, - "qwen-3-7-plus": { - "id": "qwen-3-7-plus", - "name": "Qwen 3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-11", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 131000, + "output": 98304 }, "cost": { - "input": 0.5, - "output": 2, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 1.5, - "output": 6, - "cache_read": 0.15, - "cache_write": 1.875, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.5, - "output": 6, - "cache_read": 0.15, - "cache_write": 1.875 - } + "input": 0.13, + "output": 0.85, + "cache_read": 0.025, + "cache_write": 0 } }, - "qwen3-235b-a22b-thinking-2507": { - "id": "qwen3-235b-a22b-thinking-2507", - "name": "Qwen 3 235B A22B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "kimi-k2.7-code-highspeed": { + "id": "kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -86543,11 +82095,15 @@ "field": "reasoning_content" }, "structured_output": true, - "release_date": "2025-04-29", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -86555,86 +82111,97 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.45, - "output": 3.5 + "input": 1.9, + "output": 8, + "cache_read": 0.38 } }, - "zai-org-glm-4.7-flash": { - "id": "zai-org-glm-4.7-flash", - "name": "GLM 4.7 Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-29", - "last_updated": "2026-06-11", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.06, - "output": 0.4, - "cache_read": 0.01 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-23", - "last_updated": "2026-07-24", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -86642,41 +82209,60 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 6, - "output": 30, - "cache_read": 0.6, - "cache_write": 7.5 + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "openai-gpt-6-astra": { - "id": "openai-gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "glm-4.7-flash": { + "id": "glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 200000, + "output": 131072 + }, + "cost": { + "input": 0.06, + "output": 0.4, + "cache_read": 0.01, + "cache_write": 0 + } + }, + "qwen3-vl-235b-a22b-instruct": { + "id": "qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-05", - "last_updated": "2026-09-04", + "temperature": true, + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", @@ -86686,59 +82272,36 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } + "input": 0.3, + "output": 1.5 } }, - "minimax-m27": { - "id": "minimax-m27", - "name": "MiniMax M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, + "knowledge": "2024-12", "release_date": "2026-03-18", - "last_updated": "2026-06-11", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -86747,42 +82310,49 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 198000, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.375, - "output": 1.5, - "cache_read": 0.06875 + "input": 1, + "output": 3, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "mercury-2-5": { - "id": "mercury-2-5", - "name": "Mercury 2.5", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-09", + "gpt-4o-search-preview": { + "id": "gpt-4o-search-preview", + "name": "GPT-4o Search Preview", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -86790,26 +82360,26 @@ }, "open_weights": false, "limit": { - "context": 260000, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.04999999999999999, - "output": 0.18749999999999994, - "cache_read": 0.004999999999999999 + "input": 2.5, + "output": 10 } }, - "qwen3-next-80b": { - "id": "qwen3-next-80b", - "name": "Qwen 3 Next 80b", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "llama-3.2-3b-instruct": { + "id": "llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": true, - "release_date": "2025-04-29", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2024-09-18", + "last_updated": "2024-09-18", "modalities": { "input": [ "text" @@ -86820,29 +82390,24 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 32768, + "output": 32000 }, "cost": { - "input": 0.35, - "output": 1.9 + "input": 0.03, + "output": 0.05 } }, - "minimax-m25": { - "id": "minimax-m25", - "name": "MiniMax M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, @@ -86850,8 +82415,9 @@ "field": "reasoning_content" }, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-06-11", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -86862,63 +82428,107 @@ }, "open_weights": true, "limit": { - "context": 198000, - "output": 32768 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.27, - "output": 0.95, - "cache_read": 0.03 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "qwen3-6-35b-a3b": { - "id": "qwen3-6-35b-a3b", - "name": "Qwen 3.6 35B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "gpt-5.5-pro": { + "id": "gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-07-20", - "last_updated": "2026-07-22", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 1 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "grok-4-20-multi-agent": { - "id": "grok-4-20-multi-agent", - "name": "Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning": false, + "tool_call": true, "structured_output": true, - "release_date": "2026-03-12", - "last_updated": "2026-06-11", + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -86926,36 +82536,20 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.42, - "output": 2.83, - "cache_read": 0.23, - "tiers": [ - { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45 - } + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "zai-org-glm-4.7": { - "id": "zai-org-glm-4.7", - "name": "GLM 4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -86969,14 +82563,10 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-24", - "last_updated": "2026-06-11", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -86985,44 +82575,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 198000, - "output": 16384 + "context": 131072, + "output": 32766 }, "cost": { - "input": 0.55, - "output": 2.65, - "cache_read": 0.11 + "input": 0.04, + "output": 0.15 } }, - "qwen3-5-9b": { - "id": "qwen3-5-9b", - "name": "Qwen 3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": true, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-03-05", - "last_updated": "2026-06-11", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -87030,24 +82614,28 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 32768 + "context": 203000, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.15 + "input": 0.72, + "output": 2.3, + "cache_read": 0.144, + "cache_write": 0 } }, - "qwen3-vl-235b-a22b": { - "id": "qwen3-vl-235b-a22b", - "name": "Qwen3 VL 235B", - "description": "Multimodal model for analyzing text, images, documents, and rich media", + "qwen-max-latest": { + "id": "qwen-max-latest", + "name": "Qwen Max Latest", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, - "release_date": "2026-01-16", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-01-25", + "last_updated": "2025-01-25", "modalities": { "input": [ "text", @@ -87057,42 +82645,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.21, - "output": 1.9, - "cache_read": 0.1 + "input": 1.6, + "output": 6.4 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "mistral-large-2512": { + "id": "mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-10", - "last_updated": "2026-06-11", + "temperature": true, + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", @@ -87102,68 +82676,85 @@ "text" ] }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 1.5 + } + }, + "qwen-turbo": { + "id": "qwen-turbo", + "name": "Qwen Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-11-01", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 16384 }, "cost": { - "input": 12, - "output": 60, - "cache_read": 1.2, - "cache_write": 15 + "input": 0.05, + "output": 0.2, + "reasoning": 0.5 } }, - "qwen3-5-35b-a3b": { - "id": "qwen3-5-35b-a3b", - "name": "Qwen 3.5 35B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "custom": { + "id": "custom", + "name": "Custom Model", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-25", - "last_updated": "2026-06-11", + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, + "context": 128000, "output": 16384 }, "cost": { - "input": 0.3125, - "output": 1.25, - "cache_read": 0.15625 + "input": 0, + "output": 0 } }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", - "attachment": false, + "gpt-5.6-sol": { + "id": "gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -87171,65 +82762,68 @@ "values": [ "none", "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.65, - "output": 4.95, - "cache_read": 0.165 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "openai-gpt-56-terra": { - "id": "openai-gpt-56-terra", - "name": "GPT-5.6 Terra", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-terra", + "grok-4-20-reasoning": { + "id": "grok-4-20-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-03-09", + "last_updated": "2026-03-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -87237,44 +82831,81 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125, + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, "tiers": [ { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25, + "input": 2.5, + "output": 5, + "cache_read": 0.4, "tier": { "type": "context", - "size": 272000 + "size": 200000 } } ], "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 2.5, + "output": 5, + "cache_read": 0.4 } } }, - "llama-3.3-70b": { - "id": "llama-3.3-70b", - "name": "Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 16384 + }, + "cost": { + "input": 0.26, + "output": 0.38, + "cache_read": 0.13 + } + }, + "qwen3.6-max-preview": { + "id": "qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "release_date": "2025-04-06", - "last_updated": "2026-06-11", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text" @@ -87283,29 +82914,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.7, - "output": 2.8 + "input": 1.3, + "output": 7.8, + "cache_read": 0.13, + "cache_write": 1.625 } }, - "openai-gpt-4o-mini-2024-07-18": { - "id": "openai-gpt-4o-mini-2024-07-18", - "name": "GPT-4o Mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt", + "auto": { + "id": "auto", + "name": "Auto Route", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2026-02-28", - "last_updated": "2026-06-11", + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text", @@ -87321,26 +82953,25 @@ "output": 16384 }, "cost": { - "input": 0.1875, - "output": 0.75, - "cache_read": 0.09375 + "input": 0, + "output": 0 } }, - "mistral-small-3-2-24b-instruct": { - "id": "mistral-small-3-2-24b-instruct", - "name": "Mistral Small 3.2 24B Instruct", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, + "codestral-2508": { + "id": "codestral-2508", + "name": "Codestral", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "mistral", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": true, - "release_date": "2026-01-15", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -87352,38 +82983,26 @@ "output": 16384 }, "cost": { - "input": 0.09375, - "output": 0.25 + "input": 0.3, + "output": 0.9 } }, - "inkling": { - "id": "inkling", - "name": "Inkling", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ling", - "attachment": true, + "qwen3-235b-a22b-fp8": { + "id": "qwen3-235b-a22b-fp8", + "name": "Qwen3 235B A22B FP8", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-16", - "last_updated": "2026-07-17", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -87391,118 +83010,104 @@ }, "open_weights": true, "limit": { - "context": 524288, - "output": 65536 + "context": 40960, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 5.0625, - "cache_read": 0.2125 + "input": 0.2, + "output": 0.8 } }, - "z-ai-glm-5-3": { - "id": "z-ai-glm-5-3", - "name": "GLM 5.3", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "fugu-ultra": { + "id": "fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-08-18", - "last_updated": "2026-08-18", + "release_date": "2026-06-22", + "last_updated": "2026-06-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 1000000 }, "cost": { - "input": 1.75, - "output": 5.5, - "cache_read": 0.325 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "google-gemma-4-26b-a4b-it": { - "id": "google-gemma-4-26b-a4b-it", - "name": "Google Gemma 4 26B A4B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.13, - "output": 0.4, - "cache_read": 0.05 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "kimi-k2-6": { - "id": "kimi-k2-6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "qwen-vl-plus": { + "id": "qwen-vl-plus", + "name": "Qwen-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-20", - "last_updated": "2026-06-11", + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-08-15", "modalities": { "input": [ "text", @@ -87512,44 +83117,34 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.75, - "output": 3.5, - "cache_read": 0.16 + "input": 0.21, + "output": 0.64 } }, - "openai-gpt-6-astra-pro": { - "id": "openai-gpt-6-astra-pro", - "name": "GPT-6 Astra Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-09-05", - "last_updated": "2026-09-05", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -87557,38 +83152,38 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 12.5, - "output": 62.5, - "cache_read": 1.25, - "cache_write": 15.625, + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, "tiers": [ { - "input": 25, - "output": 93.75, - "cache_read": 2.5, - "cache_write": 31.25, + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, "tier": { "type": "context", - "size": 272000 + "size": 256000 } } ], "context_over_200k": { - "input": 25, - "output": 93.75, - "cache_read": 2.5, - "cache_write": 31.25 + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 } } }, - "openai-gpt-54": { - "id": "openai-gpt-54", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", "family": "gpt", "attachment": true, "reasoning": true, @@ -87599,17 +83194,16 @@ "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -87621,75 +83215,46 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, - "output": 131072 - }, - "cost": { - "input": 3.13, - "output": 18.8, - "cache_read": 0.313 - } - }, - "gemma-4-uncensored": { - "id": "gemma-4-uncensored", - "name": "Gemma 4 Uncensored", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2026-04-13", - "last_updated": "2026-06-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.1625, - "output": 0.5 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-06-11", + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -87697,32 +83262,59 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 6, + "input": 5, "output": 30, - "cache_read": 0.6, - "cache_write": 7.5 + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } - }, - "minimax-m3-preview": { - "id": "minimax-m3-preview", - "name": "MiniMax M3 Preview", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax-m3", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + } + } + }, + "alibaba-coding-plan-cn": { + "id": "alibaba-coding-plan-cn", + "env": [ + "ALIBABA_CODING_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://coding.dashscope.aliyuncs.com/v1", + "name": "Alibaba Coding Plan (China)", + "doc": "https://help.aliyun.com/zh/model-studio/coding-plan", + "models": { + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-13", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -87730,38 +83322,36 @@ }, "open_weights": true, "limit": { - "context": 524288, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai-gpt-oss-120b": { - "id": "openai-gpt-oss-120b", - "name": "OpenAI GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-11-06", - "last_updated": "2026-06-11", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -87772,39 +83362,33 @@ }, "open_weights": true, "limit": { - "context": 128000, + "context": 202752, "output": 16384 }, "cost": { - "input": 0.07, - "output": 0.3 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-29", - "last_updated": "2026-07-01", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", @@ -87820,74 +83404,69 @@ "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "deepseek-v4-1-flash": { - "id": "deepseek-v4-1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", - "attachment": true, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 65536 }, "cost": { - "input": 0.375, - "output": 1.5, - "cache_read": 0.0075 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "gemini-3-8-flash": { - "id": "gemini-3-8-flash", - "name": "Gemini 3.8 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ @@ -87900,30 +83479,26 @@ "output": 65536 }, "cost": { - "input": 0.9375, - "output": 4.6875, - "cache_read": 0.09375 + "input": 0.1875, + "output": 1.125, + "cache_write": 0.234375 } }, - "qwen-3-6-plus": { - "id": "qwen-3-6-plus", - "name": "Qwen 3.6 Plus Uncensored", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "qwen3-max-2026-01-23": { + "id": "qwen3-max-2026-01-23", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2026-04-06", - "last_updated": "2026-06-11", + "release_date": "2026-01-23", + "last_updated": "2026-01-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -87931,58 +83506,33 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.625, - "output": 3.75, - "cache_read": 0.0625, - "cache_write": 0.78, - "tiers": [ - { - "input": 2.5, - "output": 7.5, - "cache_read": 0.0625, - "cache_write": 0.78, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.0625, - "cache_write": 0.78 - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-06-11", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", @@ -87995,41 +83545,35 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 65536 }, "cost": { - "input": 6, - "output": 30, - "cache_read": 0.6, - "cache_write": 7.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai-gpt-6-sol": { - "id": "openai-gpt-6-sol", - "name": "GPT-6 Sol", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-sol", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", @@ -88039,65 +83583,36 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 12.5, - "cache_read": 0.25, - "cache_write": 3.125, - "tiers": [ - { - "input": 5, - "output": 18.75, - "cache_read": 0.5, - "cache_write": 6.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 18.75, - "cache_read": 0.5, - "cache_write": 6.25 - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "qwen-3-8-27b": { - "id": "qwen-3-8-27b", - "name": "Qwen 3.8 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-17", - "last_updated": "2026-08-18", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -88105,71 +83620,67 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 196608, + "output": 24576 }, "cost": { - "input": 0.45, - "output": 3.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai-gpt-53-codex": { - "id": "openai-gpt-53-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-24", - "last_updated": "2026-06-11", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 2.19, - "output": 17.5, - "cache_read": 0.219 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "aion-labs-aion-3-0-mini": { - "id": "aion-labs-aion-3-0-mini", - "name": "Aion 3.0 Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -88180,42 +83691,38 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 202752, + "output": 16384 }, "cost": { - "input": 0.875, - "output": 1.75, - "cache_read": 0.225 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "grok-4-3": { - "id": "grok-4-3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-18", - "last_updated": "2026-06-11", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -88224,44 +83731,44 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 32000 + "output": 65536 }, "cost": { - "input": 1.42, - "output": 2.83, - "cache_read": 0.23, - "tiers": [ - { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.83, - "output": 5.67, - "cache_read": 0.45 - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "aion-labs-aion-3-5-mini": { - "id": "aion-labs-aion-3-5-mini", - "name": "Aion 3.5 Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, + } + } + }, + "abacus": { + "id": "abacus", + "env": [ + "ABACUS_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://routellm.abacus.ai/v1", + "name": "Abacus", + "doc": "https://abacus.ai/help/api", + "models": { + "o3": { + "id": "o3", + "name": "o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -88269,40 +83776,26 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.875, - "output": 1.75, - "cache_read": 0.225 + "input": 2, + "output": 8 } }, - "openai-gpt-56-terra-pro": { - "id": "openai-gpt-56-terra-pro", - "name": "GPT-5.6 Terra Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-terra", + "route-llm": { + "id": "route-llm", + "name": "Route LLM", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text", @@ -88314,68 +83807,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25 - } + "input": 3, + "output": 15 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini", + "grok-code-fast-1": { + "id": "grok-code-fast-1", + "name": "Grok Code Fast 1", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-19", - "last_updated": "2026-06-11", + "release_date": "2025-09-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -88384,83 +83838,32 @@ "open_weights": false, "limit": { "context": 256000, - "output": 65536 - }, - "cost": { - "input": 0.7, - "output": 3.75, - "cache_read": 0.07 - } - }, - "mercury-2": { - "id": "mercury-2", - "name": "Mercury 2", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-02-20", - "last_updated": "2026-06-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 50000 + "output": 16384 }, "cost": { - "input": 0.3125, - "output": 0.9375, - "cache_read": 0.03125 + "input": 0.2, + "output": 1.5 } }, - "grok-4-6": { - "id": "grok-4-6", - "name": "Grok 4.6", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "gpt-5.3-codex-xhigh": { + "id": "gpt-5.3-codex-xhigh", + "name": "GPT-5.3 Codex XHigh", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-10", - "last_updated": "2026-08-12", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -88468,57 +83871,26 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 200000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2.27, - "output": 6.8, - "cache_read": 0.57, - "tiers": [ - { - "input": 4.53, - "output": 13.6, - "cache_read": 1.13, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4.53, - "output": 13.6, - "cache_read": 1.13 - } + "input": 1.75, + "output": 14 } }, - "zai-org-glm-5-1": { - "id": "zai-org-glm-5-1", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "llama-3.3-70b-versatile": { + "id": "llama-3.3-70b-versatile", + "name": "Llama 3.3 70B Versatile", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-06-11", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -88529,74 +83901,62 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 80000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 1.54, - "output": 4.84, - "cache_read": 0.286 + "input": 0.59, + "output": 0.79 } }, - "xiaomi-mimo-v2-5": { - "id": "xiaomi-mimo-v2-5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-06-11", - "last_updated": "2026-06-11", + "knowledge": "2025-01", + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ "text", "image", "audio", - "video" + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08 + "input": 1.25, + "output": 10 } }, - "openai-gpt-4o-2024-11-20": { - "id": "openai-gpt-4o-2024-11-20", - "name": "GPT-4o", + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2026-02-28", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -88608,42 +83968,32 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 400000, + "output": 128000 }, "cost": { - "input": 3.125, - "output": 12.5 + "input": 1.25, + "output": 10 } }, - "openai-gpt-52": { - "id": "openai-gpt-52", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": false, + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-13", - "last_updated": "2026-06-11", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -88651,36 +84001,34 @@ }, "open_weights": false, "limit": { - "context": 256000, - "input": 272000, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 2.19, - "output": 17.5, - "cache_read": 0.219 + "input": 1, + "output": 5 } }, - "gemini-3-6-flash": { - "id": "gemini-3-6-flash", - "name": "Gemini 3.6 Flash", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-09", - "last_updated": "2026-07-21", + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-05", "modalities": { "input": [ "text", "image", "audio", - "video" + "video", + "pdf" ], "output": [ "text" @@ -88688,44 +84036,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1048576, "output": 65536 }, "cost": { - "input": 0.9375, - "output": 4.6875, - "cache_read": 0.09375 + "input": 0.3, + "output": 2.5 } }, - "gemini-3-1-pro-preview": { - "id": "gemini-3-1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini", + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" @@ -88733,61 +84068,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.5, - "cache_write": 0.5, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 0.5 - } + "input": 1.1, + "output": 4.4 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek", + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-06-11", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -88796,41 +84096,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 131072, + "output": 16384 }, "cost": { - "input": 1.65, - "output": 3.301, - "cache_read": 0.33 + "input": 1.2, + "output": 6 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "o3-pro": { + "id": "o3-pro", + "name": "o3-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", "modalities": { "input": [ "text", @@ -88842,45 +84130,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 3.6, - "output": 18, - "cache_read": 0.36, - "cache_write": 4.5 + "input": 20, + "output": 40 } }, - "claude-opus-5-fast": { - "id": "claude-opus-5-fast", - "name": "Claude Opus 5 Fast", + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "Claude Opus 4.1", "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-23", - "last_updated": "2026-07-24", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -88888,41 +84162,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 12, - "output": 60, - "cache_read": 1.2, - "cache_write": 15 + "input": 15, + "output": 75 } }, - "openai-gpt-55": { - "id": "openai-gpt-55", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "Grok 4.1 Fast (Non-Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-11-17", + "last_updated": "2025-11-17", "modalities": { "input": [ "text", @@ -88934,62 +84192,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, - "output": 131072 + "context": 2000000, + "output": 16384 }, "cost": { - "input": 6.25, - "output": 37.5, - "cache_read": 0.625, - "tiers": [ - { - "input": 12.5, - "output": 56.25, - "cache_read": 1.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 12.5, - "output": 56.25, - "cache_read": 1.25 - } + "input": 0.2, + "output": 0.5 } }, - "openai-gpt-56-sol": { - "id": "openai-gpt-56-sol", - "name": "GPT-5.6 Sol", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-sol", + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -88997,177 +84225,126 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "output": 25 } }, - "qwen3-coder-480b-a35b-instruct-turbo": { - "id": "qwen3-coder-480b-a35b-instruct-turbo", - "name": "Qwen 3 Coder 480B Turbo", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "release_date": "2026-01-27", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.35, - "output": 1.5, - "cache_read": 0.04 + "input": 1.25, + "output": 10 } }, - "z-ai-glm-5-3-flash": { - "id": "z-ai-glm-5-3-flash", - "name": "GLM 5.3 Flash", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", + "gpt-5.1-codex-max": { + "id": "gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-08-26", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 1.25, + "output": 10 } }, - "zai-org-glm-5": { - "id": "zai-org-glm-5", - "name": "GLM 5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "claude-opus-4-20250514": { + "id": "claude-opus-4-20250514", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-06-11", + "release_date": "2025-05-14", + "last_updated": "2025-05-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 198000, + "context": 200000, "output": 32000 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 15, + "output": 75 } }, - "grok-4-7": { - "id": "grok-4-7", - "name": "Grok 4.7", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "gpt-5.3-chat-latest": { + "id": "gpt-5.3-chat-latest", + "name": "GPT-5.3 Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-16", - "last_updated": "2026-09-21", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ "text", @@ -89179,55 +84356,27 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 200000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 2.27, - "output": 6.8, - "cache_read": 0.57, - "tiers": [ - { - "input": 4.53, - "output": 13.6, - "cache_read": 1.13, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4.53, - "output": 13.6, - "cache_read": 1.13 - } + "input": 1.75, + "output": 14 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-04", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ "text" @@ -89236,37 +84385,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 160000, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.33, - "output": 0.48, - "cache_read": 0.16 + "input": 1.1, + "output": 4.4 } }, - "gemini-3-5-flash-lite": { - "id": "gemini-3-5-flash-lite", - "name": "Gemini 3.5 Flash-Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "kimi-k2-turbo-preview": { + "id": "kimi-k2-turbo-preview", + "name": "Kimi K2 Turbo Preview", + "description": "Fast Kimi model for responsive chat, coding help, and agent loops", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-09", - "last_updated": "2026-07-21", + "release_date": "2025-07-08", + "last_updated": "2025-07-08", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -89274,44 +84416,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 256000, + "output": 8192 }, "cost": { - "input": 0.375, - "output": 3.125, - "cache_read": 0.0375 + "input": 0.15, + "output": 8 } }, - "gemini-3-5-flash": { - "id": "gemini-3-5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini", + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-22", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" @@ -89319,85 +84448,61 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 1.55, - "output": 9.45, - "cache_read": 0.155, - "cache_write": 0.086 + "input": 1.75, + "output": 14 } }, - "google-gemma-4-31b-it": { - "id": "google-gemma-4-31b-it", - "name": "Google Gemma 4 31B Instruct", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-03", - "last_updated": "2026-06-11", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.12, - "output": 0.36, - "cache_read": 0.09 + "input": 1.75, + "output": 14 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "grok-4-0709": { + "id": "grok-4-0709", + "name": "Grok 4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text", @@ -89409,69 +84514,60 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "output": 16384 }, "cost": { - "input": 6, - "output": 30, - "cache_read": 0.6, - "cache_write": 7.5 + "input": 3, + "output": 15 } }, - "venice-uncensored-1-2": { - "id": "venice-uncensored-1-2", - "name": "Venice Uncensored 1.2", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "venice", + "claude-sonnet-4-20250514": { + "id": "claude-sonnet-4-20250514", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "release_date": "2026-04-01", - "last_updated": "2026-06-11", + "temperature": true, + "release_date": "2025-05-14", + "last_updated": "2025-05-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.2, - "output": 0.9 + "input": 3, + "output": 15 } }, - "qwen3-5-397b-a17b": { - "id": "qwen3-5-397b-a17b", - "name": "Qwen 3.5 397B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-16", - "last_updated": "2026-06-11", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -89484,69 +84580,27 @@ }, "open_weights": true, "limit": { - "context": 128000, + "context": 262144, "output": 32768 }, "cost": { - "input": 0.75, - "output": 4.5 - } - }, - "llama-3.2-3b": { - "id": "llama-3.2-3b", - "name": "Llama 3.2 3B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-10-03", - "last_updated": "2026-06-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.6, + "output": 3 } }, - "openai-gpt-56-luna-pro": { - "id": "openai-gpt-56-luna-pro", - "name": "GPT-5.6 Luna Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-luna", + "gpt-5.1-chat-latest": { + "id": "gpt-5.1-chat-latest", + "name": "GPT-5.1 Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -89558,220 +84612,160 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, "output": 128000 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 0.3125, - "tiers": [ - { - "input": 0.5, - "output": 2.25, - "cache_read": 0.05, - "cache_write": 0.625, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.5, - "output": 2.25, - "cache_read": 0.05, - "cache_write": 0.625 - } + "input": 1.25, + "output": 10 } }, - "nvidia-nemotron-3-ultra-550b-a55b": { - "id": "nvidia-nemotron-3-ultra-550b-a55b", - "name": "NVIDIA Nemotron 3 Ultra", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, + "gpt-5.2-chat-latest": { + "id": "gpt-5.2-chat-latest", + "name": "GPT-5.2 Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-11", + "knowledge": "2024-09-30", + "release_date": "2026-01-01", + "last_updated": "2026-01-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.625, - "output": 3.125, - "cache_read": 0.1875 + "input": 1.75, + "output": 14 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash 0423", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-06-11", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, + "context": 1047576, "output": 32768 }, "cost": { - "input": 0.138, - "output": 0.275, - "cache_read": 0.028 + "input": 0.1, + "output": 0.4 } }, - "nvidia-nemotron-3-nano-30b-a3b": { - "id": "nvidia-nemotron-3-nano-30b-a3b", - "name": "NVIDIA Nemotron 3 Nano 30B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, + "gpt-4o-2024-11-20": { + "id": "gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-06-11", + "knowledge": "2024-10", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, "output": 16384 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 2.5, + "output": 10 } - } - } - }, - "alibaba-token-plan-cn": { - "id": "alibaba-token-plan-cn", - "env": [ - "ALIBABA_TOKEN_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://token-plan.cn-beijing.maas.aliyuncs.com/compatible-mode/v1", - "name": "Alibaba Token Plan (China)", - "doc": "https://www.alibabacloud.com/help/zh/model-studio/token-plan-overview", - "models": { - "happyhorse-1.1-r2v": { - "id": "happyhorse-1.1-r2v", - "name": "HappyHorse 1.1 Reference-to-Video", - "description": "Video model for reference-guided video generation", + }, + "claude-3-7-sonnet-20250219": { + "id": "claude-3-7-sonnet-20250219", + "name": "Claude Sonnet 3.7", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2026-07-17", - "last_updated": "2026-07-17", + "knowledge": "2024-10-31", + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -89779,102 +84773,66 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2.5, + "output": 15 } }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2, + "output": 8 } }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", "video", + "audio", "pdf" ], "output": [ @@ -89883,96 +84841,94 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 2, + "output": 12 } }, - "happyhorse-1.1-t2v": { - "id": "happyhorse-1.1-t2v", - "name": "HappyHorse 1.1 Text-to-Video", - "description": "Video model for prompt-driven text-to-video generation", - "attachment": false, - "reasoning": false, - "tool_call": false, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2026-07-17", - "last_updated": "2026-07-17", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25 } }, - "qwen-image-2.0": { - "id": "qwen-image-2.0", - "name": "Qwen Image 2.0", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "qwen-2.5-coder-32b": { + "id": "qwen-2.5-coder-32b", + "name": "Qwen 2.5 Coder 32B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2024-11-11", + "last_updated": "2024-11-11", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 0 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.79, + "output": 0.79 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -89980,45 +84936,27 @@ }, "open_weights": false, "limit": { - "context": 202752, - "output": 16384 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 3, + "output": 15 } }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -90028,53 +84966,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 2 } }, - "qwen3.8-max-preview": { - "id": "qwen3.8-max-preview", - "name": "Qwen3.8 Max Preview", - "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", - "family": "qwen", + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-07-19", - "last_updated": "2026-07-19", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -90082,246 +84999,193 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1047576, + "output": 32768 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.4, + "output": 1.6 } }, - "qwen-image-2.0-pro": { - "id": "qwen-image-2.0-pro", - "name": "Qwen Image 2.0 Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 8192, - "output": 0 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.05, + "output": 0.4 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "wan2.7-image": { - "id": "wan2.7-image", - "name": "Wan2.7 Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-05-29", - "last_updated": "2026-05-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, "open_weights": false, "limit": { - "context": 8192, - "output": 0 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15 } }, - "wan2.7-image-pro": { - "id": "wan2.7-image-pro", - "name": "Wan2.7 Image Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": false, + "grok-4-fast-non-reasoning": { + "id": "grok-4-fast-non-reasoning", + "name": "Grok 4 Fast (Non-Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 8192, - "output": 0 + "context": 2000000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 0.5 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", - "video" + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 98304 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.5, + "output": 3 } }, - "happyhorse-1.1-i2v": { - "id": "happyhorse-1.1-i2v", - "name": "HappyHorse 1.1 Image-to-Video", - "description": "Video model for image-to-video generation", + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-07-17", - "last_updated": "2026-07-17", + "knowledge": "2024-04", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.6 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "GPT-5 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 131072 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -90329,135 +85193,104 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.25, + "output": 10 } }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, + "gemini-3.1-flash-lite-preview": { + "id": "gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 1 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "input": 196601, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.75, + "output": 14 } }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -90465,43 +85298,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.25, + "output": 10 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { + "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", + "name": "Llama 4 Maverick 17B 128E Instruct FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -90510,35 +85330,24 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 131072 + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.14, + "output": 0.59 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo": { + "id": "meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo", + "name": "Llama 3.1 405B Instruct Turbo", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -90549,44 +85358,25 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 3.5, + "output": 3.5 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "meta-llama/Meta-Llama-3.1-8B-Instruct": { + "id": "meta-llama/Meta-Llama-3.1-8B-Instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -90597,81 +85387,56 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.02, + "output": 0.05 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "Qwen/qwen3-coder-480b-a35b-instruct": { + "id": "Qwen/qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 131072 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 262144, "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.29, + "output": 1.2 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "Qwen/QwQ-32B": { + "id": "Qwen/QwQ-32B", + "name": "QwQ 32B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-03", - "last_updated": "2025-12-05", + "release_date": "2024-11-28", + "last_updated": "2024-11-28", "modalities": { "input": [ "text" @@ -90682,39 +85447,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 0.4 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ "text" @@ -90725,45 +85477,29 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.13, + "output": 0.6 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -90771,94 +85507,68 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.09, + "output": 0.29 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "Qwen/Qwen2.5-72B-Instruct": { + "id": "Qwen/Qwen2.5-72B-Instruct", + "name": "Qwen 2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.11, + "output": 0.38 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT-OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "max" + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -90866,41 +85576,25 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.08, + "output": 0.44 } - } - } - }, - "ai21": { - "id": "ai21", - "env": [ - "AI21_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.ai21.com/studio/v1", - "name": "AI21 Labs", - "doc": "https://docs.ai21.com/docs/jamba-foundation-models", - "models": { - "jamba-large": { - "id": "jamba-large", - "name": "Jamba Large", - "description": "AI21's hybrid SSM-Transformer long-context model for enterprise agents and grounded generation", - "family": "jamba", + }, + "zai-org/glm-4.7": { + "id": "zai-org/glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-08-22", - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "release_date": "2025-06-01", + "last_updated": "2025-06-01", "modalities": { "input": [ "text" @@ -90911,27 +85605,26 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 4096 + "context": 128000, + "output": 8192 }, "cost": { - "input": 2, - "output": 8 + "input": 0.6, + "output": 2.2 } }, - "jamba-mini": { - "id": "jamba-mini", - "name": "Jamba Mini", - "description": "AI21's efficient, lightweight hybrid SSM-Transformer model for a wide range of tasks", - "family": "jamba", + "zai-org/glm-4.5": { + "id": "zai-org/glm-4.5", + "name": "GLM-4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-08-22", - "release_date": "2026-01-01", - "last_updated": "2026-01-01", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -90942,38 +85635,25 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 4096 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 0.4 + "input": 0.6, + "output": 2.2 } - } - } - }, - "inference": { - "id": "inference", - "env": [ - "INFERENCE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.net/v1", - "name": "Inference", - "doc": "https://inference.net/models", - "models": { - "meta/llama-3.1-8b-instruct": { - "id": "meta/llama-3.1-8b-instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + }, + "zai-org/glm-4.6": { + "id": "zai-org/glm-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2025-03-01", + "last_updated": "2025-03-01", "modalities": { "input": [ "text" @@ -90984,30 +85664,29 @@ }, "open_weights": true, "limit": { - "context": 16000, - "output": 4096 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.025, - "output": 0.025 + "input": 0.6, + "output": 2.2 } }, - "meta/llama-3.2-11b-vision-instruct": { - "id": "meta/llama-3.2-11b-vision-instruct", - "name": "Llama 3.2 11B Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", - "attachment": true, - "reasoning": false, + "zai-org/glm-5": { + "id": "zai-org/glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -91015,26 +85694,26 @@ }, "open_weights": true, "limit": { - "context": 16000, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.055, - "output": 0.055 + "input": 1, + "output": 3.2 } }, - "meta/llama-3.2-1b-instruct": { - "id": "meta/llama-3.2-1b-instruct", - "name": "Llama 3.2 1B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "deepseek-ai/DeepSeek-R1": { + "id": "deepseek-ai/DeepSeek-R1", + "name": "DeepSeek R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -91045,26 +85724,26 @@ }, "open_weights": true, "limit": { - "context": 16000, - "output": 4096 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.01, - "output": 0.01 + "input": 3, + "output": 7 } }, - "meta/llama-3.2-3b-instruct": { - "id": "meta/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "deepseek-ai/DeepSeek-V3.1-Terminus": { + "id": "deepseek-ai/DeepSeek-V3.1-Terminus", + "name": "DeepSeek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2025-06-01", + "last_updated": "2025-06-01", "modalities": { "input": [ "text" @@ -91075,30 +85754,29 @@ }, "open_weights": true, "limit": { - "context": 16000, - "output": 4096 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.02, - "output": 0.02 + "input": 0.27, + "output": 1 } }, - "google/gemma-3": { - "id": "google/gemma-3", - "name": "Google Gemma 3", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": false, + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2025-06-15", + "last_updated": "2025-06-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -91106,26 +85784,26 @@ }, "open_weights": true, "limit": { - "context": 125000, - "output": 4096 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 0.3 + "input": 0.27, + "output": 0.4 } }, - "osmosis/osmosis-structure-0.6b": { - "id": "osmosis/osmosis-structure-0.6b", - "name": "Osmosis Structure 0.6B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "osmosis", + "deepseek/deepseek-v3.1": { + "id": "deepseek/deepseek-v3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -91136,26 +85814,38 @@ }, "open_weights": true, "limit": { - "context": 4000, - "output": 2048 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.5 + "input": 0.55, + "output": 1.66 } - }, - "mistral/mistral-nemo-12b-instruct": { - "id": "mistral/mistral-nemo-12b-instruct", - "name": "Mistral Nemo 12B Instruct", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + } + } + }, + "cloudferro-sherlock": { + "id": "cloudferro-sherlock", + "env": [ + "CLOUDFERRO_SHERLOCK_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api-sherlock.cloudferro.com/openai/v1/", + "name": "CloudFerro Sherlock", + "doc": "https://docs.sherlock.cloudferro.com/", + "models": { + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2024-10-09", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -91166,131 +85856,66 @@ }, "open_weights": true, "limit": { - "context": 16000, - "output": 4096 + "context": 70000, + "output": 70000 }, "cost": { - "input": 0.038, - "output": 0.1 + "input": 2.92, + "output": 2.92 } }, - "qwen/qwen3-embedding-4b": { - "id": "qwen/qwen3-embedding-4b", - "name": "Qwen 3 Embedding 4B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "OpenAI GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32000, - "output": 2048 - }, - "cost": { - "input": 0.01, - "output": 0 - } - }, - "qwen/qwen-2.5-7b-vision-instruct": { - "id": "qwen/qwen-2.5-7b-vision-instruct", - "name": "Qwen 2.5 7B Vision Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 125000, - "output": 4096 - }, - "cost": { - "input": 0.2, - "output": 0.2 - } - } - } - }, - "iflowcn": { - "id": "iflowcn", - "env": [ - "IFLOW_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://apis.iflow.cn/v1", - "name": "iFlow", - "doc": "https://platform.iflow.cn/en/docs", - "models": { - "qwen3-vl-plus": { - "id": "qwen3-vl-plus", - "name": "Qwen3-VL-Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0, - "output": 0 + "input": 2.92, + "output": 2.92 } }, - "glm-4.6": { - "id": "glm-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "speakleash/Bielik-11B-v3.0-Instruct": { + "id": "speakleash/Bielik-11B-v3.0-Instruct", + "name": "Bielik 11B v3.0 Instruct", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2025-11-13", + "knowledge": "2025-03", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ "text" @@ -91299,28 +85924,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 32000, + "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.67, + "output": 0.67 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3-32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "speakleash/Bielik-11B-v2.6-Instruct": { + "id": "speakleash/Bielik-11B-v2.6-Instruct", + "name": "Bielik 11B v2.6 Instruct", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "knowledge": "2025-03", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ "text" @@ -91331,26 +85956,28 @@ }, "open_weights": true, "limit": { - "context": 128000, + "context": 32000, "output": 32000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.67, + "output": 0.67 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3-Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2026-01", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text" @@ -91359,28 +85986,51 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 196000, + "input": 180000, + "output": 16000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.2 } - }, - "deepseek-v3": { - "id": "deepseek-v3", - "name": "DeepSeek-V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + } + } + }, + "ollama-cloud": { + "id": "ollama-cloud", + "env": [ + "OLLAMA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://ollama.com/v1", + "name": "Ollama Cloud", + "doc": "https://docs.ollama.com/cloud", + "models": { + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "deepseek-v4-flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-26", - "last_updated": "2024-12-26", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -91391,27 +86041,22 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 1048576, + "output": 1048576 } }, - "qwen3-235b-a22b-thinking-2507": { - "id": "qwen3-235b-a22b-thinking-2507", - "name": "Qwen3-235B-A22B-Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "minimax-m2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "knowledge": "2025-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -91422,56 +86067,50 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 204800, + "output": 131072 } }, - "qwen3-max-preview": { - "id": "qwen3-max-preview", - "name": "Qwen3-Max-Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "devstral-small-2:24b": { + "id": "devstral-small-2:24b", + "name": "devstral-small-2:24b", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2025-12-09", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 262144, + "output": 262144 } }, - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3-Coder-Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "glm-4.7": { + "id": "glm-4.7", + "name": "glm-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "release_date": "2025-12-22", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -91482,26 +86121,25 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 202752, + "output": 131072 } }, - "qwen3-235b-a22b-instruct": { - "id": "qwen3-235b-a22b-instruct", - "name": "Qwen3-235B-A22B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "cogito-2.1:671b": { + "id": "cogito-2.1:671b", + "name": "cogito-2.1:671b", + "description": "Legacy model retained for compatibility with older integrations", + "family": "cogito", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "release_date": "2025-11-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -91512,26 +86150,22 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 163840, + "output": 32000 }, - "cost": { - "input": 0, - "output": 0 - } + "status": "deprecated" }, - "kimi-k2-0905": { - "id": "kimi-k2-0905", - "name": "Kimi-K2-0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "minimax-m2.1": { + "id": "minimax-m2.1", + "name": "minimax-m2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "release_date": "2025-12-23", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -91540,28 +86174,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 204800, + "output": 131072 } }, - "kimi-k2": { - "id": "kimi-k2", - "name": "Kimi-K2", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "gpt-oss:120b": { + "id": "gpt-oss:120b", + "name": "gpt-oss:120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "release_date": "2025-08-05", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -91570,29 +86208,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 64000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 131072, + "output": 32768 } }, - "qwen3-235b": { - "id": "qwen3-235b", - "name": "Qwen3-235B-A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "nemotron-3-nano:30b": { + "id": "nemotron-3-nano:30b", + "name": "nemotron-3-nano:30b", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "release_date": "2025-12-15", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -91603,30 +86240,24 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 1048576, + "output": 131072 } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "ministral-3:8b": { + "id": "ministral-3:8b", + "name": "ministral-3:8b", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "release_date": "2024-12-01", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -91634,26 +86265,20 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 32000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 262144, + "output": 128000 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek-V3.2-Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "rnj-1:8b": { + "id": "rnj-1:8b", + "name": "rnj-1:8b", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "rnj", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2025-12-06", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -91664,31 +86289,15 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 64000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 32768, + "output": 4096 } - } - } - }, - "minimax-cn-coding-plan": { - "id": "minimax-cn-coding-plan", - "env": [ - "MINIMAX_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://api.minimax.cn/anthropic/v1", - "name": "MiniMax Token Plan (minimax.cn)", - "doc": "https://platform.minimaxi.com/docs/token-plan/intro", - "models": { - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + }, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "kimi-k2.7-code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -91697,14 +86306,15 @@ } ], "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-25", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -91712,28 +86322,28 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 512000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "context": 262144, + "output": 262144 } }, - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "glm-5.1": { + "id": "glm-5.1", + "name": "glm-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "interleaved": { + "field": "reasoning_content" + }, + "release_date": "2026-03-27", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -91744,28 +86354,32 @@ }, "open_weights": true, "limit": { - "context": 204800, + "context": 202752, "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "deepseek-v4-pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -91776,28 +86390,25 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "context": 1048576, + "output": 1048576 } }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "glm-4.6": { + "id": "glm-4.6", + "name": "glm-4.6", + "description": "Legacy model retained for compatibility with older integrations", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2025-09-29", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -91808,28 +86419,23 @@ }, "open_weights": true, "limit": { - "context": 204800, + "context": 202752, "output": 131072 }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } + "status": "deprecated" }, - "MiniMax-M2.7-highspeed": { - "id": "MiniMax-M2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "kimi-k2-thinking", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -91840,28 +86446,27 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 262144 }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } + "status": "deprecated" }, - "MiniMax-M2": { - "id": "MiniMax-M2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "nemotron-3-super": { + "id": "nemotron-3-super", + "name": "nemotron-3-super", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "release_date": "2026-03-11", + "last_updated": "2026-03-12", "modalities": { "input": [ "text" @@ -91872,29 +86477,24 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 + "context": 262144, + "output": 65536 } }, - "MiniMax-M2.5-highspeed": { - "id": "MiniMax-M2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "ministral-3:14b": { + "id": "ministral-3:14b", + "name": "ministral-3:14b", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "release_date": "2024-12-01", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -91902,58 +86502,37 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "context": 262144, + "output": 128000 } - } - } - }, - "thinkingmachines": { - "id": "thinkingmachines", - "env": [ - "TINKER_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://tinker.thinkingmachines.dev/services/tinker-prod/anthropic/api/v1", - "name": "Thinking Machines", - "doc": "https://tinker-docs.thinkingmachines.ai/tinker/compatible-apis/anthropic/", - "models": { - "thinkingmachines/Inkling:peft:262144": { - "id": "thinkingmachines/Inkling:peft:262144", - "name": "Inkling (256K)", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, + }, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -91961,20 +86540,15 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 3.74, - "output": 9.36, - "cache_read": 0.748 + "context": 976000, + "output": 131072 } }, - "thinkingmachines/Inkling": { - "id": "thinkingmachines/Inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", + "minimax-m3": { + "id": "minimax-m3", + "name": "minimax-m3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax-m3", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -91987,19 +86561,20 @@ "low", "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "knowledge": "2025-01", + "release_date": "2026-05-31", + "last_updated": "2026-05-31", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -92007,78 +86582,72 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 65536 - }, - "cost": { - "input": 1.87, - "output": 4.68, - "cache_read": 0.374 + "context": 512000, + "output": 131072 } - } - } - }, - "stepfun-step-plan": { - "id": "stepfun-step-plan", - "env": [ - "STEPFUN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.stepfun.com/step_plan/v1", - "name": "StepFun Step Plan (China)", - "doc": "https://platform.stepfun.com/docs/zh/step-plan/integrations/reasoning-api", - "models": { - "step-5-preview": { - "id": "step-5-preview", - "name": "Step 5 Preview", - "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", - "attachment": true, + }, + "minimax-m2": { + "id": "minimax-m2", + "name": "minimax-m2", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "release_date": "2025-10-23", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "structured_output": true, - "release_date": "2026-09-16", - "last_updated": "2026-09-20", + "open_weights": true, + "limit": { + "context": 204800, + "output": 128000 + }, + "status": "deprecated" + }, + "qwen3-next:80b": { + "id": "qwen3-next:80b", + "name": "qwen3-next:80b", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "release_date": "2025-09-15", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 - } + "context": 262144, + "output": 32768 + }, + "status": "deprecated" }, - "step-router-v1": { - "id": "step-router-v1", - "name": "Step Router v1", - "description": "StepFun routing model that dispatches requests to the appropriate Step model.", + "qwen3-coder:480b": { + "id": "qwen3-coder:480b", + "name": "qwen3-coder:480b", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2025-07-22", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -92087,36 +86656,23 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 262144, + "output": 65536 } }, - "step-3.5-flash-2603": { - "id": "step-3.5-flash-2603", - "name": "Step 3.5 Flash 2603", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "kimi-k2:1t": { + "id": "kimi-k2:1t", + "name": "kimi-k2:1t", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-10", + "release_date": "2025-07-11", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -92127,34 +86683,26 @@ }, "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 - } + "context": 262144, + "output": 262144 + }, + "status": "deprecated" }, - "step-3.5-flash": { - "id": "step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "minimax-m2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -92165,40 +86713,28 @@ }, "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 196608, + "output": 196608 } }, - "step-3.7-flash": { - "id": "step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, + "deepseek-v3.1:671b": { + "id": "deepseek-v3.1:671b", + "name": "deepseek-v3.1:671b", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2025-08-21", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -92206,52 +86742,29 @@ }, "open_weights": true, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 163840, + "output": 163840 } - } - } - }, - "melious": { - "id": "melious", - "env": [ - "MELIOUS_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.melious.ai/v1", - "name": "Melious", - "doc": "https://melious.ai/docs/reference/models", - "models": { - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + }, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "kimi-k2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -92259,43 +86772,30 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.11592, - "output": 0.46368, - "cache_read": 0.023184 + "context": 262144, + "output": 262144 } }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "gemma4:31b": { + "id": "gemma4:31b", + "name": "gemma4:31b", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2025-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -92303,44 +86803,49 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.11592, - "output": 0.2898, - "cache_read": 0.023184 + "context": 262144, + "output": 262144 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "ministral-3:3b": { + "id": "ministral-3:3b", + "name": "ministral-3:3b", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "release_date": "2024-10-22", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "open_weights": true, + "limit": { + "context": 262144, + "output": 128000 + } + }, + "gemma3:12b": { + "id": "gemma3:12b", + "name": "gemma3:12b", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, + "release_date": "2024-12-01", + "last_updated": "2026-01-19", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -92348,33 +86853,24 @@ }, "open_weights": true, "limit": { - "context": 1000000, + "context": 131072, "output": 131072 - }, - "cost": { - "input": 3.1878, - "output": 15.939, - "cache_read": 0.788256 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "gemma3:4b": { + "id": "gemma3:4b", + "name": "gemma3:4b", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, + "release_date": "2024-12-01", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -92382,40 +86878,28 @@ }, "open_weights": true, "limit": { - "context": 203000, + "context": 131072, "output": 131072 - }, - "cost": { - "input": 1.10124, - "output": 3.36168, - "cache_read": 0.266616 } }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "qwen3.5:397b": { + "id": "qwen3.5:397b", + "name": "qwen3.5:397b", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, "interleaved": { - "field": "reasoning_content" + "field": "reasoning_details" }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2026-02-15", + "last_updated": "2026-02-17", "modalities": { "input": [ "text", @@ -92427,45 +86911,29 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.23184, - "output": 1.1592, - "cache_read": 0.011592 + "context": 262144, + "output": 65536 } }, "kimi-k2.6": { "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "name": "kimi-k2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -92473,32 +86941,44 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 262144, + "output": 262144 + } + }, + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "qwen3-coder-next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "release_date": "2026-02-02", + "last_updated": "2026-02-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "cost": { - "input": 0.81144, - "output": 4.0572, - "cache_read": 0.266616 + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "qwen3-vl:235b-instruct": { + "id": "qwen3-vl:235b-instruct", + "name": "qwen3-vl:235b-instruct", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-09-22", + "last_updated": "2026-01-19", "modalities": { "input": [ "text", @@ -92510,42 +86990,25 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 262144, + "output": 131072 }, - "cost": { - "input": 0.5796, - "output": 2.95596, - "cache_read": 0.139104 - } + "status": "deprecated" }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "mistral-large-3:675b": { + "id": "mistral-large-3:675b", + "name": "mistral-large-3:675b", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2025-12-02", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -92553,42 +87016,24 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.1592, - "output": 3.4776, - "cache_read": 0.11592 + "context": 262144, + "output": 262144 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "gemma3:27b": { + "id": "gemma3:27b", + "name": "gemma3:27b", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, + "release_date": "2025-07-27", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -92596,42 +87041,25 @@ }, "open_weights": true, "limit": { - "context": 1000000, + "context": 131072, "output": 131072 - }, - "cost": { - "input": 1.1592, - "output": 4.6368, - "cache_read": 0.2898 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "qwen3-vl:235b": { + "id": "qwen3-vl:235b", + "name": "qwen3-vl:235b", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-09-22", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -92639,40 +87067,27 @@ }, "open_weights": true, "limit": { - "context": 203000, - "output": 131072 + "context": 262144, + "output": 32768 }, - "cost": { - "input": 1.50696, - "output": 4.6368, - "cache_read": 0.370944 - } + "status": "deprecated" }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "nemotron-3-ultra": { + "id": "nemotron-3-ultra", + "name": "nemotron-3-ultra", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -92683,34 +87098,30 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.85472, - "output": 3.70944, - "cache_read": 0.46368 + "context": 262144, + "output": 128000 } }, - "deepseek-r1-0528": { - "id": "deepseek-r1-0528", - "name": "DeepSeek-R1", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", - "attachment": false, + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "gemini-3-flash-preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2026-04-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -92718,40 +87129,30 @@ }, "open_weights": true, "limit": { - "context": 164000, - "output": 32768 - }, - "cost": { - "input": 0.69552, - "output": 2.78208, - "cache_read": 0.185472 + "context": 1048576, + "output": 65536 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", + "gpt-oss:20b": { + "id": "gpt-oss:20b", + "name": "gpt-oss:20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2025-08-05", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -92762,39 +87163,28 @@ }, "open_weights": true, "limit": { - "context": 164000, - "output": 64000 - }, - "cost": { - "input": 0.34776, - "output": 0.5796, - "cache_read": 0.092736 + "context": 131072, + "output": 32768 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", + "glm-5": { + "id": "glm-5", + "name": "glm-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -92805,45 +87195,52 @@ }, "open_weights": true, "limit": { - "context": 1000000, + "context": 202752, "output": 131072 - }, - "cost": { - "input": 1.1592, - "output": 3.4776, - "cache_read": 0.23184 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "deepseek-v3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "release_date": "2025-06-15", + "last_updated": "2026-01-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "open_weights": true, + "limit": { + "context": 163840, + "output": 65536 + } + }, + "devstral-2:123b": { + "id": "devstral-2:123b", + "name": "devstral-2:123b", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "release_date": "2025-12-09", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -92853,119 +87250,90 @@ "limit": { "context": 262144, "output": 262144 - }, - "cost": { - "input": 0.81144, - "output": 3.4776, - "cache_read": 0.220248 } } } }, - "berget": { - "id": "berget", + "cloudflare-ai-gateway": { + "id": "cloudflare-ai-gateway", "env": [ - "BERGET_API_KEY" + "CLOUDFLARE_API_TOKEN", + "CLOUDFLARE_ACCOUNT_ID", + "CLOUDFLARE_GATEWAY_ID" ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.berget.ai/v1", - "name": "Berget.AI", - "doc": "https://api.berget.ai", + "npm": "ai-gateway-provider", + "name": "Cloudflare AI Gateway", + "doc": "https://developers.cloudflare.com/ai-gateway/", "models": { - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "workers-ai/@cf/baai/bge-m3": { + "id": "workers-ai/@cf/baai/bge-m3", + "name": "BGE M3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "bge", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "audio", - "image", - "text", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 8192 + "output": 16384 }, "cost": { - "input": 0.275, - "output": 0.55 + "input": 0.012, + "output": 0 } }, - "Qwen/Qwen3.8-27B-FP8": { - "id": "Qwen/Qwen3.8-27B-FP8", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "workers-ai/@cf/baai/bge-small-en-v1.5": { + "id": "workers-ai/@cf/baai/bge-small-en-v1.5", + "name": "BGE Small EN v1.5", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "bge", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-09-01", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.46, - "output": 3.48 + "input": 0.02, + "output": 0 } }, - "mistralai/Mistral-Small-3.2-24B-Instruct-2506": { - "id": "mistralai/Mistral-Small-3.2-24B-Instruct-2506", - "name": "Mistral Small 3.2 24B Instruct 2506", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "workers-ai/@cf/baai/bge-reranker-base": { + "id": "workers-ai/@cf/baai/bge-reranker-base", + "name": "BGE Reranker Base", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "bge", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-10-01", - "last_updated": "2025-10-01", + "release_date": "2025-04-09", + "last_updated": "2025-04-09", "modalities": { "input": [ "text" @@ -92974,81 +87342,56 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 8192 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.33, - "output": 0.33 + "input": 0.0031, + "output": 0 } }, - "moonshotai/Kimi-K3": { - "id": "moonshotai/Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-27", - "last_updated": "2026-07-28", + "workers-ai/@cf/baai/bge-base-en-v1.5": { + "id": "workers-ai/@cf/baai/bge-base-en-v1.5", + "name": "BGE Base EN v1.5", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "bge", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 327680, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 3, - "output": 15 + "input": 0.067, + "output": 0 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "workers-ai/@cf/baai/bge-large-en-v1.5": { + "id": "workers-ai/@cf/baai/bge-large-en-v1.5", + "name": "BGE Large EN v1.5", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "bge", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ "text" @@ -93057,137 +87400,59 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.54, - "output": 4.84 + "input": 0.2, + "output": 0 } }, - "zai-org/GLM-5.3-Flash": { - "id": "zai-org/GLM-5.3-Flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "workers-ai/@cf/ai4bharat/indictrans2-en-indic-1B": { + "id": "workers-ai/@cf/ai4bharat/indictrans2-en-indic-1B", + "name": "IndicTrans2 EN-Indic 1B", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "indictrans", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-09-01", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, + "context": 128000, "output": 16384 }, "cost": { - "input": 0.29, - "output": 0.58 + "input": 0.34, + "output": 0.34 } - } - } - }, - "snowflake-cortex": { - "id": "snowflake-cortex", - "env": [ - "SNOWFLAKE_ACCOUNT", - "SNOWFLAKE_CORTEX_PAT" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://${SNOWFLAKE_ACCOUNT}.snowflakecomputing.com/api/v2/cortex/v1", - "name": "Snowflake Cortex", - "doc": "https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-rest-api", - "models": { - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], - "tool_call": true, + }, + "workers-ai/@cf/ibm-granite/granite-4.0-h-micro": { + "id": "workers-ai/@cf/ibm-granite/granite-4.0-h-micro", + "name": "IBM Granite 4.0 H Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "granite", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-02-28", "release_date": "2025-10-15", "last_updated": "2025-10-15", "modalities": { "input": [ - "text", - "image", - "pdf" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 16384 - } - }, - "openai-gpt-5.5": { - "id": "openai-gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" ], "output": [ "text" @@ -93195,67 +87460,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "status": "beta" - }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] + "context": 128000, + "output": 16384 }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 + "cost": { + "input": 0.017, + "output": 0.11 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], - "tool_call": true, + "workers-ai/@cf/huggingface/distilbert-sst-2-int8": { + "id": "workers-ai/@cf/huggingface/distilbert-sst-2-int8", + "name": "DistilBERT SST-2 INT8", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "distilbert", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -93263,23 +87489,28 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 128000, "output": 16384 + }, + "cost": { + "input": 0.026, + "output": 0 } }, - "openai-gpt-5-nano": { - "id": "openai-gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "workers-ai/@cf/moonshotai/kimi-k2.5": { + "id": "workers-ai/@cf/moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", "high" @@ -93287,11 +87518,14 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", @@ -93301,27 +87535,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "output": 256000 }, - "status": "beta" + "cost": { + "input": 0.6, + "output": 3, + "cache_read": 0.1 + } }, - "openai-gpt-5": { - "id": "openai-gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", + "workers-ai/@cf/moonshotai/kimi-k2.6": { + "id": "workers-ai/@cf/moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", "high" @@ -93329,11 +87567,14 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", @@ -93343,32 +87584,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "output": 256000 }, - "status": "beta" + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } }, - "openai-gpt-4.1": { - "id": "openai-gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, + "workers-ai/@cf/mistral/mistral-7b-instruct-v0.1": { + "id": "workers-ai/@cf/mistral/mistral-7b-instruct-v0.1", + "name": "Mistral 7B Instruct v0.1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -93376,28 +87616,28 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.11, + "output": 0.19 } }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "workers-ai/@cf/google/gemma-3-12b-it": { + "id": "workers-ai/@cf/google/gemma-3-12b-it", + "name": "Gemma 3 12B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-04-11", + "last_updated": "2025-04-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -93405,66 +87645,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.35, + "output": 0.56 } }, - "mistral-large2": { - "id": "mistral-large2", - "name": "Mistral Large (latest)", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, + "workers-ai/@cf/myshell-ai/melotts": { + "id": "workers-ai/@cf/myshell-ai/melotts", + "name": "MyShell MeloTTS", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "melotts", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - } - }, - "openai-gpt-5.2": { - "id": "openai-gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" @@ -93472,29 +87674,27 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "workers-ai/@cf/openai/gpt-oss-120b": { + "id": "workers-ai/@cf/openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -93502,28 +87702,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.35, + "output": 0.75 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "workers-ai/@cf/openai/gpt-oss-20b": { + "id": "workers-ai/@cf/openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -93531,28 +87730,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.2, + "output": 0.3 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "workers-ai/@cf/mistralai/mistral-small-3.1-24b-instruct": { + "id": "workers-ai/@cf/mistralai/mistral-small-3.1-24b-instruct", + "name": "Mistral Small 3.1 24B Instruct", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-04-11", + "last_updated": "2025-04-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -93560,23 +87759,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.35, + "output": 0.56 } }, - "openai-gpt-5.6-luna": { - "id": "openai-gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, + "workers-ai/@cf/nvidia/nemotron-3-120b-a12b": { + "id": "workers-ai/@cf/nvidia/nemotron-3-120b-a12b", + "name": "Nemotron 3 Super 120B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", "high" @@ -93584,51 +87788,44 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 256000, + "output": 256000 }, - "status": "beta" + "cost": { + "input": 0.5, + "output": 1.5 + } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], - "tool_call": true, + "workers-ai/@cf/pfnet/plamo-embedding-1b": { + "id": "workers-ai/@cf/pfnet/plamo-embedding-1b", + "name": "PLaMo Embedding 1B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "plamo", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -93636,40 +87833,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.019, + "output": 0 } }, - "openai-gpt-5.6-sol": { - "id": "openai-gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "workers-ai/@cf/deepgram/aura-2-es": { + "id": "workers-ai/@cf/deepgram/aura-2-es", + "name": "Deepgram Aura 2 (ES)", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "aura", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -93677,41 +87862,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 16384 }, - "status": "beta" + "cost": { + "input": 0, + "output": 0 + } }, - "openai-gpt-5-mini": { - "id": "openai-gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "workers-ai/@cf/deepgram/aura-2-en": { + "id": "workers-ai/@cf/deepgram/aura-2-en", + "name": "Deepgram Aura 2 (EN)", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "aura", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -93719,25 +87891,25 @@ }, "open_weights": false, "limit": { - "context": 272000, - "input": 272000, - "output": 8192 + "context": 128000, + "output": 16384 }, - "status": "beta" + "cost": { + "input": 0, + "output": 0 + } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek-R1", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", + "workers-ai/@cf/deepgram/nova-3": { + "id": "workers-ai/@cf/deepgram/nova-3", + "name": "Deepgram Nova 3", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "nova", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text" @@ -93746,74 +87918,73 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 32768 + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, + "workers-ai/@cf/zai-org/glm-4.7-flash": { + "id": "workers-ai/@cf/zai-org/glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 16384 + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.06, + "output": 0.4 } }, - "openai-gpt-5.1": { - "id": "openai-gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "workers-ai/@cf/deepseek-ai/deepseek-r1-distill-qwen-32b": { + "id": "workers-ai/@cf/deepseek-ai/deepseek-r1-distill-qwen-32b", + "name": "DeepSeek R1 Distill Qwen 32B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -93821,23 +87992,25 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.5, + "output": 4.88 } }, - "snowflake-llama3.3-70b": { - "id": "snowflake-llama3.3-70b", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "workers-ai/@cf/qwen/qwen3-embedding-0.6b": { + "id": "workers-ai/@cf/qwen/qwen3-embedding-0.6b", + "name": "Qwen3 Embedding 0.6B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text" @@ -93846,42 +88019,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 16384 + }, + "cost": { + "input": 0.012, + "output": 0 } }, - "openai-gpt-5.6-terra": { - "id": "openai-gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "workers-ai/@cf/qwen/qwen3-30b-a3b-fp8": { + "id": "workers-ai/@cf/qwen/qwen3-30b-a3b-fp8", + "name": "Qwen3 30B A3B FP8", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -93889,30 +88050,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 16384 }, - "status": "beta" + "cost": { + "input": 0.051, + "output": 0.34 + } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "workers-ai/@cf/qwen/qwen2.5-coder-32b-instruct": { + "id": "workers-ai/@cf/qwen/qwen2.5-coder-32b-instruct", + "name": "Qwen 2.5 Coder 32B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-04-11", + "last_updated": "2025-04-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -93920,52 +88079,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 16384 }, - "status": "beta", - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "cost": { + "input": 0.66, + "output": 1 } }, - "gemini-3.1-pro": { - "id": "gemini-3.1-pro", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "workers-ai/@cf/qwen/qwq-32b": { + "id": "workers-ai/@cf/qwen/qwq-32b", + "name": "QwQ 32B", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2025-04-11", + "last_updated": "2025-04-11", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -93973,40 +88108,28 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.66, + "output": 1 } }, - "openai-gpt-5.4": { - "id": "openai-gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "workers-ai/@cf/pipecat-ai/smart-turn-v2": { + "id": "workers-ai/@cf/pipecat-ai/smart-turn-v2", + "name": "Pipecat Smart Turn v2", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "smart-turn", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -94014,65 +88137,25 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 16384 }, - "status": "beta", - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "cost": { + "input": 0, + "output": 0 } - } - } - }, - "sarvam": { - "id": "sarvam", - "env": [ - "SARVAM_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.sarvam.ai/v1", - "name": "Sarvam AI", - "doc": "https://docs.sarvam.ai/api-reference-docs/getting-started/models", - "models": { - "sarvam-30b": { - "id": "sarvam-30b", - "name": "Sarvam-30B", - "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", - "family": "sarvam", + }, + "workers-ai/@cf/meta/llama-3.1-8b-instruct": { + "id": "workers-ai/@cf/meta/llama-3.1-8b-instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - null, - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-02-18", - "last_updated": "2026-03-06", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ "text" @@ -94081,37 +88164,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 65536 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.28, + "output": 0.8299999999999998 } }, - "sarvam-105b": { - "id": "sarvam-105b", - "name": "Sarvam-105B", - "description": "Flagship Indian-language reasoning model for enterprise multilingual applications", - "family": "sarvam", + "workers-ai/@cf/meta/m2m100-1.2b": { + "id": "workers-ai/@cf/meta/m2m100-1.2b", + "name": "M2M100 1.2B", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "m2m", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - null, - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-02-18", - "last_updated": "2026-03-06", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ "text" @@ -94120,42 +88193,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.34, + "output": 0.34 } - } - } - }, - "nova": { - "id": "nova", - "env": [ - "NOVA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.nova.amazon.com/v1", - "name": "Nova", - "doc": "https://nova.amazon.com/dev/documentation", - "models": { - "nova-2-pro-v1": { - "id": "nova-2-pro-v1", - "name": "Nova 2 Pro", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + }, + "workers-ai/@cf/meta/llama-3.2-1b-instruct": { + "id": "workers-ai/@cf/meta/llama-3.2-1b-instruct", + "name": "Llama 3.2 1B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2026-01-03", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -94163,45 +88224,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "reasoning": 0 + "input": 0.027, + "output": 0.2 } }, - "nova-2-lite-v1": { - "id": "nova-2-lite-v1", - "name": "Nova 2 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "workers-ai/@cf/meta/llama-3.2-11b-vision-instruct": { + "id": "workers-ai/@cf/meta/llama-3.2-11b-vision-instruct", + "name": "Llama 3.2 11B Vision Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -94209,45 +88253,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "reasoning": 0 + "input": 0.049, + "output": 0.68 } - } - } - }, - "abacus": { - "id": "abacus", - "env": [ - "ABACUS_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://routellm.abacus.ai/v1", - "name": "Abacus", - "doc": "https://abacus.ai/help/api", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + }, + "workers-ai/@cf/meta/llama-4-scout-17b-16e-instruct": { + "id": "workers-ai/@cf/meta/llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout 17B 16E Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -94255,30 +88282,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 0.27, + "output": 0.85 } }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "Grok 4.1 Fast (Non-Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, + "workers-ai/@cf/meta/llama-guard-3-8b": { + "id": "workers-ai/@cf/meta/llama-guard-3-8b", + "name": "Llama Guard 3 8B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2025-11-17", - "last_updated": "2025-11-17", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -94286,26 +88311,25 @@ }, "open_weights": false, "limit": { - "context": 2000000, + "context": 128000, "output": 16384 }, "cost": { - "input": 0.2, - "output": 0.5 + "input": 0.48, + "output": 0.03 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "workers-ai/@cf/meta/llama-3-8b-instruct-awq": { + "id": "workers-ai/@cf/meta/llama-3-8b-instruct-awq", + "name": "Llama 3 8B Instruct AWQ", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ "text" @@ -94316,65 +88340,57 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 2.5, - "output": 7.5 + "input": 0.12, + "output": 0.27 } }, - "gemini-3.1-flash-image-preview": { - "id": "gemini-3.1-flash-image-preview", - "name": "Nano Banana 2 Preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "workers-ai/@cf/meta/llama-3.1-8b-instruct-awq": { + "id": "workers-ai/@cf/meta/llama-3.1-8b-instruct-awq", + "name": "Llama 3.1 8B Instruct AWQ", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.12, + "output": 0.27 } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "workers-ai/@cf/meta/llama-3.3-70b-instruct-fp8-fast": { + "id": "workers-ai/@cf/meta/llama-3.3-70b-instruct-fp8-fast", + "name": "Llama 3.3 70B Instruct FP8 Fast", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -94382,34 +88398,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.29, + "output": 2.25 } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "workers-ai/@cf/meta/llama-3-8b-instruct": { + "id": "workers-ai/@cf/meta/llama-3-8b-instruct", + "name": "Llama 3 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -94417,34 +88427,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.28, + "output": 0.83 } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "workers-ai/@cf/meta/llama-3.1-8b-instruct-fp8": { + "id": "workers-ai/@cf/meta/llama-3.1-8b-instruct-fp8", + "name": "Llama 3.1 8B Instruct FP8", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -94452,66 +88456,57 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.15, + "output": 0.29 } }, - "gemini-2.5-flash-image": { - "id": "gemini-2.5-flash-image", - "name": "Nano Banana", - "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "workers-ai/@cf/meta/llama-2-7b-chat-fp16": { + "id": "workers-ai/@cf/meta/llama-2-7b-chat-fp16", + "name": "Llama 2 7B Chat FP16", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": false, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 30 + "input": 0.56, + "output": 6.67 } }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, + "workers-ai/@cf/meta/llama-3.2-3b-instruct": { + "id": "workers-ai/@cf/meta/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -94523,28 +88518,24 @@ "output": 16384 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.051, + "output": 0.34 } }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "workers-ai/@cf/facebook/bart-large-cnn": { + "id": "workers-ai/@cf/facebook/bart-large-cnn", + "name": "BART Large CNN", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "bart", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "release_date": "2025-04-09", + "last_updated": "2025-04-09", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -94552,32 +88543,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 2, - "output": 6 + "input": 0, + "output": 0 } }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "workers-ai/@cf/aisingapore/gemma-sea-lion-v4-27b-it": { + "id": "workers-ai/@cf/aisingapore/gemma-sea-lion-v4-27b-it", + "name": "Gemma SEA-LION v4 27B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -94585,30 +88572,37 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.35, + "output": 0.56 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", @@ -94620,34 +88614,64 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "claude-3-7-sonnet-20250219": { - "id": "claude-3-7-sonnet-20250219", - "name": "Claude Sonnet 3.7", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16385, + "output": 4096 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 1.25 + } + }, + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -94655,25 +88679,28 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 3, - "output": 15 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "qwen-2.5-coder-32b": { - "id": "qwen-2.5-coder-32b", - "name": "Qwen 2.5 Coder 32B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "openai/gpt-4": { + "id": "openai/gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2024-11-11", - "last_updated": "2024-11-11", + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ "text" @@ -94682,24 +88709,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 8192, "output": 8192 }, "cost": { - "input": 0.79, - "output": 0.79 + "input": 30, + "output": 60 } }, - "o4-mini": { - "id": "o4-mini", + "openai/o4-mini": { + "id": "openai/o4-mini", "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", "family": "o-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": false, @@ -94722,26 +88758,37 @@ }, "cost": { "input": 1.1, - "output": 4.4 + "output": 4.4, + "cache_read": 0.28 } }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": false, + "openai/o3-pro": { + "id": "openai/o3-pro", + "name": "o3-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": false, "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -94753,25 +88800,33 @@ "output": 100000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 20, + "output": 80 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1 Codex", "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -94784,32 +88839,40 @@ "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.18 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -94817,33 +88880,43 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "muse-spark-1.1": { - "id": "muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", - "family": "muse", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -94851,33 +88924,45 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -94889,25 +88974,38 @@ "input": 272000, "output": 128000 }, + "provider": { + "npm": "ai-gateway-provider" + }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "gpt-5.2-chat-latest": { - "id": "gpt-5.2-chat-latest", - "name": "GPT-5.2 Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "openai/o1": { + "id": "openai/o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2026-01-01", - "last_updated": "2026-01-01", + "structured_output": true, + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text", @@ -94919,31 +89017,45 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.75, - "output": 14 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "gpt-5.1-chat-latest": { - "id": "gpt-5.1-chat-latest", - "name": "GPT-5.1 Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -94951,29 +89063,36 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1050000, + "input": 922000, "output": 128000 }, + "provider": { + "npm": "ai-gateway-provider" + }, "cost": { - "input": 1.25, - "output": 10 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -94981,31 +89100,31 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.2, - "output": 6 + "input": 10, + "output": 30 } }, - "claude-opus-4-20250514": { - "id": "claude-opus-4-20250514", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-05-14", - "last_updated": "2025-05-14", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -95013,33 +89132,44 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 15, - "output": 75 + "input": 0.15, + "output": 0.6, + "cache_read": 0.08 } }, - "gemini-3.6-flash": { - "id": "gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" @@ -95047,29 +89177,46 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "ai-gateway-provider" }, "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "kimi-k2-turbo-preview": { - "id": "kimi-k2-turbo-preview", - "name": "Kimi K2 Turbo Preview", - "description": "Fast Kimi model for responsive chat, coding help, and agent loops", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-07-08", - "last_updated": "2025-07-08", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -95077,31 +89224,45 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 8 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "grok-4.3": { - "id": "grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -95109,30 +89270,49 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "grok-4-fast-non-reasoning": { - "id": "grok-4-fast-non-reasoning", - "name": "Grok 4 Fast (Non-Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "anthropic/claude-3.5-haiku": { + "id": "anthropic/claude-3.5-haiku", + "name": "Claude Haiku 3.5 (latest)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "knowledge": "2024-07-31", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -95140,26 +89320,28 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 16384 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 0.5 + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 } }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "anthropic/claude-3.5-sonnet": { + "id": "anthropic/claude-3.5-sonnet", + "name": "Claude Sonnet 3.5 v2", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-04-30", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ "text", @@ -95173,30 +89355,42 @@ "open_weights": false, "limit": { "context": 200000, - "output": 32000 + "output": 8192 }, "cost": { - "input": 15, - "output": 75 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude Sonnet 4 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -95208,63 +89402,83 @@ "output": 64000 }, "cost": { - "input": 5, - "output": 25 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "gemini-3-pro-image-preview": { - "id": "gemini-3-pro-image-preview", - "name": "Nano Banana Pro Preview", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", + "anthropic/claude-opus-4-5": { + "id": "anthropic/claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", + "anthropic/claude-sonnet-4-5": { + "id": "anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" @@ -95272,67 +89486,95 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "gemini-3.1-flash-image": { - "id": "gemini-3.1-flash-image", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 32768 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.5, - "output": 3 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "interleaved": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" @@ -95340,32 +89582,33 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { "input": 2, - "output": 12, - "cache_read": 0.2 + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "anthropic/claude-3-sonnet": { + "id": "anthropic/claude-3-sonnet", + "name": "Claude Sonnet 3", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "knowledge": "2023-08-31", + "release_date": "2024-03-04", + "last_updated": "2024-03-04", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -95373,66 +89616,77 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 32768 + "context": 200000, + "output": 4096 }, "cost": { - "input": 2, - "output": 6 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 0.3 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": false, + "anthropic/claude-opus-4-8": { + "id": "anthropic/claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", + "anthropic/claude-3-opus": { + "id": "anthropic/claude-3-opus", + "name": "Claude Opus 3", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2023-08-31", + "release_date": "2024-02-29", + "last_updated": "2024-02-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -95440,62 +89694,77 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 4096 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "llama-3.3-70b-versatile": { - "id": "llama-3.3-70b-versatile", - "name": "Llama 3.3 70B Versatile", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "anthropic/claude-3-5-haiku": { + "id": "anthropic/claude-3-5-haiku", + "name": "Claude Haiku 3.5 (latest)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2024-07-31", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.59, - "output": 0.79 + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 } }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", + "anthropic/claude-opus-4-1": { + "id": "anthropic/claude-opus-4-1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -95503,30 +89772,43 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 5, - "output": 25 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", - "attachment": false, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -95534,32 +89816,43 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -95567,32 +89860,43 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "o3-pro": { - "id": "o3-pro", - "name": "o3-pro", - "description": "High-effort o3 tier for difficult technical reasoning and careful answers", - "family": "o-pro", + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -95600,33 +89904,51 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 20, - "output": 40 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "anthropic/claude-3-haiku": { + "id": "anthropic/claude-3-haiku", + "name": "Claude Haiku 3", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "knowledge": "2023-08-31", + "release_date": "2024-03-13", + "last_updated": "2024-03-13", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" @@ -95634,34 +89956,44 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 4096 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 + "input": 0.25, + "output": 1.25, + "cache_read": 0.03, + "cache_write": 0.3 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, + "interleaved": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ "text", "image", - "audio" + "pdf" ], "output": [ "text" @@ -95669,369 +90001,442 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "ai-gateway-provider" }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "gemini-3-pro-image": { - "id": "gemini-3-pro-image", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Claude Opus 4 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 32768 + "context": 200000, + "output": 32000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } - }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", + } + } + }, + "moonshotai-cn": { + "id": "moonshotai-cn", + "env": [ + "MOONSHOT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.moonshot.cn/v1", + "name": "Moonshot AI (China)", + "doc": "https://platform.moonshot.cn/docs/api/chat", + "models": { + "kimi-k2.7-code-highspeed": { + "id": "kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code HighSpeed", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 1.9, + "output": 8, + "cache_read": 0.38 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 10, - "output": 50 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "muse-spark-1.2": { - "id": "muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", - "attachment": true, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "kimi-k2-turbo-preview": { + "id": "kimi-k2-turbo-preview", + "name": "Kimi K2 Turbo", + "description": "Fast Kimi model for responsive chat, coding help, and agent loops", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 2.4, + "output": 10, + "cache_read": 0.6 } }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "kimi-k2-0711-preview": { + "id": "kimi-k2-0711-preview", + "name": "Kimi K2 0711", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-07-14", + "last_updated": "2025-07-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "gpt-5.1-codex-max": { - "id": "gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", - "attachment": true, + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "kimi-k2-thinking-turbo": { + "id": "kimi-k2-thinking-turbo", + "name": "Kimi K2 Thinking Turbo", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 25 + "input": 1.15, + "output": 8, + "cache_read": 0.15 } }, - "grok-code-fast-1": { - "id": "grok-code-fast-1", - "name": "Grok Code Fast 1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, + "kimi-k2-0905-preview": { + "id": "kimi-k2-0905-preview", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-09-01", - "last_updated": "2025-09-01", + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.2, - "output": 1.5 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + } + } + }, + "morph": { + "id": "morph", + "env": [ + "MORPH_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.morphllm.com/v1", + "name": "Morph", + "doc": "https://docs.morphllm.com/api-reference/introduction", + "models": { + "morph-v3-fast": { + "id": "morph-v3-fast", + "name": "Morph v3 Fast", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "morph", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "release_date": "2024-08-15", + "last_updated": "2024-08-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -96039,32 +90444,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 16000, + "output": 16000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.8, + "output": 1.2 } }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "morph-v3-large": { + "id": "morph-v3-large", + "name": "Morph v3 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "morph", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-08-15", + "last_updated": "2024-08-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -96072,31 +90473,28 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 32000, + "output": 32000 }, "cost": { - "input": 3, - "output": 15 + "input": 0.9, + "output": 1.9 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, + "auto": { + "id": "auto", + "name": "Auto", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "tool_call": false, + "temperature": false, + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -96104,28 +90502,47 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 32000, + "output": 32000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.85, + "output": 1.55 } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + } + } + }, + "sakana": { + "id": "sakana", + "env": [ + "SAKANA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.sakana.ai/v1", + "name": "Sakana AI", + "doc": "https://console.sakana.ai/models", + "models": { + "fugu-ultra-20260615": { + "id": "fugu-ultra-20260615", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ "text", @@ -96138,26 +90555,54 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 1000000 + }, + "provider": { + "shape": "responses" }, "cost": { - "input": 3, - "output": 15 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "fugu": { + "id": "fugu", + "name": "Fugu", + "description": "Multi-agent model for routing expert agents across complex analytical tasks", + "family": "fugu", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "structured_output": true, + "temperature": false, + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ "text", @@ -96170,68 +90615,37 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 1000000 }, - "cost": { - "input": 5, - "output": 25 + "provider": { + "shape": "responses" } }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", + "fugu-ultra": { + "id": "fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "gpt-5.3-codex-xhigh": { - "id": "gpt-5.3-codex-xhigh", - "name": "GPT-5.3 Codex XHigh", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2026-06-15", + "last_updated": "2026-06-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -96239,256 +90653,314 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 1000000 + }, + "provider": { + "shape": "responses" }, "cost": { - "input": 1.75, - "output": 14 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } - }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + } + } + }, + "deepinfra": { + "id": "deepinfra", + "env": [ + "DEEPINFRA_API_KEY" + ], + "npm": "@ai-sdk/deepinfra", + "name": "Deep Infra", + "doc": "https://deepinfra.com/models", + "models": { + "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { + "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", + "name": "Llama 4 Maverick 17B FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 16384 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 0.15, + "output": 0.6 } }, - "gpt-4o-2024-11-20": { - "id": "gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "meta-llama/Llama-4-Scout-17B-16E-Instruct": { + "id": "meta-llama/Llama-4-Scout-17B-16E-Instruct", + "name": "Llama 4 Scout 17B", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", + "structured_output": true, + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 327680, "output": 16384 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.1, + "output": 0.3 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "meta-llama/Llama-3.3-70B-Instruct-Turbo": { + "id": "meta-llama/Llama-3.3-70B-Instruct-Turbo", + "name": "Llama 3.3 70B Turbo", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": true, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 1, - "output": 5 + "input": 0.1, + "output": 0.32 } }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.75, + "output": 3.5, + "cache_read": 0.15 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15 + "input": 0.45, + "output": 2.25, + "cache_read": 0.07 } }, - "gpt-5.3-chat-latest": { - "id": "gpt-5.3-chat-latest", - "name": "GPT-5.3 Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.74, + "output": 3.5, + "cache_read": 0.15 } }, - "claude-sonnet-4-20250514": { - "id": "claude-sonnet-4-20250514", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-05-14", - "last_updated": "2025-05-14", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 3, - "output": 15 + "input": 0.13, + "output": 0.38 } }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", + "google/gemma-4-26B-A4B-it": { + "id": "google/gemma-4-26B-A4B-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -96498,68 +90970,64 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 0.07, + "output": 0.34 } }, - "o3": { - "id": "o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 81920 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.15, + "output": 0.95 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen3.7-Max": { + "id": "Qwen/Qwen3.7-Max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -96567,103 +91035,116 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "output": 65536 }, "cost": { - "input": 5, - "output": 25 + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "tiers": [ + { + "input": 5, + "output": 15, + "cache_read": 1, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 6.25, + "output": 18.5, + "cache_read": 1.25, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.32, + "output": 3.2 } }, - "grok-4.6": { - "id": "grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen3-Next-80B-A3B-Instruct": { + "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, + "context": 262144, "output": 32768 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 0.09, + "output": 1.1 } }, - "gemini-3.1-flash-lite-preview": { - "id": "gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen3-Max": { + "id": "Qwen/Qwen3-Max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -96671,158 +91152,185 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 256000, "output": 65536 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 1 + "input": 1.2, + "output": 6, + "cache_read": 0.24, + "tiers": [ + { + "input": 2.4, + "output": 12, + "cache_read": 0.48, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 3, + "output": 15, + "cache_read": 0.6, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "route-llm": { - "id": "route-llm", - "name": "RouteLLM", - "description": "RouteLLM routes prompts to an appropriate Abacus-backed text-generation model", - "family": "gpt", + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen 3.5 397B A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2026-07-10", + "knowledge": "2025-01", + "release_date": "2026-02-01", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 64000 + "context": 262144, + "output": 81920 }, "cost": { - "input": 3, - "output": 15 + "input": 0.45, + "output": 3, + "cache_read": 0.22 } }, - "grok-4-0709": { - "id": "grok-4-0709", - "name": "Grok 4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "Qwen/Qwen3.5-122B-A10B": { + "id": "Qwen/Qwen3.5-122B-A10B", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 3, - "output": 15 + "input": 0.29, + "output": 2.4 } }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + "Qwen/Qwen3.5-27B": { + "id": "Qwen/Qwen3.5-27B", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.26, + "output": 2.6 } }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "Qwen/Qwen3.5-9B": { + "id": "Qwen/Qwen3.5-9B", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 262144, "output": 65536 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025 + "input": 0.1, + "output": 0.15 } }, - "deepseek/deepseek-v3.1": { - "id": "deepseek/deepseek-v3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -96833,26 +91341,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 40960, + "output": 16384 }, "cost": { - "input": 0.55, - "output": 1.66 + "input": 0.08, + "output": 0.28 } }, - "meta-llama/Meta-Llama-3.3-70B-Instruct": { - "id": "meta-llama/Meta-Llama-3.3-70B-Instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo", + "name": "Qwen3 Coder 480B A35B Instruct Turbo", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -96863,28 +91372,34 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 262144, + "output": 66536 }, "cost": { - "input": 0.59, - "output": 0.79 + "input": 0.3, + "output": 1, + "cache_read": 0.1 } }, - "meta-llama/Meta-Llama-3.1-8B-Instruct": { - "id": "meta-llama/Meta-Llama-3.1-8B-Instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "Qwen/Qwen3.5-35B-A3B": { + "id": "Qwen/Qwen3.5-35B-A3B", + "name": "Qwen 3.5 35B A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-01", + "release_date": "2026-02-01", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -96892,25 +91407,37 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 81920 }, "cost": { - "input": 0.02, - "output": 0.05 + "input": 0.14, + "output": 1, + "cache_read": 0.05 } }, - "meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo": { - "id": "meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo", - "name": "Llama 3.1 405B Instruct Turbo", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -96921,30 +91448,39 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 131072, + "output": 16384 }, "cost": { - "input": 3.5, - "output": 3.5 + "input": 0.037, + "output": 0.17 } }, - "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { - "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", - "name": "Llama 4 Maverick 17B Instruct", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", - "attachment": true, - "reasoning": false, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -96952,30 +91488,41 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 8192 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.14, - "output": 0.59 + "input": 0.03, + "output": 0.14 } }, - "thinkingmachines/Inkling": { - "id": "thinkingmachines/Inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", + "XiaomiMiMo/MiMo-V2.5": { + "id": "XiaomiMiMo/MiMo-V2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -96984,31 +91531,39 @@ "open_weights": true, "limit": { "context": 262144, - "output": 131072 + "output": 16384 }, "cost": { - "input": 3.74, - "output": 9.36, - "cache_read": 0.748 + "input": 0.4, + "output": 2, + "cache_read": 0.08 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "XiaomiMiMo/MiMo-V2.5-Pro": { + "id": "XiaomiMiMo/MiMo-V2.5-Pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "audio" ], "output": [ "text" @@ -97016,29 +91571,34 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 1048576, + "output": 16384 }, "cost": { - "input": 0.14, - "output": 0.4 + "input": 1, + "output": 3, + "cache_read": 0.2 } }, - "Qwen/QwQ-32B": { - "id": "Qwen/QwQ-32B", - "name": "QwQ 32B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning": { + "id": "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning", + "name": "Nemotron 3 Nano Omni 30B A3B Reasoning", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-11-28", - "last_updated": "2024-11-28", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -97046,26 +91606,30 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 0.4 + "input": 0.2, + "output": 0.8 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "nvidia/Nemotron-3-Nano-30B-A3B": { + "id": "nvidia/Nemotron-3-Nano-30B-A3B", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "release_date": "2025-07-01", - "last_updated": "2025-07-01", + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ "text" @@ -97077,24 +91641,26 @@ "open_weights": true, "limit": { "context": 262144, - "output": 8192 + "output": 262144 }, "cost": { - "input": 0.13, - "output": 0.6 + "input": 0.05, + "output": 0.2 } }, - "Qwen/Qwen2.5-72B-Instruct": { - "id": "Qwen/Qwen2.5-72B-Instruct", - "name": "Qwen 2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "nvidia/Llama-3.3-Nemotron-Super-49B-v1.5": { + "id": "nvidia/Llama-3.3-Nemotron-Super-49B-v1.5", + "name": "Llama 3.3 Nemotron Super 49B v1.5", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ "text" @@ -97105,27 +91671,31 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.11, - "output": 0.38 + "input": 0.4, + "output": 0.4 } }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", + "zai-org/GLM-4.7-Flash": { + "id": "zai-org/GLM-4.7-Flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -97136,26 +91706,36 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 202752, + "output": 16384 }, "cost": { - "input": 0.09, - "output": 0.29 + "input": 0.06, + "output": 0.4, + "cache_read": 0.01 } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "name": "Qwen3-Coder 480B-A35B Instruct", - "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", - "family": "qwen", + "zai-org/GLM-4.6": { + "id": "zai-org/GLM-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -97166,27 +91746,76 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 0.43, + "output": 1.74, + "cache_read": 0.08 + } + }, + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-12", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, "output": 16384 }, "cost": { - "input": 0.29, - "output": 1.2 + "input": 0.6, + "output": 2.08, + "cache_read": 0.12 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "zai-org/GLM-4.7": { + "id": "zai-org/GLM-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -97197,26 +91826,119 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 8192 + "context": 202752, + "output": 16384 }, "cost": { - "input": 0.32, - "output": 3.2 + "input": 0.4, + "output": 1.75, + "cache_read": 0.08 } }, - "deepseek-ai/DeepSeek-R1": { - "id": "deepseek-ai/DeepSeek-R1", - "name": "DeepSeek R1", + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1048576, + "output": 32768 + }, + "cost": { + "input": 0.93, + "output": 3, + "cache_read": 0.18 + } + }, + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 16384 + }, + "cost": { + "input": 1.05, + "output": 3.5, + "cache_read": 0.205 + } + }, + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", + "name": "DeepSeek-R1-0528", "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "knowledge": "2024-07", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -97225,14 +91947,15 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 163840, + "output": 64000 }, "cost": { - "input": 3, - "output": 7 + "input": 0.5, + "output": 2.15, + "cache_read": 0.35 } }, "deepseek-ai/DeepSeek-V4-Flash": { @@ -97242,8 +91965,24 @@ "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, "knowledge": "2025-05", @@ -97259,13 +91998,13 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 1048576, + "output": 16384 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.03 + "input": 0.09, + "output": 0.18, + "cache_read": 0.018 } }, "deepseek-ai/DeepSeek-V4-Pro": { @@ -97275,8 +92014,24 @@ "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, "knowledge": "2025-05", @@ -97292,27 +92047,35 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 1048576, + "output": 16384 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.15 + "input": 1.3, + "output": 2.6, + "cache_read": 0.1 } }, "deepseek-ai/DeepSeek-V3.2": { "id": "deepseek-ai/DeepSeek-V3.2", - "name": "DeepSeek V3.2", + "name": "DeepSeek-V3.2", "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-06-15", - "last_updated": "2025-06-15", + "knowledge": "2024-12", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text" @@ -97321,28 +92084,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 163840, + "output": 64000 }, "cost": { - "input": 0.27, - "output": 0.4 + "input": 0.26, + "output": 0.38, + "cache_read": 0.13 } }, - "deepseek-ai/DeepSeek-V3.1-Terminus": { - "id": "deepseek-ai/DeepSeek-V3.1-Terminus", - "name": "DeepSeek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-06-01", - "last_updated": "2025-06-01", + "knowledge": "2025-06", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -97353,12 +92121,13 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 196608, + "output": 131072 }, "cost": { - "input": 0.27, - "output": 1 + "input": 0.15, + "output": 1.15, + "cache_read": 0.03 } }, "MiniMaxAI/MiniMax-M3": { @@ -97385,12 +92154,13 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 524288, + "output": 128000 }, "cost": { "input": 0.3, - "output": 1.2 + "output": 1.2, + "cache_read": 0.06 } }, "MiniMaxAI/MiniMax-M2.7": { @@ -97415,572 +92185,763 @@ }, "open_weights": true, "limit": { - "context": 204800, + "context": 196608, "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.25, + "output": 1, + "cache_read": 0.05 } - }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + } + } + }, + "google-vertex-anthropic": { + "id": "google-vertex-anthropic", + "env": [ + "GOOGLE_VERTEX_PROJECT", + "GOOGLE_VERTEX_LOCATION", + "GOOGLE_APPLICATION_CREDENTIALS" + ], + "npm": "@ai-sdk/google-vertex/anthropic", + "name": "Vertex (Anthropic)", + "doc": "https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude", + "models": { + "claude-haiku-4-5@20251001": { + "id": "claude-haiku-4-5@20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "moonshotai/Kimi-K3": { - "id": "moonshotai/Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "claude-opus-4@20250514": { + "id": "claude-opus-4@20250514", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 32000 }, + "status": "deprecated", "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "claude-opus-4-1@20250805": { + "id": "claude-opus-4-1@20250805", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 32000 }, + "status": "deprecated", "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "claude-opus-4-5@20251101": { + "id": "claude-opus-4-5@20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "claude-3-5-haiku@20241022": { + "id": "claude-3-5-haiku@20241022", + "name": "Claude Haiku 3.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-07-31", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 } }, - "zai-org/GLM-4.7": { - "id": "zai-org/GLM-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, + "claude-sonnet-4@20250514": { + "id": "claude-sonnet-4@20250514", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "output": 64000 }, + "status": "deprecated", "cost": { - "input": 0.6, - "output": 2.2 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, + "claude-opus-4-7@default": { + "id": "claude-opus-4-7@default", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 3.2 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "zai-org/GLM-4.6": { - "id": "zai-org/GLM-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, + "claude-sonnet-4-5@20250929": { + "id": "claude-sonnet-4-5@20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "zai-org/GLM-4.5": { - "id": "zai-org/GLM-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, + "claude-sonnet-5@default": { + "id": "claude-sonnet-5@default", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 96000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "claude-opus-4-6@default": { + "id": "claude-opus-4-6@default", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.08, - "output": 0.44 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } - } - } - }, - "novita-ai": { - "id": "novita-ai", - "env": [ - "NOVITA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.novita.ai/openai", - "name": "NovitaAI", - "doc": "https://novita.ai/docs/guides/introduction", - "models": { - "kwaipilot/kat-coder-pro": { - "id": "kwaipilot/kat-coder-pro", - "name": "Kat Coder Pro", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, - "reasoning": false, + }, + "claude-opus-4-8@default": { + "id": "claude-opus-4-8@default", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01-05", - "last_updated": "2026-01-05", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "deepseek/deepseek-prover-v2-671b": { - "id": "deepseek/deepseek-prover-v2-671b", - "name": "Deepseek Prover V2 671B", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "attachment": false, - "reasoning": false, - "tool_call": false, + "claude-sonnet-4-6@default": { + "id": "claude-sonnet-4-6@default", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-04-30", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 160000, - "output": 160000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.7, - "output": 2.5 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } - }, - "deepseek/deepseek-r1-distill-qwen-14b": { - "id": "deepseek/deepseek-r1-distill-qwen-14b", - "name": "DeepSeek R1 Distill Qwen 14B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": false, - "tool_call": false, + } + } + }, + "v0": { + "id": "v0", + "env": [ + "V0_API_KEY" + ], + "npm": "@ai-sdk/vercel", + "name": "v0", + "doc": "https://sdk.vercel.ai/providers/ai-sdk-providers/vercel", + "models": { + "v0-1.0-md": { + "id": "v0-1.0-md", + "name": "v0-1.0-md", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "v0", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 128000, + "output": 32000 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 3, + "output": 15 } }, - "deepseek/deepseek-v3-0324": { - "id": "deepseek/deepseek-v3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-03-25", - "last_updated": "2025-03-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 163840 - }, - "cost": { - "input": 0.27, - "output": 1.12, - "cache_read": 0.135 - } - }, - "deepseek/deepseek-r1-distill-llama-70b": { - "id": "deepseek/deepseek-r1-distill-llama-70b", - "name": "DeepSeek R1 Distill LLama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "v0-1.5-lg": { + "id": "v0-1.5-lg", + "name": "v0-1.5-lg", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "v0", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-01-27", - "last_updated": "2025-01-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.8, - "output": 0.8 - } - }, - "deepseek/deepseek-v3-turbo": { - "id": "deepseek/deepseek-v3-turbo", - "name": "DeepSeek V3 (Turbo)\t", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "attachment": false, - "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-03-05", - "last_updated": "2025-03-05", + "release_date": "2025-06-09", + "last_updated": "2025-06-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 64000, - "output": 16000 + "context": 512000, + "output": 32000 }, "cost": { - "input": 0.4, - "output": 1.3 + "input": 15, + "output": 75 } }, - "deepseek/deepseek-r1-0528-qwen3-8b": { - "id": "deepseek/deepseek-r1-0528-qwen3-8b", - "name": "DeepSeek R1 0528 Qwen3 8B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "attachment": false, + "v0-1.5-md": { + "id": "v0-1.5-md", + "name": "v0-1.5-md", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "v0", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-05-29", - "last_updated": "2025-05-29", + "release_date": "2025-06-09", + "last_updated": "2025-06-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, "output": 32000 }, "cost": { - "input": 0.06, - "output": 0.09 + "input": 3, + "output": 15 } - }, - "deepseek/deepseek-v3.1-terminus": { - "id": "deepseek/deepseek-v3.1-terminus", - "name": "Deepseek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + } + } + }, + "azure": { + "id": "azure", + "env": [ + "AZURE_RESOURCE_NAME", + "AZURE_API_KEY" + ], + "npm": "@ai-sdk/azure", + "name": "Azure", + "doc": "https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models", + "models": { + "codex-mini": { + "id": "codex-mini", + "name": "Codex Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "temperature": false, + "knowledge": "2024-04", + "release_date": "2025-05-16", + "last_updated": "2025-05-16", "modalities": { "input": [ "text" @@ -97989,30 +92950,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.27, - "output": 1, - "cache_read": 0.135 + "input": 1.5, + "output": 6, + "cache_read": 0.375 } }, - "deepseek/deepseek-ocr-2": { - "id": "deepseek/deepseek-ocr-2", - "name": "deepseek/deepseek-ocr-2", - "description": "OCR model for extracting structured text from documents and screenshots", - "attachment": true, + "phi-3.5-moe-instruct": { + "id": "phi-3.5-moe-instruct", + "name": "Phi-3.5-MoE-instruct", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", + "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "temperature": true, + "knowledge": "2023-10", + "release_date": "2024-08-20", + "last_updated": "2024-08-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -98020,55 +92983,57 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.03, - "output": 0.03 + "input": 0.16, + "output": 0.64 } }, - "deepseek/deepseek-ocr": { - "id": "deepseek/deepseek-ocr", - "name": "DeepSeek-OCR", - "description": "OCR model for extracting structured text from documents and screenshots", - "attachment": true, + "gpt-3.5-turbo-instruct": { + "id": "gpt-3.5-turbo-instruct", + "name": "GPT-3.5 Turbo Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, "temperature": true, - "release_date": "2025-10-24", - "last_updated": "2025-10-24", + "knowledge": "2021-08", + "release_date": "2023-09-21", + "last_updated": "2023-09-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 4096, + "output": 4096 }, "cost": { - "input": 0.03, - "output": 0.03 + "input": 1.5, + "output": 2 } }, - "deepseek/deepseek-r1-turbo": { - "id": "deepseek/deepseek-r1-turbo", - "name": "DeepSeek R1 (Turbo)\t", + "deepseek-r1-0528": { + "id": "deepseek-r1-0528", + "name": "DeepSeek-R1-0528", "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-03-05", - "last_updated": "2025-03-05", + "knowledge": "2024-07", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -98079,30 +93044,28 @@ }, "open_weights": true, "limit": { - "context": 64000, - "output": 16000 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0.7, - "output": 2.5 + "input": 1.35, + "output": 5.4 } }, - "deepseek/deepseek-v3.2-exp": { - "id": "deepseek/deepseek-v3.2-exp", - "name": "Deepseek V3.2 Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek-V4-Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -98113,142 +93076,114 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 65536 - }, - "cost": { - "input": 0.27, - "output": 0.41 - } - }, - "deepseek/deepseek-r1-distill-qwen-32b": { - "id": "deepseek/deepseek-r1-distill-qwen-32b", - "name": "DeepSeek R1 Distill Qwen 32B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-01-20", - "last_updated": "2025-01-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 384000 }, - "open_weights": true, - "limit": { - "context": 64000, - "output": 32000 + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 0.3, - "output": 0.3 + "input": 0.19, + "output": 0.51 } }, - "deepseek/deepseek-v3.1": { - "id": "deepseek/deepseek-v3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "gpt-5.2-chat": { + "id": "gpt-5.2-chat", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "medium" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.27, - "output": 1, - "cache_read": 0.135 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "o3": { + "id": "o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 393216 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.6, - "output": 3.2, - "cache_read": 0.135 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "deepseek/deepseek-r1-0528": { - "id": "deepseek/deepseek-r1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "deepseek-v3-0324": { + "id": "deepseek-v3-0324", + "name": "DeepSeek-V3-0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2024-07", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ "text" @@ -98259,35 +93194,26 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 32768 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.7, - "output": 2.5, - "cache_read": 0.35 + "input": 1.14, + "output": 4.56 } }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "Deepseek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "phi-3-small-128k-instruct": { + "id": "phi-3-small-128k-instruct", + "name": "Phi-3-small-instruct (128k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -98298,46 +93224,26 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 65536 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.269, - "output": 0.4, - "cache_read": 0.1345 + "input": 0.15, + "output": 0.6 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "meta-llama-3-8b-instruct": { + "id": "meta-llama-3-8b-instruct", + "name": "Meta-Llama-3-8B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-12", + "release_date": "2024-04-18", + "last_updated": "2024-04-18", "modalities": { "input": [ "text" @@ -98348,25 +93254,26 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 8192, + "output": 2048 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0.3, + "output": 0.61 } }, - "meta-llama/llama-4-maverick-17b-128e-instruct-fp8": { - "id": "meta-llama/llama-4-maverick-17b-128e-instruct-fp8", - "name": "Llama 4 Maverick Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", + "mistral-small-2503": { + "id": "mistral-small-2503", + "name": "Mistral Small 3.1", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-04-06", - "last_updated": "2025-04-06", + "knowledge": "2024-09", + "release_date": "2025-03-01", + "last_updated": "2025-03-01", "modalities": { "input": [ "text", @@ -98376,28 +93283,26 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 8192 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.27, - "output": 0.85 + "input": 0.1, + "output": 0.3 } }, - "meta-llama/llama-3-70b-instruct": { - "id": "meta-llama/llama-3-70b-instruct", - "name": "Llama3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "text-embedding-3-large": { + "id": "text-embedding-3-large", + "name": "text-embedding-3-large", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2024-04-25", - "last_updated": "2024-04-25", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -98406,28 +93311,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8000 + "context": 8191, + "output": 3072 }, "cost": { - "input": 0.51, - "output": 0.74 + "input": 0.13, + "output": 0 } }, - "meta-llama/llama-3.3-70b-instruct": { - "id": "meta-llama/llama-3.3-70b-instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "o1-mini": { + "id": "o1-mini", + "name": "o1-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-07", - "last_updated": "2024-12-07", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-09-12", + "last_updated": "2024-09-12", "modalities": { "input": [ "text" @@ -98436,27 +93342,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 120000 + "context": 128000, + "output": 65536 }, "cost": { - "input": 0.135, - "output": 0.4 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "meta-llama/llama-3.1-8b-instruct": { - "id": "meta-llama/llama-3.1-8b-instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "phi-3.5-mini-instruct": { + "id": "phi-3.5-mini-instruct", + "name": "Phi-3.5-mini-instruct", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2024-07-24", - "last_updated": "2024-07-24", + "knowledge": "2023-10", + "release_date": "2024-08-20", + "last_updated": "2024-08-20", "modalities": { "input": [ "text" @@ -98467,54 +93375,26 @@ }, "open_weights": true, "limit": { - "context": 16384, - "output": 16384 - }, - "cost": { - "input": 0.02, - "output": 0.05 - } - }, - "meta-llama/llama-4-scout-17b-16e-instruct": { - "id": "meta-llama/llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-06", - "last_updated": "2025-04-06", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.18, - "output": 0.59 + "input": 0.13, + "output": 0.52 } }, - "meta-llama/llama-3.2-3b-instruct": { - "id": "meta-llama/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "mistral-nemo": { + "id": "mistral-nemo", + "name": "Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2024-09-18", + "knowledge": "2024-07", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text" @@ -98525,28 +93405,29 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 32000 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.03, - "output": 0.05 + "input": 0.15, + "output": 0.15 } }, - "meta-llama/llama-3-8b-instruct": { - "id": "meta-llama/llama-3-8b-instruct", - "name": "Llama 3 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "cohere-embed-v-4-0": { + "id": "cohere-embed-v-4-0", + "name": "Embed v4", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-04-25", - "last_updated": "2024-04-25", + "temperature": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -98554,96 +93435,126 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 128000, + "output": 1536 }, "cost": { - "input": 0.04, - "output": 0.04 + "input": 0.12, + "output": 0 } }, - "nousresearch/hermes-2-pro-llama-3-8b": { - "id": "nousresearch/hermes-2-pro-llama-3-8b", - "name": "Hermes 2 Pro Llama 3 8B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, "temperature": true, - "release_date": "2024-06-27", - "last_updated": "2024-06-27", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-08-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.14, - "output": 0.14 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "xiaomimimo/mimo-v2-flash": { - "id": "xiaomimimo/mimo-v2-flash", - "name": "XiaomiMiMo/MiMo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-19", - "last_updated": "2025-12-19", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32000 + "context": 272000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.3 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "xiaomimimo/mimo-v2-pro": { - "id": "xiaomimimo/mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", - "attachment": false, + "gpt-5-chat": { + "id": "gpt-5-chat", + "name": "GPT-5 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-05-27", + "tool_call": false, + "temperature": false, + "knowledge": "2024-10-24", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -98651,48 +93562,28 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "xiaomimimo/mimo-v2.5-pro": { - "id": "xiaomimimo/mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "cohere-command-a": { + "id": "cohere-command-a", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-27", + "knowledge": "2024-06-01", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ "text" @@ -98703,42 +93594,26 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 256000, + "output": 8000 }, "cost": { - "input": 0.522, - "output": 1.044, - "cache_read": 0.0043, - "tiers": [ - { - "input": 0.522, - "output": 1.044, - "cache_read": 0.0043, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.522, - "output": 1.044, - "cache_read": 0.0043 - } + "input": 2.5, + "output": 10 } }, - "baidu/ernie-4.5-vl-28b-a3b": { - "id": "baidu/ernie-4.5-vl-28b-a3b", - "name": "ERNIE 4.5 VL 28B A3B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "llama-3.2-11b-vision-instruct": { + "id": "llama-3.2-11b-vision-instruct", + "name": "Llama-3.2-11B-Vision-Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2026-06-14", + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text", @@ -98750,59 +93625,66 @@ }, "open_weights": true, "limit": { - "context": 30000, - "output": 8000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.14, - "output": 0.56 + "input": 0.37, + "output": 0.37 } }, - "baidu/ernie-4.5-vl-28b-a3b-thinking": { - "id": "baidu/ernie-4.5-vl-28b-a3b-thinking", - "name": "ERNIE-4.5-VL-28B-A3B-Thinking", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-11-26", - "last_updated": "2025-11-26", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 400000, + "output": 272000 }, "cost": { - "input": 0.39, - "output": 0.39 + "input": 15, + "output": 120 } }, - "baidu/ernie-4.5-21B-a3b-thinking": { - "id": "baidu/ernie-4.5-21B-a3b-thinking", - "name": "ERNIE-4.5-21B-A3B-Thinking", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "ernie", + "cohere-command-r-08-2024": { + "id": "cohere-command-r-08-2024", + "name": "Command R", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-09-19", - "last_updated": "2025-09-19", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -98813,124 +93695,133 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 0.15, + "output": 0.6 } }, - "baidu/ernie-4.5-21B-a3b": { - "id": "baidu/ernie-4.5-21B-a3b", - "name": "ERNIE 4.5 21B A3B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "ernie", - "attachment": false, + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 120000, - "output": 8000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "baidu/ernie-4.5-vl-424b-a47b": { - "id": "baidu/ernie-4.5-vl-424b-a47b", - "name": "ERNIE 4.5 VL 424B A47B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "gpt-4": { + "id": "gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": true, "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "knowledge": "2023-11", + "release_date": "2023-03-14", + "last_updated": "2023-03-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 123000, - "output": 16000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.42, - "output": 1.25 + "input": 60, + "output": 120 } }, - "baidu/ernie-4.5-300b-a47b-paddle": { - "id": "baidu/ernie-4.5-300b-a47b-paddle", - "name": "ERNIE 4.5 300B A47B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 123000, - "output": 12000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.28, - "output": 1.1 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "phi-3-medium-128k-instruct": { + "id": "phi-3-medium-128k-instruct", + "name": "Phi-3-medium-instruct (128k)", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -98938,65 +93829,59 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.14, - "output": 0.4 + "input": 0.17, + "output": 0.68 } }, - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma 3 27B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "gpt-4-32k": { + "id": "gpt-4-32k", + "name": "GPT-4 32K", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "knowledge": "2023-11", + "release_date": "2023-03-14", + "last_updated": "2023-03-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 98304, - "output": 16384 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.119, - "output": 0.2 + "input": 60, + "output": 120 } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "meta-llama-3.1-405b-instruct": { + "id": "meta-llama-3.1-405b-instruct", + "name": "Meta-Llama-3.1-405B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -99004,29 +93889,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 5.33, + "output": 16 } }, - "google/gemma-3-12b-it": { - "id": "google/gemma-3-12b-it", - "name": "Gemma 3 12B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "cohere-command-r-plus-08-2024": { + "id": "cohere-command-r-plus-08-2024", + "name": "Command R+", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -99034,26 +93919,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.05, - "output": 0.1 + "input": 2.5, + "output": 10 } }, - "minimaxai/minimax-m1-80k": { - "id": "minimaxai/minimax-m1-80k", - "name": "MiniMax M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "phi-4-mini": { + "id": "phi-4-mini", + "name": "Phi-4-mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -99064,26 +93949,26 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 40000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.55, - "output": 2.2 + "input": 0.075, + "output": 0.3 } }, - "mistralai/mistral-nemo": { - "id": "mistralai/mistral-nemo", - "name": "Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + "gpt-3.5-turbo-1106": { + "id": "gpt-3.5-turbo-1106", + "name": "GPT-3.5 Turbo 1106", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": true, "temperature": true, - "release_date": "2024-07-30", - "last_updated": "2024-07-30", + "knowledge": "2021-08", + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" @@ -99092,60 +93977,59 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 60288, - "output": 16000 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0.04, - "output": 0.17 + "input": 1, + "output": 2 } }, - "inclusionai/ring-2.6-1t": { - "id": "inclusionai/ring-2.6-1t", - "name": "Ring-2.6-1T", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "ring", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "llama-4-scout-17b-16e-instruct": { + "id": "llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout 17B 16E Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-05-08", - "last_updated": "2026-05-27", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.06 + "input": 0.2, + "output": 0.78 } }, - "inclusionai/ling-2.6-1t": { - "id": "inclusionai/ling-2.6-1t", - "name": "Ling-2.6-1T", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "ling", + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-23", - "last_updated": "2026-06-29", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -99156,27 +94040,26 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 128000, "output": 32768 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.06 + "input": 0.71, + "output": 0.71 } }, - "inclusionai/ling-2.6-flash": { - "id": "inclusionai/ling-2.6-flash", - "name": "Ling-2.6-flash", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", + "grok-4-20-non-reasoning": { + "id": "grok-4-20-non-reasoning", + "name": "Grok 4.20 (Non-Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-09", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ "text" @@ -99185,89 +94068,78 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 262000, + "output": 8192 }, + "status": "beta", "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 2, + "output": 6 } }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "gpt-5.1-chat": { + "id": "gpt-5.1-chat", + "name": "GPT-5.1 Chat", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "high", - "max" + "medium" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text", "image", - "video" + "audio" ], "output": [ - "text" + "text", + "image", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 128000, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek-V4-Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning_options": [], + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -99275,17 +94147,21 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 384000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 0.8, - "output": 3.4, - "cache_read": 0.16 + "input": 1.74, + "output": 3.48 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", "name": "Kimi K2 Thinking", "description": "Kimi reasoning model for long-horizon research, planning, and tool use", "family": "kimi-thinking", @@ -99293,13 +94169,11 @@ "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "interleaved": true, "temperature": true, - "release_date": "2025-11-07", - "last_updated": "2026-06-29", + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-12-02", "modalities": { "input": [ "text" @@ -99319,89 +94193,119 @@ "cache_read": 0.15 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "knowledge": "2025-07-31", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "moonshotai/kimi-k2-instruct": { - "id": "moonshotai/kimi-k2-instruct", - "name": "Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "attachment": false, - "reasoning": false, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-07-11", - "last_updated": "2025-07-11", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 128000 + }, + "status": "beta", + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.57, - "output": 2.3 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -99412,111 +94316,106 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0.6, - "output": 2.5 + "input": 1.35, + "output": 5.4 } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "Grok 4.1 Fast (Non-Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2025-06-27", + "last_updated": "2025-06-27", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "input": 128000, + "output": 8192 }, + "status": "beta", "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "zai-org/glm-4.6v": { - "id": "zai-org/glm-4.6v", - "name": "GLM 4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glmv", + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 Nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "video", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.055 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "zai-org/glm-4.5": { - "id": "zai-org/glm-4.5", - "name": "GLM-4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "phi-3-small-8k-instruct": { + "id": "phi-3-small-8k-instruct", + "name": "Phi-3-small-instruct (8k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -99527,74 +94426,67 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 8192, + "output": 2048 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0.15, + "output": 0.6 } }, - "zai-org/glm-4.6": { - "id": "zai-org/glm-4.6", - "name": "GLM 4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "gpt-5.3-chat": { + "id": "gpt-5.3-chat", + "name": "GPT-5.3 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "medium" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.55, - "output": 2.2, - "cache_read": 0.11 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "zai-org/glm-5": { - "id": "zai-org/glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "gpt-3.5-turbo-0125": { + "id": "gpt-3.5-turbo-0125", + "name": "GPT-3.5 Turbo 0125", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-12", + "knowledge": "2021-08", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -99603,40 +94495,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202800, - "output": 131072 + "context": 16384, + "output": 16384 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.5, + "output": 1.5 } }, - "zai-org/glm-4.5v": { - "id": "zai-org/glm-4.5v", - "name": "GLM 4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glmv", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "cohere-embed-v3-multilingual": { + "id": "cohere-embed-v3-multilingual", + "name": "Embed v3 Multilingual", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2023-11-07", + "last_updated": "2023-11-07", "modalities": { "input": [ - "text", - "video", - "image" + "text" ], "output": [ "text" @@ -99644,33 +94526,26 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 16384 + "context": 512, + "output": 1024 }, "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11 + "input": 0.1, + "output": 0 } }, - "zai-org/glm-4.7-flash": { - "id": "zai-org/glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "gpt-3.5-turbo-0613": { + "id": "gpt-3.5-turbo-0613", + "name": "GPT-3.5 Turbo 0613", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2021-08", + "release_date": "2023-06-13", + "last_updated": "2023-06-13", "modalities": { "input": [ "text" @@ -99679,81 +94554,119 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.01 + "input": 3, + "output": 4 } }, - "zai-org/glm-4.7": { - "id": "zai-org/glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt-codex", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ - "text" + "text", + "image", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "zai-org/glm-5.2": { - "id": "zai-org/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "gpt-5.1-codex-max": { + "id": "gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.125 + } + }, + "phi-4-reasoning": { + "id": "phi-4-reasoning", + "name": "Phi-4-reasoning", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -99764,64 +94677,93 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 32000, + "output": 4096 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.125, + "output": 0.5 } }, - "zai-org/glm-5.1": { - "id": "zai-org/glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", + "knowledge": "2025-12-31", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 1.38, - "output": 4.4, - "cache_read": 0.26 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "zai-org/autoglm-phone-9b-multilingual": { - "id": "zai-org/autoglm-phone-9b-multilingual", - "name": "AutoGLM-Phone-9B-Multilingual", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "mistral-medium-2505": { + "id": "mistral-medium-2505", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-12-10", - "last_updated": "2025-12-10", + "knowledge": "2025-05", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ "text", @@ -99831,33 +94773,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 65536 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.035, - "output": 0.138 + "input": 0.4, + "output": 2 } }, - "zai-org/glm-4.5-air": { - "id": "zai-org/glm-4.5-air", - "name": "GLM 4.5 Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", + "meta-llama-3.1-70b-instruct": { + "id": "meta-llama-3.1-70b-instruct", + "name": "Meta-Llama-3.1-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-10-13", - "last_updated": "2025-10-13", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -99868,25 +94805,24 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.13, - "output": 0.85, - "cache_read": 0.025 + "input": 2.68, + "output": 3.54 } }, - "gryphe/mythomax-l2-13b": { - "id": "gryphe/mythomax-l2-13b", - "name": "Mythomax L2 13B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "text-embedding-ada-002": { + "id": "text-embedding-ada-002", + "name": "text-embedding-ada-002", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-04-25", - "last_updated": "2024-04-25", + "release_date": "2022-12-15", + "last_updated": "2022-12-15", "modalities": { "input": [ "text" @@ -99895,32 +94831,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 4096, - "output": 3200 + "context": 8192, + "output": 1536 }, "cost": { - "input": 0.09, - "output": 0.09 + "input": 0.1, + "output": 0 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ "text" @@ -99931,63 +94873,103 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 131100 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax-m2.7", - "attachment": false, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.125 + } + }, + "gpt-4-turbo-vision": { + "id": "gpt-4-turbo-vision", + "name": "GPT-4 Turbo Vision", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 10, + "output": 30 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", + "mistral-large-2411": { + "id": "mistral-large-2411", + "name": "Mistral Large 24.11", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-05-27", + "knowledge": "2024-09", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", "modalities": { "input": [ "text" @@ -99996,37 +94978,45 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 2, + "output": 6 } }, - "minimax/minimax-m2.5-highspeed": { - "id": "minimax/minimax-m2.5-highspeed", - "name": "MiniMax M2.5 Highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax-m2.5", + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -100034,30 +95024,25 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 131100 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.03 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "Minimax M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "text-embedding-3-small": { + "id": "text-embedding-3-small", + "name": "text-embedding-3-small", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "tool_call": false, + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -100066,32 +95051,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 8191, + "output": 1536 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.02, + "output": 0 } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "deepseek-v3.2-speciale": { + "id": "deepseek-v3.2-speciale", + "name": "DeepSeek-V3.2-Speciale", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -100102,60 +95084,59 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.58, + "output": 1.68 } }, - "paddlepaddle/paddleocr-vl": { - "id": "paddlepaddle/paddleocr-vl", - "name": "PaddleOCR-VL", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "gpt-3.5-turbo-0301": { + "id": "gpt-3.5-turbo-0301", + "name": "GPT-3.5 Turbo 0301", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2025-10-22", - "last_updated": "2025-10-22", + "knowledge": "2021-08", + "release_date": "2023-03-01", + "last_updated": "2023-03-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 4096, + "output": 4096 }, "cost": { - "input": 0.02, - "output": 0.02 + "input": 1.5, + "output": 2 } }, - "qwen/qwen3-vl-235b-a22b-instruct": { - "id": "qwen/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "meta-llama-3.1-8b-instruct": { + "id": "meta-llama-3.1-8b-instruct", + "name": "Meta-Llama-3.1-8B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -100163,34 +95144,38 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 128000, "output": 32768 }, "cost": { "input": 0.3, - "output": 1.5 + "output": 0.61 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7-Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-27", + "knowledge": "2025-03-31", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -100198,53 +95183,69 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 200000, + "output": 32000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 1.25, - "output": 3.75, - "cache_read": 0.25, - "cache_write": 1.5625 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "qwen/qwen3-vl-235b-a22b-thinking": { - "id": "qwen/qwen3-vl-235b-a22b-thinking", - "name": "Qwen3 VL 235B A22B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.98, - "output": 3.95 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5-27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -100252,15 +95253,16 @@ } ], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "knowledge": "2025-01", + "release_date": "2026-02-06", + "last_updated": "2026-02-06", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -100269,26 +95271,30 @@ "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 0.3, - "output": 2.4 + "input": 0.6, + "output": 3 } }, - "qwen/qwen3-235b-a22b-instruct-2507": { - "id": "qwen/qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "phi-4": { + "id": "phi-4", + "name": "Phi-4", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -100299,162 +95305,187 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.09, - "output": 0.58 + "input": 0.125, + "output": 0.5 } }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "attachment": false, + "grok-4-fast-reasoning": { + "id": "grok-4-fast-reasoning", + "name": "Grok 4 Fast (Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-09-10", - "last_updated": "2025-09-10", + "knowledge": "2025-07", + "release_date": "2025-09-19", + "last_updated": "2025-09-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "gpt-image-1.5": { + "id": "gpt-image-1.5", + "name": "GPT-Image-1.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-03", - "last_updated": "2026-02-03", + "tool_call": false, + "temperature": false, + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 0, + "output": 0 }, "cost": { - "input": 0.2, - "output": 1.5 + "input": 5, + "output": 32, + "cache_read": 1.25 } }, - "qwen/qwen3-omni-30b-a3b-instruct": { - "id": "qwen/qwen3-omni-30b-a3b-instruct", - "name": "Qwen3 Omni 30B A3B Instruct", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2024-04", - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "video", - "audio", "image" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 0.97, - "input_audio": 2.2, - "output_audio": 1.788 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "qwen/qwen3-omni-30b-a3b-thinking": { - "id": "qwen/qwen3-omni-30b-a3b-thinking", - "name": "Qwen3 Omni 30B A3B Thinking", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", - "audio", - "video", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.25, - "output": 0.97, - "input_audio": 2.2, - "output_audio": 1.788 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "o1": { + "id": "o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -100462,56 +95493,28 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 2.11, - "output": 8.45 + "input": 15, + "output": 60, + "cache_read": 7.5 } }, - "qwen/qwen3-235b-a22b-fp8": { - "id": "qwen/qwen3-235b-a22b-fp8", - "name": "Qwen3 235B A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "deepseek-v3.1": { + "id": "deepseek-v3.1", + "name": "DeepSeek-V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "output": 20000 - }, - "cost": { - "input": 0.2, - "output": 0.8 - } - }, - "qwen/qwen3-coder-480b-a35b-instruct": { - "id": "qwen/qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "knowledge": "2024-07", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ "text" @@ -100522,28 +95525,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.38, - "output": 1.55 + "input": 0.56, + "output": 1.68 } }, - "qwen/qwen3-coder-30b-a3b-instruct": { - "id": "qwen/qwen3-coder-30b-a3b-instruct", - "name": "Qwen3 Coder 30b A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "attachment": false, + "llama-3.2-90b-vision-instruct": { + "id": "llama-3.2-90b-vision-instruct", + "name": "Llama-3.2-90B-Vision-Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-10-09", - "last_updated": "2025-10-09", + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -100551,58 +95556,71 @@ }, "open_weights": true, "limit": { - "context": 160000, - "output": 32768 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.07, - "output": 0.27 + "input": 2.04, + "output": 2.04 } }, - "qwen/qwen3-vl-30b-a3b-instruct": { - "id": "qwen/qwen3-vl-30b-a3b-instruct", - "name": "qwen/qwen3-vl-30b-a3b-instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-11", - "last_updated": "2025-10-11", + "knowledge": "2025-02-31", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", - "video", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.2, - "output": 0.7 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "qwen/qwen3-235b-a22b-thinking-2507": { - "id": "qwen/qwen3-235b-a22b-thinking-2507", - "name": "Qwen3 235B A22b Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "ministral-3b": { + "id": "ministral-3b", + "name": "Ministral 3B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "knowledge": "2024-03", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ "text" @@ -100613,94 +95631,147 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 3 + "input": 0.04, + "output": 0.04 } }, - "qwen/qwen2.5-7b-instruct": { - "id": "qwen/qwen2.5-7b-instruct", - "name": "Qwen2.5 7B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": false, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 32000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.07, - "output": 0.07 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5-122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 Mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 3.2 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "qwen/qwen3-30b-a3b-fp8": { - "id": "qwen/qwen3-30b-a3b-fp8", - "name": "Qwen3 30B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "knowledge": "2025-01", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -100708,309 +95779,348 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 20000 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 0.09, - "output": 0.45 + "input": 0.95, + "output": 4 } }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-09-10", - "last_updated": "2025-09-10", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 1047576, "output": 32768 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "qwen/qwen3-vl-30b-a3b-thinking": { - "id": "qwen/qwen3-vl-30b-a3b-thinking", - "name": "qwen/qwen3-vl-30b-a3b-thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-10-11", - "last_updated": "2025-10-11", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.2, - "output": 1 - } - }, - "qwen/qwen-mt-plus": { - "id": "qwen/qwen-mt-plus", - "name": "Qwen MT Plus", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-09-03", - "last_updated": "2025-09-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 200000, + "output": 128000 }, - "open_weights": true, - "limit": { - "context": 16384, - "output": 8192 + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.25, - "output": 0.75 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "qwen/qwen3-32b-fp8": { - "id": "qwen/qwen3-32b-fp8", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "Grok 4.1 Fast (Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "release_date": "2025-06-27", + "last_updated": "2025-06-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 20000 + "context": 128000, + "input": 128000, + "output": 8192 }, + "status": "beta", "cost": { - "input": 0.1, - "output": 0.45 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "qwen/qwen3-8b-fp8": { - "id": "qwen/qwen3-8b-fp8", - "name": "Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 20000 + "context": 272000, + "output": 128000 }, "cost": { - "input": 0.035, - "output": 0.138 + "input": 0.25, + "output": 2, + "cache_read": 0.03 } }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5-35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-image-1": { + "id": "gpt-image-1", + "name": "GPT-Image-1", + "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", + "family": "gpt-image", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-04-24", + "last_updated": "2025-04-24", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 0, + "output": 0 }, "cost": { - "input": 0.25, - "output": 2 + "input": 5, + "output": 40, + "cache_read": 1.25 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5-397B-A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 64000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "qwen/qwen3-vl-8b-instruct": { - "id": "qwen/qwen3-vl-8b-instruct", - "name": "qwen/qwen3-vl-8b-instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "gpt-4-turbo": { + "id": "gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-10-17", - "last_updated": "2025-10-17", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.08, - "output": 0.5 + "input": 10, + "output": 30 } }, - "qwen/qwen2.5-vl-72b-instruct": { - "id": "qwen/qwen2.5-vl-72b-instruct", - "name": "Qwen2.5 VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 272000, + "output": 128000 }, "cost": { - "input": 0.8, - "output": 0.8 + "input": 0.05, + "output": 0.4, + "cache_read": 0.01 } }, - "qwen/qwen-2.5-72b-instruct": { - "id": "qwen/qwen-2.5-72b-instruct", - "name": "Qwen 2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "phi-3-medium-4k-instruct": { + "id": "phi-3-medium-4k-instruct", + "name": "Phi-3-medium-instruct (4k)", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-10-15", - "last_updated": "2024-10-15", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -101021,56 +96131,82 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 8192 + "context": 4096, + "output": 1024 }, "cost": { - "input": 0.38, - "output": 0.4 + "input": 0.17, + "output": 0.68 } }, - "qwen/qwen3-4b-fp8": { - "id": "qwen/qwen3-4b-fp8", - "name": "Qwen3 4B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 20000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.03, - "output": 0.03 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "baichuan/baichuan-m2-32b": { - "id": "baichuan/baichuan-m2-32b", - "name": "baichuan-m2-32b", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "baichuan", + "cohere-embed-v3-english": { + "id": "cohere-embed-v3-english", + "name": "Embed v3 English", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", "attachment": false, "reasoning": false, "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-08-13", - "last_updated": "2025-08-13", + "temperature": false, + "release_date": "2023-11-07", + "last_updated": "2023-11-07", "modalities": { "input": [ "text" @@ -101081,18 +96217,19 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 512, + "output": 1024 }, "cost": { - "input": 0.07, - "output": 0.07 + "input": 0.1, + "output": 0 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "OpenAI: GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -101101,84 +96238,90 @@ "values": [ "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.04, - "output": 0.15 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "OpenAI GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "codestral-2501": { + "id": "codestral-2501", + "name": "Codestral 25.01", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "codestral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "knowledge": "2024-03", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.05, - "output": 0.25 + "input": 0.3, + "output": 0.9 } }, - "microsoft/wizardlm-2-8x22b": { - "id": "microsoft/wizardlm-2-8x22b", - "name": "Wizardlm 2 8x22B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "phi-4-reasoning-plus": { + "id": "phi-4-reasoning-plus", + "name": "Phi-4-reasoning-plus", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, "temperature": true, - "release_date": "2024-04-24", - "last_updated": "2024-04-24", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -101189,27 +96332,31 @@ }, "open_weights": true, "limit": { - "context": 65535, - "output": 8000 + "context": 32000, + "output": 4096 }, "cost": { - "input": 0.62, - "output": 0.62 + "input": 0.125, + "output": 0.5 } }, - "sao10K/l3-70b-euryale-v2.1": { - "id": "sao10K/l3-70b-euryale-v2.1", - "name": "L3 70B Euryale V2.1\t", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, + "phi-4-multimodal": { + "id": "phi-4-multimodal", + "name": "Phi-4-multimodal", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "phi", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2024-06-18", - "last_updated": "2024-06-18", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" @@ -101217,82 +96364,89 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.48, - "output": 1.48 + "input": 0.08, + "output": 0.32, + "input_audio": 4 } }, - "sao10K/L3-8B-stheno-v3.2": { - "id": "sao10K/L3-8B-stheno-v3.2", - "name": "L3 8B Stheno V3.2", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-11-29", - "last_updated": "2024-11-29", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 32000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.05 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "sao10K/l3-8b-lunaris": { - "id": "sao10K/l3-8b-lunaris", - "name": "Sao10k L3 8B Lunaris\t", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, + "model-router": { + "id": "model-router", + "name": "Model Router", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "model-router", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2024-11-28", - "last_updated": "2024-11-28", + "tool_call": true, + "release_date": "2025-05-19", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.05 + "input": 0.14, + "output": 0 } }, - "sao10K/l31-70b-euryale-v2.2": { - "id": "sao10K/l31-70b-euryale-v2.2", - "name": "L31 70B Euryale V2.2", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "grok-4-20-reasoning": { + "id": "grok-4-20-reasoning", + "name": "Grok 4.20 (Reasoning)", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "knowledge": "2025-09", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ "text" @@ -101301,40 +96455,39 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, + "context": 262000, "output": 8192 }, + "status": "beta", "cost": { - "input": 1.48, - "output": 1.48 + "input": 2, + "output": 6 } - } - } - }, - "302ai": { - "id": "302ai", - "env": [ - "302AI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.302.ai/v1", - "name": "302.AI", - "doc": "https://doc.302.ai", - "models": { - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "grok-4-1-fast-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": true, + }, + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", @@ -101346,36 +96499,31 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.5 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "glm-4.6v": { - "id": "glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -101384,25 +96532,55 @@ "open_weights": true, "limit": { "context": 128000, - "output": 32768 + "output": 128000 }, "cost": { - "input": 0.145, - "output": 0.43 + "input": 0.58, + "output": 1.68 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "gpt-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "llama-4-maverick-17b-128e-instruct-fp8": { + "id": "llama-4-maverick-17b-128e-instruct-fp8", + "name": "Llama 4 Maverick 17B 128E Instruct FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.25, + "output": 1 + } + }, + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -101414,13 +96592,12 @@ "structured_output": true, "temperature": false, "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -101428,48 +96605,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, "output": 128000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 0, - "tiers": [ - { - "input": 5, - "output": 22.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5 - } + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "glm-4.5": { - "id": "glm-4.5", - "name": "GLM-4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "phi-4-mini-reasoning": { + "id": "phi-4-mini-reasoning", + "name": "Phi-4-mini-reasoning", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -101480,75 +96637,87 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.286, - "output": 1.142 + "input": 0.075, + "output": 0.3 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "claude-haiku-4-5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, + "phi-3-mini-128k-instruct": { + "id": "phi-3-mini-128k-instruct", + "name": "Phi-3-mini-instruct (128k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-16", - "last_updated": "2025-10-16", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.13, + "output": 0.52 + } + }, + "gpt-image-2": { + "id": "gpt-image-2", + "name": "GPT-Image-2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 0, + "output": 0 }, "cost": { - "input": 1, - "output": 5 + "input": 5, + "output": 30, + "cache_read": 1.25 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "phi-3-mini-4k-instruct": { + "id": "phi-3-mini-4k-instruct", + "name": "Phi-3-mini-instruct (4k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -101557,89 +96726,98 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 4096, + "output": 1024 }, "cost": { - "input": 1.8, - "output": 5.3 + "input": 0.13, + "output": 0.52 } }, - "gemini-3.1-flash-image-preview": { - "id": "gemini-3.1-flash-image-preview", - "name": "gemini-3.1-flash-image-preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, + "meta-llama-3-70b-instruct": { + "id": "meta-llama-3-70b-instruct", + "name": "Meta-Llama-3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-27", - "last_updated": "2026-02-27", + "knowledge": "2023-12", + "release_date": "2024-04-18", + "last_updated": "2024-04-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 8192, + "output": 2048 }, "cost": { - "input": 0.5, - "output": 60 + "input": 2.68, + "output": 3.54 } }, - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text", "image", - "video" + "audio" ], "output": [ - "text" + "text", + "image", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 512000 + "context": 272000, + "output": 128000 }, "cost": { - "input": 0.72, - "output": 2.88 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "gpt-5.4-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -101657,13 +96835,14 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-19", - "last_updated": "2026-03-19", + "knowledge": "2025-12-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -101671,170 +96850,241 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.25 - } - }, - "gemini-3.5-flash-thinking": { - "id": "gemini-3.5-flash-thinking", - "name": "gemini-3.5-flash-thinking", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } + } + } + } + }, + "cerebras": { + "id": "cerebras", + "env": [ + "CEREBRAS_API_KEY" + ], + "npm": "@ai-sdk/cerebras", + "name": "Cerebras", + "doc": "https://inference-docs.cerebras.ai/models/overview", + "models": { + "gemma-4-31b": { + "id": "gemma-4-31b", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-04-02", + "last_updated": "2026-07-01", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 40960 }, + "status": "beta", "cost": { - "input": 1.5, - "output": 9 + "input": 0.99, + "output": 1.49 } }, - "gemini-3-pro-preview": { - "id": "gemini-3-pro-preview", - "name": "gemini-3-pro-preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "attachment": true, - "reasoning": false, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "release_date": "2025-08-05", + "last_updated": "2026-06-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 131072, + "output": 40960 }, "cost": { - "input": 2, - "output": 12 + "input": 0.35, + "output": 0.75 } }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, + "zai-glm-4.7": { + "id": "zai-glm-4.7", + "name": "Z.AI GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-01-07", + "last_updated": "2026-06-10", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 131072, + "output": 40960 }, + "status": "beta", "cost": { - "input": 0.72, - "output": 3.2 + "input": 2.25, + "output": 2.75, + "cache_read": 0, + "cache_write": 0 } - }, - "gemini-2.5-flash-image": { - "id": "gemini-2.5-flash-image", - "name": "gemini-2.5-flash-image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, + } + } + }, + "zai-coding-plan": { + "id": "zai-coding-plan", + "env": [ + "ZHIPU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.z.ai/api/coding/paas/v4", + "name": "Z.AI Coding Plan", + "doc": "https://docs.z.ai/devpack/overview", + "models": { + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-10-08", - "last_updated": "2025-10-08", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 30 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "gpt-4o": { - "id": "gpt-4o", - "name": "gpt-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -101842,26 +97092,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 200000, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "high", "max" ] @@ -101873,14 +97124,11 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -101892,16 +97140,18 @@ "output": 131072 }, "cost": { - "input": 0.075, - "output": 0.25 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "glm-4.6": { - "id": "glm-4.6", - "name": "glm-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", "family": "glm", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -101909,39 +97159,52 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, + "context": 200000, "output": 131072 }, "cost": { - "input": 0.286, - "output": 1.142 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "qwen3-235b-a22b-instruct-2507", - "description": "Tool-capable chat model for instruction following and agentic application workflows", + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-07-30", - "last_updated": "2025-07-30", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -101950,52 +97213,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 65536 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.29, - "output": 1.143 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, + "glm-5-turbo": { + "id": "glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -102003,47 +97255,72 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 200000, "output": 131072 }, "cost": { - "input": 2.16, - "output": 6.36 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + } + } + }, + "nvidia": { + "id": "nvidia", + "env": [ + "NVIDIA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://integrate.api.nvidia.com/v1", + "name": "Nvidia", + "doc": "https://docs.api.nvidia.com/nim/", + "models": { + "baai/bge-m3": { + "id": "baai/bge-m3", + "name": "BGE M3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "bge", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-01-30", + "last_updated": "2026-04-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 1024 + }, + "cost": { + "input": 0, + "output": 0 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "moonshotai/kimi-k2-instruct-0905": { + "id": "moonshotai/kimi-k2-instruct-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -102051,19 +97328,19 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 3, - "output": 15 + "input": 0, + "output": 0 } }, - "gpt-5.6-sol-pro": { - "id": "gpt-5.6-sol-pro", - "name": "gpt-5.6-sol-pro", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -102071,6 +97348,7 @@ "type": "effort", "values": [ "none", + "minimal", "low", "medium", "high", @@ -102080,167 +97358,162 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 30 + "input": 0, + "output": 0 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "gpt-5-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "minimaxai/minimax-m3": { + "id": "minimaxai/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 2 + "input": 0, + "output": 0 } }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "minimaxai/minimax-m2.7": { + "id": "minimaxai/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "release_date": "2026-03-18", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.12, - "output": 0.69 + "input": 0, + "output": 0 } }, - "claude-sonnet-4-6-thinking": { - "id": "claude-sonnet-4-6-thinking", - "name": "claude-sonnet-4-6-thinking", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "stepfun-ai/step-3.7-flash": { + "id": "stepfun-ai/step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2025-08", - "release_date": "2026-02-18", - "last_updated": "2026-03-13", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 256000, + "output": 16384 }, "cost": { - "input": 3, - "output": 15 + "input": 0, + "output": 0 } }, - "glm-5": { - "id": "glm-5", - "name": "glm-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "stepfun-ai/step-3.5-flash": { + "id": "stepfun-ai/step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-02-02", + "last_updated": "2026-02-02", "modalities": { "input": [ "text" @@ -102251,27 +97524,26 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.6 + "input": 0, + "output": 0 } }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "gpt-4.1-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "google/gemma-3n-e4b-it": { + "id": "google/gemma-3n-e4b-it", + "name": "Gemma 3n E4b It", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2024-06", + "release_date": "2025-06-03", + "last_updated": "2025-06-03", "modalities": { "input": [ "text", @@ -102281,28 +97553,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 0, + "output": 0 } }, - "grok-4-fast-reasoning": { - "id": "grok-4-fast-reasoning", - "name": "grok-4-fast-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "google/gemma-3n-e2b-it": { + "id": "google/gemma-3n-e2b-it", + "name": "Gemma 3n E2b It", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2024-06", + "release_date": "2025-06-12", + "last_updated": "2025-06-12", "modalities": { "input": [ "text", @@ -102312,42 +97584,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 0.5 + "input": 0, + "output": 0 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "google/google-paligemma": { + "id": "google/google-paligemma", + "name": "paligemma", + "description": "Gemini multimodal model for text, image, audio, video, and document tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2024-05-14", + "last_updated": "2024-08-26", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -102355,106 +97615,92 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 4 + "input": 0, + "output": 0 } }, - "gpt-5.2-chat-latest": { - "id": "gpt-5.2-chat-latest", - "name": "gpt-5.2-chat-latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma-4-31B-IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-12", - "last_updated": "2025-12-12", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, + "context": 256000, "output": 16384 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0, + "output": 0 } }, - "gpt-5.1-chat-latest": { - "id": "gpt-5.1-chat-latest", - "name": "gpt-5.1-chat-latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium" - ] - } - ], + "google/gemma-2-2b-it": { + "id": "google/gemma-2-2b-it", + "name": "Gemma 2 2b It", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "temperature": true, + "release_date": "2024-07-16", + "last_updated": "2024-07-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 16384 + "output": 4096 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0, + "output": 0 } }, - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", + "microsoft/phi-4-mini-instruct": { + "id": "microsoft/phi-4-mini-instruct", + "name": "Phi-4-Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2024-12", + "release_date": "2024-12-01", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -102465,29 +97711,27 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0, + "output": 0 } }, - "gemini-2.5-flash-preview-09-2025": { - "id": "gemini-2.5-flash-preview-09-2025", - "name": "gemini-2.5-flash-preview-09-2025", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "attachment": true, + "microsoft/phi-4-multimodal-instruct": { + "id": "microsoft/phi-4-multimodal-instruct", + "name": "Phi 4 Multimodal", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-09-26", - "last_updated": "2025-09-26", + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -102495,20 +97739,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 0, + "output": 0 } }, - "glm-4.5v": { - "id": "glm-4.5v", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", - "attachment": true, + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -102516,15 +97761,16 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-12", - "last_updated": "2025-08-12", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -102532,20 +97778,20 @@ }, "open_weights": true, "limit": { - "context": 64000, - "output": 16384 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.29, - "output": 0.86 + "input": 0, + "output": 0 } }, - "claude-fable-5-1": { - "id": "claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT-OSS-120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -102553,49 +97799,75 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-08", + "release_date": "2025-08-04", + "last_updated": "2025-08-14", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "openai/whisper-large-v3": { + "id": "openai/whisper-large-v3", + "name": "Whisper Large v3", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "knowledge": "2023-09", + "release_date": "2023-09-01", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 0, + "output": 4096 }, "cost": { - "input": 10, - "output": 50 + "input": 0, + "output": 0 } }, - "gemini-3.6-flash": { - "id": "gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -102605,49 +97877,41 @@ "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 7.5 + "input": 0, + "output": 0 } }, - "claude-opus-5-thinking": { - "id": "claude-opus-5-thinking", - "name": "claude-opus-5-thinking", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "bytedance/seed-oss-36b-instruct": { + "id": "bytedance/seed-oss-36b-instruct", + "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "seed", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "structured_output": true, + "temperature": true, + "release_date": "2025-09-04", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -102655,25 +97919,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 5, - "output": 25 + "input": 0, + "output": 0 } }, - "qwen3-coder-480b-a35b-instruct": { - "id": "qwen3-coder-480b-a35b-instruct", - "name": "qwen3-coder-480b-a35b-instruct", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "mistralai/mistral-7b-instruct-v03": { + "id": "mistralai/mistral-7b-instruct-v03", + "name": "Mistral-7B-Instruct-v0.3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "release_date": "2025-04-01", + "last_updated": "2025-04-01", "modalities": { "input": [ "text" @@ -102682,30 +97946,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, + "context": 65536, "output": 65536 }, "cost": { - "input": 0.86, - "output": 3.43 + "input": 0, + "output": 0 } }, - "doubao-seed-1-6-vision-250815": { - "id": "doubao-seed-1-6-vision-250815", - "name": "doubao-seed-1-6-vision-250815", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "mistralai/magistral-small-2506": { + "id": "mistralai/magistral-small-2506", + "name": "Magistral Small 2506", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -102713,31 +97976,58 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.114, - "output": 1.143 + "input": 0, + "output": 0 } }, - "claude-sonnet-4-5-20250929-thinking": { - "id": "claude-sonnet-4-5-20250929-thinking", - "name": "claude-sonnet-4-5-20250929-thinking", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistralai/mixtral-8x7b-instruct": { + "id": "mistralai/mixtral-8x7b-instruct", + "name": "Mistral: Mixtral 8x7B Instruct", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2023-12-10", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "mistralai/mistral-medium-3-instruct": { + "id": "mistralai/mistral-medium-3-instruct", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -102745,19 +98035,19 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 3, - "output": 15 + "input": 0, + "output": 0 } }, - "grok-4.3": { - "id": "grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "mistralai/mistral-small-4-119b-2603": { + "id": "mistralai/mistral-small-4-119b-2603", + "name": "mistral-small-4-119b-2603", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -102765,8 +98055,6 @@ "type": "effort", "values": [ "none", - "low", - "medium", "high" ] } @@ -102774,40 +98062,38 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 30000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 2.5 + "input": 0, + "output": 0 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "kimi-k2-thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "mistralai/mistral-nemotron": { + "id": "mistralai/mistral-nemotron", + "name": "mistral-nemotron", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "nemotron", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "release_date": "2025-06-11", + "last_updated": "2025-06-12", "modalities": { "input": [ "text" @@ -102816,27 +98102,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.575, - "output": 2.3 + "input": 0, + "output": 0 } }, - "grok-4-fast-non-reasoning": { - "id": "grok-4-fast-non-reasoning", - "name": "grok-4-fast-non-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "mistralai/mistral-large-3-675b-instruct-2512": { + "id": "mistralai/mistral-large-3-675b-instruct-2512", + "name": "Mistral Large 3 675B Instruct 2512", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2025-01", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", @@ -102846,109 +98134,115 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.2, - "output": 0.5 + "input": 0, + "output": 0 } }, - "grok-4.20-beta-0309-reasoning": { - "id": "grok-4.20-beta-0309-reasoning", - "name": "grok-4.20-beta-0309-reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistralai/mixtral-8x22b-instruct": { + "id": "mistralai/mixtral-8x22b-instruct", + "name": "Mistral: Mixtral 8x22B Instruct", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "release_date": "2024-04-17", + "last_updated": "2024-04-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 65536, + "output": 13108 }, "cost": { - "input": 2, - "output": 6 + "input": 0, + "output": 0 } }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "claude-opus-4-1-20250805", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "nvidia/cosmos-transfer1-7b": { + "id": "nvidia/cosmos-transfer1-7b", + "name": "cosmos-transfer1-7b", + "description": "Video model for prompt-guided generation, editing, and motion workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-13", + "last_updated": "2025-06-30", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ - "text" + "video" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 0, + "output": 4096 }, "cost": { - "input": 15, - "output": 75 + "input": 0, + "output": 0 } }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "gpt-5-pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "nvidia/cosmos-transfer2_5-2b": { + "id": "nvidia/cosmos-transfer2_5-2b", + "name": "cosmos-transfer2.5-2b", + "description": "Video model for prompt-guided generation, editing, and motion workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-08", - "last_updated": "2025-10-08", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "video" + ] + }, + "open_weights": true, + "limit": { + "context": 0, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/llama-nemotron-embed-vl-1b-v2": { + "id": "nvidia/llama-nemotron-embed-vl-1b-v2", + "name": "llama-nemotron-embed-vl-1b-v2", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "nemotron", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-02-10", + "last_updated": "2026-02-10", "modalities": { "input": [ "text", @@ -102958,186 +98252,163 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 272000 + "context": 32768, + "output": 2048 }, "cost": { - "input": 15, - "output": 120 + "input": 0, + "output": 0 } }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "claude-opus-4-5-20251101", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning": { + "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning", + "name": "Nemotron 3 Nano Omni", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, { "type": "budget_tokens", - "min": 1024, - "max": 63999 + "min": -1, + "max": 32768 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 256000, + "output": 65536 }, "cost": { - "input": 5, - "output": 25 + "input": 0, + "output": 0 } }, - "gemini-3-pro-image-preview": { - "id": "gemini-3-pro-image-preview", - "name": "gemini-3-pro-image-preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "nvidia/magpie-tts-zeroshot": { + "id": "nvidia/magpie-tts-zeroshot", + "name": "magpie-tts-zeroshot", + "description": "Speech generation model for controllable voice, narration, and audio delivery", "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "temperature": false, + "release_date": "2025-05-22", + "last_updated": "2025-06-12", "modalities": { "input": [ "text", - "image" + "audio" ], "output": [ - "text" + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 64000 + "context": 0, + "output": 4096 }, "cost": { - "input": 2, - "output": 120 + "input": 0, + "output": 0 } }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, + "nvidia/nvidia-nemotron-nano-9b-v2": { + "id": "nvidia/nvidia-nemotron-nano-9b-v2", + "name": "nvidia-nemotron-nano-9b-v2", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2024-09", + "release_date": "2025-08-18", + "last_updated": "2025-08-18", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 0, + "output": 0 } }, - "gemini-2.0-flash-lite": { - "id": "gemini-2.0-flash-lite", - "name": "gemini-2.0-flash-lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "nvidia/synthetic-video-detector": { + "id": "nvidia/synthetic-video-detector", + "name": "synthetic-video-detector", + "description": "Video model for prompt-guided generation, editing, and motion workflows", "attachment": true, "reasoning": false, "tool_call": false, "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-06-16", - "last_updated": "2025-06-16", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 8192 + "context": 0, + "output": 4096 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 0, + "output": 0 } }, - "qwen3-235b-a22b": { - "id": "qwen3-235b-a22b", - "name": "Qwen3-235B-A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "nvidia/nemotron-content-safety-reasoning-4b": { + "id": "nvidia/nemotron-content-safety-reasoning-4b", + "name": "nemotron-content-safety-reasoning-4b", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "nemotron", "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": false, + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ "text" @@ -103146,92 +98417,57 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 16384 + "output": 4096 }, "cost": { - "input": 0.29, - "output": 2.86 + "input": 0, + "output": 0 } }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "nvidia/nv-embed-v1": { + "id": "nvidia/nv-embed-v1", + "name": "nv-embed-v1", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "release_date": "2024-06-07", + "last_updated": "2025-07-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 32768, + "output": 2048 }, "cost": { - "input": 10, - "output": 50 + "input": 0, + "output": 0 } }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "nvidia/usdcode": { + "id": "nvidia/usdcode", + "name": "usdcode", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-01-01", + "last_updated": "2026-01-01", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -103239,127 +98475,110 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 4096 }, "cost": { - "input": 2, - "output": 12 + "input": 0, + "output": 0 } }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "nvidia/riva-translate-4b-instruct-v1_1": { + "id": "nvidia/riva-translate-4b-instruct-v1_1", + "name": "riva-translate-4b-instruct-v1_1", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-12-12", + "last_updated": "2025-12-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/sparsedrive": { + "id": "nvidia/sparsedrive", + "name": "sparsedrive", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "release_date": "2025-03-18", + "last_updated": "2025-07-20", "modalities": { "input": [ - "text", - "image" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 2, - "output": 6 + "input": 0, + "output": 0 } }, - "gpt-5.6-luna-pro": { - "id": "gpt-5.6-luna-pro", - "name": "gpt-5.6-luna-pro", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "nvidia/rerank-qa-mistral-4b": { + "id": "nvidia/rerank-qa-mistral-4b", + "name": "rerank-qa-mistral-4b", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2024-03-17", + "last_updated": "2025-01-17", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 1.2 + "input": 0, + "output": 0 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "nvidia/streampetr": { + "id": "nvidia/streampetr", + "name": "streampetr", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", - "image", "video" ], "output": [ @@ -103368,121 +98587,81 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.66, - "output": 3.3 + "input": 0, + "output": 0 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "gpt-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "nvidia/active-speaker-detection": { + "id": "nvidia/active-speaker-detection", + "name": "Active Speaker Detection", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 0, + "output": 4096 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0, + "output": 0 } }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, + "nvidia/llama-3_1-nemotron-safety-guard-8b-v3": { + "id": "nvidia/llama-3_1-nemotron-safety-guard-8b-v3", + "name": "llama-3.1-nemotron-safety-guard-8b-v3", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "release_date": "2025-10-28", + "last_updated": "2025-10-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 5, - "output": 25 + "input": 0, + "output": 0 } }, - "qwen3.7-max-2026-06-08": { - "id": "qwen3.7-max-2026-06-08", - "name": "qwen3.7-max-2026-06-08", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "nvidia/llama-3_2-nemoretriever-300m-embed-v1": { + "id": "nvidia/llama-3_2-nemoretriever-300m-embed-v1", + "name": "llama-3_2-nemoretriever-300m-embed-v1", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-07-24", + "last_updated": "2025-07-24", "modalities": { "input": [ "text" @@ -103491,41 +98670,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 32768, + "output": 2048 }, "cost": { - "input": 1.8, - "output": 5.3 + "input": 0, + "output": 0 } }, - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "nvidia/nemotron-voicechat": { + "id": "nvidia/nemotron-voicechat", + "name": "nemotron-voicechat", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text", - "image", - "video", "audio" ], "output": [ @@ -103534,25 +98702,24 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.283, - "output": 1.705 + "input": 0, + "output": 0 } }, - "kimi-k2-0905-preview": { - "id": "kimi-k2-0905-preview", - "name": "kimi-k2-0905-preview", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "nvidia/nv-embedcode-7b-v1": { + "id": "nvidia/nv-embedcode-7b-v1", + "name": "nv-embedcode-7b-v1", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", + "tool_call": false, + "temperature": false, + "release_date": "2025-03-17", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" @@ -103561,49 +98728,36 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 32768, + "output": 2048 }, "cost": { - "input": 0.632, - "output": 2.53 + "input": 0, + "output": 0 } }, - "deepseek-flash": { - "id": "deepseek-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "nvidia/nemotron-3-ultra-550b-a55b": { + "id": "nvidia/nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -103612,278 +98766,174 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 384000 + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 + "input": 0.5, + "output": 2.5, + "cache_read": 0.15 } }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "nvidia/nemotron-3-nano-30b-a3b": { + "id": "nvidia/nemotron-3-nano-30b-a3b", + "name": "nemotron-3-nano-30b-a3b", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "knowledge": "2024-09", + "release_date": "2024-12", + "last_updated": "2024-12", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 9 + "input": 0, + "output": 0 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "gemini-2.5-pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "nvidia/cosmos-predict1-5b": { + "id": "nvidia/cosmos-predict1-5b", + "name": "cosmos-predict1-5b", + "description": "Video model for prompt-guided generation, editing, and motion workflows", "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "tool_call": false, + "temperature": false, + "release_date": "2025-03-18", + "last_updated": "2025-03-18", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "text" + "video" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 0, + "output": 4096 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0, + "output": 0 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "nvidia/bevformer": { + "id": "nvidia/bevformer", + "name": "bevformer", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2025-03-18", + "last_updated": "2025-07-20", "modalities": { "input": [ - "text", - "image", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.188, - "output": 1.133 + "input": 0, + "output": 0 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "gemini-2.5-flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "nvidia/studiovoice": { + "id": "nvidia/studiovoice", + "name": "studiovoice", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2024-10-03", + "last_updated": "2025-06-13", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5 - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 10, - "output": 50 + "input": 0, + "output": 0 } }, - "doubao-seed-1-6-thinking-250715": { - "id": "doubao-seed-1-6-thinking-250715", - "name": "doubao-seed-1-6-thinking-250715", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "nvidia/gliner-pii": { + "id": "nvidia/gliner-pii", + "name": "gliner-pii", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 16000 - }, - "cost": { - "input": 0.121, - "output": 1.21 - } - }, - "gpt-5-thinking": { - "id": "gpt-5-thinking", - "name": "gpt-5-thinking", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-08-08", - "last_updated": "2025-08-08", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.25, - "output": 10 + "input": 0, + "output": 0 } }, - "MiniMax-M1": { - "id": "MiniMax-M1", - "name": "MiniMax-M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "nvidia/nemotron-mini-4b-instruct": { + "id": "nvidia/nemotron-mini-4b-instruct", + "name": "nemotron-mini-4b-instruct", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-06-16", - "last_updated": "2025-06-16", + "release_date": "2024-08-21", + "last_updated": "2024-08-26", "modalities": { "input": [ "text" @@ -103892,31 +98942,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.132, - "output": 1.254 + "input": 0, + "output": 0 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "nvidia/llama-nemotron-rerank-vl-1b-v2": { + "id": "nvidia/llama-nemotron-rerank-vl-1b-v2", + "name": "llama-nemotron-rerank-vl-1b-v2", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "nemotron", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -103924,79 +98974,59 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0, + "output": 0 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "gpt-5.4-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "Nemotron 3 Super", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-19", - "last_updated": "2026-03-19", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.75, - "output": 4.5 + "input": 0.2, + "output": 0.8 } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "glm-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "nvidia/usdvalidate": { + "id": "nvidia/usdvalidate", + "name": "usdvalidate", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-07-24", + "last_updated": "2025-01-08", "modalities": { "input": [ "text" @@ -104007,73 +99037,53 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 0, + "output": 4096 }, "cost": { - "input": 0.286, - "output": 1.142 + "input": 0, + "output": 0 } }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "nvidia/nemotron-3-content-safety": { + "id": "nvidia/nemotron-3-content-safety", + "name": "nemotron-3-content-safety", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 1.2 + "input": 0, + "output": 0 } }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "abacusai/dracarys-llama-3_1-70b-instruct": { + "id": "abacusai/dracarys-llama-3_1-70b-instruct", + "name": "dracarys-llama-3.1-70b-instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2024-09-11", + "last_updated": "2025-05-22", "modalities": { "input": [ "text" @@ -104084,35 +99094,29 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0, + "output": 0 } }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "deepseek-ai/deepseek-v4-flash": { + "id": "deepseek-ai/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", - "xhigh" + "none", + "high", + "max" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, @@ -104120,85 +99124,86 @@ "field": "reasoning_content" }, "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 0.18, - "output": 0.564 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "gpt-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "deepseek-ai/deepseek-v4-pro": { + "id": "deepseek-ai/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-12", - "last_updated": "2025-12-12", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 393216 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "deepseek-v3.2-thinking": { - "id": "deepseek-v3.2-thinking", - "name": "DeepSeek-V3.2-Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next-80B-A3B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, "knowledge": "2024-12", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2024-12-01", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -104209,131 +99214,91 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.29, - "output": 0.43 + "input": 0, + "output": 0 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen/qwen-image-edit": { + "id": "qwen/qwen-image-edit", + "name": "Qwen Image Edit", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-19", + "last_updated": "2025-08-19", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 0, + "output": 0 }, "cost": { - "input": 5, - "output": 25 + "input": 0, + "output": 0 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "qwen/qwen-image": { + "id": "qwen/qwen-image", + "name": "Qwen Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 0, + "output": 0 }, "cost": { - "input": 5, - "output": 30 + "input": 0, + "output": 0 } }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "claude-sonnet-4-5-20250929", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "qwen/qwen3-coder-480b-a35b-instruct": { + "id": "qwen/qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -104341,97 +99306,83 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 66536 }, "cost": { - "input": 3, - "output": 15 + "input": 0, + "output": 0 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "gpt-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "qwen/qwen2.5-coder-32b-instruct": { + "id": "qwen/qwen2.5-coder-32b-instruct", + "name": "Qwen2.5 Coder 32b Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2024-11-06", + "last_updated": "2024-11-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 128000, + "output": 4096 }, "cost": { - "input": 2, - "output": 8 + "input": 0, + "output": 0 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5-397B-A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "structured_output": true, + "temperature": true, + "knowledge": "2026-01", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 8192 }, "cost": { - "input": 2, - "output": 10 + "input": 0, + "output": 0 } }, - "qwen3.5-35b-a3b": { - "id": "qwen3.5-35b-a3b", - "name": "Qwen3.5 35B-A3B", + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, @@ -104439,9 +99390,6 @@ "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens" } ], "tool_call": true, @@ -104466,22 +99414,20 @@ "output": 65536 }, "cost": { - "input": 0.06, - "output": 0.46 + "input": 0, + "output": 0 } }, - "qwen3-30b-a3b": { - "id": "qwen3-30b-a3b", - "name": "Qwen3-30B-A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "sarvamai/sarvam-m": { + "id": "sarvamai/sarvam-m", + "name": "sarvam-m", + "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-29", - "last_updated": "2025-04-29", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ "text" @@ -104490,85 +99436,87 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, "output": 8192 }, "cost": { - "input": 0.11, - "output": 1.08 + "input": 0, + "output": 0 } }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "meta/llama-3.1-8b-instruct": { + "id": "meta/llama-3.1-8b-instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "knowledge": "2023-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 16000, + "output": 4096 }, "cost": { - "input": 0.75, - "output": 3.75 + "input": 0, + "output": 0 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "meta/llama-3.1-70b-instruct": { + "id": "meta/llama-3.1-70b-instruct", + "name": "Llama 3.1 70b Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "structured_output": true, + "temperature": true, + "release_date": "2024-07-16", + "last_updated": "2024-07-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4096 }, + "cost": { + "input": 0, + "output": 0 + } + }, + "meta/llama-3.2-1b-instruct": { + "id": "meta/llama-3.2-1b-instruct", + "name": "Llama 3.2 1b Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2023-12", + "release_date": "2024-09-18", + "last_updated": "2024-09-18", "modalities": { "input": [ "text" @@ -104579,26 +99527,26 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.4, - "output": 4.4 + "input": 0, + "output": 0 } }, - "gemini-2.5-flash-nothink": { - "id": "gemini-2.5-flash-nothink", - "name": "gemini-2.5-flash-nothink", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "meta/llama-3.2-11b-vision-instruct": { + "id": "meta/llama-3.2-11b-vision-instruct", + "name": "Llama 3.2 11b Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-24", - "last_updated": "2025-06-24", + "knowledge": "2023-12", + "release_date": "2024-09-18", + "last_updated": "2024-09-18", "modalities": { "input": [ "text", @@ -104608,57 +99556,56 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 0, + "output": 0 } }, - "doubao-seed-1-8-251215": { - "id": "doubao-seed-1-8-251215", - "name": "doubao-seed-1-8-251215", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "meta/llama-3.3-70b-instruct": { + "id": "meta/llama-3.3-70b-instruct", + "name": "Llama 3.3 70b Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "release_date": "2024-11-26", + "last_updated": "2024-11-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 224000, - "output": 64000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.114, - "output": 0.286 + "input": 0, + "output": 0 } }, - "claude-opus-4-1-20250805-thinking": { - "id": "claude-opus-4-1-20250805-thinking", - "name": "claude-opus-4-1-20250805-thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "meta/llama-guard-4-12b": { + "id": "meta/llama-guard-4-12b", + "name": "Llama Guard 4 12B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-05-27", - "last_updated": "2025-05-27", + "release_date": "2025-04-05", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", @@ -104668,36 +99615,26 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 15, - "output": 75 + "input": 0, + "output": 0 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "meta/esmfold": { + "id": "meta/esmfold", + "name": "esmfold", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2024-03-15", + "last_updated": "2025-06-12", "modalities": { "input": [ "text" @@ -104708,97 +99645,85 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.4, - "output": 4.4 + "input": 0, + "output": 0 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "gemini-3-flash-preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "attachment": true, + "meta/esm2-650m": { + "id": "meta/esm2-650m", + "name": "esm2-650m", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "release_date": "2024-08-29", + "last_updated": "2025-03-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0, + "output": 0 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "claude-haiku-4-5-20251001", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "meta/llama-3.2-90b-vision-instruct": { + "id": "meta/llama-3.2-90b-vision-instruct", + "name": "Llama-3.2-90B-Vision-Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-16", - "last_updated": "2025-10-16", + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1, - "output": 5 + "input": 0, + "output": 0 } }, - "mistral-large-2512": { - "id": "mistral-large-2512", - "name": "mistral-large-2512", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "meta/llama-4-maverick-17b-128e-instruct": { + "id": "meta/llama-4-maverick-17b-128e-instruct", + "name": "Llama 4 Maverick 17b 128e Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "knowledge": "2024-02", + "release_date": "2025-04-01", + "last_updated": "2025-04-01", "modalities": { "input": [ "text", @@ -104808,26 +99733,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 262144 + "output": 4096 }, "cost": { - "input": 1.1, - "output": 3.3 + "input": 0, + "output": 0 } }, - "MiniMax-M2": { - "id": "MiniMax-M2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "meta/llama-3.2-3b-instruct": { + "id": "meta/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-10-26", - "last_updated": "2025-10-26", + "release_date": "2024-09-18", + "last_updated": "2024-09-18", "modalities": { "input": [ "text" @@ -104836,249 +99763,232 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 32768, + "output": 32000 }, "cost": { - "input": 0.33, - "output": 1.32 + "input": 0, + "output": 0 } }, - "gemini-2.5-flash-lite-preview-09-2025": { - "id": "gemini-2.5-flash-lite-preview-09-2025", - "name": "gemini-2.5-flash-lite-preview-09-2025", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "attachment": true, + "upstage/solar-10_7b-instruct": { + "id": "upstage/solar-10_7b-instruct", + "name": "solar-10.7b-instruct", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-09-26", - "last_updated": "2025-09-26", + "release_date": "2024-06-05", + "last_updated": "2025-04-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 0, + "output": 0 } }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "gpt-4.1-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "black-forest-labs/flux_1-schnell": { + "id": "black-forest-labs/flux_1-schnell", + "name": "FLUX.1-schnell", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "tool_call": false, + "structured_output": false, + "temperature": false, + "knowledge": "2024-07", + "release_date": "2024-08-01", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "image" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 77, + "input": 77, + "output": 0 }, "cost": { - "input": 0.4, - "output": 1.6 + "input": 0, + "output": 0 } }, - "gpt-5.6-terra-pro": { - "id": "gpt-5.6-terra-pro", - "name": "gpt-5.6-terra-pro", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", + "black-forest-labs/flux_1-kontext-dev": { + "id": "black-forest-labs/flux_1-kontext-dev", + "name": "FLUX.1-Kontext-dev", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2025-08-12", + "last_updated": "2025-08-12", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 40960, + "output": 40960 }, "cost": { - "input": 2, - "output": 12 + "input": 0, + "output": 0 } }, - "gemini-3.8-flash": { - "id": "gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "black-forest-labs/flux.1-dev": { + "id": "black-forest-labs/flux.1-dev", + "name": "FLUX.1-dev", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2024-08", + "release_date": "2024-08-01", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 4096, + "output": 0 }, "cost": { - "input": 0.75, - "output": 3.75 + "input": 0, + "output": 0 } }, - "ministral-14b-2512": { - "id": "ministral-14b-2512", - "name": "ministral-14b-2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "attachment": true, + "black-forest-labs/flux_2-klein-4b": { + "id": "black-forest-labs/flux_2-klein-4b", + "name": "FLUX.2 Klein 4B", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "flux", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "knowledge": "2025-06", + "release_date": "2026-01-14", + "last_updated": "2026-01-31", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ - "text" + "image" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 40960, + "output": 40960 }, "cost": { - "input": 0.33, - "output": 0.33 + "input": 0, + "output": 0 } - }, - "glm-5-turbo": { - "id": "glm-5-turbo", - "name": "glm-5-turbo", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, + } + } + }, + "evroc": { + "id": "evroc", + "env": [ + "EVROC_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://models.think.evroc.com/v1", + "name": "evroc", + "doc": "https://docs.evroc.com/products/think/overview.html", + "models": { + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.72, - "output": 3.2 + "input": 1.4375, + "output": 5.75 } }, - "grok-4.1": { - "id": "grok-4.1", - "name": "grok-4.1", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "google/gemma-4-26B-A4B-it": { + "id": "google/gemma-4-26B-A4B-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -105088,309 +99998,289 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 2, - "output": 10 + "input": 0.144, + "output": 0.575 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "claude-sonnet-4-6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-18", - "last_updated": "2026-03-13", + "Qwen/Qwen3-Embedding-8B": { + "id": "Qwen/Qwen3-Embedding-8B", + "name": "Qwen3 Embedding 8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 40960, + "output": 4096 }, "cost": { - "input": 3, - "output": 15 + "input": 0.115, + "output": 0.115 } }, - "gpt-5.3-chat-latest": { - "id": "gpt-5.3-chat-latest", - "name": "GPT-5.3 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "Qwen/Qwen3-Reranker-4B": { + "id": "Qwen/Qwen3-Reranker-4B", + "name": "Qwen3 Reranker 4B", + "description": "Reranking model for improving retrieval quality in search and recommendation systems", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "tool_call": false, + "release_date": "2025-07-30", + "last_updated": "2025-07-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 32000, + "output": 4096 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.0575, + "output": 0 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "Qwen/Qwen3.6-35B-A3B-FP8": { + "id": "Qwen/Qwen3.6-35B-A3B-FP8", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 262144, "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.8 + "input": 0.345, + "output": 1.38 } }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, + "Qwen/Qwen3-VL-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", + "name": "Qwen3 VL 30B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, + "release_date": "2025-07-30", + "last_updated": "2025-07-30", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 100000, + "output": 100000 + }, + "cost": { + "input": 0.23, + "output": 0.92 + } + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 2, - "output": 12 + "input": 0.23, + "output": 0.92 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "deepseek-v3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "openai/whisper-large-v3-turbo": { + "id": "openai/whisper-large-v3-turbo", + "name": "Whisper Large v3 Turbo", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "tool_call": false, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", "modalities": { "input": [ - "text" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 448, + "output": 448 }, "cost": { - "input": 0.29, - "output": 0.43 + "input": 0.0023, + "output": 0.0023, + "output_audio": 2.3 } }, - "claude-opus-4-7-thinking": { - "id": "claude-opus-4-7-thinking", - "name": "claude-opus-4-7-thinking", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "openai/whisper-large-v3": { + "id": "openai/whisper-large-v3", + "name": "Whisper 3 Large", + "description": "Open Whisper checkpoint for robust multilingual transcription and captioning", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 448, + "output": 4096 }, "cost": { - "input": 5, - "output": 25 + "input": 0.0023, + "output": 0.0023, + "output_audio": 2.3 } }, - "o3": { - "id": "o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "mistralai/Mistral-Medium-3.5-128B": { + "id": "mistralai/Mistral-Medium-3.5-128B", + "name": "Mistral Medium 3.5", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "none", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2, - "output": 8 + "input": 1.725, + "output": 6.9 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "mistralai/Voxtral-Small-24B-2507": { + "id": "mistralai/Voxtral-Small-24B-2507", + "name": "Voxtral Small 24B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "voxtral", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "reasoning": false, + "tool_call": false, + "release_date": "2025-03-01", + "last_updated": "2025-03-01", "modalities": { "input": [ + "audio", "text" ], "output": [ @@ -105399,98 +100289,65 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 32000, + "output": 32000 }, "cost": { - "input": 1.4, - "output": 4.4 + "input": 0.0023, + "output": 0.0023, + "output_audio": 2.3 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "claude-opus-4-7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "nvidia/Llama-3.3-70B-Instruct-FP8": { + "id": "nvidia/Llama-3.3-70B-Instruct-FP8", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 1.15, + "output": 1.15 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "evroc/roc": { + "id": "evroc/roc", + "name": "roc", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2026-01", + "release_date": "2026-06-06", + "last_updated": "2026-06-06", "modalities": { "input": [ "text", @@ -105501,133 +100358,193 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, "output": 262144 }, "cost": { - "input": 0.95, - "output": 4 + "input": 2.875, + "output": 11.516 } }, - "gpt-5": { - "id": "gpt-5", - "name": "gpt-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "KBLab/kb-whisper-large": { + "id": "KBLab/kb-whisper-large", + "name": "KB Whisper", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2024-10-01", + "last_updated": "2024-10-01", + "modalities": { + "input": [ + "audio" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 448, + "output": 448 + }, + "cost": { + "input": 0.0023, + "output": 0.0023, + "output_audio": 2.3 + } + }, + "intfloat/multilingual-e5-large-instruct": { + "id": "intfloat/multilingual-e5-large-instruct", + "name": "E5 Multi-Lingual Large Embeddings 0.6B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2024-06-01", + "last_updated": "2024-06-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 512, + "output": 512 + }, + "cost": { + "input": 0.114, + "output": 0.114 + } + } + } + }, + "xiaomi": { + "id": "xiaomi", + "env": [ + "XIAOMI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.xiaomimimo.com/v1", + "name": "Xiaomi", + "doc": "https://platform.xiaomimimo.com/#/docs", + "models": { + "mimo-v2.5-pro-ultraspeed": { + "id": "mimo-v2.5-pro-ultraspeed", + "name": "MiMo-V2.5-Pro-UltraSpeed", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-06-08", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 131072 }, + "status": "beta", "cost": { - "input": 1.25, - "output": 10 + "input": 1.305, + "output": 2.61, + "cache_read": 0.0108 } }, - "grok-4.6": { - "id": "grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-06-24", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 2, - "output": 6 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "gemini-3.1-flash-lite-preview": { - "id": "gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "mimo-v2-omni": { + "id": "mimo-v2-omni", + "name": "MiMo-V2-Omni", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-06-24", "modalities": { "input": [ "text", "image", - "video", "audio", + "video", "pdf" ], "output": [ @@ -105636,25 +100553,36 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.25, - "output": 1.5 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "qwen3-max-2025-09-23": { - "id": "qwen3-max-2025-09-23", - "name": "qwen3-max-2025-09-23", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "mimo-v2-flash": { + "id": "mimo-v2-flash", + "name": "MiMo-V2-Flash", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "knowledge": "2024-12-01", + "release_date": "2025-12-16", + "last_updated": "2026-06-24", "modalities": { "input": [ "text" @@ -105663,41 +100591,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 258048, + "context": 262144, "output": 65536 }, + "status": "deprecated", "cost": { - "input": 0.86, - "output": 3.43 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-06-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -105705,224 +100633,78 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.285, - "output": 1.15 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-06-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 5, - "output": 30 - } - }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5 + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } } } }, - "openrouter": { - "id": "openrouter", + "inception": { + "id": "inception", "env": [ - "OPENROUTER_API_KEY" + "INCEPTION_API_KEY" ], - "npm": "@openrouter/ai-sdk-provider", - "api": "https://openrouter.ai/api/v1", - "name": "OpenRouter", - "doc": "https://openrouter.ai/models", + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.inceptionlabs.ai/v1/", + "name": "Inception", + "doc": "https://platform.inceptionlabs.ai/docs", "models": { - "sao10k/l3-lunaris-8b": { - "id": "sao10k/l3-lunaris-8b", - "name": "Llama 3 8B Lunaris", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-08-13", - "last_updated": "2024-08-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 7372 - }, - "cost": { - "input": 0.04, - "output": 0.05 - } - }, - "sao10k/l3.3-euryale-70b": { - "id": "sao10k/l3.3-euryale-70b", - "name": "Llama 3.3 Euryale 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "mercury-edit-2": { + "id": "mercury-edit-2", + "name": "Mercury Edit 2", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-12-18", - "last_updated": "2024-12-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.65, - "output": 0.75 - } - }, - "sao10k/l3.1-euryale-70b": { - "id": "sao10k/l3.1-euryale-70b", - "name": "Llama 3.1 Euryale 70B v2.2", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-08-28", - "last_updated": "2024-08-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.85, - "output": 0.85 - } - }, - "kwaipilot/kat-coder-pro-v2.5": { - "id": "kwaipilot/kat-coder-pro-v2.5", - "name": "KAT-Coder-Pro V2.5", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "kat-coder", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-10", - "last_updated": "2026-07-10", + "release_date": "2026-03-30", + "last_updated": "2026-03-30", "modalities": { "input": [ "text" @@ -105933,37 +100715,41 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.74, - "output": 2.96, - "cache_read": 0.15 + "input": 0.25, + "output": 0.75, + "cache_read": 0.025 } }, - "bytedance-seed/seed-1.6-flash": { - "id": "bytedance-seed/seed-1.6-flash", - "name": "Seed 1.6 Flash", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, + "mercury-2": { + "id": "mercury-2", + "name": "Mercury 2", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2025-01-01", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" @@ -105971,55 +100757,56 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "output": 50000 }, "cost": { - "input": 0.075, - "output": 0.3, - "tiers": [ - { - "input": 0.1, - "output": 0.8, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.25, + "output": 0.75, + "cache_read": 0.025 } - }, - "bytedance-seed/seed-2.0-lite": { - "id": "bytedance-seed/seed-2.0-lite", - "name": "Seed 2.0 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + } + } + }, + "anthropic": { + "id": "anthropic", + "env": [ + "ANTHROPIC_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "name": "Anthropic", + "doc": "https://docs.anthropic.com/en/docs/about-claude/models", + "models": { + "claude-opus-5-5": { + "id": "claude-opus-5-5", + "name": "Claude Opus 5.5", + "description": "Claude model for long-running agentic coding and knowledge work", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "temperature": false, + "knowledge": "2026-06", + "release_date": "2026-09-22", + "last_updated": "2026-09-22", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -106027,46 +100814,68 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1000000, + "output": 128000 }, - "cost": { - "input": 0.25, - "output": 2, - "tiers": [ - { - "input": 0.5, - "output": 4, - "tier": { - "type": "context", - "size": 128000 + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 8, + "output": 40, + "cache_read": 0.4, + "cache_write": 10 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } } } - ] + } + }, + "cost": { + "input": 4, + "output": 20, + "cache_read": 0.2, + "cache_write": 5 } }, - "bytedance-seed/seed-2-1-turbo": { - "id": "bytedance-seed/seed-2-1-turbo", - "name": "Seed 2.1 Turbo", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -106074,45 +100883,40 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.5, - "output": 2.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "bytedance-seed/seed-2.0-mini": { - "id": "bytedance-seed/seed-2.0-mini", - "name": "Seed 2.0 Mini", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -106120,54 +100924,40 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.4, - "tiers": [ - { - "input": 0.2, - "output": 0.8, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "bytedance-seed/seed-2.0-code": { - "id": "bytedance-seed/seed-2.0-code", - "name": "Seed 2.0 Code", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "seed", + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -106175,46 +100965,41 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 200000, + "output": 32000 }, + "status": "deprecated", "cost": { - "input": 0.5, - "output": 3, - "tiers": [ - { - "input": 1, - "output": 6, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "bytedance-seed/seed-1.6": { - "id": "bytedance-seed/seed-1.6", - "name": "Seed 1.6", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "image", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -106222,54 +101007,46 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 2, - "tiers": [ - { - "input": 0.5, - "output": 4, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "~moonshotai/kimi-latest": { - "id": "~moonshotai/kimi-latest", - "name": "Kimi Latest", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-14", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -106277,161 +101054,41 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 - }, - "cost": { - "input": 1.4989, - "output": 10.758, - "cache_read": 0.3 - } - }, - "poolside/laguna-s-2.1:free": { - "id": "poolside/laguna-s-2.1:free", - "name": "Laguna S 2.1 (free)", - "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", - "family": "laguna-s", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "poolside/laguna-s-2.1": { - "id": "poolside/laguna-s-2.1", - "name": "Laguna S 2.1", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "laguna-s", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.09, - "output": 0.18, - "cache_read": 0.009 - } - }, - "poolside/laguna-xs-2.1:free": { - "id": "poolside/laguna-xs-2.1:free", - "name": "Laguna XS 2.1 (free)", - "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", - "family": "laguna", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-02", - "last_updated": "2026-07-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, - "cost": { - "input": 0, - "output": 0 - } - }, - "poolside/laguna-xs-2.1": { - "id": "poolside/laguna-xs-2.1", - "name": "Laguna XS 2.1", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-02", - "last_updated": "2026-07-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 }, "cost": { - "input": 0.06, - "output": 0.12, - "cache_read": 0.03 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "prism-ml/ternary-bonsai-2-27b": { - "id": "prism-ml/ternary-bonsai-2-27b", - "name": "Ternary Bonsai 2 27B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -106441,53 +101098,20 @@ { "type": "effort", "values": [ + "low", "medium", - "xhigh" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-09-18", - "last_updated": "2026-09-18", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.075, - "output": 0.5 - } - }, - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-29", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", @@ -106500,26 +101124,35 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude Opus 4.5 (latest)", + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "Claude Opus 4.5", "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, @@ -106527,7 +101160,7 @@ "temperature": true, "knowledge": "2025-05", "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", @@ -106550,36 +101183,31 @@ "cache_write": 6.25 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", "high", + "xhigh", "max" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -106595,59 +101223,52 @@ "context": 1000000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } + }, "cost": { "input": 5, "output": 25, "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "cache_write": 6.25 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -106659,80 +101280,26 @@ ] }, "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "anthropic/claude-3-haiku": { - "id": "anthropic/claude-3-haiku", - "name": "Claude 3 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-03-13", - "last_updated": "2024-03-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, "limit": { "context": 200000, - "output": 4096 + "output": 32000 }, + "status": "deprecated", "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.03, - "cache_write": 0.3 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -106746,10 +101313,9 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": false, + "release_date": "2026-06-07", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", @@ -106766,14 +101332,14 @@ "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", "name": "Claude Haiku 4.5 (latest)", "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", "family": "claude-haiku", @@ -106781,7 +101347,8 @@ "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, @@ -106812,34 +101379,34 @@ "cache_write": 1.25 } }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", "high", - "xhigh", "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "knowledge": "2025-05-31", + "release_date": "2026-02-04", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -106855,6 +101422,26 @@ "context": 1000000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } + }, "cost": { "input": 5, "output": 25, @@ -106862,31 +101449,25 @@ "cache_write": 6.25 } }, - "anthropic/claude-fable-5.1": { - "id": "anthropic/claude-fable-5.1", - "name": "Claude Fable 5.1", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", @@ -106900,33 +101481,43 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 64000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -106940,63 +101531,118 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 128000 }, "cost": { "input": 3, "output": 15, "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "cache_write": 3.75 + } + } + } + }, + "tencent-coding-plan": { + "id": "tencent-coding-plan", + "env": [ + "TENCENT_CODING_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.lkeap.cloud.tencent.com/coding/v3", + "name": "Tencent Coding Plan (China)", + "doc": "https://cloud.tencent.com/document/product/1772/128947", + "models": { + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", "image", - "pdf" + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "hunyuan-turbos": { + "id": "hunyuan-turbos", + "name": "Hunyuan-TurboS", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-03-08", + "last_updated": "2026-03-08", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -107004,22 +101650,88 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "hunyuan-t1": { + "id": "hunyuan-t1", + "name": "Hunyuan-T1", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-03-08", + "last_updated": "2026-03-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "tc-code-latest": { + "id": "tc-code-latest", + "name": "Auto", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "auto", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-03-08", + "last_updated": "2026-03-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -107027,15 +101739,132 @@ } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 202752, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "hunyuan-2.0-instruct": { + "id": "hunyuan-2.0-instruct", + "name": "Tencent HY 2.0 Instruct", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-03-08", + "last_updated": "2026-03-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "hunyuan-2.0-thinking": { + "id": "hunyuan-2.0-thinking", + "name": "Tencent HY 2.0 Think", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-03-08", + "last_updated": "2026-03-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + } + } + }, + "freemodel": { + "id": "freemodel", + "env": [ + "FREEMODEL_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://cc.freemodel.dev/v1", + "name": "FreeModel", + "doc": "https://freemodel.dev", + "models": { + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "image", "text", + "image", "pdf" ], "output": [ @@ -107048,41 +101877,20 @@ "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -107095,11 +101903,10 @@ } ], "tool_call": true, - "structured_output": true, "temperature": false, "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -107116,39 +101923,36 @@ "output": 128000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -107165,55 +101969,37 @@ "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "anthropic/claude-opus-5.5": { - "id": "anthropic/claude-opus-5.5", - "name": "Claude Opus 5.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", @@ -107226,63 +102012,72 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.freemodel.dev/v1" + }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 + "input": 1.75, + "output": 14, + "cache_read": 0.175, + "cache_write": 1.75 } }, - "nex-agi/nex-n2.5-mini": { - "id": "nex-agi/nex-n2.5-mini", - "name": "Nex-N2.5-Mini", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "agi", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "tool_call": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.025, - "output": 0.1, - "cache_read": 0.0025 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "nex-agi/nex-n2.5-mini:free": { - "id": "nex-agi/nex-n2.5-mini:free", - "name": "Nex-N2.5-Mini (free)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "agi", + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -107290,40 +102085,51 @@ "type": "effort", "values": [ "none", + "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.freemodel.dev/v1" }, "cost": { - "input": 0, - "output": 0 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 2.5 } }, - "nex-agi/nex-n2.5-pro:free": { - "id": "nex-agi/nex-n2.5-pro:free", - "name": "Nex-N2.5-Pro (free)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "agi", + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -107331,16 +102137,19 @@ "type": "effort", "values": [ "none", + "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -107350,110 +102159,151 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.freemodel.dev/v1" }, "cost": { - "input": 0, - "output": 0 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075, + "cache_write": 0.75 } }, - "nex-agi/nex-n2.5-pro": { - "id": "nex-agi/nex-n2.5-pro", - "name": "Nex-N2.5-Pro", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "agi", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.075, - "output": 0.25, - "cache_read": 0.015 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "cohere/command-r-08-2024": { - "id": "cohere/command-r-08-2024", - "name": "Command R", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "cohere/command-a-plus": { - "id": "cohere/command-a-plus", - "name": "Command A+", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -107461,251 +102311,317 @@ }, "open_weights": false, "limit": { - "context": 192000, - "output": 64000 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.freemodel.dev/v1" }, "cost": { - "input": 0.3, - "output": 1.5, - "cache_read": 0.15 + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 5 } - }, - "cohere/command-a": { - "id": "cohere/command-a", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", - "attachment": false, - "reasoning": false, - "tool_call": false, + } + } + }, + "sap-ai-core": { + "id": "sap-ai-core", + "env": [ + "AICORE_SERVICE_KEY" + ], + "npm": "@jerome-benoit/sap-ai-provider-v2", + "name": "SAP AI Core", + "doc": "https://help.sap.com/docs/sap-ai-core", + "models": { + "anthropic--claude-4.8-opus": { + "id": "anthropic--claude-4.8-opus", + "name": "anthropic--claude-4.8-opus", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "cohere/north-mini-code:free": { - "id": "cohere/north-mini-code:free", - "name": "North Mini Code (free)", - "description": "Cohere coding model for practical software engineering and agentic edits", - "family": "north", - "attachment": false, + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "gemini-3.1-flash-lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "cohere/command-r7b-12-2024": { - "id": "cohere/command-r7b-12-2024", - "name": "Command R7B", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "anthropic--claude-4.6-sonnet": { + "id": "anthropic--claude-4.6-sonnet", + "name": "anthropic--claude-4.6-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-12-02", - "last_updated": "2024-12-02", + "knowledge": "2025-08", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.0375, - "output": 0.15 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "cohere/command-r-plus-08-2024": { - "id": "cohere/command-r-plus-08-2024", - "name": "Command R+", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", - "attachment": false, + "anthropic--claude-3-sonnet": { + "id": "anthropic--claude-3-sonnet", + "name": "anthropic--claude-3-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2023-08-31", + "release_date": "2024-03-04", + "last_updated": "2024-03-04", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4000 + "context": 200000, + "output": 4096 }, "cost": { - "input": 2.5, - "output": 10 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek/deepseek-chat-v3.1": { - "id": "deepseek/deepseek-chat-v3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "anthropic--claude-4-sonnet": { + "id": "anthropic--claude-4-sonnet", + "name": "anthropic--claude-4-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-03-31", - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 0.95, - "cache_read": 0.13 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek/deepseek-v4-flash-0731": { - "id": "deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "gemini-2.5-pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2025-01", + "release_date": "2025-03-25", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1310720, - "output": 943718 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.04, - "output": 0.64, - "cache_read": 0.016 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "deepseek/deepseek-v4.1-flash": { - "id": "deepseek/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", + "gpt-5": { + "id": "gpt-5", + "name": "gpt-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "minimal", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -107715,22 +102631,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "deepseek/deepseek-v4-flash-vision-exp": { - "id": "deepseek/deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "gemini-2.5-flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -107738,384 +102655,425 @@ "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "budget_tokens", + "min": 0, + "max": 24576 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", - "modalities": { - "input": [ + "knowledge": "2025-01", + "release_date": "2025-04-17", + "last_updated": "2025-06-05", + "modalities": { + "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 943718 + "output": 65536 }, "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "deepseek/deepseek-r1-distill-llama-70b": { - "id": "deepseek/deepseek-r1-distill-llama-70b", - "name": "R1 Distill Llama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "gemini-3.5-flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2025-01-23", - "last_updated": "2025-01-23", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 7372 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.8, - "output": 0.8 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 } }, - "deepseek/deepseek-v3.1-terminus": { - "id": "deepseek/deepseek-v3.1-terminus", - "name": "DeepSeek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "anthropic--claude-4.5-haiku": { + "id": "anthropic--claude-4.5-haiku", + "name": "anthropic--claude-4.5-haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.27, - "output": 1, - "cache_read": 0.135 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "deepseek/deepseek-chat": { - "id": "deepseek/deepseek-chat", - "name": "DeepSeek Chat", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "anthropic--claude-3-haiku": { + "id": "anthropic--claude-3-haiku", + "name": "anthropic--claude-3-haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "knowledge": "2023-08-31", + "release_date": "2024-03-13", + "last_updated": "2024-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 16384 + "context": 200000, + "output": 4096 }, "cost": { - "input": 0.32, - "output": 0.89 + "input": 0.25, + "output": 1.25, + "cache_read": 0.03, + "cache_write": 0.3 } }, - "deepseek/deepseek-v4-pro-0813": { - "id": "deepseek/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, + "anthropic--claude-4-opus": { + "id": "anthropic--claude-4-opus", + "name": "anthropic--claude-4-opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.462, - "output": 1.386, - "cache_read": 0.0154 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "deepseek/deepseek-v3.2-exp": { - "id": "deepseek/deepseek-v3.2-exp", - "name": "DeepSeek V3.2 Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "anthropic--claude-4.5-sonnet": { + "id": "anthropic--claude-4.5-sonnet", + "name": "anthropic--claude-4.5-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2025-07-31", "release_date": "2025-09-29", "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.27, - "output": 0.41 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek/deepseek-chat-v3-0324": { - "id": "deepseek/deepseek-chat-v3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "anthropic--claude-3.5-sonnet": { + "id": "anthropic--claude-3.5-sonnet", + "name": "anthropic--claude-3.5-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "knowledge": "2024-04-30", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 147456 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.25, - "output": 1 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek/deepseek-r1": { - "id": "deepseek/deepseek-r1", - "name": "DeepSeek-R1", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", - "attachment": false, + "anthropic--claude-4.6-opus": { + "id": "anthropic--claude-4.6-opus", + "name": "anthropic--claude-4.6-opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "knowledge": "2025-05", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 64000, - "output": 16000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.7, - "output": 2.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "gemini-2.5-flash-lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 512, + "max": 24576 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 384000 + "output": 65536 }, "cost": { - "input": 0.951432, - "output": 1.902864, - "cache_read": 0.079286 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "input_audio": 0.3 } }, - "deepseek/deepseek-r1-0528": { - "id": "deepseek/deepseek-r1-0528", - "name": "R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek", - "attachment": false, + "anthropic--claude-3.7-sonnet": { + "id": "anthropic--claude-3.7-sonnet", + "name": "anthropic--claude-3.7-sonnet", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "knowledge": "2024-10-31", + "release_date": "2025-02-24", + "last_updated": "2025-02-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.5, - "output": 2.15, - "cache_read": 0.35 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "sonar": { + "id": "sonar", + "name": "sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text" @@ -108124,79 +103082,112 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 65536 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.269, - "output": 0.4, - "cache_read": 0.1345 + "input": 1, + "output": 1 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "sonar-pro": { + "id": "sonar-pro", + "name": "sonar-pro", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 8192 + }, + "cost": { + "input": 3, + "output": 15 + } + }, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "gpt-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "low", + "medium", "high", "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.08246, - "output": 0.16492, - "cache_read": 0.016492 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "unbiased/pareto": { - "id": "unbiased/pareto", - "name": "Pareto", - "description": "Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer", + "gpt-4.1": { + "id": "gpt-4.1", + "name": "gpt-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -108204,134 +103195,151 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.25 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "tencent/hy3-preview": { - "id": "tencent/hy3-preview", - "name": "Hy3 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], + "anthropic--claude-3-opus": { + "id": "anthropic--claude-3-opus", + "name": "anthropic--claude-3-opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "knowledge": "2023-08-31", + "release_date": "2024-02-29", + "last_updated": "2024-02-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 200000, + "output": 4096 }, "cost": { - "input": 0.18, - "output": 0.6, - "cache_read": 0.06 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "tencent/hunyuan-a13b-instruct": { - "id": "tencent/hunyuan-a13b-instruct", - "name": "Hunyuan A13B Instruct", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", - "attachment": false, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "gpt-5-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], - "tool_call": false, + "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-07-08", - "last_updated": "2025-07-08", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "tencent/hy-mt2-30b-a3b": { - "id": "tencent/hy-mt2-30b-a3b", - "name": "Hy-MT2-30B-A3B", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "gpt-4.1-mini": { + "id": "gpt-4.1-mini", + "name": "gpt-4.1-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-20", - "last_updated": "2026-08-20", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.074, - "output": 0.295 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "tencent/hy-mt2-1.8b": { - "id": "tencent/hy-mt2-1.8b", - "name": "Hy-MT2-1.8B", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + "sonar-deep-research": { + "id": "sonar-deep-research", + "name": "sonar-deep-research", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar-deep-research", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-20", - "last_updated": "2026-08-20", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-02-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text" @@ -108340,176 +103348,234 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.044, - "output": 0.177 + "input": 2, + "output": 8, + "reasoning": 3 } }, - "tencent/hy3": { - "id": "tencent/hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "gpt-5-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", + "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 192000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 0.0825, - "output": 0.33, - "cache_read": 0.020625 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "tencent/hy-mt2-7b": { - "id": "tencent/hy-mt2-7b", - "name": "Hy-MT2-7B", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "anthropic--claude-4.5-opus": { + "id": "anthropic--claude-4.5-opus", + "name": "anthropic--claude-4.5-opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, "temperature": true, - "release_date": "2026-08-19", - "last_updated": "2026-08-19", + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.074, - "output": 0.295 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "tencent/hy4-preview": { - "id": "tencent/hy4-preview", - "name": "Hy4 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "anthropic--claude-4.7-opus": { + "id": "anthropic--claude-4.7-opus", + "name": "anthropic--claude-4.7-opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "high" + "medium", + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.834, - "output": 2.501, - "cache_read": 0.042 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "anthracite-org/magnum-v4-72b": { - "id": "anthracite-org/magnum-v4-72b", - "name": "Magnum v4 72B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "gpt-5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 4096 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 5 + "input": 5, + "output": 30, + "cache_read": 0.5 } - }, - "meta-llama/llama-4-scout": { - "id": "meta-llama/llama-4-scout", - "name": "Llama 4 Scout", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": true, - "reasoning": false, + } + } + }, + "opencode": { + "id": "opencode", + "env": [ + "OPENCODE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://opencode.ai/zen/v1", + "name": "OpenCode Zen", + "doc": "https://opencode.ai/docs/zen", + "models": { + "ring-2.6-1t-free": { + "id": "ring-2.6-1t-free", + "name": "Ring 2.6 1T Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "ring-1t-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-06", + "release_date": "2026-05-08", + "last_updated": "2026-05-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -108517,30 +103583,33 @@ }, "open_weights": true, "limit": { - "context": 1310720, - "output": 16384 + "context": 262000, + "output": 66000 }, + "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.3 + "input": 0, + "output": 0 } }, - "meta-llama/llama-guard-4-12b": { - "id": "meta-llama/llama-guard-4-12b", - "name": "Llama Guard 4 12B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", + "mimo-v2-pro-free": { + "id": "mimo-v2-pro-free", + "name": "MiMo V2 Pro Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo-pro-free", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-04-30", - "last_updated": "2025-04-30", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -108549,31 +103618,47 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 16384 + "context": 1048576, + "output": 64000 }, + "status": "deprecated", "cost": { - "input": 0.18, - "output": 0.18 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "meta-llama/llama-4-maverick": { - "id": "meta-llama/llama-4-maverick", - "name": "Llama 4 Maverick", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": true, - "reasoning": false, + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -108581,27 +103666,31 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 16384 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.1875, - "output": 0.6525 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "meta-llama/llama-3.3-70b-instruct": { - "id": "meta-llama/llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -108612,27 +103701,35 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.32 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "meta-llama/llama-3.1-8b-instruct": { - "id": "meta-llama/llama-3.1-8b-instruct", - "name": "Llama-3.1-8B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Legacy model retained for compatibility with older integrations", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -108643,31 +103740,38 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 117964 + "context": 204800, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.05, - "output": 0.08, - "cache_read": 0.025 + "input": 0.6, + "output": 2.2, + "cache_read": 0.1 } }, - "meta-llama/llama-3.2-1b-instruct": { - "id": "meta-llama/llama-3.2-1b-instruct", - "name": "Llama 3.2 1B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "mimo-v2.5-free": { + "id": "mimo-v2.5-free", + "name": "MiMo V2.5 Free", + "description": "MiMo omni model for text, image, video, audio, and agents", + "family": "mimo-v2.5-free", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2024-12", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" @@ -108675,27 +103779,27 @@ }, "open_weights": true, "limit": { - "context": 60000, - "output": 54000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.027, - "output": 0.201 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "meta-llama/llama-3.2-3b-instruct": { - "id": "meta-llama/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "kimi-k2": { + "id": "kimi-k2", + "name": "Kimi K2", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-k2", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -108706,27 +103810,32 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 117964 + "context": 262144, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 0.05, - "output": 0.33 + "input": 0.4, + "output": 2.5, + "cache_read": 0.4 } }, - "meta-llama/llama-3.1-70b-instruct": { - "id": "meta-llama/llama-3.1-70b-instruct", - "name": "Llama-3.1-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "minimax-m2.1": { + "id": "minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-01", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -108737,147 +103846,122 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 204800, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.4, - "output": 0.4 + "input": 0.3, + "output": 1.2, + "cache_read": 0.1 } }, - "~google/gemini-flash-latest": { - "id": "~google/gemini-flash-latest", - "name": "Gemini Flash Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "nemotron-3-ultra-free": { + "id": "nemotron-3-ultra-free", + "name": "Nemotron 3 Ultra Free", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron-free", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2026-02", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image", - "video", - "pdf", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "~google/gemini-pro-latest": { - "id": "~google/gemini-pro-latest", - "name": "Gemini Pro Latest", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "glm-4.7-free": { + "id": "glm-4.7-free", + "name": "GLM-4.7 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "glm-free", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "audio", - "pdf", - "image", - "text", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 204800, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0, + "output": 0, + "cache_read": 0 } }, - "perceptron/perceptron-mk1": { - "id": "perceptron/perceptron-mk1", - "name": "Perceptron Mk1", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "gemini-3-flash": { + "id": "gemini-3-flash", + "name": "Gemini 3 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -108885,37 +103969,49 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google" }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "z-ai/glm-4.6v": { - "id": "z-ai/glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, + "deepseek-v4-flash-free": { + "id": "deepseek-v4-flash-free", + "name": "DeepSeek V4 Flash Free", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash-free", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -108923,74 +104019,107 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.055 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "z-ai/glm-4.5": { - "id": "z-ai/glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, + "claude-sonnet-4": { + "id": "claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "z-ai/glm-5v-turbo": { - "id": "z-ai/glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "image", "text", - "video" + "image", + "pdf" ], "output": [ "text" @@ -108998,100 +104127,123 @@ }, "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "z-ai/glm-5.3-flash": { - "id": "z-ai/glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1310720, - "output": 943718 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.05 + "input": 1.07, + "output": 8.5, + "cache_read": 0.107 } }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 16384 + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google" }, "cost": { - "input": 0.43, - "output": 1.75, - "cache_read": 0.08 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "minimax-m3-free": { + "id": "minimax-m3-free", + "name": "MiniMax-M3 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax-m3-free", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -109100,16 +104252,15 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-01", + "release_date": "2026-05-31", + "last_updated": "2026-05-31", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -109117,75 +104268,104 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 128000 + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.6, - "output": 1.92, - "cache_read": 0.12 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "z-ai/glm-4.5v": { - "id": "z-ai/glm-4.5v", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "gemini-3-pro": { + "id": "gemini-3-pro", + "name": "Gemini 3 Pro", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "high" + ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 1048576, + "output": 65536 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/google" }, "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "z-ai/glm-5.3-flashx": { - "id": "z-ai/glm-5.3-flashx", - "name": "GLM 5.3 FlashX", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "kimi-k2.5-free": { + "id": "kimi-k2.5-free", + "name": "Kimi K2.5 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-free", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-09-18", - "last_updated": "2026-09-18", + "knowledge": "2024-10", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", @@ -109196,41 +104376,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 0.37, - "output": 1.25, - "cache_read": 0.075 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "z-ai/glm-4.7-flash": { - "id": "z-ai/glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { - "field": "reasoning_details" + "field": "reasoning_content" }, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -109238,18 +104417,19 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 117964 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.0605, - "output": 0.4 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, @@ -109260,13 +104440,12 @@ ], "tool_call": true, "interleaved": { - "field": "reasoning_details" + "field": "reasoning_content" }, - "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -109281,16 +104460,16 @@ "output": 131072 }, "cost": { - "input": 0.4, - "output": 1.75, - "cache_read": 0.08 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -109301,7 +104480,7 @@ "type": "effort", "values": [ "high", - "xhigh" + "max" ] } ], @@ -109311,8 +104490,9 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -109323,19 +104503,19 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.6496, - "output": 2.0416, - "cache_read": 0.12064 + "input": 1.74, + "output": 3.84, + "cache_read": 0.145 } }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "glm-4.6": { + "id": "glm-4.6", + "name": "GLM-4.6", + "description": "Legacy model retained for compatibility with older integrations", "family": "glm", "attachment": false, "reasoning": true, @@ -109345,13 +104525,10 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -109363,34 +104540,31 @@ "open_weights": true, "limit": { "context": 204800, - "output": 128000 + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.966, - "output": 3.036, - "cache_read": 0.1794 + "input": 0.6, + "output": 2.2, + "cache_read": 0.1 } }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": false, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "knowledge": "2024-10", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -109399,178 +104573,240 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 262144, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 0.4, + "output": 2.5, + "cache_read": 0.4 } }, - "z-ai/glm-5.2:free": { - "id": "z-ai/glm-5.2:free", - "name": "GLM 5.2 (free)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 1024 } ], - "tool_call": false, - "structured_output": false, + "tool_call": true, + "interleaved": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 29491 + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "z-ai/glm-4.5-air": { - "id": "z-ai/glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", - "attachment": false, + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.13, - "output": 0.85, - "cache_read": 0.025 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "z-ai/glm-5.3": { - "id": "z-ai/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1310720, - "output": 131072 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.84, - "output": 2.64, - "cache_read": 0.156 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "inference-net/schematron-v2-small": { - "id": "inference-net/schematron-v2-small", - "name": "Schematron V2 Small", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-09-12", - "last_updated": "2026-09-12", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.05, - "output": 0.23, - "cache_read": 0.05 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "inference-net/schematron-v2-turbo": { - "id": "inference-net/schematron-v2-turbo", - "name": "Schematron V2 Turbo", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "qwen3-coder": { + "id": "qwen3-coder", + "name": "Qwen3 Coder", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "release_date": "2026-09-12", - "last_updated": "2026-09-12", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -109581,45 +104817,42 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 262144, + "output": 65536 }, + "status": "deprecated", "cost": { - "input": 0.03, - "output": 0.15, - "cache_read": 0.03 + "input": 0.45, + "output": 1.8 } }, - "thinkingmachines/inkling-small:free": { - "id": "thinkingmachines/inkling-small:free", - "name": "Inkling Small (free)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ling", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", - "low", - "medium", "high", "max" ] } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -109627,44 +104860,31 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 262144 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "thinkingmachines/inkling-small": { - "id": "thinkingmachines/inkling-small", - "name": "Inkling Small", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ling", - "attachment": true, + "hy3-preview-free": { + "id": "hy3-preview-free", + "name": "Hy3 preview Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "hy3-free", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "knowledge": "2025-06", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -109672,45 +104892,36 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 262144 + "context": 256000, + "output": 64000 }, + "status": "deprecated", "cost": { - "input": 0.45, - "output": 1.2, - "cache_read": 0.1 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "thinkingmachines/inkling:free": { - "id": "thinkingmachines/inkling:free", - "name": "Inkling (free)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ling", + "minimax-m3": { + "id": "minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", "image", - "audio" + "video" ], "output": [ "text" @@ -109718,110 +104929,86 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 262144 + "context": 512000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "thinkingmachines/inkling": { - "id": "thinkingmachines/inkling", - "name": "Inkling", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ling", + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 471859 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.17 + "input": 1.07, + "output": 8.5, + "cache_read": 0.107 } }, - "meituan/longcat-2.0": { - "id": "meituan/longcat-2.0", - "name": "LongCat 2.0", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "longcat", + "gpt-5.3-codex-spark": { + "id": "gpt-5.3-codex-spark", + "name": "GPT-5.3 Codex Spark", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex-spark", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-20", - "last_updated": "2026-07-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048756, - "output": 262144 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.006 - } - }, - "openrouter/bodybuilder": { - "id": "openrouter/bodybuilder", - "name": "Body Builder (beta)", - "description": "Preview model for early access evaluation, prototyping, and compatibility testing", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "structured_output": true, "temperature": false, - "release_date": "2025-12-05", - "last_updated": "2025-12-05", + "knowledge": "2025-08-31", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -109833,21 +105020,42 @@ "open_weights": false, "limit": { "context": 128000, + "input": 128000, "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openrouter/free": { - "id": "openrouter/free", - "name": "Free Models Router", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "gpt-5.1-codex-max": { + "id": "gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -109859,29 +105067,48 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 200000, - "output": 8000 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openrouter/pareto-code": { - "id": "openrouter/pareto-code", - "name": "Pareto Code Router", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -109889,21 +105116,31 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 200000 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openrouter/fusion": { - "id": "openrouter/fusion", - "name": "Fusion", - "description": "General-purpose chat model for instruction following, writing, and analysis", + "ling-2.6-flash-free": { + "id": "ling-2.6-flash-free", + "name": "Ling 2.6 Flash Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "ling-flash-free", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "tool_call": true, + "temperature": true, + "knowledge": "2025-06", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text" @@ -109912,60 +105149,80 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262100, + "output": 32800 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0 } }, - "openrouter/auto": { - "id": "openrouter/auto", - "name": "Auto Router", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "auto", + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.5", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2023-11-08", - "last_updated": "2023-11-08", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", "image", - "audio", - "pdf", "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 262144, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0.2, + "output": 1.2, + "cache_read": 0.02, + "cache_write": 0.25 } }, - "perplexity/sonar-pro": { - "id": "perplexity/sonar-pro", - "name": "Sonar Pro", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", + "claude-3-5-haiku": { + "id": "claude-3-5-haiku", + "name": "Claude Haiku 3.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-haiku", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2025-03-07", - "last_updated": "2025-03-07", + "knowledge": "2024-07-31", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -109974,61 +105231,44 @@ "open_weights": false, "limit": { "context": 200000, - "output": 8000 + "output": 8192 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 3, - "output": 15 + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 } }, - "perplexity/sonar-deep-research": { - "id": "perplexity/sonar-deep-research", - "name": "Sonar Deep Research", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar-deep-research", - "attachment": false, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-03-07", - "last_updated": "2025-03-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 115200 - }, - "cost": { - "input": 2, - "output": 8, - "reasoning": 3 - } - }, - "perplexity/sonar": { - "id": "perplexity/sonar", - "name": "Sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-01-27", - "last_updated": "2025-01-27", + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -110040,35 +105280,49 @@ }, "open_weights": false, "limit": { - "context": 127072, - "output": 114364 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 1, - "output": 1 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "perplexity/sonar-reasoning-pro": { - "id": "perplexity/sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded reasoning model for multi-step research and cited answers", - "family": "sonar-reasoning", + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-03-07", - "last_updated": "2025-03-07", + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -110076,78 +105330,80 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 115200 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 2, - "output": 8 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "perplexity/sonar-pro-search": { - "id": "perplexity/sonar-pro-search", - "name": "Sonar Pro Search", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", - "attachment": true, + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": true, + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-10-30", - "last_updated": "2025-10-30", + "knowledge": "2025-01", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 3, - "output": 15 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "meta/muse-spark-1.3": { - "id": "meta/muse-spark-1.3", - "name": "Muse Spark 1.3", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "muse", + "grok-4.5": { + "id": "grok-4.5", + "name": "Grok 4.5", + "description": "xAI's latest Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text", - "image", - "video", - "pdf", - "audio" + "image" ], "output": [ "text" @@ -110155,43 +105411,59 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 500000, + "output": 500000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 2, + "output": 6, + "cache_read": 0.5, + "tiers": [ + { + "input": 4, + "output": 12, + "cache_read": 1, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } }, - "meta/muse-glimmer-30b": { - "id": "meta/muse-glimmer-30b", - "name": "Muse Glimmer 30B", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, + "north-mini-code-free": { + "id": "north-mini-code-free", + "name": "North Mini Code Free", + "description": "Cohere coding model for practical software engineering and agentic edits", + "family": "north-free", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high", - "xhigh" + "none", + "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "knowledge": "2025-09-23", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -110199,46 +105471,70 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.04 + "input": 0, + "output": 0 } }, - "meta/muse-spark-1.1": { - "id": "meta/muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "muse", + "grok-code": { + "id": "grok-code", + "name": "Grok Code Fast 1", + "description": "Legacy model retained for compatibility with older integrations", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-08-20", + "last_updated": "2025-08-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 256000 + }, + "status": "deprecated", + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", - "video", - "pdf", - "audio" + "pdf" ], "output": [ "text" @@ -110246,46 +105542,46 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 200000, + "output": 32000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "meta/muse-spark-1.2": { - "id": "meta/muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "video", - "pdf", - "audio" + "image" ], "output": [ "text" @@ -110293,74 +105589,71 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "meta/muse-spark-1.2-contributor": { - "id": "meta/muse-spark-1.2-contributor", - "name": "Muse Spark 1.2 Contributor", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "muse", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-08-21", + "knowledge": "2024-10", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", "image", - "video", - "pdf", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 943718 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 + "input": 0.6, + "output": 3, + "cache_read": 0.08 } }, - "meta/muse-spark-1.3-contributor": { - "id": "meta/muse-spark-1.3-contributor", - "name": "Muse Spark 1.3 Contributor", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "muse", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high", @@ -110370,17 +105663,15 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", "image", - "video", - "pdf", - "audio" + "pdf" ], "output": [ "text" @@ -110388,59 +105679,31 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 - }, - "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 - } - }, - "nousresearch/hermes-3-llama-3.1-70b": { - "id": "nousresearch/hermes-3-llama-3.1-70b", - "name": "Hermes 3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "nousresearch", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-08-18", - "last_updated": "2024-08-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1000000, + "output": 128000 }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.7, - "output": 0.7 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "nousresearch/hermes-3-llama-3.1-405b": { - "id": "nousresearch/hermes-3-llama-3.1-405b", - "name": "Hermes 3 405B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "nousresearch", + "hy3-free": { + "id": "hy3-free", + "name": "Hy3 Free", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hy3-free", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-08-16", - "last_updated": "2024-08-16", + "release_date": "2026-06-26", + "last_updated": "2026-06-26", "modalities": { "input": [ "text" @@ -110451,77 +105714,88 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 256000, + "output": 64000 }, "cost": { - "input": 1, - "output": 1 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "nousresearch/hermes-4-405b": { - "id": "nousresearch/hermes-4-405b", - "name": "Hermes 4 405B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "hermes", - "attachment": false, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { "input": 1, - "output": 3 + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "~z-ai/glm-flash-latest": { - "id": "~z-ai/glm-flash-latest", - "name": "GLM Flash Latest", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -110529,40 +105803,65 @@ }, "open_weights": false, "limit": { - "context": 1310720, - "output": 131072 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.075, - "output": 0.25, - "cache_read": 0.015 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "~z-ai/glm-latest": { - "id": "~z-ai/glm-latest", - "name": "GLM Latest", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-19", - "last_updated": "2026-08-19", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -110570,28 +105869,32 @@ }, "open_weights": false, "limit": { - "context": 1310720, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.5614, - "output": 1.7644, - "cache_read": 0.10426 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "arcee-ai/trinity-large-thinking": { - "id": "arcee-ai/trinity-large-thinking", - "name": "Trinity Large Thinking", - "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", - "family": "trinity", + "minimax-m2.5-free": { + "id": "minimax-m2.5-free", + "name": "MiniMax-M2.5 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax-free", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-05-28", + "knowledge": "2025-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -110602,112 +105905,92 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 80000 + "context": 204800, + "output": 131072 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.25, - "output": 0.8, - "cache_read": 0.06 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "~openai/gpt-terra-latest": { - "id": "~openai/gpt-terra-latest", - "name": "GPT Terra Latest", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-terra", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "output": 128000 - }, + "context": 262144, + "output": 65536 + }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "~openai/gpt-luna-latest": { - "id": "~openai/gpt-luna-latest", - "name": "GPT Luna Latest", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-luna", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh", "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -110715,65 +105998,83 @@ }, "open_weights": false, "limit": { - "context": 1050000, + "context": 1000000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/anthropic" + }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 272000 - } - } + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 + } + }, + "minimax-m2.1-free": { + "id": "minimax-m2.1-free", + "name": "MiniMax-M2.1 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax-free", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0 } }, - "~openai/gpt-sol-latest": { - "id": "~openai/gpt-sol-latest", - "name": "GPT Sol Latest", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-sol", + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "pdf", - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -110781,63 +106082,47 @@ }, "open_weights": false, "limit": { - "context": 1050000, + "context": 400000, + "input": 272000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai" + }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "~openai/gpt-astra-latest": { - "id": "~openai/gpt-astra-latest", - "name": "GPT Astra Latest", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-astra", + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": false, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -110846,63 +106131,51 @@ "open_weights": false, "limit": { "context": 1050000, + "input": 922000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai" + }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } + "input": 30, + "output": 180, + "cache_read": 30 } }, - "~openai/gpt-mini-latest": { - "id": "~openai/gpt-mini-latest", - "name": "GPT Mini Latest", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, - "temperature": false, + "interleaved": true, + "temperature": true, "knowledge": "2025-08-31", - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -110910,28 +106183,35 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "cognitivecomputations/dolphin-mistral-24b-venice-edition": { - "id": "cognitivecomputations/dolphin-mistral-24b-venice-edition", - "name": "Uncensored", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", + "mimo-v2-flash-free": { + "id": "mimo-v2-flash-free", + "name": "MiMo V2 Flash Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo-flash-free", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "knowledge": "2024-12", + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ "text" @@ -110942,57 +106222,93 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 262144, + "output": 65536 }, + "status": "deprecated", "cost": { - "input": 0.2, - "output": 0.9 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "thedrummer/skyfall-36b-v2": { - "id": "thedrummer/skyfall-36b-v2", - "name": "Skyfall 36B V2", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gemini-3.1-pro": { + "id": "gemini-3.1-pro", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2025-03-10", - "last_updated": "2025-03-10", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 29491 + "context": 1048576, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/google" }, "cost": { - "input": 0.55, - "output": 0.8, - "cache_read": 0.25 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "thedrummer/unslopnemo-12b": { - "id": "thedrummer/unslopnemo-12b", - "name": "UnslopNemo 12B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "trinity-large-preview-free": { + "id": "trinity-large-preview-free", + "name": "Trinity Large Preview", + "description": "Legacy model retained for compatibility with older integrations", + "family": "trinity", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2024-11-08", - "last_updated": "2024-11-08", + "knowledge": "2025-06", + "release_date": "2026-01-28", + "last_updated": "2026-01-28", "modalities": { "input": [ "text" @@ -111003,26 +106319,32 @@ }, "open_weights": true, "limit": { - "context": 1024000, - "output": 819200 + "context": 131072, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.4, - "output": 0.4 + "input": 0, + "output": 0 } }, - "thedrummer/cydonia-24b-v4.1": { - "id": "thedrummer/cydonia-24b-v4.1", - "name": "Cydonia 24B V4.1", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "big-pickle": { + "id": "big-pickle", + "name": "Big Pickle", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "big-pickle", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2025-09-27", - "last_updated": "2025-09-27", + "knowledge": "2025-01", + "release_date": "2025-10-17", + "last_updated": "2025-10-17", "modalities": { "input": [ "text" @@ -111031,83 +106353,93 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 200000, + "input": 160000, + "output": 32000 }, "cost": { - "input": 0.3, - "output": 0.5, - "cache_read": 0.15 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "baidu/ernie-4.5-vl-424b-a47b": { - "id": "baidu/ernie-4.5-vl-424b-a47b", - "name": "ERNIE 4.5 VL 424B A47B ", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ernie", + "gpt-5.5-pro": { + "id": "gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] } ], - "tool_call": false, + "tool_call": true, "structured_output": false, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-06-30", - "last_updated": "2025-06-30", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 123000, - "output": 16000 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 0.42, - "output": 1.25 + "input": 30, + "output": 180, + "cache_read": 30 } }, - "x-ai/grok-4.20-multi-agent": { - "id": "x-ai/grok-4.20-multi-agent", - "name": "Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "qwen3.6-plus-free": { + "id": "qwen3.6-plus-free", + "name": "Qwen3.6 Plus Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen-free", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -111115,157 +106447,102 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 1800000 + "context": 262144, + "output": 65536 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0, + "output": 0, + "cache_read": 0 } }, - "x-ai/grok-4.7": { - "id": "x-ai/grok-4.7", - "name": "Grok 4.7", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "knowledge": "2025-04", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 450000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.6, - "output": 4.8, - "cache_read": 0.4, - "tiers": [ - { - "input": 3.2, - "output": 9.6, - "cache_read": 0.8, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 3.2, - "output": 9.6, - "cache_read": 0.8 - } + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "mimo-v2-omni-free": { + "id": "mimo-v2-omni-free", + "name": "MiMo V2 Omni Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "mimo-omni-free", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text", "image", + "audio", "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 900000 + "context": 262144, + "output": 64000 }, + "status": "deprecated", "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0, + "output": 0, + "cache_read": 0 } }, - "x-ai/grok-4.5": { - "id": "x-ai/grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "GPT-5 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -111280,14 +106557,14 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -111295,44 +106572,43 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 450000 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 0.6 - } + "input": 1.07, + "output": 8.5, + "cache_read": 0.107 } }, - "x-ai/grok-build-0.1": { - "id": "x-ai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ "text", @@ -111345,37 +106621,25 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 230400 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai" }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 4, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 4, - "cache_read": 0.4 - } + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "x-ai/grok-4.20": { - "id": "x-ai/grok-4.20", - "name": "Grok 4.20", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "glm-5-free": { + "id": "glm-5-free", + "name": "GLM-5 Free", + "description": "Legacy model retained for compatibility with older integrations", + "family": "glm-free", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -111383,77 +106647,59 @@ } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "knowledge": "2025-04", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 1800000 + "context": 204800, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0, + "output": 0, + "cache_read": 0 } }, - "x-ai/grok-4.6": { - "id": "x-ai/grok-4.6", - "name": "Grok 4.6", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -111461,99 +106707,72 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 450000 + "context": 262144, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/anthropic" }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625 } }, - "~anthropic/claude-opus-latest": { - "id": "~anthropic/claude-opus-latest", - "name": "Claude Opus Latest", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "nemotron-3-super-free": { + "id": "nemotron-3-super-free", + "name": "Nemotron 3 Super Free", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron-free", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2026-02", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 204800, "output": 128000 }, + "status": "deprecated", "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "~anthropic/claude-haiku-latest": { - "id": "~anthropic/claude-haiku-latest", - "name": "Claude Haiku Latest", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "grok-build-0.1": { + "id": "grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Grok coding model for agentic engineering, edits, and codebase workflows", + "family": "grok-build", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2026-05-20", + "last_updated": "2026-05-20", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -111561,49 +106780,43 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 256000, + "output": 256000 }, "cost": { "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "output": 2, + "cache_read": 0.2 } }, - "~anthropic/claude-sonnet-latest": { - "id": "~anthropic/claude-sonnet-latest", - "name": "Claude Sonnet Latest", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -111611,40 +106824,44 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai" + }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 1.07, + "output": 8.5, + "cache_read": 0.107 } }, - "~anthropic/claude-fable-latest": { - "id": "~anthropic/claude-fable-latest", - "name": "Claude Fable Latest", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -111657,40 +106874,59 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai" + }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } - }, - "upstage/solar-pro-3": { - "id": "upstage/solar-pro-3", - "name": "Solar Pro 3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "solar-pro", + } + } + }, + "inference": { + "id": "inference", + "env": [ + "INFERENCE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.net/v1", + "name": "Inference", + "doc": "https://inference.net/models", + "models": { + "mistral/mistral-nemo-12b-instruct": { + "id": "mistral/mistral-nemo-12b-instruct", + "name": "Mistral Nemo 12B Instruct", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -111699,43 +106935,59 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 117964 + "context": 16000, + "output": 4096 + }, + "cost": { + "input": 0.038, + "output": 0.1 + } + }, + "google/gemma-3": { + "id": "google/gemma-3", + "name": "Google Gemma 3", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 125000, + "output": 4096 }, "cost": { "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "output": 0.3 } }, - "upstage/solar-pro4": { - "id": "upstage/solar-pro4", - "name": "Solar Pro 4", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "solar", + "osmosis/osmosis-structure-0.6b": { + "id": "osmosis/osmosis-structure-0.6b", + "name": "Osmosis Structure 0.6B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "osmosis", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -111744,43 +106996,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 524288, - "output": 131072 + "context": 4000, + "output": 2048 }, "cost": { - "input": 0.09, - "output": 0.36, - "cache_read": 0.018 + "input": 0.1, + "output": 0.5 } }, - "upstage/solar-mini4": { - "id": "upstage/solar-mini4", - "name": "Solar Mini 4", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "solar", + "qwen/qwen3-embedding-4b": { + "id": "qwen/qwen3-embedding-4b", + "name": "Qwen 3 Embedding 4B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -111789,42 +107026,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 524288, - "output": 131072 + "context": 32000, + "output": 2048 }, "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.005 + "input": 0.01, + "output": 0 } }, - "~deepseek/deepseek-flash-latest": { - "id": "~deepseek/deepseek-flash-latest", - "name": "DeepSeek Flash Latest", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", + "qwen/qwen-2.5-7b-vision-instruct": { + "id": "qwen/qwen-2.5-7b-vision-instruct", + "name": "Qwen 2.5 7B Vision Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-14", - "last_updated": "2026-09-14", + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", @@ -111834,42 +107057,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 943718 + "context": 125000, + "output": 4096 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01 + "input": 0.2, + "output": 0.2 } }, - "~deepseek/deepseek-pro-latest": { - "id": "~deepseek/deepseek-pro-latest", - "name": "DeepSeek Pro Latest", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", + "meta/llama-3.1-8b-instruct": { + "id": "meta/llama-3.1-8b-instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-14", - "last_updated": "2026-09-14", + "knowledge": "2023-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -111878,42 +107087,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 16000, + "output": 4096 }, "cost": { - "input": 0.39996, - "output": 1.19988, - "cache_read": 0.012726 + "input": 0.025, + "output": 0.025 } }, - "~deepseek/deepseek-v4-flash-latest": { - "id": "~deepseek/deepseek-v4-flash-latest", - "name": "DeepSeek V4 Flash Latest", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek", + "meta/llama-3.2-1b-instruct": { + "id": "meta/llama-3.2-1b-instruct", + "name": "Llama 3.2 1B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-01", - "last_updated": "2026-08-01", + "knowledge": "2023-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text" @@ -111922,92 +107117,61 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1310720, - "output": 943718 + "context": 16000, + "output": 4096 }, "cost": { - "input": 0.038, - "output": 0.55, - "cache_read": 0.0228 + "input": 0.01, + "output": 0.01 } }, - "~x-ai/grok-latest": { - "id": "~x-ai/grok-latest", - "name": "Grok Latest", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "meta/llama-3.2-11b-vision-instruct": { + "id": "meta/llama-3.2-11b-vision-instruct", + "name": "Llama 3.2 11B Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "knowledge": "2023-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 450000 + "context": 16000, + "output": 4096 }, "cost": { - "input": 1.6, - "output": 4.8, - "cache_read": 0.4, - "tiers": [ - { - "input": 3.2, - "output": 9.6, - "cache_read": 0.8, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 3.2, - "output": 9.6, - "cache_read": 0.8 - } + "input": 0.055, + "output": 0.055 } }, - "bytedance/ui-tars-1.5-7b": { - "id": "bytedance/ui-tars-1.5-7b", - "name": "UI-TARS 7B ", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, + "meta/llama-3.2-3b-instruct": { + "id": "meta/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "knowledge": "2025-01-31", - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "knowledge": "2023-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -112016,31 +107180,85 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 2048 + "context": 16000, + "output": 4096 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.1 + "input": 0.02, + "output": 0.02 } - }, - "liquid/lfm-2.5-2.6b:free": { - "id": "liquid/lfm-2.5-2.6b:free", - "name": "LFM2.5-2.6B (free)", - "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", - "family": "liquid", - "attachment": false, + } + } + }, + "inceptron": { + "id": "inceptron", + "env": [ + "INCEPTRON_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.inceptron.io/v1", + "name": "Inceptron", + "doc": "https://docs.inceptron.io", + "models": { + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.66, + "output": 3.5, + "cache_read": 0.2, + "cache_write": 0 + } + }, + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" @@ -112048,36 +107266,82 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.75, + "output": 3.5, + "cache_read": 0.2, + "cache_write": 0 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "moonshotai/Kimi-K2.6-Fast": { + "id": "moonshotai/Kimi-K2.6-Fast", + "name": "Kimi K2.6 Fast", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "status": "alpha", + "cost": { + "input": 1.32, + "output": 7, + "cache_read": 0.4, + "cache_write": 0 + } + }, + "zai-org/GLM-5.1-FP8": { + "id": "zai-org/GLM-5.1-FP8", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" @@ -112085,105 +107349,126 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 16384 + "context": 202752, + "output": 202752 }, "cost": { - "input": 0.09, - "output": 0.34, - "cache_read": 0.05 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26, + "cache_write": 0 } }, - "google/gemini-3.1-flash-image-preview": { - "id": "google/gemini-3.1-flash-image-preview", - "name": "Nano Banana 2 Preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", - "high" + "high", + "max" ] } ], - "tool_call": false, + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "image", "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 58982 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 3 + "input": 1.2, + "output": 4.2, + "cache_read": 0.26, + "cache_write": 0 } }, - "google/lyria-3-pro-preview": { - "id": "google/lyria-3-pro-preview", - "name": "Lyria 3 Pro Preview", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "lyria", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-25", - "last_updated": "2026-03-25", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 196608, + "output": 196608 }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.9, + "cache_read": 0.05, + "cache_write": 0 } - }, - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma 3 27B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + } + } + }, + "llama": { + "id": "llama", + "env": [ + "LLAMA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.llama.com/compat/v1/", + "name": "Llama", + "doc": "https://llama.developer.meta.com/docs/models", + "models": { + "llama-4-scout-17b-16e-instruct-fp8": { + "id": "llama-4-scout-17b-16e-instruct-fp8", + "name": "Llama-4-Scout-17B-16E-Instruct-FP8", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -112195,37 +107480,29 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 117964 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.08, - "output": 0.45, - "cache_read": 0.04 + "input": 0, + "output": 0 } }, - "google/gemma-4-31b-it:free": { - "id": "google/gemma-4-31b-it:free", - "name": "Gemma 4 31B (free)", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "cerebras-llama-4-maverick-17b-128e-instruct": { + "id": "cerebras-llama-4-maverick-17b-128e-instruct", + "name": "Cerebras-Llama-4-Maverick-17B-128E-Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-01", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" @@ -112233,71 +107510,59 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "output": 4096 }, "cost": { "input": 0, "output": 0 } }, - "google/gemini-2.5-flash-image": { - "id": "google/gemini-2.5-flash-image", - "name": "Nano Banana", - "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", - "family": "gemini-flash", + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 8192 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333 + "input": 0, + "output": 0 } }, - "google/gemma-4-26b-a4b-it:free": { - "id": "google/gemma-4-26b-a4b-it:free", - "name": "Gemma 4 26B A4B (free)", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "groq-llama-4-maverick-17b-128e-instruct": { + "id": "groq-llama-4-maverick-17b-128e-instruct", + "name": "Groq-Llama-4-Maverick-17B-128E-Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-01", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "image", - "text", - "video" + "text" ], "output": [ "text" @@ -112305,27 +107570,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "output": 4096 }, "cost": { "input": 0, "output": 0 } }, - "google/gemma-2-27b-it": { - "id": "google/gemma-2-27b-it", - "name": "Gemma 2 27B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, + "cerebras-llama-4-scout-17b-16e-instruct": { + "id": "cerebras-llama-4-scout-17b-16e-instruct", + "name": "Cerebras-Llama-4-Scout-17B-16E-Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2024-07-13", - "last_updated": "2024-07-13", + "knowledge": "2025-01", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text" @@ -112336,85 +107600,60 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 2048 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.65, - "output": 0.65 + "input": 0, + "output": 0 } }, - "google/gemini-3.6-flash": { - "id": "google/gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "llama-3.3-8b-instruct": { + "id": "llama-3.3-8b-instruct", + "name": "Llama-3.3-8B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667 + "input": 0, + "output": 0 } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "llama-4-maverick-17b-128e-instruct-fp8": { + "id": "llama-4-maverick-17b-128e-instruct-fp8", + "name": "Llama-4-Maverick-17B-128E-Instruct-FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "image", "text", - "video" + "image" ], "output": [ "text" @@ -112422,83 +107661,67 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 235929 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.09, - "output": 0.3, - "cache_read": 0.05 + "input": 0, + "output": 0 } - }, - "google/gemini-3-pro-image-preview": { - "id": "google/gemini-3-pro-image-preview", - "name": "Nano Banana Pro Preview", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + } + } + }, + "llmtr": { + "id": "llmtr", + "env": [ + "LLMTR_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://llmtr.com/v1", + "name": "LLMTR", + "doc": "https://llmtr.com/docs", + "models": { + "sincap": { + "id": "sincap", + "name": "Sincap", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "release_date": "2026-05-05", + "last_updated": "2026-05-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 32768 + "context": 128000, + "output": 8192 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "input": 0, + "output": 0 } }, - "google/gemini-3.5-flash-lite": { - "id": "google/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "magibu-11b-v8": { + "id": "magibu-11b-v8", + "name": "Magibu 11B v8", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "release_date": "2026-06-05", + "last_updated": "2026-06-05", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -112506,178 +107729,123 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 8192, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333 + "input": 0, + "output": 0 } }, - "google/gemini-3.1-flash-image": { - "id": "google/gemini-3.1-flash-image", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], + "gemma-4": { + "id": "gemma-4", + "name": "Gemma 4", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "image", "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.5, - "output": 3 + "input": 5, + "output": 10 } }, - "google/lyria-3-clip-preview": { - "id": "google/lyria-3-clip-preview", - "name": "Lyria 3 Clip Preview", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "lyria", + "medgemma-4b": { + "id": "medgemma-4b", + "name": "MedGemma 4B", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "release_date": "2026-03-25", - "last_updated": "2026-03-25", + "release_date": "2026-04-26", + "last_updated": "2026-04-26", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 8192, + "output": 4096 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 5 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "qwen3-6-35b": { + "id": "qwen3-6-35b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", "video", - "audio", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 16384, "output": 65536 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 5, + "output": 10 } }, - "google/gemma-3-12b-it": { - "id": "google/gemma-3-12b-it", - "name": "Gemma 3 12B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "trendyol-7b": { + "id": "trendyol-7b", + "name": "Trendyol 7B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "release_date": "2026-06-06", + "last_updated": "2026-06-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -112685,369 +107853,255 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.05, - "output": 0.15 + "input": 0, + "output": 0 } - }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + } + } + }, + "cohere": { + "id": "cohere", + "env": [ + "COHERE_API_KEY" + ], + "npm": "@ai-sdk/cohere", + "name": "Cohere", + "doc": "https://docs.cohere.com/docs/models", + "models": { + "c4ai-aya-expanse-32b": { + "id": "c4ai-aya-expanse-32b", + "name": "Aya Expanse 32B", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2024-10-24", + "last_updated": "2024-10-24", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "reasoning": 9, - "cache_read": 0.15, - "cache_write": 0.083333 + "context": 128000, + "output": 4000 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "command-a-03-2025": { + "id": "command-a-03-2025", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-06-01", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 8000 }, "cost": { - "input": 1.25, - "output": 10, - "reasoning": 10, - "cache_read": 0.125, - "cache_write": 0.375, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 2.5, + "output": 10 } }, - "google/gemini-3-pro-image": { - "id": "google/gemini-3-pro-image", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", + "c4ai-aya-vision-32b": { + "id": "c4ai-aya-vision-32b", + "name": "Aya Vision 32B", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "release_date": "2025-03-04", + "last_updated": "2025-05-14", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375 + "context": 16000, + "output": 4000 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "command-r7b-arabic-02-2025": { + "id": "command-r7b-arabic-02-2025", + "name": "Command R7B Arabic", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-06-01", + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65535 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.3, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333 + "input": 0.0375, + "output": 0.15 } }, - "google/gemini-3.1-flash-lite-image": { - "id": "google/gemini-3.1-flash-lite-image", - "name": "Nano Banana 2 Lite", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash-lite", + "c4ai-aya-vision-8b": { + "id": "c4ai-aya-vision-8b", + "name": "Aya Vision 8B", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], + "reasoning": false, "tool_call": false, - "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "release_date": "2025-03-04", + "last_updated": "2025-05-14", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 58982 - }, - "cost": { - "input": 0.25, - "output": 1.5 + "context": 16000, + "output": 4000 } }, - "google/gemini-3.7-flash": { - "id": "google/gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "command-r-08-2024": { + "id": "command-r-08-2024", + "name": "Command R", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667 + "input": 0.15, + "output": 0.6 } }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "command-r7b-12-2024": { + "id": "command-r7b-12-2024", + "name": "Command R7B", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_details" + "temperature": true, + "knowledge": "2024-06-01", + "release_date": "2024-12-02", + "last_updated": "2024-12-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "structured_output": true, + "open_weights": true, + "limit": { + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.0375, + "output": 0.15 + } + }, + "command-a-vision-07-2025": { + "id": "command-a-vision-07-2025", + "name": "Command A Vision", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "knowledge": "2024-06-01", + "release_date": "2025-07-31", + "last_updated": "2025-07-31", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 8000 }, "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 2.5, + "output": 10 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "command-a-plus-05-2026": { + "id": "command-a-plus-05-2026", + "name": "Command A Plus", + "description": "Cohere's stronger command model for multilingual agents and enterprise workflows", + "family": "command-a", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -113055,113 +108109,80 @@ "type": "toggle" }, { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2025-04-01", + "release_date": "2026-05-20", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 64000 }, "cost": { - "input": 0.5, - "output": 3, - "reasoning": 3, - "cache_read": 0.05, - "cache_write": 0.083333 + "input": 2.5, + "output": 10 } }, - "google/gemini-3.8-flash": { - "id": "google/gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "command-a-translate-08-2025": { + "id": "command-a-translate-08-2025", + "name": "Command A Translate", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "family": "command-a", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2024-06-01", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 8000, + "output": 8000 }, "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667 + "input": 2.5, + "output": 10 } }, - "google/gemma-3-4b-it": { - "id": "google/gemma-3-4b-it", - "name": "Gemma 3 4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "command-r-plus-08-2024": { + "id": "command-r-plus-08-2024", + "name": "Command R+", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -113169,132 +108190,148 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.05, - "output": 0.1 + "input": 2.5, + "output": 10 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, + "command-a-reasoning-08-2025": { + "id": "command-a-reasoning-08-2025", + "name": "Command A Reasoning", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-06-01", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65535 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.1, - "output": 0.4, - "reasoning": 0.4, - "cache_read": 0.01, - "cache_write": 0.083333 + "input": 2.5, + "output": 10 } }, - "google/gemini-2.5-pro-preview": { - "id": "google/gemini-2.5-pro-preview", - "name": "Gemini 2.5 Pro Preview 06-05", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini", - "attachment": true, + "north-mini-code-1-0": { + "id": "north-mini-code-1-0", + "name": "North Mini Code", + "description": "Cohere coding model for practical software engineering and agentic edits", + "family": "north", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01-31", - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "knowledge": "2025-09-23", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "pdf", - "image", - "text", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.cohere.ai/compatibility/v1" }, "cost": { - "input": 1.25, - "output": 10, - "reasoning": 10, - "cache_read": 0.125, - "cache_write": 0.375, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0, + "output": 0 } }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "c4ai-aya-expanse-8b": { + "id": "c4ai-aya-expanse-8b", + "name": "Aya Expanse 8B", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-10-24", + "last_updated": "2024-10-24", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8000, + "output": 4000 + } + } + } + }, + "sarvam": { + "id": "sarvam", + "env": [ + "SARVAM_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.sarvam.ai/v1", + "name": "Sarvam AI", + "doc": "https://docs.sarvam.ai/api-reference-docs/getting-started/models", + "models": { + "sarvam-105b": { + "id": "sarvam-105b", + "name": "Sarvam-105B", + "description": "Flagship Indian-language reasoning model for enterprise multilingual applications", + "family": "sarvam", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", + null, "low", "medium", "high" @@ -113302,51 +108339,38 @@ } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2026-02-18", + "last_updated": "2026-03-06", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.083333 + "context": 131072, + "output": 131072 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "sarvam-30b": { + "id": "sarvam-30b", + "name": "Sarvam-30B", + "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", + "family": "sarvam", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", + null, "low", "medium", "high" @@ -113354,48 +108378,50 @@ } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-02-18", + "last_updated": "2026-03-06", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 65536, "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.083333 } - }, - "writer/palmyra-x5": { - "id": "writer/palmyra-x5", - "name": "Palmyra X5", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "palmyra", + } + } + }, + "stepfun": { + "id": "stepfun", + "env": [ + "STEPFUN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.stepfun.com/v1", + "name": "StepFun", + "doc": "https://platform.stepfun.com/docs/zh/overview/concept", + "models": { + "step-1-32k": { + "id": "step-1-32k", + "name": "Step 1 (32K)", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2026-01-21", - "last_updated": "2026-01-21", + "knowledge": "2024-06", + "release_date": "2025-01-01", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -113406,29 +108432,44 @@ }, "open_weights": false, "limit": { - "context": 1040000, - "output": 8192 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 6 + "input": 2.05, + "output": 9.59, + "cache_read": 0.41 } }, - "ibm-granite/granite-4.0-h-micro": { - "id": "ibm-granite/granite-4.0-h-micro", - "name": "Granite 4.0 Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "granite", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "step-3.7-flash": { + "id": "step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-10-20", - "last_updated": "2025-10-20", + "knowledge": "2026-01-01", + "release_date": "2026-05-29", + "last_updated": "2026-06-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -113436,36 +108477,39 @@ }, "open_weights": true, "limit": { - "context": 131000, - "output": 117900 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.017, - "output": 0.112 + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 } }, - "ibm-granite/granite-4.2-8b": { - "id": "ibm-granite/granite-4.2-8b", - "name": "Granite 4.2 8B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "granite", + "step-3.5-flash-2603": { + "id": "step-3.5-flash-2603", + "name": "Step 3.5 Flash 2603", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "high" ] } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-31", - "last_updated": "2026-08-31", + "knowledge": "2025-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" @@ -113476,91 +108520,81 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 117964 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.06, - "output": 0.25, - "cache_read": 0.015 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "mistralai/mistral-nemo": { - "id": "mistralai/mistral-nemo", - "name": "Mistral Nemo", - "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", - "family": "mistral-nemo", + "stepaudio-2.5-tts": { + "id": "stepaudio-2.5-tts", + "name": "StepAudio 2.5 TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "step", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", + "tool_call": false, + "temperature": false, + "release_date": "2026-04-16", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.019, - "output": 0.03 + "context": 0, + "output": 0 } }, - "mistralai/ministral-8b-2512": { - "id": "mistralai/ministral-8b-2512", - "name": "Ministral 3 8B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, + "stepaudio-2.5-asr": { + "id": "stepaudio-2.5-asr", + "name": "StepAudio 2.5 ASR", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "step", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "tool_call": false, + "temperature": false, + "release_date": "2026-04-24", + "last_updated": "2026-07-02", "modalities": { "input": [ - "text", - "image" + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 209715 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.015 + "context": 0, + "output": 0 } }, - "mistralai/mistral-small-24b-instruct-2501": { - "id": "mistralai/mistral-small-24b-instruct-2501", - "name": "Mistral Small 3", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "step-3.5-flash": { + "id": "step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2023-10-31", - "release_date": "2025-01-30", - "last_updated": "2025-01-30", + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-06-15", "modalities": { "input": [ "text" @@ -113571,73 +108605,56 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 16384 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.05, - "output": 0.08 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "mistralai/mistral-saba": { - "id": "mistralai/mistral-saba", - "name": "Saba", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": true, + "step-tts-2": { + "id": "step-tts-2", + "name": "Step TTS 2", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "step", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-02-17", - "last_updated": "2025-02-17", + "tool_call": false, + "temperature": false, + "release_date": "2026-03-01", + "last_updated": "2026-07-02", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 26214 - }, - "cost": { - "input": 0.2, - "output": 0.6, - "cache_read": 0.02 + "context": 0, + "output": 0 } }, - "mistralai/mistral-medium-3-5": { - "id": "mistralai/mistral-medium-3-5", - "name": "Mistral Medium 3.5", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, + "step-2-16k": { + "id": "step-2-16k", + "name": "Step 2 (16K)", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "knowledge": "2024-06", + "release_date": "2025-01-01", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -113645,64 +108662,92 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 209715 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 1.5, - "output": 7.5 + "input": 5.21, + "output": 16.44, + "cache_read": 1.04 } - }, - "mistralai/mistral-medium-3.1": { - "id": "mistralai/mistral-medium-3.1", - "name": "Mistral Medium 3.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, + } + } + }, + "hpc-ai": { + "id": "hpc-ai", + "env": [ + "HPC_AI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.hpc-ai.com/inference/v1", + "name": "HPC-AI", + "doc": "https://www.hpc-ai.com/doc/docs/quickstart/", + "models": { + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-08-13", - "last_updated": "2025-08-13", + "temperature": false, + "knowledge": "2025-01-01", + "release_date": "2026-01-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 104857 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.04 + "input": 0.3, + "output": 1.5, + "cache_read": 0.05 } }, - "mistralai/mistral-small-3.2-24b-instruct": { - "id": "mistralai/mistral-small-3.2-24b-instruct", - "name": "Mistral Small 3.2 24B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, - "reasoning": false, + "zai-org/glm-5.1": { + "id": "zai-org/glm-5.1", + "name": "GLM 5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2023-10-31", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", + "release_date": "2026-04-08", + "last_updated": "2026-06-01", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -113711,73 +108756,74 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 202000, + "output": 202000 }, "cost": { - "input": 0.09375, - "output": 0.25 + "input": 0.615, + "output": 2.46, + "cache_read": 0.133 } }, - "mistralai/mistral-large": { - "id": "mistralai/mistral-large", - "name": "Mistral Large", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, - "reasoning": false, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax-m2.5", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-11-30", - "release_date": "2024-02-26", - "last_updated": "2024-02-26", + "release_date": "2026-02-12", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 102400 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } - }, - "mistralai/mistral-small-2603": { - "id": "mistralai/mistral-small-2603", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, + } + } + }, + "minimax-cn": { + "id": "minimax-cn", + "env": [ + "MINIMAX_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://api.minimaxi.com/anthropic/v1", + "name": "MiniMax (minimaxi.com)", + "doc": "https://platform.minimaxi.com/docs/guides/quickstart", + "models": { + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -113785,66 +108831,61 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 209715 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 0.3, + "output": 1.2 } }, - "mistralai/mistral-medium-3": { - "id": "mistralai/mistral-medium-3", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, + "MiniMax-M2.5-highspeed": { + "id": "MiniMax-M2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 104857 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.04 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "mistralai/voxtral-small-24b-2507": { - "id": "mistralai/voxtral-small-24b-2507", - "name": "Voxtral Small 24B 2507", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "voxtral", - "attachment": true, - "reasoning": false, + "MiniMax-M2.7-highspeed": { + "id": "MiniMax-M2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -113852,64 +108893,61 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 26214 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "mistralai/mistral-large-2407": { - "id": "mistralai/mistral-large-2407", - "name": "Mistral Large 2407", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, - "reasoning": false, + "MiniMax-M2": { + "id": "MiniMax-M2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-03-31", - "release_date": "2024-11-19", - "last_updated": "2024-11-19", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 104857 + "context": 196608, + "output": 128000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 0.3, + "output": 1.2 } }, - "mistralai/ministral-14b-2512": { - "id": "mistralai/ministral-14b-2512", - "name": "Ministral 3 14B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, - "reasoning": false, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -113917,32 +108955,37 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 209715 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.02 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "mistralai/mistral-small-3.1-24b-instruct": { - "id": "mistralai/mistral-small-3.1-24b-instruct", - "name": "Mistral Small 3.1 24B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "knowledge": "2023-10-31", - "release_date": "2025-03-17", - "last_updated": "2025-03-17", + "release_date": "2026-06-01", + "last_updated": "2026-06-25", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -113950,30 +108993,46 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 102400 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.351, - "output": 0.555 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "tiers": [ + { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12, + "tier": { + "type": "context", + "size": 512000 + } + } + ], + "context_over_200k": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 + } } }, - "mistralai/ministral-3b-2512": { - "id": "mistralai/ministral-3b-2512", - "name": "Ministral 3 3B 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, - "reasoning": false, + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -113981,32 +109040,43 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 104857 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.01 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 } - }, - "mistralai/mixtral-8x22b-instruct": { - "id": "mistralai/mixtral-8x22b-instruct", - "name": "Mixtral 8x22B Instruct", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": true, + } + } + }, + "alibaba-coding-plan": { + "id": "alibaba-coding-plan", + "env": [ + "ALIBABA_CODING_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://coding-intl.dashscope.aliyuncs.com/v1", + "name": "Alibaba Coding Plan", + "doc": "https://www.alibabacloud.com/help/en/model-studio/coding-plan", + "models": { + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-01-31", - "release_date": "2024-04-17", - "last_updated": "2024-04-17", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -114014,75 +109084,77 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 52428 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "mistralai/codestral-2508": { - "id": "mistralai/codestral-2508", - "name": "Codestral 2508", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "codestral", - "attachment": true, - "reasoning": false, + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-01", - "last_updated": "2025-08-01", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 204800 + "context": 202752, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.03 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "sakana/fugu-max": { - "id": "sakana/fugu-max", - "name": "Fugu Max", - "description": "Multi-agent model for routing expert agents across complex analytical tasks", - "family": "fugu", - "attachment": true, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -114091,42 +109163,34 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 64000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "sakana/fugu-ultra-v2": { - "id": "sakana/fugu-ultra-v2", - "name": "Fugu Ultra v2", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-08-28", - "release_date": "2026-09-11", - "last_updated": "2026-09-11", + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -114135,56 +109199,37 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 65536 }, "cost": { - "input": 5, - "output": 30, + "input": 2.5, + "output": 7.5, "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "cache_write": 3.125 } }, - "sakana/fugu-ultra": { - "id": "sakana/fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -114193,56 +109238,29 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 65536 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.1875, + "output": 1.125, + "cache_write": 0.234375 } }, - "sakana/sakana-namazu": { - "id": "sakana/sakana-namazu", - "name": "Sakana Namazu", - "description": "Multi-agent model for routing expert agents across complex analytical tasks", - "family": "sakana-namazu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "qwen3-max-2026-01-23": { + "id": "qwen3-max-2026-01-23", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-23", + "last_updated": "2026-01-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -114251,19 +109269,20 @@ "open_weights": false, "limit": { "context": 262144, - "output": 65536 + "output": 32768 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.15 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "inclusionai/ling-3.0-flash-fin:free": { - "id": "inclusionai/ling-3.0-flash-fin:free", - "name": "Ling 3.0 Flash Fin (free)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -114272,13 +109291,15 @@ } ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -114286,20 +109307,22 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { "input": 0, - "output": 0 + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "inclusionai/ling-3.0-flash-sante:free": { - "id": "inclusionai/ling-3.0-flash-sante:free", - "name": "Ling 3.0 Flash Sante (free)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", - "attachment": false, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -114307,45 +109330,50 @@ } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "knowledge": "2025-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 262144, "output": 32768 }, "cost": { "input": 0, - "output": 0 + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "inclusionai/ling-3.0-flash": { - "id": "inclusionai/ling-3.0-flash", - "name": "Ling 3.0 Flash", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-07-23", - "last_updated": "2026-07-23", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -114356,37 +109384,32 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 196608, + "input": 196601, + "output": 24576 }, "cost": { - "input": 0.021, - "output": 0.063, - "cache_read": 0.0042 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "inclusionai/ling-3.0-flash-vl": { - "id": "inclusionai/ling-3.0-flash-vl", - "name": "Ling 3.0 Flash VL", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen3-coder-next": { + "id": "qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -114395,19 +109418,20 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 65536 }, "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.012 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "inclusionai/ling-3.0-flash-fin": { - "id": "inclusionai/ling-3.0-flash-fin", - "name": "Ling 3.0 Flash Fin", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -114416,10 +109440,12 @@ } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -114430,40 +109456,33 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 202752, + "output": 16384 }, "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.012 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k3", - "attachment": true, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -114474,23 +109493,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 943718 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + } + } + }, + "longcat": { + "id": "longcat", + "env": [ + "LONGCAT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.longcat.chat/openai", + "name": "LongCat", + "doc": "https://longcat.chat/platform/docs/", + "models": { + "LongCat-2.0": { + "id": "LongCat-2.0", + "name": "LongCat-2.0", + "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", + "family": "longcat", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -114499,50 +109531,79 @@ ], "tool_call": true, "interleaved": { - "field": "reasoning_details" + "field": "reasoning_content" }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.75, + "output": 2.95, + "cache_read": 0.015 + } + } + } + }, + "poe": { + "id": "poe", + "env": [ + "POE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.poe.com/v1", + "name": "Poe", + "doc": "https://creator.poe.com/docs/external-applications/openai-compatible-api", + "models": { + "trytako/tako": { + "id": "trytako/tako", + "name": "Tako", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "tako", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": false, + "release_date": "2024-08-15", + "last_updated": "2024-08-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2048, + "output": 0 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", - "attachment": false, + "xai/grok-code-fast-1": { + "id": "xai/grok-code-fast-1", + "name": "Grok Code Fast 1", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "temperature": false, + "release_date": "2025-08-22", + "last_updated": "2025-08-22", "modalities": { "input": [ "text" @@ -114551,38 +109612,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 98304 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 0.2, + "output": 1.5, + "cache_read": 0.02 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "xai/grok-4.1-fast-reasoning": { + "id": "xai/grok-4.1-fast-reasoning", + "name": "Grok-4.1-Fast-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ "text", @@ -114592,30 +109644,33 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 - }, - "cost": { - "input": 0.45, - "output": 2.25, - "cache_read": 0.07 + "context": 2000000, + "output": 30000 } }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "xai/grok-3-mini": { + "id": "xai/grok-3-mini", + "name": "Grok 3 Mini", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-12-31", - "release_date": "2025-09-04", - "last_updated": "2025-09-04", + "temperature": false, + "release_date": "2025-04-11", + "last_updated": "2025-04-11", "modalities": { "input": [ "text" @@ -114624,198 +109679,178 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 98304 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 2.5 + "input": 0.3, + "output": 0.5, + "cache_read": 0.075 } }, - "moonshotai/kimi-k2": { - "id": "moonshotai/kimi-k2", - "name": "Kimi K2 0711", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, + "xai/grok-4.1-fast-non-reasoning": { + "id": "xai/grok-4.1-fast-non-reasoning", + "name": "Grok-4.1-Fast-Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-12-31", - "release_date": "2025-07-11", - "last_updated": "2025-07-11", + "temperature": false, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 - }, - "cost": { - "input": 0.57, - "output": 2.3 + "context": 2000000, + "output": 30000 } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "xai/grok-3": { + "id": "xai/grok-3", + "name": "Grok 3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": false, + "release_date": "2025-04-11", + "last_updated": "2025-04-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.7062, - "output": 3.3, - "cache_read": 0.18 + "input": 3, + "output": 15, + "cache_read": 0.75 } }, - "rekaai/reka-flash-3": { - "id": "rekaai/reka-flash-3", - "name": "Reka Flash 3", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "reka", - "attachment": false, + "xai/grok-4-fast-reasoning": { + "id": "xai/grok-4-fast-reasoning", + "name": "Grok-4-Fast-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01-31", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "tool_call": true, + "temperature": false, + "release_date": "2025-09-16", + "last_updated": "2025-09-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 58982 + "context": 2000000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.2 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "rekaai/reka-edge": { - "id": "rekaai/reka-edge", - "name": "Reka Edge", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "reka", + "xai/grok-4": { + "id": "xai/grok-4", + "name": "Grok-4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "temperature": false, + "release_date": "2025-07-10", + "last_updated": "2025-07-10", "modalities": { "input": [ - "image", "text", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 14745 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 3, + "output": 15, + "cache_read": 0.75 } }, - "nvidia/nemotron-3-nano-30b-a3b": { - "id": "nvidia/nemotron-3-nano-30b-a3b", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "xai/grok-4-fast-non-reasoning": { + "id": "xai/grok-4-fast-non-reasoning", + "name": "Grok-4-Fast-Non-Reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "temperature": false, + "release_date": "2025-09-16", + "last_updated": "2025-09-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 2000000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.03 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "nvidia/nemotron-3.5-content-safety:free": { - "id": "nvidia/nemotron-3.5-content-safety:free", - "name": "Nemotron 3.5 Content Safety (free)", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "nemotron", + "xai/grok-4.20-multi-agent": { + "id": "xai/grok-4.20-multi-agent", + "name": "Grok-4.20-Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "reasoning": false, + "tool_call": true, + "temperature": false, + "release_date": "2026-03-13", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -114825,653 +109860,560 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 8192 + "output": 0 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "nvidia/nemotron-3.5-lightning": { - "id": "nvidia/nemotron-3.5-lightning", - "name": "Nemotron 3.5 Lightning 30B A3B", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "topazlabs-co/topazlabs": { + "id": "topazlabs-co/topazlabs", + "name": "TopazLabs", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "topazlabs", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "temperature": false, + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.08, - "output": 0.2, - "cache_read": 0.04 + "context": 204, + "output": 0 } }, - "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free": { - "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free", - "name": "Nemotron 3 Nano Omni (free)", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", + "fireworks-ai/kimi-k2.5-fw": { + "id": "fireworks-ai/kimi-k2.5-fw", + "name": "Kimi-K2.5-FW", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "temperature": false, + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 262144, + "input": 245760, + "output": 16384 }, "cost": { "input": 0, "output": 0 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium" - ] - }, - { - "type": "budget_tokens" - } - ], + "google/veo-3.1-fast": { + "id": "google/veo-3.1-fast", + "name": "Veo-3.1-Fast", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "temperature": false, + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 - }, - "cost": { - "input": 0.08, - "output": 0.45 + "context": 480, + "output": 0 } }, - "nvidia/nemotron-3-ultra-550b-a55b:free": { - "id": "nvidia/nemotron-3-ultra-550b-a55b:free", - "name": "Nemotron 3 Ultra (free)", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "medium", - "high" - ] - }, - { - "type": "budget_tokens" - } - ], + "google/imagen-3": { + "id": "google/imagen-3", + "name": "Imagen-3", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "temperature": false, + "release_date": "2024-10-15", + "last_updated": "2024-10-15", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0, + "context": 480, "output": 0 } }, - "nvidia/nemotron-3-ultra-550b-a55b": { - "id": "nvidia/nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "medium", - "high" - ] - }, - { - "type": "budget_tokens" - } - ], + "google/nano-banana-pro": { + "id": "google/nano-banana-pro", + "name": "Nano-Banana-Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "nano-banana", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "temperature": false, + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 182520 + "context": 65536, + "output": 0 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "nvidia/nemotron-3-super-120b-a12b:free": { - "id": "nvidia/nemotron-3-super-120b-a12b:free", - "name": "Nemotron 3 Super (free)", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium" - ] - }, - { - "type": "budget_tokens" - } - ], + "google/lyria": { + "id": "google/lyria", + "name": "Lyria", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "lyria", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "temperature": false, + "release_date": "2025-06-04", + "last_updated": "2025-06-04", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 - }, - "cost": { - "input": 0, + "context": 0, "output": 0 } }, - "nvidia/nemotron-3.5-content-safety": { - "id": "nvidia/nemotron-3.5-content-safety", - "name": "Nemotron 3.5 Content Safety", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "nemotron", + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini-3.1-Flash-Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "high" + ] } ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "tool_call": true, + "temperature": false, + "release_date": "2026-02-18", + "last_updated": "2026-02-18", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 0.25, + "output": 1.5 } }, - "nvidia/nemotron-3.5-lightning:free": { - "id": "nvidia/nemotron-3.5-lightning:free", - "name": "Nemotron 3.5 Lightning (free)", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "google/nano-banana": { + "id": "google/nano-banana", + "name": "Nano-Banana", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "nano-banana", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "temperature": false, + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 65536, + "output": 0 }, "cost": { - "input": 0, - "output": 0 + "input": 0.21, + "output": 1.8, + "cache_read": 0.021 } }, - "xiaomi/mimo-v2.6-pro": { - "id": "xiaomi/mimo-v2.6-pro", - "name": "MiMo-V2.6-Pro", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", - "family": "mimo", + "google/gemini-deep-research": { + "id": "google/gemini-deep-research", + "name": "gemini-deep-research", + "description": "Legacy model retained for compatibility with older integrations", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 + "output": 0 }, + "status": "deprecated", "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "input": 1.6, + "output": 9.6 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "google/veo-3": { + "id": "google/veo-3", + "name": "Veo-3", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "temperature": false, + "release_date": "2025-05-21", + "last_updated": "2025-05-21", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1050000, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "context": 480, + "output": 0 } }, - "xiaomi/mimo-v2.6-pro-ultraspeed": { - "id": "xiaomi/mimo-v2.6-pro-ultraspeed", - "name": "MiMo-V2.6-Pro-UltraSpeed", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", - "family": "mimo", + "google/gemini-3-flash": { + "id": "google/gemini-3-flash", + "name": "Gemini-3-Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "temperature": false, + "release_date": "2025-10-07", + "last_updated": "2025-10-07", "modalities": { "input": [ "text", "image", - "audio", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 + "output": 65536 }, "cost": { - "input": 4.35, - "output": 8.7, - "cache_read": 0.036 + "input": 0.4, + "output": 2.4, + "cache_read": 0.04 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "google/imagen-3-fast": { + "id": "google/imagen-3-fast", + "name": "Imagen-3-Fast", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "temperature": false, + "release_date": "2024-10-17", + "last_updated": "2024-10-17", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1050000, - "output": 131072 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "context": 480, + "output": 0 } }, - "xiaomi/mimo-v2.6-flash": { - "id": "xiaomi/mimo-v2.6-flash", - "name": "MiMo-V2.6-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini-2.5-Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 0, + "max": 32768 } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": false, + "release_date": "2025-02-05", + "last_updated": "2025-02-05", "modalities": { "input": [ "text", "image", - "audio", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1065535, + "output": 65535 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.87, + "output": 7, + "cache_read": 0.087 } }, - "undi95/remm-slerp-l2-13b": { - "id": "undi95/remm-slerp-l2-13b", - "name": "ReMM SLERP 13B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, + "google/veo-3.1": { + "id": "google/veo-3.1", + "name": "Veo-3.1", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2023-06-30", - "release_date": "2023-07-22", - "last_updated": "2023-07-22", + "tool_call": true, + "temperature": false, + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text" ], + "output": [ + "video" + ] + }, + "open_weights": false, + "limit": { + "context": 480, + "output": 0 + } + }, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini-2.5-Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], + "tool_call": true, + "temperature": false, + "release_date": "2025-04-26", + "last_updated": "2025-04-26", + "modalities": { + "input": [ + "text", + "image", + "video", + "audio" + ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 6144, - "output": 5529 + "context": 1065535, + "output": 65535 }, "cost": { - "input": 0.35, - "output": 0.65 + "input": 0.21, + "output": 1.8, + "cache_read": 0.021 } }, - "gryphe/mythomax-l2-13b": { - "id": "gryphe/mythomax-l2-13b", - "name": "MythoMax 13B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, + "google/imagen-4-fast": { + "id": "google/imagen-4-fast", + "name": "Imagen-4-Fast", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2023-06-30", - "release_date": "2023-07-02", - "last_updated": "2023-07-02", + "tool_call": true, + "temperature": false, + "release_date": "2025-06-25", + "last_updated": "2025-06-25", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 3686 - }, - "cost": { - "input": 0.08, - "output": 0.11 + "context": 480, + "output": 0 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini-3.5-Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.27, - "output": 1.08, - "cache_read": 0.027 + "input": 1.5152, + "output": 9.0909, + "cache_read": 0.1515 } }, - "minimax/minimax-01": { - "id": "minimax/minimax-01", - "name": "MiniMax-01", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + "google/gemma-4-31b": { + "id": "google/gemma-4-31b", + "name": "Gemma-4-31B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-03-31", - "release_date": "2025-01-15", - "last_updated": "2025-01-15", + "tool_call": true, + "temperature": false, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -115481,100 +110423,93 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000192, - "output": 900172 + "context": 262144, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 1.1 + "input": 0, + "output": 0 } }, - "minimax/minimax-m2-her": { - "id": "minimax/minimax-m2-her", - "name": "MiniMax-M2 Her", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "google/veo-3-fast": { + "id": "google/veo-3-fast", + "name": "Veo-3-Fast", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-01-23", - "last_updated": "2026-01-23", + "tool_call": true, + "temperature": false, + "release_date": "2025-10-13", + "last_updated": "2025-10-13", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 2048 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "context": 480, + "output": 0 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, + "google/gemini-3-pro": { + "id": "google/gemini-3-pro", + "name": "Gemini-3-Pro", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": false, + "release_date": "2025-10-22", + "last_updated": "2025-10-22", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1048576, + "output": 65536 }, + "status": "deprecated", "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 1.6, + "output": 9.6, + "cache_read": 0.16 } }, - "minimax/minimax-m1": { - "id": "minimax/minimax-m1", - "name": "MiniMax M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "google/gemini-2.0-flash": { + "id": "google/gemini-2.0-flash", + "name": "Gemini-2.0-Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "temperature": false, + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -115582,137 +110517,128 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 40000 + "context": 990000, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 2.2 + "input": 0.1, + "output": 0.42 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini-2.5-Flash-Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 0, + "max": 24576 } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "temperature": false, + "release_date": "2025-06-19", + "last_updated": "2025-06-19", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 512000 + "context": 1024000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.07, + "output": 0.28 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "google/imagen-4": { + "id": "google/imagen-4", + "name": "Imagen-4", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "temperature": false, + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "context": 480, + "output": 0 } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "google/imagen-4-ultra": { + "id": "google/imagen-4-ultra", + "name": "Imagen-4-Ultra", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "temperature": false, + "release_date": "2025-05-24", + "last_updated": "2025-05-24", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.255, - "output": 1.02 + "context": 480, + "output": 0 } }, - "mancer/weaver": { - "id": "mancer/weaver", - "name": "Weaver (alpha)", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "alpha", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2023-06-30", - "release_date": "2023-08-02", - "last_updated": "2023-08-02", + "google/gemini-3.1-pro": { + "id": "google/gemini-3.1-pro", + "name": "Gemini-3.1-Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -115720,193 +110646,163 @@ }, "open_weights": false, "limit": { - "context": 8000, - "output": 6000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 0.75 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "stepfun/step-3.5-flash": { - "id": "stepfun/step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "google/veo-2": { + "id": "google/veo-2", + "name": "Veo-2", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", + "temperature": false, + "release_date": "2024-12-02", + "last_updated": "2024-12-02", "modalities": { "input": [ "text" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.3 + "context": 480, + "output": 0 } }, - "stepfun/step-3.7-flash": { - "id": "stepfun/step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "google/gemini-2.0-flash-lite": { + "id": "google/gemini-2.0-flash-lite", + "name": "Gemini-2.0-Flash-Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "temperature": false, + "release_date": "2025-02-05", + "last_updated": "2025-02-05", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 256000, - "output": 230400 + "context": 990000, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "input": 0.052, + "output": 0.21 } }, - "dots-studio/dots-3-note-preview:free": { - "id": "dots-studio/dots-3-note-preview:free", - "name": "Dots3-Note Preview (free)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "elevenlabs/elevenlabs-music": { + "id": "elevenlabs/elevenlabs-music", + "name": "ElevenLabs-Music", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "elevenlabs", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "release_date": "2025-08-29", + "last_updated": "2025-08-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 512000, - "output": 460800 - }, - "cost": { - "input": 0, + "context": 2000, "output": 0 } }, - "inception/mercury-2.5": { - "id": "inception/mercury-2.5", - "name": "Mercury 2.5", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "elevenlabs/elevenlabs-v2.5-turbo": { + "id": "elevenlabs/elevenlabs-v2.5-turbo", + "name": "ElevenLabs-v2.5-Turbo", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "elevenlabs", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "temperature": false, + "release_date": "2024-10-28", + "last_updated": "2024-10-28", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 260000, - "output": 65536 - }, - "cost": { - "input": 0.04, - "output": 0.15, - "cache_read": 0.004 + "context": 128000, + "output": 0 } }, - "inception/mercury-2": { - "id": "inception/mercury-2", - "name": "Mercury 2", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "elevenlabs/elevenlabs-v3": { + "id": "elevenlabs/elevenlabs-v3", + "name": "ElevenLabs-v3", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "elevenlabs", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-04", - "last_updated": "2026-03-04", + "temperature": false, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ "text" ], + "output": [ + "audio" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 0 + } + }, + "openai/chatgpt-4o-latest": { + "id": "openai/chatgpt-4o-latest", + "name": "ChatGPT-4o-Latest", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": false, + "release_date": "2024-08-14", + "last_updated": "2024-08-14", + "modalities": { + "input": [ + "text", + "image" + ], "output": [ "text" ] @@ -115914,27 +110810,25 @@ "open_weights": false, "limit": { "context": 128000, - "output": 50000 + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.025 + "input": 4.5, + "output": 14 } }, - "amazon/nova-lite-v1": { - "id": "amazon/nova-lite-v1", - "name": "Nova Lite 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", + "openai/gpt-3.5-turbo-instruct": { + "id": "openai/gpt-3.5-turbo-instruct", + "name": "GPT-3.5-Turbo-Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "temperature": false, + "release_date": "2023-09-20", + "last_updated": "2023-09-20", "modalities": { "input": [ "text", @@ -115946,37 +110840,39 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 5120 + "context": 3500, + "output": 1024 }, "cost": { - "input": 0.06, - "output": 0.24 + "input": 1.4, + "output": 1.8 } }, - "amazon/nova-2-lite-v1": { - "id": "amazon/nova-2-lite-v1", - "name": "Nova 2 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova", + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "temperature": false, + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" @@ -115984,27 +110880,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65535 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.3, - "output": 2.5 + "input": 1.8, + "output": 7.2, + "cache_read": 0.45 } }, - "amazon/nova-pro-v1": { - "id": "amazon/nova-pro-v1", - "name": "Nova Pro 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", + "openai/gpt-5.2-instant": { + "id": "openai/gpt-5.2-instant", + "name": "GPT-5.2-Instant", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "temperature": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -116016,30 +110910,29 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 5120 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.8, - "output": 3.2 + "input": 1.6, + "output": 13, + "cache_read": 0.16 } }, - "amazon/nova-premier-v1": { - "id": "amazon/nova-premier-v1", - "name": "Nova Premier 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova", + "openai/gpt-4o-search": { + "id": "openai/gpt-4o-search", + "name": "GPT-4o-Search", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-10-31", - "last_updated": "2025-10-31", + "temperature": false, + "release_date": "2025-03-11", + "last_updated": "2025-03-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -116047,31 +110940,38 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 2.5, - "output": 12.5, - "cache_read": 0.625 + "input": 2.2, + "output": 9 } }, - "amazon/nova-micro-v1": { - "id": "amazon/nova-micro-v1", - "name": "Nova Micro 1.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2-Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "temperature": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116079,28 +110979,29 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 5120 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.035, - "output": 0.14 + "input": 19, + "output": 150 } }, - "relace/relace-search": { - "id": "relace/relace-search", - "name": "Relace Search", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, + "openai/gpt-4-classic-0314": { + "id": "openai/gpt-4-classic-0314", + "name": "GPT-4-Classic-0314", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "temperature": false, + "release_date": "2024-08-26", + "last_updated": "2024-08-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116108,28 +111009,41 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 + "context": 8192, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 1, - "output": 3 + "input": 27, + "output": 54 } }, - "relace/relace-apply-3": { - "id": "relace/relace-apply-3", - "name": "Relace Apply 3", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "temperature": false, - "release_date": "2025-09-26", - "last_updated": "2025-09-26", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116137,29 +111051,30 @@ }, "open_weights": false, "limit": { - "context": 256000, + "context": 400000, "output": 128000 }, "cost": { - "input": 0.85, - "output": 1.25 + "input": 1.1, + "output": 9, + "cache_read": 0.11 } }, - "aion-labs/aion-2.0": { - "id": "aion-labs/aion-2.0", - "name": "Aion-2.0", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-5-chat": { + "id": "openai/gpt-5-chat", + "name": "GPT-5-Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116167,31 +111082,30 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.8, - "output": 1.6, - "cache_read": 0.2 + "input": 1.1, + "output": 9, + "cache_read": 0.11 } }, - "aion-labs/aion-rp-llama-3.1-8b": { - "id": "aion-labs/aion-rp-llama-3.1-8b", - "name": "Aion-RP 1.0 (8B)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5-Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2025-02-04", - "last_updated": "2025-02-04", + "tool_call": true, + "temperature": false, + "release_date": "2023-09-13", + "last_updated": "2023-09-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116199,29 +111113,30 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 29491 + "context": 16384, + "output": 2048 }, "cost": { - "input": 0.8, - "output": 1.6 + "input": 0.45, + "output": 1.4 } }, - "aion-labs/aion-3.0": { - "id": "aion-labs/aion-3.0", - "name": "Aion-3.0", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5-Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-07", - "last_updated": "2026-07-07", + "temperature": false, + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116229,39 +111144,29 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 3, - "output": 6, - "cache_read": 0.75 + "input": 14, + "output": 110 } }, - "aion-labs/aion-3.5": { - "id": "aion-labs/aion-3.5", - "name": "Aion 3.5", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "temperature": false, + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116269,39 +111174,35 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 3, - "output": 6, - "cache_read": 0.75 + "context": 128000, + "output": 8192 } }, - "aion-labs/aion-3.5-mini": { - "id": "aion-labs/aion-3.5-mini", - "name": "Aion 3.5 Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", + "temperature": false, + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116309,91 +111210,92 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.7, - "output": 1.4, - "cache_read": 0.18 + "input": 0.99, + "output": 4, + "cache_read": 0.25 } }, - "aion-labs/aion-3.0-mini": { - "id": "aion-labs/aion-3.0-mini", - "name": "Aion-3.0-Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/sora-2": { + "id": "openai/sora-2", + "name": "Sora-2", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "sora", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-07", - "last_updated": "2026-07-07", + "temperature": false, + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.7, - "output": 1.4, - "cache_read": 0.18 + "context": 0, + "output": 0 } }, - "qwen/qwen3-235b-a22b-2507": { - "id": "qwen/qwen3-235b-a22b-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/o3-pro": { + "id": "openai/o3-pro", + "name": "o3-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-07-21", - "last_updated": "2025-07-21", + "temperature": false, + "release_date": "2025-06-10", + "last_updated": "2025-06-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.0875, - "output": 0.35, - "cache_read": 0.0175 + "input": 18, + "output": 72 } }, - "qwen/qwen3-vl-235b-a22b-instruct": { - "id": "qwen/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-4-classic": { + "id": "openai/gpt-4-classic", + "name": "GPT-4-Classic", + "description": "Legacy model retained for compatibility with older integrations", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "temperature": false, + "release_date": "2024-03-25", + "last_updated": "2024-03-25", "modalities": { "input": [ "text", @@ -116403,37 +111305,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 8192, + "output": 4096 }, + "status": "deprecated", "cost": { - "input": 0.21, - "output": 1.9, - "cache_read": 0.1 + "input": 27, + "output": 54 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "openai/gpt-4o-aug": { + "id": "openai/gpt-4o-aug", + "name": "GPT-4o-Aug", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "temperature": false, + "release_date": "2024-11-21", + "last_updated": "2024-11-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116441,101 +111338,90 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.475, - "output": 4.425, - "cache_read": 0.295, - "cache_write": 1.84375 + "input": 2.2, + "output": 9, + "cache_read": 1.1 } }, - "qwen/qwen3.8-27b": { - "id": "qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4-Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", + "high", "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.42, - "output": 3, - "cache_read": 0.085 + "input": 0.18, + "output": 1.1, + "cache_read": 0.018 } }, - "qwen/qwen3-vl-235b-a22b-thinking": { - "id": "qwen/qwen3-vl-235b-a22b-thinking", - "name": "Qwen3 VL 235B A22B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/sora-2-pro": { + "id": "openai/sora-2-pro", + "name": "Sora-2-Pro", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "sora", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "temperature": false, + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 4 + "context": 0, + "output": 0 } }, - "qwen/qwen3.8-2.4t-a95b": { - "id": "qwen/qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -116543,85 +111429,76 @@ "values": [ "low", "medium", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "temperature": false, + "release_date": "2025-11-12", + "last_updated": "2025-11-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 - } + "input": 1.1, + "output": 9, + "cache_read": 0.11 + } }, - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.3-instant": { + "id": "openai/gpt-5.3-instant", + "name": "GPT-5.3-Instant", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "input": 111616, + "output": 16384 }, "cost": { - "input": 0.195, - "output": 1.56 + "input": 1.6, + "output": 13, + "cache_read": 0.16 } }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", - "attachment": false, + "openai/gpt-5.3-codex-spark": { + "id": "openai/gpt-5.3-codex-spark", + "name": "GPT-5.3-Codex-Spark", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "temperature": false, + "release_date": "2026-03-04", + "last_updated": "2026-03-04", "modalities": { "input": [ "text" @@ -116630,132 +111507,148 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.15, - "output": 1.2 + "input": 0, + "output": 0 } }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT-5.1-Codex-Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", + "temperature": false, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.12, - "output": 0.8, - "cache_read": 0.07 + "input": 1.1, + "output": 9, + "cache_read": 0.11 } }, - "qwen/qwen3-coder": { - "id": "qwen/qwen3-coder", - "name": "Qwen3 Coder 480B A35B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "openai/dall-e-3": { + "id": "openai/dall-e-3", + "name": "DALL-E-3", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "dall-e", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "temperature": false, + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 1, - "cache_read": 0.1 + "context": 800, + "output": 0 } }, - "qwen/qwen3-32b": { - "id": "qwen/qwen3-32b", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.08, - "output": 0.28 + "input": 0.99, + "output": 4 } }, - "qwen/qwen-plus": { - "id": "qwen/qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", + "temperature": false, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -116763,59 +111656,40 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.26, - "output": 0.78, - "cache_read": 0.052, - "cache_write": 0.325, - "tiers": [ - { - "input": 0.78, - "output": 2.34, - "cache_read": 0.156, - "cache_write": 0.975, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.78, - "output": 2.34, - "cache_read": 0.156, - "cache_write": 0.975 - } + "input": 1.6, + "output": 13, + "cache_read": 0.16 } }, - "qwen/qwen3.7-flash": { - "id": "qwen/qwen3.7-flash", - "name": "Qwen3.7 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "temperature": false, + "release_date": "2026-02-10", + "last_updated": "2026-02-10", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -116823,52 +111697,36 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 991000, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.03, - "output": 0.13, - "cache_read": 0.006, - "cache_write": 0.038, - "tiers": [ - { - "input": 0.1, - "output": 0.4, - "cache_read": 0.02, - "cache_write": 0.125, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.2, - "output": 0.8, - "cache_read": 0.04, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 256000 - } - } - ] + "input": 1.6, + "output": 13, + "cache_read": 0.16 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1-Codex-Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "temperature": false, + "release_date": "2025-11-12", + "last_updated": "2025-11-12", "modalities": { "input": [ "text" @@ -116879,51 +111737,27 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.78, - "output": 3.9, - "cache_read": 0.156, - "cache_write": 0.975, - "tiers": [ - { - "input": 1.56, - "output": 7.8, - "cache_read": 0.312, - "cache_write": 1.95, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.95, - "output": 9.75, - "cache_read": 0.39, - "cache_write": 2.4375, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.22, + "output": 1.8, + "cache_read": 0.022 } }, - "qwen/qwen3-coder-30b-a3b-instruct": { - "id": "qwen/qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/o4-mini-deep-research": { + "id": "openai/o4-mini-deep-research", + "name": "o4-mini-deep-research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "family": "o-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2025-06-27", + "last_updated": "2025-06-27", "modalities": { "input": [ "text" @@ -116932,71 +111766,53 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 1.8, + "output": 7.2, + "cache_read": 0.45 } }, - "qwen/qwen3.8-omni-flash": { - "id": "qwen/qwen3.8-omni-flash", - "name": "Qwen3.8 Omni Flash", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", + "openai/gpt-image-1.5": { + "id": "openai/gpt-image-1.5", + "name": "gpt-image-1.5", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016 + "context": 128000, + "output": 0 } }, - "qwen/qwen3-vl-30b-a3b-instruct": { - "id": "qwen/qwen3-vl-30b-a3b-instruct", - "name": "Qwen3 VL 30B A3B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT-4.1-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "temperature": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text", @@ -117006,290 +111822,302 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1047576, "output": 32768 }, "cost": { - "input": 0.13, - "output": 0.52 + "input": 0.09, + "output": 0.36, + "cache_read": 0.022 } }, - "qwen/qwen3-235b-a22b-thinking-2507": { - "id": "qwen/qwen3-235b-a22b-thinking-2507", - "name": "Qwen3 235B A22B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-3.5-turbo-raw": { + "id": "openai/gpt-3.5-turbo-raw", + "name": "GPT-3.5-Turbo-Raw", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "temperature": false, + "release_date": "2023-09-27", + "last_updated": "2023-09-27", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 117964 + "context": 4524, + "output": 2048 }, "cost": { - "input": 0.23, - "output": 2.3 + "input": 0.45, + "output": 1.4 } }, - "qwen/qwen3-235b-a22b": { - "id": "qwen/qwen3-235b-a22b", - "name": "Qwen3 235B-A22B", - "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", - "family": "qwen", - "attachment": false, + "openai/o1": { + "id": "openai/o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2024-12-18", + "last_updated": "2024-12-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.455, - "output": 1.82 + "input": 14, + "output": 54 } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/o1-pro": { + "id": "openai/o1-pro", + "name": "o1-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "release_date": "2025-03-19", + "last_updated": "2025-03-19", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.26, - "output": 2.08 + "input": 140, + "output": 540 } }, - "qwen/qwen-plus-2025-07-28": { - "id": "qwen/qwen-plus-2025-07-28", - "name": "Qwen Plus 0728", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-08", - "last_updated": "2025-09-08", + "temperature": false, + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.26, - "output": 0.78, - "tiers": [ - { - "input": 0.78, - "output": 2.34, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.78, - "output": 2.34 - } + "input": 2.2, + "output": 14, + "cache_read": 0.22 } }, - "qwen/qwen3.6-35b-a3b": { - "id": "qwen/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4-Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": false, + "release_date": "2026-03-12", + "last_updated": "2026-03-12", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 1, - "cache_read": 0.05 + "input": 0.68, + "output": 4, + "cache_read": 0.068 } }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "temperature": false, + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.09, - "output": 1.1 + "input": 1.8, + "output": 7.2, + "cache_read": 0.45 } }, - "qwen/qwen3-vl-30b-a3b-thinking": { - "id": "qwen/qwen3-vl-30b-a3b-thinking", - "name": "Qwen3 VL 30B A3B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/o3-deep-research": { + "id": "openai/o3-deep-research", + "name": "o3-deep-research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "temperature": false, + "release_date": "2025-06-27", + "last_updated": "2025-06-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.2, - "output": 2.4 + "input": 9, + "output": 36, + "cache_read": 2.2 } }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "temperature": false, + "release_date": "2025-06-25", + "last_updated": "2025-06-25", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -117297,47 +112125,56 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.1875, - "output": 1.125, - "cache_write": 0.234375, - "tiers": [ - { - "input": 0.75, - "output": 3, - "cache_write": 0.9375, - "tier": { - "type": "context", - "size": 256000 - } - } + "input": 0.22, + "output": 1.8, + "cache_read": 0.022 + } + }, + "openai/gpt-image-1": { + "id": "openai/gpt-image-1", + "name": "GPT-Image-1", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": false, + "release_date": "2025-03-31", + "last_updated": "2025-03-31", + "modalities": { + "input": [ + "text", + "image" ], - "context_over_200k": { - "input": 0.75, - "output": 3, - "cache_write": 0.9375 - } + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 0 } }, - "qwen/qwen3-coder-flash": { - "id": "qwen/qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "temperature": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -117345,126 +112182,97 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.195, - "output": 0.975, - "cache_read": 0.039, - "cache_write": 0.24375, - "tiers": [ - { - "input": 0.325, - "output": 1.625, - "cache_read": 0.065, - "cache_write": 0.40625, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.52, - "output": 2.6, - "cache_read": 0.104, - "cache_write": 0.65, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.36, + "output": 1.4, + "cache_read": 0.09 } }, - "qwen/qwen-2.5-coder-32b-instruct": { - "id": "qwen/qwen-2.5-coder-32b-instruct", - "name": "Qwen2.5 Coder 32B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4-Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2024-11-11", - "last_updated": "2024-11-11", + "tool_call": true, + "temperature": false, + "release_date": "2023-09-13", + "last_updated": "2023-09-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 29491 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.66, - "output": 1 + "input": 9, + "output": 27 } }, - "qwen/qwen3.5-flash-02-23": { - "id": "qwen/qwen3.5-flash-02-23", - "name": "Qwen3.5-Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-image-1-mini": { + "id": "openai/gpt-image-1-mini", + "name": "GPT-Image-1-Mini", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-25", - "last_updated": "2026-02-25", + "temperature": false, + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.065, - "output": 0.26 + "context": 0, + "output": 0 } }, - "qwen/qwen3-max-thinking": { - "id": "qwen/qwen3-max-thinking", - "name": "Qwen3 Max Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-09", - "last_updated": "2026-02-09", + "temperature": false, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -117472,203 +112280,144 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.78, - "output": 3.9, - "tiers": [ - { - "input": 1.56, - "output": 7.8, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.95, - "output": 9.75, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.045, + "output": 0.36, + "cache_read": 0.0045 } }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4-Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "temperature": false, + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.65, - "output": 3.25, - "cache_read": 0.13, - "cache_write": 0.8125, - "tiers": [ - { - "input": 1.17, - "output": 5.85, - "cache_read": 0.234, - "cache_write": 1.4625, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.95, - "output": 9.75, - "cache_read": 0.39, - "cache_write": 2.4375, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 27, + "output": 160 } }, - "qwen/qwen3.5-9b": { - "id": "qwen/qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "openai/o3-mini-high": { + "id": "openai/o3-mini-high", + "name": "o3-mini-high", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.1, - "output": 0.15 + "input": 0.99, + "output": 4 } }, - "qwen/qwen3.5-plus-20260420": { - "id": "qwen/qwen3.5-plus-20260420", - "name": "Qwen3.5 Plus 2026-04-20", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.5", + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5-Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "structured_output": false, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.8, - "cache_write": 0.375, - "tiers": [ - { - "input": 0.375, - "output": 2.25, - "cache_write": 0.46875, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.375, - "output": 2.25, - "cache_write": 0.46875 - } + "input": 27.2727, + "output": 163.6364 } }, - "qwen/qwen3.8-flash": { - "id": "qwen/qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": false, + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -117676,75 +112425,69 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 124096, + "output": 4096 }, "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016, - "cache_write": 0.2 + "input": 0.14, + "output": 0.54, + "cache_read": 0.068 } }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-4o-mini-search": { + "id": "openai/gpt-4o-mini-search", + "name": "GPT-4o-mini-Search", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "temperature": false, + "release_date": "2025-03-11", + "last_updated": "2025-03-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.3125, - "output": 1.25, - "cache_read": 0.15625 + "input": 0.14, + "output": 0.54 } }, - "qwen/qwen3.6-max-preview": { - "id": "qwen/qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "temperature": false, + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -117752,147 +112495,139 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 1.027, - "output": 6.162, - "cache_write": 1.28375, - "tiers": [ - { - "input": 1.58, - "output": 9.48, - "cache_write": 1.975, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 1.1, + "output": 9 } }, - "qwen/qwen3-30b-a3b": { - "id": "qwen/qwen3-30b-a3b", - "name": "Qwen3 30B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "temperature": false, + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.12, - "output": 0.5 + "input": 1.6, + "output": 13, + "cache_read": 0.16 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "openai/gpt-5.1-instant": { + "id": "openai/gpt-5.1-instant", + "name": "GPT-5.1-Instant", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "temperature": false, + "release_date": "2025-11-12", + "last_updated": "2025-11-12", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.55, - "output": 3.5, - "cache_read": 0.225 + "input": 1.1, + "output": 9, + "cache_read": 0.11 } }, - "qwen/qwen3-8b": { - "id": "qwen/qwen3-8b", - "name": "Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "openai/gpt-image-2": { + "id": "openai/gpt-image-2", + "name": "GPT-Image-2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 0, + "output": 0 }, "cost": { - "input": 0.117, - "output": 0.455 + "input": 5.0505, + "output": 32.3232, + "cache_read": 1.2626 } }, - "qwen/qwen3-vl-32b-instruct": { - "id": "qwen/qwen3-vl-32b-instruct", - "name": "Qwen3 VL 32B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-23", - "last_updated": "2025-10-23", + "temperature": false, + "release_date": "2025-11-12", + "last_updated": "2025-11-12", "modalities": { "input": [ "text", @@ -117902,124 +112637,125 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.104, - "output": 0.416 + "input": 1.1, + "output": 9, + "cache_read": 0.11 } }, - "qwen/qwen3-vl-8b-instruct": { - "id": "qwen/qwen3-vl-8b-instruct", - "name": "Qwen3 VL 8B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-10-14", - "last_updated": "2025-10-14", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ - "image", - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.117, - "output": 0.455 + "input": 4.5455, + "output": 27.2727, + "cache_read": 0.4545 } }, - "qwen/qwen3-30b-a3b-instruct-2507": { - "id": "qwen/qwen3-30b-a3b-instruct-2507", - "name": "Qwen3 30B A3B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "runwayml/runway": { + "id": "runwayml/runway", + "name": "Runway", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "runway", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-07-29", - "last_updated": "2025-07-29", + "temperature": false, + "release_date": "2024-10-11", + "last_updated": "2024-10-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32000 - }, - "cost": { - "input": 0.04815, - "output": 0.19305 + "context": 256, + "output": 0 } }, - "qwen/qwen2.5-vl-72b-instruct": { - "id": "qwen/qwen2.5-vl-72b-instruct", - "name": "Qwen2.5 VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "runwayml/runway-gen-4-turbo": { + "id": "runwayml/runway-gen-4-turbo", + "name": "Runway-Gen-4-Turbo", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "runway", "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2025-02-01", - "last_updated": "2025-02-01", + "tool_call": true, + "temperature": false, + "release_date": "2025-05-09", + "last_updated": "2025-05-09", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 115200 - }, - "cost": { - "input": 0.8, - "output": 1, - "cache_read": 0.4 + "context": 256, + "output": 0 } }, - "qwen/qwen-2.5-7b-instruct": { - "id": "qwen/qwen-2.5-7b-instruct", - "name": "Qwen2.5 7B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "cerebras/llama-3.1-8b-cs": { + "id": "cerebras/llama-3.1-8b-cs", + "name": "Llama-3.1-8B-CS", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2024-10-16", - "last_updated": "2024-10-16", + "temperature": false, + "release_date": "2025-05-13", + "last_updated": "2025-05-13", "modalities": { "input": [ "text" @@ -118028,39 +112764,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 29491 + "context": 128000, + "output": 0 }, "cost": { "input": 0.1, - "output": 0.2 + "output": 0.1 } }, - "qwen/qwen3.5-plus-02-15": { - "id": "qwen/qwen3.5-plus-02-15", - "name": "Qwen3.5 Plus 2026-02-15", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "cerebras/qwen3-235b-2507-cs": { + "id": "cerebras/qwen3-235b-2507-cs", + "name": "qwen3-235b-2507-cs", + "description": "Legacy model retained for compatibility with older integrations", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "temperature": false, + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -118068,41 +112795,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 0, + "output": 0 }, - "cost": { - "input": 0.26, - "output": 1.56, - "tiers": [ - { - "input": 0.325, - "output": 1.95, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.325, - "output": 1.95 - } - } + "status": "deprecated" }, - "qwen/qwen-2.5-72b-instruct": { - "id": "qwen/qwen-2.5-72b-instruct", - "name": "Qwen2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "cerebras/gpt-oss-120b-cs": { + "id": "cerebras/gpt-oss-120b-cs", + "name": "GPT-OSS-120B-CS", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "temperature": false, + "release_date": "2025-08-06", + "last_updated": "2025-08-06", "modalities": { "input": [ "text" @@ -118111,30 +112828,26 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 128000, + "output": 0 }, "cost": { - "input": 0.36, - "output": 0.4 + "input": 0.35, + "output": 0.75 } }, - "qwen/qwen3-30b-a3b-thinking-2507": { - "id": "qwen/qwen3-30b-a3b-thinking-2507", - "name": "Qwen3 30B A3B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "cerebras/llama-3.3-70b-cs": { + "id": "cerebras/llama-3.3-70b-cs", + "name": "llama-3.3-70b-cs", + "description": "Legacy model retained for compatibility with older integrations", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-05-13", + "last_updated": "2025-05-13", "modalities": { "input": [ "text" @@ -118143,77 +112856,55 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 81920, - "output": 32768 + "context": 0, + "output": 0 }, - "cost": { - "input": 0.2, - "output": 2.4 - } + "status": "deprecated" }, - "qwen/qwen3.6-27b": { - "id": "qwen/qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "cerebras/qwen3-32b-cs": { + "id": "cerebras/qwen3-32b-cs", + "name": "qwen3-32b-cs", + "description": "Legacy model retained for compatibility with older integrations", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "temperature": false, + "release_date": "2025-05-15", + "last_updated": "2025-05-15", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262140 + "context": 0, + "output": 0 }, - "cost": { - "input": 0.32, - "output": 2.7, - "cache_read": 0.15 - } + "status": "deprecated" }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", + "anthropic/claude-sonnet-3.5": { + "id": "anthropic/claude-sonnet-3.5", + "name": "Claude-Sonnet-3.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-sonnet", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "release_date": "2024-06-05", + "last_updated": "2024-06-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -118221,121 +112912,108 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 189096, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0.325, - "output": 1.95, - "cache_write": 0.40625, - "tiers": [ - { - "input": 1.3, - "output": 3.9, - "cache_write": 1.625, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.3, - "output": 3.9, - "cache_write": 1.625 - } + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "qwen/qwen3-14b": { - "id": "qwen/qwen3-14b", - "name": "Qwen3 14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "anthropic/claude-sonnet-3.7": { + "id": "anthropic/claude-sonnet-3.7", + "name": "Claude-Sonnet-3.7", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "temperature": false, + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 196608, + "output": 128000 }, "cost": { - "input": 0.12, - "output": 0.24 + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "qwen/qwen3-vl-8b-thinking": { - "id": "qwen/qwen3-vl-8b-thinking", - "name": "Qwen3 VL 8B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-sonnet-3.5-june": { + "id": "anthropic/claude-sonnet-3.5-june", + "name": "Claude-Sonnet-3.5-June", + "description": "Legacy model retained for compatibility with older integrations", + "family": "claude-sonnet", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-14", - "last_updated": "2025-10-14", + "temperature": false, + "release_date": "2024-11-18", + "last_updated": "2024-11-18", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 189096, + "output": 8192 }, + "status": "deprecated", "cost": { - "input": 0.18, - "output": 2.1 + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude-Sonnet-4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 0, + "max": 31999 } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "temperature": false, + "release_date": "2025-09-26", + "last_updated": "2025-09-26", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -118343,140 +113021,73 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 983040, + "output": 32768 }, "cost": { - "input": 0.32, - "output": 1.28, - "cache_read": 0.064, - "cache_write": 0.4, - "tiers": [ - { - "input": 0.96, - "output": 3.84, - "cache_read": 0.192, - "cache_write": 1.2, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.96, - "output": 3.84, - "cache_read": 0.192, - "cache_write": 1.2 - } + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "qwen/qwen3.8-27b:free": { - "id": "qwen/qwen3.8-27b:free", - "name": "Qwen3.8 27B (free)", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude-Sonnet-4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "release_date": "2025-05-21", + "last_updated": "2025-05-21", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 235929 + "context": 983040, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "qwen/qwen3.8-max-0902": { - "id": "qwen/qwen3.8-max-0902", - "name": "Qwen3.8 Max 0902", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude-Haiku-4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 0, + "max": 63999 } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "temperature": false, + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 - } - }, - "morph/morph-v3-large": { - "id": "morph/morph-v3-large", - "name": "Morph V3 Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "morph", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-07", - "last_updated": "2025-07-07", - "modalities": { - "input": [ - "text" + "pdf" ], "output": [ "text" @@ -118484,29 +113095,32 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 192000, + "output": 64000 }, "cost": { - "input": 0.9, - "output": 1.9 + "input": 0.85, + "output": 4.3, + "cache_read": 0.085, + "cache_write": 1.1 } }, - "morph/morph-v3-fast": { - "id": "morph/morph-v3-fast", - "name": "Morph V3 Fast", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "morph", - "attachment": false, + "anthropic/claude-haiku-3": { + "id": "anthropic/claude-haiku-3", + "name": "Claude-Haiku-3", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-07-07", - "last_updated": "2025-07-07", + "tool_call": true, + "temperature": false, + "release_date": "2024-03-09", + "last_updated": "2024-03-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -118514,19 +113128,20 @@ }, "open_weights": false, "limit": { - "context": 81920, - "output": 38000 + "context": 189096, + "output": 8192 }, "cost": { - "input": 0.8, - "output": 1.2 + "input": 0.21, + "output": 1.1, + "cache_read": 0.021, + "cache_write": 0.26 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude-Opus-4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -118537,16 +113152,15 @@ "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2026-04-15", + "last_updated": "2026-04-15", "modalities": { "input": [ "text", @@ -118559,101 +113173,74 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1048576, "output": 128000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 4.3, + "output": 21, + "cache_read": 0.43, + "cache_write": 5.4 } }, - "openai/gpt-5.4-image-2": { - "id": "openai/gpt-5.4-image-2", - "name": "GPT-5.4 Image 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "anthropic/claude-haiku-3.5": { + "id": "anthropic/claude-haiku-3.5", + "name": "Claude-Haiku-3.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": false, - "structured_output": true, + "reasoning": false, + "tool_call": true, "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2024-10-01", + "last_updated": "2024-10-01", "modalities": { "input": [ - "image", "text", + "image", "pdf" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { - "context": 272000, - "output": 128000 + "context": 189096, + "output": 8192 }, "cost": { - "input": 8, - "output": 15, - "cache_read": 2 + "input": 0.68, + "output": 3.4, + "cache_read": 0.068, + "cache_write": 0.85 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude-Opus-4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -118666,77 +113253,85 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1048576, "output": 128000 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 4.2929, + "output": 21.4646 } }, - "openai/gpt-audio": { - "id": "openai/gpt-audio", - "name": "GPT Audio", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude-Opus-4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 0, + "max": 31999 + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "temperature": false, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "audio" + "image", + "pdf" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 196608, + "output": 32000 }, "cost": { - "input": 2.5, - "output": 10 + "input": 13, + "output": 64, + "cache_read": 1.3, + "cache_write": 16 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude-Opus-4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 63999 + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "temperature": false, + "release_date": "2025-11-21", + "last_updated": "2025-11-21", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -118744,43 +113339,43 @@ }, "open_weights": false, "limit": { - "context": 16385, - "output": 4096 + "context": 196608, + "output": 64000 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 4.3, + "output": 21, + "cache_read": 0.43, + "cache_write": 5.3 } }, - "openai/gpt-6-astra-pro": { - "id": "openai/gpt-6-astra-pro", - "name": "GPT-6 Astra Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude-Sonnet-4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -118788,57 +113383,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, + "context": 983040, "output": 128000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } + "input": 2.6, + "output": 13, + "cache_read": 0.26, + "cache_write": 3.2 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Claude-Opus-4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "release_date": "2025-05-21", + "last_updated": "2025-05-21", "modalities": { "input": [ "text", @@ -118851,34 +113417,20 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 192512, + "output": 28672 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 13, + "output": 64, + "cache_read": 1.3, + "cache_write": 16 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude-Opus-4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -118889,21 +113441,19 @@ "low", "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-02-04", + "last_updated": "2026-02-04", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -118911,67 +113461,56 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 983040, "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 4.3, + "output": 21, + "cache_read": 0.43, + "cache_write": 5.3 } }, - "openai/gpt-4o-2024-05-13": { - "id": "openai/gpt-4o-2024-05-13", - "name": "GPT-4o (2024-05-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "lumalabs/ray2": { + "id": "lumalabs/ray2", + "name": "Ray2", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "ray", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", + "temperature": false, + "release_date": "2025-02-20", + "last_updated": "2025-02-20", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "video" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 5, - "output": 15 + "context": 5000, + "output": 0 } }, - "openai/gpt-4o-mini-2024-07-18": { - "id": "openai/gpt-4o-mini-2024-07-18", - "name": "GPT-4o-mini (2024-07-18)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "o-mini", + "poetools/claude-code": { + "id": "poetools/claude-code", + "name": "claude-code", + "description": "Claude model for careful reasoning, writing, coding, and tool use", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-10-31", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "temperature": false, + "release_date": "2025-11-27", + "last_updated": "2025-11-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -118979,344 +113518,210 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "context": 0, + "output": 0 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", + "empiriolabs/deepseek-v4-pro-el": { + "id": "empiriolabs/deepseek-v4-pro-el", + "name": "DeepSeek-V4-Pro-EL", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-04-24", + "last_updated": "2026-05-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "input": 1000000, + "output": 384000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 1.67, + "output": 3.33 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", + "empiriolabs/deepseek-v4-flash-el": { + "id": "empiriolabs/deepseek-v4-flash-el", + "name": "DeepSeek-V4-Flash-EL", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2026-04-24", + "last_updated": "2026-05-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "input": 1000000, + "output": 384000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.14, + "output": 0.28 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "stabilityai/stablediffusionxl": { + "id": "stabilityai/stablediffusionxl", + "name": "StableDiffusionXL", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "stable-diffusion", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "temperature": false, + "release_date": "2023-07-09", + "last_updated": "2023-07-09", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "context": 200, + "output": 0 } }, - "openai/gpt-chat-latest": { - "id": "openai/gpt-chat-latest", - "name": "GPT Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "ideogramai/ideogram": { + "id": "ideogramai/ideogram", + "name": "Ideogram", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ideogram", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": false, - "release_date": "2026-05-05", - "last_updated": "2026-05-05", + "release_date": "2024-04-03", + "last_updated": "2024-04-03", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "context": 150, + "output": 0 } }, - "openai/gpt-audio-mini": { - "id": "openai/gpt-audio-mini", - "name": "GPT Audio Mini", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "o-mini", + "ideogramai/ideogram-v2a": { + "id": "ideogramai/ideogram-v2a", + "name": "Ideogram-v2a", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ideogram", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "temperature": false, + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ - "text", - "audio" + "text" ], "output": [ - "text", - "audio" + "image" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.6, - "output": 2.4 + "context": 150, + "output": 0 } }, - "openai/gpt-5.6-sol-pro": { - "id": "openai/gpt-5.6-sol-pro", - "name": "GPT-5.6 Sol Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-sol", + "ideogramai/ideogram-v2": { + "id": "ideogramai/ideogram-v2", + "name": "Ideogram-v2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ideogram", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2024-08-21", + "last_updated": "2024-08-21", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } + "context": 150, + "output": 0 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", + "ideogramai/ideogram-v2a-turbo": { + "id": "ideogramai/ideogram-v2a-turbo", + "name": "Ideogram-v2a-Turbo", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "ideogram", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "context": 150, + "output": 0 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2 Pro", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", + "novita/glm-4.7": { + "id": "novita/glm-4.7", + "name": "glm-4.7", + "description": "Legacy model retained for compatibility with older integrations", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "image", - "text", - "pdf" + "text" ], "output": [ "text" @@ -119324,20 +113729,15 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 205000, + "output": 131072 }, - "cost": { - "input": 21, - "output": 168 - } + "status": "deprecated" }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "novita/minimax-m2.1": { + "id": "novita/minimax-m2.1", + "name": "minimax-m2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -119346,16 +113746,12 @@ } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "release_date": "2025-12-26", + "last_updated": "2025-12-26", "modalities": { "input": [ - "image", - "text", - "pdf" + "text" ], "output": [ "text" @@ -119363,37 +113759,24 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "context": 205000, + "output": 131072 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "novita/glm-4.6": { + "id": "novita/glm-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "pdf" + "text" ], "output": [ "text" @@ -119401,28 +113784,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "context": 0, + "output": 0 } }, - "openai/gpt-4": { - "id": "openai/gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, + "novita/kimi-k2-thinking": { + "id": "novita/kimi-k2-thinking", + "name": "kimi-k2-thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "temperature": false, + "release_date": "2025-11-07", + "last_updated": "2025-11-07", "modalities": { "input": [ "text" @@ -119433,44 +113814,30 @@ }, "open_weights": false, "limit": { - "context": 8191, - "output": 4096 - }, - "cost": { - "input": 30, - "output": 60 + "context": 256000, + "output": 0 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "novita/kimi-k2.5": { + "id": "novita/kimi-k2.5", + "name": "Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -119478,34 +113845,34 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 262144 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", + "novita/glm-4.6v": { + "id": "novita/glm-4.6v", + "name": "glm-4.6v", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "temperature": false, + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text" @@ -119513,66 +113880,52 @@ }, "open_weights": false, "limit": { - "context": 1047576, + "context": 131000, "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "novita/kimi-k2.6": { + "id": "novita/kimi-k2.6", + "name": "Kimi-K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-05-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 262144 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 0.96, + "output": 4.04, + "cache_read": 0.16 } }, - "openai/o1": { - "id": "openai/o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "novita/glm-4.7-flash": { + "id": "novita/glm-4.7-flash", + "name": "glm-4.7-flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -119581,16 +113934,12 @@ } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -119599,40 +113948,27 @@ "open_weights": false, "limit": { "context": 200000, - "output": 100000 - }, - "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "output": 65500 } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", + "novita/glm-4.7-n": { + "id": "novita/glm-4.7-n", + "name": "glm-4.7-n", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "image", - "text", - "pdf" + "text" ], "output": [ "text" @@ -119640,28 +113976,25 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 15, - "output": 120 + "context": 205000, + "output": 131072 } }, - "openai/gpt-3.5-turbo-instruct": { - "id": "openai/gpt-3.5-turbo-instruct", - "name": "GPT-3.5 Turbo Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "novita/glm-5": { + "id": "novita/glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2021-09-30", - "release_date": "2023-09-28", - "last_updated": "2023-09-28", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text" @@ -119672,356 +114005,331 @@ }, "open_weights": false, "limit": { - "context": 4095, - "output": 3685 + "context": 205000, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 2 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "openai/gpt-4o-2024-08-06": { - "id": "openai/gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "novita/deepseek-v3.2": { + "id": "novita/deepseek-v3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-08-06", - "last_updated": "2024-08-06", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 16384 + "output": 0 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.27, + "output": 0.4, + "cache_read": 0.13 } - }, - "openai/gpt-6-astra": { - "id": "openai/gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-astra", + } + } + }, + "kimi-for-coding": { + "id": "kimi-for-coding", + "env": [ + "KIMI_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://api.kimi.com/coding/v1", + "name": "Kimi For Coding", + "doc": "https://www.kimi.com/code/docs/en/third-party-tools/other-coding-agents.html", + "models": { + "k2p7": { + "id": "k2p7", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 + } + }, + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-11", + "last_updated": "2025-12", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-5.6-luna-pro": { - "id": "openai/gpt-5.6-luna-pro", - "name": "GPT-5.6 Luna Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-luna", - "attachment": true, + "k2p5": { + "id": "k2p5", + "name": "Kimi K2.5", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "k2p6": { + "id": "k2p6", + "name": "Kimi K2.6", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04", + "last_updated": "2026-04", "modalities": { "input": [ - "image", "text", - "pdf" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "openai/gpt-5-image-mini": { - "id": "openai/gpt-5-image-mini", - "name": "GPT-5 Image Mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", - "attachment": true, + } + } + }, + "dinference": { + "id": "dinference", + "env": [ + "DINFERENCE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.dinference.com/v1", + "name": "DInference", + "doc": "https://dinference.com", + "models": { + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": true, + "tool_call": true, "temperature": true, - "release_date": "2025-10-16", - "last_updated": "2025-10-16", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "pdf", - "image", "text" ], "output": [ - "image", "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 2.5, - "output": 2, - "cache_read": 0.25 + "input": 0.22, + "output": 0.88 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.45, + "output": 1.65 } }, - "openai/o3-pro": { - "id": "openai/o3-pro", - "name": "o3-pro", - "description": "High-effort o3 tier for difficult technical reasoning and careful answers", - "family": "o-pro", - "attachment": true, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "pdf", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 200000, - "output": 100000 + "output": 128000 }, "cost": { - "input": 20, - "output": 80 + "input": 1.25, + "output": 3.89 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-08", + "last_updated": "2025-08", "modalities": { "input": [ "text" @@ -120036,32 +114344,25 @@ "output": 32768 }, "cost": { - "input": 0.018, - "output": 0.09 + "input": 0.0675, + "output": 0.27 } }, - "openai/gpt-oss-safeguard-20b": { - "id": "openai/gpt-oss-safeguard-20b", - "name": "GPT OSS Safeguard 20B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -120072,40 +114373,42 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.075, - "output": 0.3, - "cache_read": 0.0375 + "input": 0.75, + "output": 2.4 } - }, - "openai/o3-mini-high": { - "id": "openai/o3-mini-high", - "name": "o3 Mini High", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + } + } + }, + "perplexity-agent": { + "id": "perplexity-agent", + "env": [ + "PERPLEXITY_API_KEY" + ], + "npm": "@ai-sdk/openai", + "api": "https://api.perplexity.ai/v1", + "name": "Perplexity Agent", + "doc": "https://docs.perplexity.ai/docs/agent-api/models", + "models": { + "xai/grok-4-1-fast-non-reasoning": { + "id": "xai/grok-4-1-fast-non-reasoning", + "name": "Grok 4.1 Fast (Non-Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-10-31", - "release_date": "2025-02-12", - "last_updated": "2025-02-12", + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-11-19", + "last_updated": "2025-11-19", "modalities": { "input": [ "text", - "pdf" + "image" ], "output": [ "text" @@ -120113,31 +114416,46 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 2000000, + "output": 30000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "openai/gpt-3.5-turbo-16k": { - "id": "openai/gpt-3.5-turbo-16k", - "name": "GPT-3.5 Turbo 16k", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2021-09-30", - "release_date": "2023-08-28", - "last_updated": "2023-08-28", + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -120145,32 +114463,62 @@ }, "open_weights": false, "limit": { - "context": 16385, - "output": 4096 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 4 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "openai/gpt-5.2-chat": { - "id": "openai/gpt-5.2-chat", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-10", - "last_updated": "2025-12-10", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-05", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "audio", + "video", + "pdf" ], "output": [ "text" @@ -120178,27 +114526,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03 } }, - "openai/gpt-6-sol-pro": { - "id": "openai/gpt-6-sol-pro", - "name": "GPT-6 Sol Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -120208,15 +114555,17 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -120224,100 +114573,122 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { "input": 2, - "output": 10, + "output": 12, "cache_read": 0.2, - "cache_write": 2.5, "tiers": [ { "input": 4, - "output": 15, + "output": 18, "cache_read": 0.4, - "cache_write": 5, "tier": { "type": "context", - "size": 272000 + "size": 200000 } } ], "context_over_200k": { "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 + "output": 18, + "cache_read": 0.4 } } }, - "openai/gpt-5-image": { - "id": "openai/gpt-5-image", - "name": "GPT-5 Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2024-10-01", - "release_date": "2025-10-14", - "last_updated": "2025-10-14", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "image", "text", + "image", + "video", + "audio", "pdf" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 10, - "output": 10, - "cache_read": 1.25 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "tiers": [ + { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 0.5, + "output": 3, + "cache_read": 0.05 + } } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "pdf", - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -120330,23 +114701,22 @@ "output": 128000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-luna", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -120356,16 +114726,14 @@ } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -120378,35 +114746,16 @@ "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -120416,16 +114765,16 @@ "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -120442,15 +114791,15 @@ "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "family": "gpt", "attachment": true, "reasoning": true, @@ -120458,25 +114807,23 @@ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "pdf", - "image", - "text" + "text", + "image" ], "output": [ "text" @@ -120489,15 +114836,15 @@ "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, "openai/gpt-5.5": { "id": "openai/gpt-5.5", "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "family": "gpt", "attachment": true, "reasoning": true, @@ -120505,16 +114852,15 @@ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, "knowledge": "2025-12-01", "release_date": "2026-04-23", @@ -120538,72 +114884,75 @@ "cost": { "input": 5, "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "cache_read": 0.5 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "Nemotron 3 Super 120B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2026-02", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.25, + "output": 2.5 } }, - "openai/gpt-4o-2024-11-20": { - "id": "openai/gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "anthropic/claude-opus-4-5": { + "id": "anthropic/claude-opus-4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", @@ -120616,59 +114965,39 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "openai/gpt-3.5-turbo-0613": { - "id": "openai/gpt-3.5-turbo-0613", - "name": "GPT-3.5 Turbo (older v0613)", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2021-09-30", - "release_date": "2024-01-25", - "last_updated": "2024-01-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 4095, - "output": 3685 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1, - "output": 2 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", + "anthropic/claude-sonnet-4-5": { + "id": "anthropic/claude-sonnet-4-5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", @@ -120681,27 +115010,26 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "openai/gpt-5.6-terra-pro": { - "id": "openai/gpt-5.6-terra-pro", - "name": "GPT-5.6 Terra Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-terra", + "anthropic/claude-opus-4-7": { + "id": "anthropic/claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -120711,11 +115039,10 @@ } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -120728,47 +115055,26 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "openai/gpt-6-luna": { - "id": "openai/gpt-6-luna", - "name": "GPT-6 Luna", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-luna", + "anthropic/claude-haiku-4-5": { + "id": "anthropic/claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -120778,11 +115084,10 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", @@ -120794,90 +115099,27 @@ ] }, "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25 - } - } - }, - "openai/o4-mini-high": { - "id": "openai/o4-mini-high", - "name": "o4 Mini High", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-06-30", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", - "modalities": { - "input": [ - "image", - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, "limit": { "context": 200000, - "output": 100000 + "output": 64000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 1, + "output": 5, + "cache_read": 0.1 } }, - "openai/gpt-6-luna-pro": { - "id": "openai/gpt-6-luna-pro", - "name": "GPT-6 Luna Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "anthropic/claude-opus-4-6": { + "id": "anthropic/claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -120887,15 +115129,15 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "pdf", + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -120903,46 +115145,26 @@ }, "open_weights": false, "limit": { - "context": 1050000, + "context": 200000, "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25 - } + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-terra", + "anthropic/claude-sonnet-4-6": { + "id": "anthropic/claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -120952,11 +115174,10 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ "text", @@ -120969,52 +115190,30 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "perplexity/sonar": { + "id": "perplexity/sonar", + "name": "Sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -121023,38 +115222,52 @@ "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 8192 }, "cost": { - "input": 10, - "output": 30 + "input": 0.25, + "output": 2.5, + "cache_read": 0.0625 } - }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + } + } + }, + "siliconflow": { + "id": "siliconflow", + "env": [ + "SILICONFLOW_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.siliconflow.com/v1", + "name": "SiliconFlow", + "doc": "https://cloud.siliconflow.com/models", + "models": { + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "moonshotai/Kimi-K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-21", + "last_updated": "2026-06-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -121062,41 +115275,37 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.77, + "output": 4, + "cache_read": 0.2 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "moonshotai/Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", @@ -121106,41 +115315,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.03 + "input": 0.45, + "output": 2.25 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "baidu/ERNIE-4.5-300B-A47B": { + "id": "baidu/ERNIE-4.5-300B-A47B", + "name": "baidu/ERNIE-4.5-300B-A47B", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "ernie", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2025-07-02", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -121148,44 +115347,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.28, + "output": 1.1 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "ByteDance-Seed/Seed-OSS-36B-Instruct": { + "id": "ByteDance-Seed/Seed-OSS-36B-Instruct", + "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "seed", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2025-09-04", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -121193,47 +115377,30 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.21, + "output": 0.57 } }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-sol", - "attachment": true, + "stepfun-ai/Step-3.5-Flash": { + "id": "stepfun-ai/Step-3.5-Flash", + "name": "stepfun-ai/Step-3.5-Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "family": "step", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -121241,66 +115408,29 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 0.1, + "output": 0.3 } }, - "openai/gpt-6-sol": { - "id": "openai/gpt-6-sol", - "name": "GPT-6 Sol", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -121308,58 +115438,59 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } + "input": 0.13, + "output": 0.4 + } + }, + "google/gemma-4-26B-A4B-it": { + "id": "google/gemma-4-26B-A4B-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.12, + "output": 0.4 } }, - "openai/o1-pro": { - "id": "openai/o1-pro", - "name": "o1-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": false, + "inclusionAI/Ling-flash-2.0": { + "id": "inclusionAI/Ling-flash-2.0", + "name": "inclusionAI/Ling-flash-2.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, + "reasoning": false, + "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2025-03-19", - "last_updated": "2025-03-19", + "temperature": true, + "release_date": "2025-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -121367,27 +115498,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 131000, + "output": 131000 }, "cost": { - "input": 150, - "output": 600 + "input": 0.14, + "output": 0.57 } }, - "microsoft/phi-4": { - "id": "microsoft/phi-4", - "name": "Phi 4", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-06-30", - "release_date": "2025-01-10", - "last_updated": "2025-01-10", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text" @@ -121396,28 +115526,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 14745 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.07, - "output": 0.14 + "input": 0.2, + "output": 1.6 } }, - "microsoft/wizardlm-2-8x22b": { - "id": "microsoft/wizardlm-2-8x22b", - "name": "WizardLM-2 8x22B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "Qwen/Qwen2.5-7B-Instruct": { + "id": "Qwen/Qwen2.5-7B-Instruct", + "name": "Qwen/Qwen2.5-7B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2024-04-16", - "last_updated": "2024-04-16", + "release_date": "2024-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -121426,39 +115556,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65535, - "output": 8000 + "context": 33000, + "output": 4000 }, "cost": { - "input": 0.62, - "output": 0.62 + "input": 0.05, + "output": 0.05 } - } - } - }, - "perplexity": { - "id": "perplexity", - "env": [ - "PERPLEXITY_API_KEY" - ], - "npm": "@ai-sdk/perplexity", - "name": "Perplexity", - "doc": "https://docs.perplexity.ai", - "models": { - "sonar-pro": { - "id": "sonar-pro", - "name": "Sonar Pro", - "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", - "family": "sonar-pro", + }, + "Qwen/Qwen3-VL-235B-A22B-Instruct": { + "id": "Qwen/Qwen3-VL-235B-A22B-Instruct", + "name": "Qwen/Qwen3-VL-235B-A22B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "release_date": "2025-10-04", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -121470,36 +115589,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 262000, + "output": 262000 }, "cost": { - "input": 3, - "output": 15 + "input": 0.3, + "output": 1.5 } }, - "sonar-deep-research": { - "id": "sonar-deep-research", - "name": "Perplexity Sonar Deep Research", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-02-01", - "last_updated": "2025-09-01", + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -121510,27 +115619,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2, - "output": 8, - "reasoning": 3 + "input": 0.3, + "output": 3.2 } }, - "sonar": { - "id": "sonar", - "name": "Sonar", - "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", - "family": "sonar", + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text" @@ -121541,41 +115649,36 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1, - "output": 1 + "input": 0.39, + "output": 2.34 } }, - "sonar-reasoning-pro": { - "id": "sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", - "family": "sonar-reasoning", - "attachment": true, + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "release_date": "2025-07-28", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -121583,40 +115686,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262000, + "output": 262000 }, "cost": { - "input": 2, - "output": 8 + "input": 0.13, + "output": 0.6 } - } - } - }, - "iteracompute": { - "id": "iteracompute", - "env": [ - "ITERACOMPUTE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.iteracompute.com/v1", - "name": "IteraCompute", - "doc": "https://iteracompute.com/docs.html", - "models": { - "deepseek/deepseek-v4-flash-0731": { - "id": "deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + }, + "Qwen/Qwen3-Coder-480B-A35B-Instruct": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "name": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2025-07-31", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -121625,30 +115714,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 970000, - "output": 393216 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.34, - "output": 1.05, - "cache_read": 0.035 + "input": 0.25, + "output": 1 } }, - "deepseek/deepseek-v4-pro-0813": { - "id": "deepseek/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "Qwen/Qwen3.5-122B-A10B": { + "id": "Qwen/Qwen3.5-122B-A10B", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text" @@ -121657,72 +115744,58 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 393216 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.1, - "output": 3.3, - "cache_read": 0.11 + "input": 0.26, + "output": 2.08 } }, - "z-ai/glm-5.3-flash": { - "id": "z-ai/glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "Qwen/Qwen3-Coder-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "name": "Qwen/Qwen3-Coder-30B-A3B-Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2025-08-01", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.14, - "output": 0.49, - "cache_read": 0.03 + "input": 0.07, + "output": 0.28 } }, - "z-ai/glm-5.3": { - "id": "z-ai/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "Qwen/Qwen3.5-27B": { + "id": "Qwen/Qwen3.5-27B", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text" @@ -121731,62 +115804,59 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.2, - "output": 3.5, - "cache_read": 0.26 + "input": 0.25, + "output": 2 } }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen3-30B-A3B-Instruct-2507": { + "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "structured_output": true, + "temperature": true, + "release_date": "2025-07-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 999999 + "context": 262000, + "output": 262000 }, "cost": { - "input": 3, - "output": 14.9, - "cache_read": 0.29 + "input": 0.09, + "output": 0.3 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "Qwen/Qwen3-VL-30B-A3B-Thinking": { + "id": "Qwen/Qwen3-VL-30B-A3B-Thinking", + "name": "Qwen/Qwen3-VL-30B-A3B-Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2025-10-11", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -121796,37 +115866,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 524288 + "context": 262000, + "output": 262000 }, "cost": { "input": 0.29, - "output": 1.2, - "cache_read": 0.08 + "output": 1 } }, - "ornith-ai/ornith-1.5-35b-a3b": { - "id": "ornith-ai/ornith-1.5-35b-a3b", - "name": "Ornith 1.5 35B A3B", - "description": "Mixture-of-experts coding-reasoning model for agentic software tasks, tool use, and image understanding", - "family": "ornith", + "Qwen/Qwen3-VL-8B-Instruct": { + "id": "Qwen/Qwen3-VL-8B-Instruct", + "name": "Qwen/Qwen3-VL-8B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-18", - "last_updated": "2026-08-23", + "release_date": "2025-10-15", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -121836,41 +115897,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 327680, - "input": 262144, - "output": 65536 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.3, - "output": 3, - "cache_read": 0.03 + "input": 0.18, + "output": 0.68 } }, - "qwen/qwen3.8-27b": { - "id": "qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "Qwen/Qwen3-VL-235B-A22B-Thinking": { + "id": "Qwen/Qwen3-VL-235B-A22B-Thinking", + "name": "Qwen/Qwen3-VL-235B-A22B-Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2025-10-04", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -121880,91 +115929,68 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 327680, - "input": 262144, - "output": 65536 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 0.45, + "output": 3.5 } }, - "qwen/qwen3.8-2.4t-a95b": { - "id": "qwen/qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", + "Qwen/Qwen3-8B": { + "id": "Qwen/Qwen3-8B", + "name": "Qwen/Qwen3-8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 970000, - "output": 131072 + "context": 131000, + "output": 131000 }, "cost": { - "input": 1.95, - "output": 5.95, - "cache_read": 0.2 + "input": 0.06, + "output": 0.06 } - } - } - }, - "the-grid-ai": { - "id": "the-grid-ai", - "env": [ - "THEGRID_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.thegrid.ai/v1", - "name": "The Grid AI", - "doc": "https://thegrid.ai/docs", - "models": { - "agent-prime": { - "id": "agent-prime", - "name": "Agent Prime", - "description": "Reliable models for dependable agentic applications, multi-step tool use, and reasoning workflows. Any model that meets the contract spec can serve your request.", + }, + "Qwen/Qwen3.5-9B": { + "id": "Qwen/Qwen3.5-9B", + "name": "Qwen/Qwen3.5-9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-05-04", - "last_updated": "2026-07-15", + "release_date": "2026-03-03", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -121975,45 +116001,39 @@ }, "open_weights": false, "limit": { - "context": 196608, - "input": 120000, - "output": 30000 + "context": 262144, + "output": 262144 }, - "status": "beta" + "cost": { + "input": 0.1, + "output": 0.15 + } }, - "code-max": { - "id": "code-max", - "name": "Code Max", - "description": "Frontier models for complex research, architectural decisions, debugging, and multi-file development. Any model that meets the contract spec can serve your request.", - "attachment": true, + "Qwen/Qwen3-32B": { + "id": "Qwen/Qwen3-32B", + "name": "Qwen/Qwen3-32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-05-04", - "last_updated": "2026-07-24", + "temperature": true, + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -122021,41 +116041,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 + "context": 131000, + "output": 131000 }, - "status": "beta" + "cost": { + "input": 0.14, + "output": 0.57 + } }, - "agent-max": { - "id": "agent-max", - "name": "Agent Max", - "description": "Frontier models for autonomous research, deep multi-step tool chains, and complex long-horizon tasks. Any model that meets the contract spec can serve your request.", + "Qwen/Qwen3-VL-32B-Instruct": { + "id": "Qwen/Qwen3-VL-32B-Instruct", + "name": "Qwen/Qwen3-VL-32B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-05-04", - "last_updated": "2026-07-24", + "temperature": true, + "release_date": "2025-10-21", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -122067,45 +116072,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 + "context": 262000, + "output": 262000 }, - "status": "beta" + "cost": { + "input": 0.2, + "output": 0.6 + } }, - "text-max": { - "id": "text-max", - "name": "Text Max", - "description": "Frontier models for deep reasoning, long context, and complex workflows. Any model that meets the contract spec can serve your request.", - "attachment": true, + "Qwen/Qwen3-14B": { + "id": "Qwen/Qwen3-14B", + "name": "Qwen/Qwen3-14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-02-26", - "last_updated": "2026-07-24", + "temperature": true, + "release_date": "2025-04-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -122113,35 +116112,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 + "context": 131000, + "output": 131000 + }, + "cost": { + "input": 0.07, + "output": 0.28 } }, - "code-prime": { - "id": "code-prime", - "name": "Code Prime", - "description": "Reliable models for everyday software tasks, code completion, review, and standard debugging. Any model that meets the contract spec can serve your request.", + "Qwen/Qwen2.5-72B-Instruct": { + "id": "Qwen/Qwen2.5-72B-Instruct", + "name": "Qwen/Qwen2.5-72B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-05-04", - "last_updated": "2026-07-15", + "release_date": "2024-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -122152,39 +116142,30 @@ }, "open_weights": false, "limit": { - "context": 196608, - "input": 120000, - "output": 30000 + "context": 33000, + "output": 4000 }, - "status": "beta" + "cost": { + "input": 0.59, + "output": 0.59 + } }, - "text-prime": { - "id": "text-prime", - "name": "Text Prime", - "description": "Reliable models for everyday text generation, editing, and analysis across diverse workflows. Any model that meets the contract spec can serve your request.", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "Qwen/Qwen3-VL-30B-A3B-Instruct": { + "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", + "name": "Qwen/Qwen3-VL-30B-A3B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-07-15", + "release_date": "2025-10-05", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -122192,38 +116173,31 @@ }, "open_weights": false, "limit": { - "context": 196608, - "input": 120000, - "output": 30000 + "context": 262000, + "output": 262000 + }, + "cost": { + "input": 0.29, + "output": 1 } }, - "code-standard": { - "id": "code-standard", - "name": "Code Standard", - "description": "Price-optimized models for rapid autocomplete, linting, high-frequency suggestions, and batch edits. Any model that meets the contract spec can serve your request.", - "attachment": false, + "Qwen/Qwen3-VL-32B-Thinking": { + "id": "Qwen/Qwen3-VL-32B-Thinking", + "name": "Qwen/Qwen3-VL-32B-Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-05-04", - "last_updated": "2026-07-15", + "release_date": "2025-10-21", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -122231,36 +116205,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 120000, - "output": 16000 + "context": 262000, + "output": 262000 }, - "status": "beta" + "cost": { + "input": 0.2, + "output": 1.5 + } }, - "text-standard": { - "id": "text-standard", - "name": "Text Standard", - "description": "Price-optimized models with low-latency, high-throughput and shorter maximum outputs. Any model that meets the contract spec can serve your request.", + "Qwen/Qwen3.5-35B-A3B": { + "id": "Qwen/Qwen3.5-35B-A3B", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-02-26", - "last_updated": "2026-07-15", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text" @@ -122271,35 +116235,33 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 120000, - "output": 16000 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.24, + "output": 1.8 } }, - "agent-standard": { - "id": "agent-standard", - "name": "Agent Standard", - "description": "Price-optimized models for fast tool calls, simple agent loops, high-throughput automation, and orchestration. Any model that meets the contract spec can serve your request.", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "openai/gpt-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-05-04", - "last_updated": "2026-07-15", + "release_date": "2025-08-13", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -122310,56 +116272,29 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 120000, - "output": 16000 + "context": 131000, + "output": 8000 }, - "status": "beta" - } - } - }, - "meta": { - "id": "meta", - "env": [ - "META_MODEL_API_KEY" - ], - "npm": "@ai-sdk/openai", - "api": "https://api.meta.ai/v1", - "name": "Meta", - "doc": "https://dev.meta.ai/docs", - "models": { - "muse-spark-1.3": { - "id": "muse-spark-1.3", - "name": "Muse Spark 1.3", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "cost": { + "input": 0.05, + "output": 0.45 + } + }, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "openai/gpt-oss-20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "release_date": "2025-08-13", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "video", - "pdf", - "audio" + "text" ], "output": [ "text" @@ -122367,45 +116302,35 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 131000, + "output": 8000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 0.04, + "output": 0.18 } }, - "muse-spark-1.1": { - "id": "muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", - "family": "muse", - "attachment": true, + "tencent/Hy3-preview": { + "id": "tencent/Hy3-preview", + "name": "Hy3 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text", - "image", - "pdf", - "video" + "text" ], "output": [ "text" @@ -122413,46 +116338,40 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 0.066, + "output": 0.26, + "cache_read": 0.029 } }, - "muse-spark-1.2": { - "id": "muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", - "attachment": true, + "tencent/Hunyuan-A13B-Instruct": { + "id": "tencent/Hunyuan-A13B-Instruct", + "name": "tencent/Hunyuan-A13B-Instruct", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "release_date": "2025-06-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "video", - "pdf", - "audio" + "text" ], "output": [ "text" @@ -122460,46 +116379,42 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 131000, + "output": 131000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 0.14, + "output": 0.57 } }, - "muse-spark-1.2-contributor": { - "id": "muse-spark-1.2-contributor", - "name": "Muse Spark 1.2 Contributor", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", + "zai-org/GLM-5V-Turbo": { + "id": "zai-org/GLM-5V-Turbo", + "name": "zai-org/GLM-5V-Turbo", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", - "image", - "video", - "pdf", - "audio" + "image" ], "output": [ "text" @@ -122507,103 +116422,82 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 200000, "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 + "input": 1.2, + "output": 4, + "cache_write": 0 } }, - "muse-spark-1.3-contributor": { - "id": "muse-spark-1.3-contributor", - "name": "Muse Spark 1.3 Contributor", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", - "family": "muse", - "attachment": true, + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "zai-org/GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "release_date": "2026-02-12", + "last_updated": "2026-06-15", "modalities": { "input": [ - "text", - "image", - "video", - "pdf", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 205000, + "output": 205000 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 + "input": 0.95, + "output": 2.55 } - } - } - }, - "cline-pass": { - "id": "cline-pass", - "env": [ - "CLINE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.cline.bot/api/v1", - "name": "ClinePass", - "doc": "https://docs.cline.bot/getting-started/clinepass", - "models": { - "cline-pass/mimo-v2.6-pro": { - "id": "cline-pass/mimo-v2.6-pro", - "name": "MiMo-V2.6-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": true, + }, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", - "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -122611,31 +116505,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 1049000, + "output": 262000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "input": 1.4, + "output": 4.4, + "cache_write": 0 } }, - "cline-pass/qwen3.7-max": { - "id": "cline-pass/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "zai-org/GLM-4.5-Air": { + "id": "zai-org/GLM-4.5-Air", + "name": "zai-org/GLM-4.5-Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2025-07-28", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -122646,39 +116536,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 131000, + "output": 131000 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 0.14, + "output": 0.86 } }, - "cline-pass/mimo-v2.5": { - "id": "cline-pass/mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "zai-org/GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-04-08", + "last_updated": "2026-04-08", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -122686,46 +116576,37 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 205000, + "output": 205000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 1.4, + "output": 4.4, + "cache_write": 0 } }, - "cline-pass/glm-5.3-flash": { - "id": "cline-pass/glm-5.3-flash", - "name": "cline-pass/glm-5.3-flash", - "description": "Latest natively multimodal model in the GLM-5 series", - "family": "glm-flash", - "attachment": true, + "deepseek-ai/DeepSeek-R1": { + "id": "deepseek-ai/DeepSeek-R1", + "name": "deepseek-ai/DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2025-05-28", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -122733,44 +116614,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 0.5, + "output": 2.18 } }, - "cline-pass/qwen3.8-max": { - "id": "cline-pass/qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, + "deepseek-ai/DeepSeek-V3.1-Terminus": { + "id": "deepseek-ai/DeepSeek-V3.1-Terminus", + "name": "deepseek-ai/DeepSeek-V3.1-Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "release_date": "2025-09-29", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -122778,127 +116654,120 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 164000, + "output": 164000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 + "input": 0.27, + "output": 1 } }, - "cline-pass/kimi-k3": { - "id": "cline-pass/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "deepseek-ai/DeepSeek-V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "release_date": "2025-08-25", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 164000, + "output": 164000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.27, + "output": 1 } }, - "cline-pass/deepseek-v4.1-flash": { - "id": "cline-pass/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "deepseek-ai/DeepSeek-V3.2-Exp": { + "id": "deepseek-ai/DeepSeek-V3.2-Exp", + "name": "deepseek-ai/DeepSeek-V3.2-Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2025-10-10", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 + "input": 0.27, + "output": 0.41 } }, - "cline-pass/kimi-k2.6": { - "id": "cline-pass/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -122906,32 +116775,38 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "cline-pass/mimo-v2.5-pro": { - "id": "cline-pass/mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -122942,119 +116817,97 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { "input": 1.74, "output": 3.48, - "cache_read": 0.0145 + "cache_read": 0.145 } }, - "cline-pass/mimo-v2.6-flash": { - "id": "cline-pass/mimo-v2.6-flash", - "name": "MiMo-V2.6-Flash", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "deepseek-ai/DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "release_date": "2025-12-03", + "last_updated": "2025-12-03", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.27, + "output": 0.42 } }, - "cline-pass/minimax-m3": { - "id": "cline-pass/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "deepseek-ai/DeepSeek-V3": { + "id": "deepseek-ai/DeepSeek-V3", + "name": "deepseek-ai/DeepSeek-V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2024-12-26", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 512000 + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.25, + "output": 1 } }, - "cline-pass/glm-5.2": { - "id": "cline-pass/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMaxAI/MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-02-15", + "last_updated": "2026-06-15", "modalities": { "input": [ "text" @@ -123063,43 +116916,50 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 197000, + "output": 131000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.3, + "output": 1.2 } - }, - "cline-pass/deepseek-v4-pro": { - "id": "cline-pass/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + } + } + }, + "umans-ai-coding-plan": { + "id": "umans-ai-coding-plan", + "env": [ + "UMANS_AI_CODING_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.code.umans.ai/v1", + "name": "Umans AI Coding Plan", + "doc": "https://app.umans.ai/offers/code/docs", + "models": { + "umans-kimi-k2.7": { + "id": "umans-kimi-k2.7", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -123107,85 +116967,78 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.0145 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "cline-pass/muse-spark-1.3-contributor": { - "id": "cline-pass/muse-spark-1.3-contributor", - "name": "Muse Spark 1.3 Contributor", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", - "family": "muse", - "attachment": true, + "umans-glm-5.1": { + "id": "umans-glm-5.1", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text", - "image", - "video", - "pdf", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 204800, "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "cline-pass/glm-5.3": { - "id": "cline-pass/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "umans-coder": { + "id": "umans-coder", + "name": "Umans Coder", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -123193,34 +117046,48 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "cline-pass/kimi-k2.7-code": { - "id": "cline-pass/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "umans-flash": { + "id": "umans-flash", + "name": "Umans Flash", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -123232,95 +117099,92 @@ "output": 262144 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "cline-pass/qwen3.7-plus": { - "id": "cline-pass/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, + "umans-glm-5.2": { + "id": "umans-glm-5.2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 405504, + "output": 131072 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.04, - "cache_write": 0.5, - "tiers": [ - { - "input": 1.2, - "output": 4.8, - "cache_read": 0.12, - "cache_write": 1.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.2, - "output": 4.8, - "cache_read": 0.12, - "cache_write": 1.5 - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "cline-pass/deepseek-v4-flash": { - "id": "cline-pass/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "umans-qwen3.6-35b-a3b": { + "id": "umans-qwen3.6-35b-a3b", + "name": "Qwen3.6 35B A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -123328,48 +117192,44 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } } } }, - "modal": { - "id": "modal", + "io-net": { + "id": "io-net", "env": [ - "MODAL_PROXY_TOKEN" + "IOINTELLIGENCE_API_KEY" ], "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.us-west.modal.direct/v1", - "name": "Modal", - "doc": "https://modal.com/docs/guide/endpoints", + "api": "https://api.intelligence.io.solutions/api/v1", + "name": "IO.NET", + "doc": "https://io.net/docs/guides/intelligence/io-intelligence", "models": { - "thinkingmachines/Inkling-NVFP4": { - "id": "thinkingmachines/Inkling-NVFP4", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { + "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", + "name": "Llama 4 Maverick 17B 128E Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "knowledge": "2024-12", + "release_date": "2025-01-15", + "last_updated": "2025-01-15", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" @@ -123377,40 +117237,28 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 262144 + "context": 430000, + "output": 4096 }, "cost": { - "input": 1.2, - "output": 5, - "cache_read": 0.27 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075, + "cache_write": 0.3 } }, - "Qwen/Qwen3.8-2.4T-A95B": { - "id": "Qwen/Qwen3.8-2.4T-A95B", - "name": "Qwen3.8-Max", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -123421,31 +117269,28 @@ }, "open_weights": true, "limit": { - "context": 1010000, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 0.13, + "output": 0.38, + "cache_read": 0.065, + "cache_write": 0.26 } }, - "moonshotai/Kimi-K3": { - "id": "moonshotai/Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "meta-llama/Llama-3.2-90B-Vision-Instruct": { + "id": "meta-llama/Llama-3.2-90B-Vision-Instruct", + "name": "Llama 3.2 90B Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text", @@ -123457,97 +117302,61 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 16000, + "output": 4096 }, "cost": { - "input": 3, - "output": 15, - "reasoning": 15, - "cache_read": 0.3 + "input": 0.35, + "output": 0.4, + "cache_read": 0.175, + "cache_write": 0.7 } }, - "zai-org/GLM-5.3-Flash": { - "id": "zai-org/GLM-5.3-Flash", - "name": "GLM 5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + "moonshotai/Kimi-K2-Thinking": { + "id": "moonshotai/Kimi-K2-Thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2024-08", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 32768, + "output": 4096 }, "cost": { - "input": 0.45, - "output": 1.5, - "cache_read": 0.09 + "input": 0.55, + "output": 2.25, + "cache_read": 0.275, + "cache_write": 1.1 } - } - } - }, - "coralbricks": { - "id": "coralbricks", - "env": [ - "CORAL_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.coralbricks.ai/v1", - "name": "CoralBricks", - "doc": "https://www.coralbricks.ai/docs", - "models": { - "glm-5.3-fp4": { - "id": "glm-5.3-fp4", - "name": "GLM 5.3 FP4", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + }, + "moonshotai/Kimi-K2-Instruct-0905": { + "id": "moonshotai/Kimi-K2-Instruct-0905", + "name": "Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2024-08", + "release_date": "2024-09-05", + "last_updated": "2024-09-05", "modalities": { "input": [ "text" @@ -123556,46 +117365,34 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 32768, + "output": 4096 }, "cost": { - "input": 1.12, - "output": 4.4, - "cache_read": 0, - "cache_write": 1.68 + "input": 0.39, + "output": 1.9, + "cache_read": 0.195, + "cache_write": 0.78 } }, - "glm-5.3-flash-fp4": { - "id": "glm-5.3-flash-fp4", - "name": "GLM 5.3 Flash FP4", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "Qwen/Qwen2.5-VL-32B-Instruct": { + "id": "Qwen/Qwen2.5-VL-32B-Instruct", + "name": "Qwen 2.5 VL 32B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2024-09", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -123603,43 +117400,28 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 32000, + "output": 4096 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0, - "cache_write": 0.23 + "input": 0.05, + "output": 0.22, + "cache_read": 0.025, + "cache_write": 0.1 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "Qwen/Qwen3-Next-80B-A3B-Instruct": { + "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", + "name": "Qwen 3 Next 80B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-12", + "release_date": "2025-01-10", + "last_updated": "2025-01-10", "modalities": { "input": [ "text" @@ -123650,41 +117432,29 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 262144, + "output": 4096 }, "cost": { - "input": 0.12, - "output": 0.6, - "cache_read": 0, - "cache_write": 0.18 + "input": 0.1, + "output": 0.8, + "cache_read": 0.05, + "cache_write": 0.2 } }, - "deepseek-v4.1-flash-fast-fp4": { - "id": "deepseek-v4.1-flash-fast-fp4", - "name": "DeepSeek V4.1 Flash FP4", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen 3 235B Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2024-12", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ "text" @@ -123695,43 +117465,28 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0, - "cache_write": 0.09 + "input": 0.11, + "output": 0.6, + "cache_read": 0.055, + "cache_write": 0.22 } - } - } - }, - "routing-run": { - "id": "routing-run", - "env": [ - "ROUTING_RUN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.routing.run/v1", - "name": "routing.run", - "doc": "https://docs.routing.run/api-reference/models", - "models": { - "nemotron-3-ultra": { - "id": "nemotron-3-ultra", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT-OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ "text" @@ -123743,35 +117498,30 @@ "open_weights": true, "limit": { "context": 131072, - "output": 32000 + "output": 4096 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 0.04, + "output": 0.4, + "cache_read": 0.02, + "cache_write": 0.08 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT-OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -123779,72 +117529,63 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 64000, + "output": 4096 }, "cost": { - "input": 0.275, - "output": 1.1 + "input": 0.03, + "output": 0.14, + "cache_read": 0.015, + "cache_write": 0.06 } }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistralai/Devstral-Small-2505": { + "id": "mistralai/Devstral-Small-2505", + "name": "Devstral Small 2505", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2024-12", + "release_date": "2025-05-01", + "last_updated": "2025-05-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.16, - "output": 0.48 + "input": 0.05, + "output": 0.22, + "cache_read": 0.025, + "cache_write": 0.1 } }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistralai/Magistral-Small-2506": { + "id": "mistralai/Magistral-Small-2506", + "name": "Magistral Small 2506", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral-small", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-06-01", + "last_updated": "2025-06-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -123852,36 +117593,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.7, - "output": 4.2 + "input": 0.5, + "output": 1.5, + "cache_read": 0.25, + "cache_write": 1 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistralai/Mistral-Large-Instruct-2411": { + "id": "mistralai/Mistral-Large-Instruct-2411", + "name": "Mistral Large Instruct 2411", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -123889,30 +117626,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 5, - "output": 25 + "input": 2, + "output": 6, + "cache_read": 1, + "cache_write": 4 } }, - "glm-5.2-nitro": { - "id": "glm-5.2-nitro", - "name": "GLM 5.2 Nitro", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "mistralai/Mistral-Nemo-Instruct-2407": { + "id": "mistralai/Mistral-Nemo-Instruct-2407", + "name": "Mistral Nemo Instruct 2407", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-05", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ "text" @@ -123923,30 +117658,28 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.8, - "output": 2.4 + "input": 0.02, + "output": 0.04, + "cache_read": 0.01, + "cache_write": 0.04 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar": { + "id": "Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar", + "name": "Qwen 3 Coder 480B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-12", + "release_date": "2025-01-15", + "last_updated": "2025-01-15", "modalities": { "input": [ "text" @@ -123957,68 +117690,61 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 106000, + "output": 4096 }, "cost": { - "input": 0.8, - "output": 2.4 + "input": 0.22, + "output": 0.95, + "cache_read": 0.11, + "cache_write": 0.44 } }, - "kimi-k2.6-nitro": { - "id": "kimi-k2.6-nitro", - "name": "Kimi K2.6 Nitro", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "zai-org/GLM-4.6": { + "id": "zai-org/GLM-4.6", + "name": "GLM 4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2024-10", + "release_date": "2024-11-15", + "last_updated": "2024-11-15", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 200000, - "output": 32000 + "output": 4096 }, "cost": { - "input": 0.275, - "output": 1.1 + "input": 0.4, + "output": 1.75, + "cache_read": 0.2, + "cache_write": 0.8 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", + "name": "DeepSeek R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -124029,19 +117755,33 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.348, - "output": 0.696 + "input": 2, + "output": 8.75, + "cache_read": 1, + "cache_write": 4 } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + } + } + }, + "gmicloud": { + "id": "gmicloud", + "env": [ + "GMICLOUD_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.gmi-serving.com/v1", + "name": "GMI Cloud", + "doc": "https://docs.gmicloud.ai/inference-engine/api-reference/llm-api-reference", + "models": { + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], @@ -124049,34 +117789,34 @@ "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 3, - "output": 15 + "input": 0.855, + "output": 3.6, + "cache_read": 0.144 } }, - "kimi-k2.7-code-nitro": { - "id": "kimi-k2.7-code-nitro", - "name": "Kimi K2.7 Code Nitro", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "moonshotai/kimi-k2.7-code-highspeed": { + "id": "moonshotai/kimi-k2.7-code-highspeed", + "name": "Kimi K2.7 Code Highspeed", + "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", "family": "kimi-k2", "attachment": true, "reasoning": true, @@ -124092,9 +117832,7 @@ "last_updated": "2026-06-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -124102,111 +117840,30 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.275, - "output": 1.1 - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 1.5, - "output": 9 + "input": 1.9, + "output": 8, + "cache_read": 0.38 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "Qwen/Qwen3.7-Max": { + "id": "Qwen/Qwen3.7-Max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "video" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 0.275, - "output": 1.1 - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" ], "output": [ "text" @@ -124215,66 +117872,21 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 922000, - "output": 128000 + "output": 65536 }, "cost": { "input": 2.5, - "output": 15 + "output": 7.5, + "cache_read": 0.25, + "cache_write": 3.125 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.112, - "output": 0.224 - } - } - } - }, - "echo": { - "id": "echo", - "env": [ - "ECHO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://echo.tracerml.ai/v1", - "name": "Echo", - "doc": "https://echo.tracerml.ai/docs/api", - "models": { - "echo": { - "id": "echo", - "name": "Echo", - "description": "Adaptive model for coding, reasoning, and tool-driven agent workflows through one OpenAI-compatible endpoint", - "attachment": false, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -124283,18 +117895,22 @@ "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-19", - "last_updated": "2026-08-16", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -124302,53 +117918,40 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "status": "beta", "cost": { - "input": 10, - "output": 50 + "input": 5, + "output": 30, + "cache_read": 0.5 } - } - } - }, - "neuralwatt": { - "id": "neuralwatt", - "env": [ - "NEURALWATT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.neuralwatt.com/v1", - "name": "Neuralwatt", - "doc": "https://portal.neuralwatt.com/docs", - "models": { - "glm-5.2-short-fast": { - "id": "glm-5.2-short-fast", - "name": "GLM 5.2 Short Fast", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, + }, + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "low", + "medium", "high", + "xhigh", "max" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, - "interleaved": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text" @@ -124357,89 +117960,101 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 199984, - "output": 32000 + "context": 409600, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, - "status": "deprecated", "cost": { - "input": 1.45, - "output": 4.5, - "cache_read": 0.145 + "input": 4.5, + "output": 22.5, + "cache_read": 0.45 } }, - "kimi-k3-flex": { - "id": "kimi-k3-flex", - "name": "Kimi K3 Flex", - "description": "Kimi K3 on the flex tier: discounted, best-effort latency, requests may be held under load", - "family": "kimi-k3", + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", + "medium", "high", + "xhigh", "max" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048560, - "output": 1048560 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.95, - "output": 9.75, - "cache_read": 0.195 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "deepseek-v4-flash-flex": { - "id": "deepseek-v4-flash-flex", - "name": "DeepSeek V4 Flash Flex", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text" @@ -124448,89 +118063,71 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048560, - "output": 393216 + "context": 409600, + "output": 64000 }, "cost": { - "input": 0.091, - "output": 0.182, - "cache_read": 0.0182 + "input": 3, + "output": 15, + "cache_read": 0.3 } }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - }, - { - "type": "budget_tokens" + "type": "budget_tokens", + "min": 1024, + "max": 127999 } ], "tool_call": true, - "interleaved": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048560, - "output": 1048560 + "context": 409600, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 5, + "output": 25, + "cache_read": 0.5 } }, - "glm-5.2-short-flex": { - "id": "glm-5.2-short-flex", - "name": "GLM 5.2 Short Flex", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "zai-org/GLM-5.1-FP8": { + "id": "zai-org/GLM-5.1-FP8", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -124541,45 +118138,33 @@ }, "open_weights": true, "limit": { - "context": 199984, - "output": 32000 + "context": 202752, + "output": 131072 }, - "status": "deprecated", "cost": { - "input": 0.9425, - "output": 2.925, - "cache_read": 0.09425 + "input": 0.98, + "output": 3.08, + "cache_read": 0.182 } }, - "qwen3.6-35b": { - "id": "qwen3.6-35b", - "name": "Qwen3.6 35B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, + "zai-org/GLM-5-FP8": { + "id": "zai-org/GLM-5-FP8", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -124587,19 +118172,19 @@ }, "open_weights": true, "limit": { - "context": 131056, - "output": 131056 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0.29, - "output": 1.15, - "cache_read": 0.029 + "input": 0.6, + "output": 1.92, + "cache_read": 0.12 } }, - "glm-5.2-fast": { - "id": "glm-5.2-fast", - "name": "GLM 5.2 Fast", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "zai-org/GLM-5.2-FP8": { + "id": "zai-org/GLM-5.2-FP8", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", "attachment": false, "reasoning": true, @@ -124607,21 +118192,19 @@ { "type": "effort", "values": [ - "none", "high", "max" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, - "interleaved": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -124632,47 +118215,35 @@ }, "open_weights": true, "limit": { - "context": 1048560, - "output": 1048560 + "context": 1000000, + "output": 131072 }, - "status": "deprecated", "cost": { - "input": 1.45, - "output": 4.5, - "cache_read": 0.145 + "input": 0.979, + "output": 3.08, + "cache_read": 0.182 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -124680,45 +118251,35 @@ }, "open_weights": true, "limit": { - "context": 1048560, - "output": 1048560 + "context": 1048575, + "output": 384000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.112, + "output": 0.224, + "cache_read": 0.022 } }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -124726,123 +118287,109 @@ }, "open_weights": true, "limit": { - "context": 1048560, - "output": 393216 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 1.392, + "output": 2.784, + "cache_read": 0.116 } - }, - "kimi-k3-fast": { - "id": "kimi-k3-fast", - "name": "Kimi K3 Fast", - "description": "Kimi K3 with thinking disabled for low-latency tool calling, vision, and JSON work", - "family": "kimi-k3", - "attachment": true, + } + } + }, + "xiaomi-token-plan-cn": { + "id": "xiaomi-token-plan-cn", + "env": [ + "XIAOMI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://token-plan-cn.xiaomimimo.com/v1", + "name": "Xiaomi Token Plan (China)", + "doc": "https://platform.xiaomimimo.com/#/docs", + "models": { + "mimo-v2.5-tts-voiceclone": { + "id": "mimo-v2.5-tts-voiceclone", + "name": "MiMo-V2.5-TTS-VoiceClone", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "audio" ] }, "open_weights": true, "limit": { - "context": 1048560, - "output": 1048560 + "context": 8192, + "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0, + "output": 0 } }, - "qwen3.6-35b-flex": { - "id": "qwen3.6-35b-flex", - "name": "Qwen3.6 35B Flex", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "mimo-v2.5-tts-voicedesign": { + "id": "mimo-v2.5-tts-voicedesign", + "name": "MiMo-V2.5-TTS-VoiceDesign", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "audio" ] }, "open_weights": true, "limit": { - "context": 131056, - "output": 131056 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.1885, - "output": 0.7475, - "cache_read": 0.01885 + "input": 0, + "output": 0 } }, - "qwen-3.8-27b": { - "id": "qwen-3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, - "interleaved": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ "text" @@ -124850,123 +118397,106 @@ }, "open_weights": true, "limit": { - "context": 262128, + "context": 1048576, "output": 131072 }, "cost": { - "input": 0.45, - "output": 3.2, - "cache_read": 0.25 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "glm-5.2-short": { - "id": "glm-5.2-short", - "name": "GLM 5.2 Short", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "mimo-v2-tts": { + "id": "mimo-v2-tts", + "name": "MiMo-V2-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "reasoning": false, + "tool_call": false, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": true, "limit": { - "context": 199984, - "output": 32000 + "context": 8192, + "output": 8192 }, - "status": "deprecated", "cost": { - "input": 1.45, - "output": 4.5, - "cache_read": 0.145 + "input": 0, + "output": 0 } }, - "glm-5.3-flash-flex": { - "id": "glm-5.3-flash-flex", - "name": "GLM-5.3 Flash Flex", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - }, - { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, - "interleaved": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048560, - "output": 1048560 + "context": 1048576, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.0975, - "output": 0.325, - "cache_read": 0.0195 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "qwen3.6-35b-fast": { - "id": "qwen3.6-35b-fast", - "name": "Qwen3.6 35B Fast", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "qwen3.6", - "attachment": true, - "reasoning": false, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -124974,85 +118504,79 @@ }, "open_weights": true, "limit": { - "context": 131056, - "output": 131056 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.29, - "output": 1.15, - "cache_read": 0.029 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "glm-5.2-flex": { - "id": "glm-5.2-flex", - "name": "GLM 5.2 Flex", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "mimo-v2.5-tts": { + "id": "mimo-v2.5-tts", + "name": "MiMo-V2.5-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, "open_weights": true, "limit": { - "context": 1048560, - "output": 1048560 + "context": 8192, + "output": 8192 }, - "status": "deprecated", "cost": { - "input": 0.9425, - "output": 2.925, - "cache_read": 0.09425 + "input": 0, + "output": 0 } - }, - "deepseek-v4-flash-speed": { - "id": "deepseek-v4-flash-speed", - "name": "DeepSeek V4 Flash (Speed)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + } + } + }, + "zeldoc": { + "id": "zeldoc", + "env": [ + "ZELDOC_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.zeldoc.ai/v1", + "name": "Zeldoc", + "doc": "https://docs.zeldoc.ai", + "models": { + "z-code": { + "id": "z-code", + "name": "Z-Code", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "high", "max" ] } ], "tool_call": true, - "interleaved": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2025-01", + "release_date": "2026-04-15", + "last_updated": "2026-04-15", "modalities": { "input": [ "text" @@ -125061,43 +118585,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048560, - "output": 393216 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0, + "output": 0 } - }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM 5.2", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + } + } + }, + "scaleway": { + "id": "scaleway", + "env": [ + "SCALEWAY_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.scaleway.ai/v1", + "name": "Scaleway", + "doc": "https://www.scaleway.com/en/docs/generative-apis/", + "models": { + "qwen3-235b-a22b-instruct-2507": { + "id": "qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, - "structured_output": false, "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "knowledge": "2025-04", + "release_date": "2025-07-01", + "last_updated": "2026-03-17", "modalities": { "input": [ "text" @@ -125108,42 +118630,27 @@ }, "open_weights": true, "limit": { - "context": 1048560, - "output": 1048560 + "context": 260000, + "output": 16384 }, - "status": "deprecated", "cost": { - "input": 1.45, - "output": 4.5, - "cache_read": 0.145 + "input": 0.75, + "output": 2.25, + "reasoning": 8.4 } }, - "glm-5.2-short-fast-flex": { - "id": "glm-5.2-short-fast-flex", - "name": "GLM 5.2 Short Fast Flex", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": true, - "structured_output": false, "temperature": true, - "release_date": "2026-06-17", - "last_updated": "2026-06-17", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2026-03-17", "modalities": { "input": [ "text" @@ -125154,60 +118661,77 @@ }, "open_weights": true, "limit": { - "context": 199984, - "output": 32000 + "context": 128000, + "output": 32768 }, - "status": "deprecated", "cost": { - "input": 0.9425, - "output": 2.925, - "cache_read": 0.09425 + "input": 0.2, + "output": 0.8 } }, - "kimi-k2.7-code-fast": { - "id": "kimi-k2.7-code-fast", - "name": "Kimi K2.7 Code Fast", - "description": "Kimi K2.7 Code with reasoning capped to a short budget for lower latency; reasoning cannot be disabled on this model", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, + "qwen3-embedding-8b": { + "id": "qwen3-embedding-8b", + "name": "Qwen3 Embedding 8B", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2025-25-11", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262128, - "output": 262128 + "context": 32768, + "output": 4096 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.095 + "input": 0.1, + "output": 0 } }, - "gemma-4-31b": { - "id": "gemma-4-31b", - "name": "Gemma 4 31B", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "bge-multilingual-gemma2": { + "id": "bge-multilingual-gemma2", + "name": "BGE Multilingual Gemma2", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2024-07-26", + "last_updated": "2025-06-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 8191, + "output": 3072 + }, + "cost": { + "input": 0.1, + "output": 0 + } + }, + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -125215,16 +118739,18 @@ "type": "effort", "values": [ "none", - "max" + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-04", + "release_date": "2026-05-01", + "last_updated": "2026-05-22", "modalities": { "input": [ "text", @@ -125236,41 +118762,27 @@ }, "open_weights": true, "limit": { - "context": 262128, + "context": 128000, "output": 16384 }, + "status": "beta", "cost": { - "input": 0.144, - "output": 0.42, - "cache_read": 0.0144 + "input": 0.25, + "output": 1.5 } }, - "glm-5.3-flex": { - "id": "glm-5.3-flex", - "name": "GLM 5.3 Flex", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2026-03-17", "modalities": { "input": [ "text" @@ -125281,20 +118793,19 @@ }, "open_weights": true, "limit": { - "context": 1048560, - "output": 1048560 + "context": 100000, + "output": 16384 }, "cost": { - "input": 0.9425, - "output": 2.925, - "cache_read": 0.09425 + "input": 0.9, + "output": 0.9 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -125303,21 +118814,16 @@ "values": [ "none", "low", - "high", - "max" + "medium", + "high" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -125328,43 +118834,26 @@ }, "open_weights": true, "limit": { - "context": 1048560, - "output": 393216 + "context": 256000, + "output": 16384 }, - "status": "deprecated", "cost": { - "input": 1, - "output": 3, - "cache_read": 0.1 + "input": 1.8, + "output": 5.5 } }, - "qwen-3.8-27b-flex": { - "id": "qwen-3.8-27b-flex", - "name": "Qwen3.8 27B Flex", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", + "pixtral-12b-2409": { + "id": "pixtral-12b-2409", + "name": "Pixtral 12B 2409", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "pixtral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2024-09", + "release_date": "2024-09-25", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -125376,34 +118865,26 @@ }, "open_weights": true, "limit": { - "context": 262128, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.2925, - "output": 2.08, - "cache_read": 0.1625 + "input": 0.2, + "output": 0.2 } }, - "kimi-k2.7-code-flex": { - "id": "kimi-k2.7-code-flex", - "name": "Kimi K2.7 Code Flex", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], + "mistral-small-3.2-24b-instruct-2506": { + "id": "mistral-small-3.2-24b-instruct-2506", + "name": "Mistral Small 3.2 24B Instruct (2506)", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-03", + "release_date": "2025-06-20", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -125415,41 +118896,35 @@ }, "open_weights": true, "limit": { - "context": 262128, - "output": 262128 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.6175, - "output": 2.6, - "cache_read": 0.06175 + "input": 0.15, + "output": 0.35 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM 5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT-OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] - }, - { - "type": "budget_tokens" } ], "tool_call": true, - "interleaved": true, - "structured_output": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2024-01-01", + "last_updated": "2026-03-17", "modalities": { "input": [ "text" @@ -125460,34 +118935,38 @@ }, "open_weights": true, "limit": { - "context": 1048560, - "output": 1048560 + "context": 128000, + "output": 32768 }, "cost": { - "input": 1.45, - "output": 4.5, - "cache_read": 0.145 + "input": 0.15, + "output": 0.6 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "gemma-4-26b-a4b-it": { + "id": "gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-01", + "last_updated": "2026-05-22", "modalities": { "input": [ "text", @@ -125499,20 +118978,20 @@ }, "open_weights": true, "limit": { - "context": 262128, - "output": 262128 + "context": 256000, + "output": 16384 }, + "status": "beta", "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.095 + "input": 0.25, + "output": 0.5 } }, - "deepseek-v4.1-flash-flex": { - "id": "deepseek-v4.1-flash-flex", - "name": "DeepSeek V4.1 Flash Flex", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "mistral-medium-3.5-128b": { + "id": "mistral-medium-3.5-128b", + "name": "Mistral Medium 3.5 128B", + "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", + "family": "mistral-medium", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -125520,20 +118999,15 @@ "type": "effort", "values": [ "none", - "low", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ "text", @@ -125545,179 +119019,105 @@ }, "open_weights": true, "limit": { - "context": 1048560, - "output": 393216 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0.0975, - "output": 0.39, - "cache_read": 0.00975 + "input": 1.5, + "output": 7.5 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "high", - "max" + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048560, - "output": 393216 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 - } - } - } - }, - "gitlab": { - "id": "gitlab", - "env": [ - "GITLAB_TOKEN" - ], - "npm": "gitlab-ai-provider", - "name": "GitLab Duo", - "doc": "https://docs.gitlab.com/user/duo_agent_platform/", - "models": { - "duo-chat-gpt-5-4-nano": { - "id": "duo-chat-gpt-5-4-nano", - "name": "Agentic Chat (GPT-5.4 Nano)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", + "knowledge": "2025-04", "release_date": "2026-03-17", "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.6, + "output": 3.6 } }, - "duo-chat-opus-5-5": { - "id": "duo-chat-opus-5-5", - "name": "Agentic Chat (Claude Opus 5.5)", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, + "whisper-large-v3": { + "id": "whisper-large-v3", + "name": "Whisper Large v3", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "whisper", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "knowledge": "2023-09", + "release_date": "2023-09-01", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text", - "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 0, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.003, + "output": 0 } }, - "duo-chat-gpt-6-astra": { - "id": "duo-chat-gpt-6-astra", - "name": "Agentic Chat (GPT-6 Astra)", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", + "gemma-3-27b-it": { + "id": "gemma-3-27b-it", + "name": "Gemma-3-27B-IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "temperature": true, + "knowledge": "2024-12", + "release_date": "2024-12-01", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -125725,361 +119125,292 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 40000, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.25, + "output": 0.5 } }, - "duo-chat-opus-4-5": { - "id": "duo-chat-opus-4-5", - "name": "Agentic Chat (Claude Opus 4.5)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "voxtral-small-24b-2507": { + "id": "voxtral-small-24b-2507", + "name": "Voxtral Small 24B 2507", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "voxtral", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2026-01-08", - "last_updated": "2026-01-08", + "release_date": "2025-07-01", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 32000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.15, + "output": 0.35 } }, - "duo-chat-gpt-5-mini": { - "id": "duo-chat-gpt-5-mini", - "name": "Agentic Chat (GPT-5 Mini)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "devstral-2-123b-instruct-2512": { + "id": "devstral-2-123b-instruct-2512", + "name": "Devstral 2 123B Instruct (2512)", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "temperature": true, + "knowledge": "2025-12", + "release_date": "2026-01-07", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "output": 16384 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 2 } - }, - "duo-chat-sonnet-4-5": { - "id": "duo-chat-sonnet-4-5", - "name": "Agentic Chat (Claude Sonnet 4.5)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + } + } + }, + "ovhcloud": { + "id": "ovhcloud", + "env": [ + "OVHCLOUD_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://oai.endpoints.kepler.ai.cloud.ovh.net/v1", + "name": "OVHcloud AI Endpoints", + "doc": "https://www.ovhcloud.com/en/public-cloud/ai-endpoints/catalog//", + "models": { + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder-30B-A3B-Instruct", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2026-01-08", - "last_updated": "2026-01-08", + "release_date": "2025-10-28", + "last_updated": "2025-10-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.07, + "output": 0.26 } }, - "duo-chat-gpt-5-6-sol": { - "id": "duo-chat-gpt-5-6-sol", - "name": "Agentic Chat (GPT-5.6 Sol)", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3-32B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2025-07-16", + "last_updated": "2025-07-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.09, + "output": 0.25 } }, - "duo-chat-sonnet-4-6": { - "id": "duo-chat-sonnet-4-6", - "name": "Agentic Chat (Claude Sonnet 4.6)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "qwen3guard-gen-8b": { + "id": "qwen3guard-gen-8b", + "name": "Qwen3Guard-Gen-8B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "context": 32768, + "output": 16384 } }, - "duo-chat-gpt-5-4-mini": { - "id": "duo-chat-gpt-5-4-mini", - "name": "Agentic Chat (GPT-5.4 Mini)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "qwen3guard-gen-0.6b": { + "id": "qwen3guard-gen-0.6b", + "name": "Qwen3Guard-Gen-0.6B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 + "context": 32768, + "output": 16384 } }, - "duo-chat-haiku-4-5": { - "id": "duo-chat-haiku-4-5", - "name": "Agentic Chat (Claude Haiku 4.5)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "meta-llama-3_3-70b-instruct": { + "id": "meta-llama-3_3-70b-instruct", + "name": "Meta-Llama-3_3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2026-01-08", - "last_updated": "2026-01-08", + "release_date": "2025-04-01", + "last_updated": "2025-04-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.74, + "output": 0.74 } }, - "duo-chat-fable-5-1": { - "id": "duo-chat-fable-5-1", - "name": "Agentic Chat (Claude Fable 5.1)", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", + "mistral-small-3.2-24b-instruct-2506": { + "id": "mistral-small-3.2-24b-instruct-2506", + "name": "Mistral-Small-3.2-24B-Instruct-2506", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "structured_output": true, + "temperature": true, + "release_date": "2025-07-16", + "last_updated": "2025-07-16", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.1, + "output": 0.31 } }, - "duo-chat-gpt-5-5": { - "id": "duo-chat-gpt-5-5", - "name": "Agentic Chat (GPT-5.5)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "qwen2.5-vl-72b-instruct": { + "id": "qwen2.5-vl-72b-instruct", + "name": "Qwen2.5-VL-72B-Instruct", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2025-03-31", + "last_updated": "2025-03-31", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 1.01, + "output": 1.01 } }, - "duo-chat-fable-5": { - "id": "duo-chat-fable-5", - "name": "Agentic Chat (Claude Fable 5)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -126087,157 +119418,154 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "structured_output": true, + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.09, + "output": 0.47 } }, - "duo-chat-gpt-5-3-codex": { - "id": "duo-chat-gpt-5-3-codex", - "name": "Agentic Chat (GPT-5.3 Codex)", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistral-7b-instruct-v0.3": { + "id": "mistral-7b-instruct-v0.3", + "name": "Mistral-7B-Instruct-v0.3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "temperature": true, + "release_date": "2025-04-01", + "last_updated": "2025-04-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.11, + "output": 0.11 } }, - "duo-chat-gpt-5-1": { - "id": "duo-chat-gpt-5-1", - "name": "Agentic Chat (GPT-5.1)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistral-nemo-instruct-2407": { + "id": "mistral-nemo-instruct-2407", + "name": "Mistral-Nemo-Instruct-2407", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "temperature": true, + "release_date": "2024-11-20", + "last_updated": "2024-11-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.14, + "output": 0.14 } }, - "duo-chat-opus-4-7": { - "id": "duo-chat-opus-4-7", - "name": "Agentic Chat (Claude Opus 4.7)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen3.6-27b": { + "id": "qwen3.6-27b", + "name": "Qwen3.6-27B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.47, + "output": 3.19 } }, - "duo-chat-gpt-5-2": { - "id": "duo-chat-gpt-5-2", - "name": "Agentic Chat (GPT-5.2)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5-9B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-23", - "last_updated": "2026-01-23", + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", @@ -126247,161 +119575,124 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.12, + "output": 0.18 } }, - "duo-chat-opus-5": { - "id": "duo-chat-opus-5", - "name": "Agentic Chat (Claude Opus 5)", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5-397B-A17B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "structured_output": true, + "temperature": true, + "release_date": "2026-05-18", + "last_updated": "2026-05-18", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.71, + "output": 4.25 } }, - "duo-chat-opus-4-6": { - "id": "duo-chat-opus-4-6", - "name": "Agentic Chat (Claude Opus 4.6)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "duo-chat-gpt-5-6-luna": { - "id": "duo-chat-gpt-5-6-luna", - "name": "Agentic Chat (GPT-5.6 Luna)", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "gpt-oss-20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.05, + "output": 0.18 } - }, - "duo-chat-gpt-5-codex": { - "id": "duo-chat-gpt-5-codex", - "name": "Agentic Chat (GPT-5 Codex)", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + } + } + }, + "friendli": { + "id": "friendli", + "env": [ + "FRIENDLI_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.friendli.ai/serverless/v1", + "name": "Friendli", + "doc": "https://friendli.ai/docs/guides/serverless_endpoints/introduction", + "models": { + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -126411,465 +119702,387 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.14, + "output": 0.4 } }, - "duo-chat-sonnet-5": { - "id": "duo-chat-sonnet-5", - "name": "Agentic Chat (Claude Sonnet 5)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-29", + "last_updated": "2026-01-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.2, + "output": 0.8 } }, - "duo-chat-opus-4-8": { - "id": "duo-chat-opus-4-8", - "name": "Agentic Chat (Claude Opus 4.8)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "duo-chat-gpt-5-2-codex": { - "id": "duo-chat-gpt-5-2-codex", - "name": "Agentic Chat (GPT-5.2 Codex)", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 202752, + "output": 202752 }, "cost": { - "input": 0, - "output": 0 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "duo-chat-gpt-5-4": { - "id": "duo-chat-gpt-5-4", - "name": "Agentic Chat (GPT-5.4)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0, - "output": 0 + "input": 0.5, + "output": 1.5, + "cache_read": 0.25 } }, - "duo-chat-gpt-5-6-terra": { - "id": "duo-chat-gpt-5-6-terra", - "name": "Agentic Chat (GPT-5.6 Terra)", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 196608, + "output": 196608 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } } } }, - "infer": { - "id": "infer", + "tencent-tokenhub": { + "id": "tencent-tokenhub", "env": [ - "INFER_API_KEY" + "TENCENT_TOKENHUB_API_KEY" ], - "npm": "@ai-sdk/openai", - "api": "https://infer.flow7.org/v1", - "name": "Infer by Flow7", - "doc": "https://infer.flow7.org/opencode", + "npm": "@ai-sdk/openai-compatible", + "api": "https://tokenhub.tencentmaas.com/v1", + "name": "Tencent TokenHub", + "doc": "https://cloud.tencent.com/document/product/1823/130050", "models": { - "infer/gpt-6-astra:official": { - "id": "infer/gpt-6-astra:official", - "name": "GPT-6 Astra (Official API)", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, + "hy3": { + "id": "hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "temperature": true, + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 271999, - "output": 128000 - }, - "provider": { - "shape": "responses" + "context": 256000, + "output": 64000 }, "cost": { - "input": 12.5, - "output": 62.5, - "cache_read": 1.25, - "cache_write": 15.625 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "infer/gpt-5.6-sol:official": { - "id": "infer/gpt-5.6-sol:official", - "name": "GPT-5.6 Sol (Official API)", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, + "hy3-preview": { + "id": "hy3-preview", + "name": "Hy3 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 271999, - "output": 128000 - }, - "provider": { - "shape": "responses" + "context": 256000, + "output": 64000 }, "cost": { - "input": 2.5, - "output": 12.5, - "cache_read": 0.25, - "cache_write": 3.125 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } } } }, - "azure": { - "id": "azure", + "wandb": { + "id": "wandb", "env": [ - "AZURE_RESOURCE_NAME", - "AZURE_API_KEY" + "WANDB_API_KEY" ], - "npm": "@ai-sdk/azure", - "name": "Azure", - "doc": "https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models", + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.inference.wandb.ai/v1", + "name": "Weights & Biases", + "doc": "https://docs.wandb.ai/guides/integrations/inference/", "models": { - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "Grok 4.1 Fast (Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "ibm-granite/granite-4.1-8b": { + "id": "ibm-granite/granite-4.1-8b", + "name": "Granite 4.1 8B", + "description": "Granite 4.1 8B is a long-context instruct model capable of enhanced tool calling, instruction following, and chat capabilities.", + "family": "granite", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-06-27", - "last_updated": "2025-06-27", + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.5, + "input": 0.05, + "output": 0.1, "cache_read": 0.05 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama 3.3 70B", + "description": "Multilingual model excelling in conversational tasks, detailed instruction-following, and coding.", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2023-12", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, + "context": 128000, "output": 128000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0.71, + "output": 0.71, + "cache_read": 0.71 } }, - "deepseek-v3.2-speciale": { - "id": "deepseek-v3.2-speciale", - "name": "DeepSeek-V3.2-Speciale", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "meta-llama/Llama-3.1-70B-Instruct": { + "id": "meta-llama/Llama-3.1-70B-Instruct", + "name": "Llama 3.1 70B", + "description": "Efficient conversational model optimized for responsive multilingual chatbot interactions.", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -126884,78 +120097,57 @@ "output": 128000 }, "cost": { - "input": 0.58, - "output": 1.68 + "input": 0.8, + "output": 0.8, + "cache_read": 0.8 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "meta-llama/Llama-3.1-8B-Instruct": { + "id": "meta-llama/Llama-3.1-8B-Instruct", + "name": "Llama 3.1 8B", + "description": "Efficient conversational model optimized for responsive multilingual chatbot interactions.", + "family": "llama", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 128000, + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.22, + "output": 0.22, + "cache_read": 0.22 } }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Kimi K2.6 is a multimodal Mixture-of-Experts language model featuring 32 billion activated parameters and a total of 1 trillion parameters.", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text", @@ -126965,42 +120157,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "Grok 4.1 Fast (Non-Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi K2.5", + "description": "Kimi K2.5 is a multimodal Mixture-of-Experts language model featuring 32 billion activated parameters and a total of 1 trillion parameters.", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-06-27", - "last_updated": "2025-06-27", + "knowledge": "2025-01", + "release_date": "2026-02-02", + "last_updated": "2026-02-02", "modalities": { "input": [ "text", @@ -127010,34 +120191,35 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 128000, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "phi-4-reasoning": { - "id": "phi-4-reasoning", - "name": "Phi-4-reasoning", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", - "attachment": false, + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Kimi K2.7 Code is a 1T-parameter MoE model with 32B active parameters purpose-built for long-horizon agentic coding and software engineering.", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -127045,113 +120227,102 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 4096 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.125, - "output": 0.5 + "input": 0.94, + "output": 4, + "cache_read": 0.19 } }, - "text-embedding-3-small": { - "id": "text-embedding-3-small", - "name": "text-embedding-3-small", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B", + "description": "Gemma 4 31B Dense is designed for advanced reasoning, agentic workflows, and longer context and is natively trained on 140+ languages.", + "family": "gemma", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8191, - "output": 1536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.02, - "output": 0 + "input": 0.12, + "output": 0.35, + "cache_read": 0.09 } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 Nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "microsoft/Phi-4-mini-instruct": { + "id": "microsoft/Phi-4-mini-instruct", + "name": "Phi 4 Mini 3.8B", + "description": "Compact, efficient model ideal for fast responses in resource-constrained environments.", + "family": "phi", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "knowledge": "2023-10", + "release_date": "2025-02-01", + "last_updated": "2025-02-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, + "context": 128000, "output": 128000 }, + "status": "deprecated", "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.08, + "output": 0.35, + "cache_read": 0.08 } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "Qwen/Qwen3.6-35B-A3B": { + "id": "Qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B A3B", + "description": "Qwen3.6-35B-A3B is an MoE multimodal model with 262K context optimized for agentic coding workflows.", + "family": "qwen3.6", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "temperature": true, + "release_date": "2026-04-15", + "last_updated": "2026-04-15", "modalities": { "input": [ "text", @@ -127161,110 +120332,101 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.25, + "output": 1.25, + "cache_read": 0.25 } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt-codex", - "attachment": false, + "Qwen/Qwen3.6-27B": { + "id": "Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", + "description": "Qwen3.6-27B is a 27B dense multimodal model with 262K context built for flagship-level agentic coding.", + "family": "qwen3.6", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ - "text", - "image", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.6, + "output": 3.6, + "cache_read": 0.12 } }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, - "reasoning": false, + "Qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen3 235B A22B Thinking-2507", + "description": "High-performance Mixture-of-Experts model optimized for structured reasoning, math, and long-form generation.", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "knowledge": "2025-04", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 262144 }, "status": "deprecated", "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.1, + "output": 0.1, + "cache_read": 0.1 } }, - "ministral-3b": { - "id": "ministral-3b", - "name": "Ministral 3B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "Qwen/Qwen3-Coder-480B-A35B-Instruct": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", + "name": "Qwen3 Coder 480B A35B", + "description": "Mixture-of-Experts model optimized for agentic coding tasks such as function calling, tool use, and long-context reasoning.", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "knowledge": "2025-04", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ "text" @@ -127275,236 +120437,202 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.04, - "output": 0.04 + "input": 1, + "output": 1.5, + "cache_read": 1 } }, - "gpt-chat-latest": { - "id": "gpt-chat-latest", - "name": "GPT Chat Latest", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "Qwen/Qwen3.5-27B": { + "id": "Qwen/Qwen3.5-27B", + "name": "Qwen3.5-27B", + "description": "Qwen3.5-27B is a dense model from the Qwen3.5 family built for high performance across a large range of benchmarks.", + "family": "qwen3.5", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-05-05", - "last_updated": "2026-05-28", + "temperature": true, + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "input": 111616, - "output": 16384 + "context": 262144, + "output": 262144 }, - "status": "beta", + "status": "deprecated", "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.39, + "output": 3.12, + "cache_read": 0.08 } }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "Qwen/Qwen3-30B-A3B-Instruct-2507": { + "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen3 30B A3B Instruct 2507", + "description": "Qwen3-30B-A3B-Instruct-2507 is a 30.5B MoE instruction-tuned model with enhanced reasoning, coding, and long-context understanding.", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "structured_output": true, + "temperature": true, + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.1, + "output": 0.3, + "cache_read": 0.1 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B-2507", + "description": "Efficient multilingual, Mixture-of-Experts, instruction-tuned model, optimized for logical reasoning.", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.03 + "input": 0.1, + "output": 0.1, + "cache_read": 0.1 } }, - "gpt-image-2": { - "id": "gpt-image-2", - "name": "GPT-Image-2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "Qwen/Qwen3.5-35B-A3B": { + "id": "Qwen/Qwen3.5-35B-A3B", + "name": "Qwen3.5-35B-A3B", + "description": "Qwen3.5-35B-A3B is an open-weights multimodal MoE model built for efficient, high-throughput inference across chat, reasoning, and agentic tasks.", + "family": "qwen3.5", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", "image" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 1.25 + "input": 0.25, + "output": 1.25, + "cache_read": 0.25 } }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Efficient Mixture-of-Experts model designed for high-reasoning, agentic and general-purpose use cases.", + "family": "gpt-oss", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 131072 }, - "status": "deprecated", "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 0.04, + "output": 0.14, + "cache_read": 0.04 } }, - "llama-4-maverick-17b-128e-instruct-fp8": { - "id": "llama-4-maverick-17b-128e-instruct-fp8", - "name": "Llama 4 Maverick 17B 128E Instruct FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": true, - "reasoning": false, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "gpt-oss-20b", + "description": "Lower latency Mixture-of-Experts model trained on OpenAI's Harmony response format with reasoning capabilities.", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -127512,36 +120640,32 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 16384 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.03, + "output": 0.13, + "cache_read": 0.03 } }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B": { + "id": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B", + "name": "Nemotron 3 Ultra", + "description": "Nemotron 3 Ultra is a powerful MoE model designed for long-running agents across coding, deep research, and enterprise automation.", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "structured_output": true, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -127550,80 +120674,64 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 262144 }, - "status": "deprecated", "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.75, + "output": 2.75, + "cache_read": 0.15 } }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8": { + "id": "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8", + "name": "Nemotron 3 Super", + "description": "Nemotron 3 is a LatentMoE model designed to deliver strong agentic, reasoning, and conversational capabilities.", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-08-01", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 262144, + "output": 262144 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.2, + "output": 0.8, + "cache_read": 0.2 } }, - "text-embedding-ada-002": { - "id": "text-embedding-ada-002", - "name": "text-embedding-ada-002", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "JetBrains/Mellum2-12B-A2.5B-Instruct": { + "id": "JetBrains/Mellum2-12B-A2.5B-Instruct", + "name": "Mellum2 12B A2.5B", + "description": "Mellum2-12B-A2.5B-Instruct is a fast MoE model with 131K context built for coding, tool use, and low-latency AI workflows.", "attachment": false, "reasoning": false, - "tool_call": false, - "release_date": "2022-12-15", - "last_updated": "2022-12-15", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text" @@ -127632,156 +120740,135 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0 + "input": 0.05, + "output": 0.1, + "cache_read": 0.05 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "OpenPipe/Qwen3-14B-Instruct": { + "id": "OpenPipe/Qwen3-14B-Instruct", + "name": "Qwen3 14B Instruct", + "description": "An efficient multilingual, dense, instruction-tuned model, optimized by OpenPipe for building agents with finetuning.", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "temperature": true, + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.05, + "output": 0.22, + "cache_read": 0.05 } }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM 5.2", + "description": "GLM-5.2 is a Mixture-of-Experts language model featuring 40 billion activated parameters and a total of 744 billion parameters.", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 262144, + "output": 262144 }, - "status": "deprecated", "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 1.39, + "output": 4.4, + "cache_read": 0.26 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM 5.1", + "description": "Powerful MoE model for long-horizon agentic engineering and advanced reasoning.", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "structured_output": true, + "temperature": true, + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 202752, + "output": 202752 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "phi-4-multimodal": { - "id": "phi-4-multimodal", - "name": "Phi-4-multimodal", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "phi", - "attachment": true, + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek V3.1", + "description": "A large hybrid model that supports both thinking and non-thinking modes via prompt templates.", + "family": "deepseek", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -127789,20 +120876,20 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 161000, + "output": 161000 }, "cost": { - "input": 0.08, - "output": 0.32, - "input_audio": 4 + "input": 0.55, + "output": 1.65, + "cache_read": 0.55 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "deepseek-ai/DeepSeek-V4-Flash": { + "id": "deepseek-ai/DeepSeek-V4-Flash", + "name": "DeepSeek V4 Flash", + "description": "DeepSeek V4-Flash is an MoE model with 1M context length great for coding, reasoning, and agentic workloads.", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -127811,16 +120898,14 @@ } ], "tool_call": true, - "interleaved": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -127828,171 +120913,195 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.95, - "output": 4 + "input": 0.14, + "output": 0.28, + "cache_read": 0.07 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "DeepSeek V4-Pro is a 1.6T-parameter MoE model with 49B active parameters excelling at advanced reasoning, coding, and complex agentic workloads.", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 1.74, + "output": 3.48, + "cache_read": 0.14 } }, - "claude-opus-5-5": { - "id": "claude-opus-5-5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax M2.5", + "description": "MoE model with a highly sparse architecture designed for high-throughput and low latency with strong coding capabilities.", + "family": "minimax-m2.5", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 196608 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.3 + } + } + } + }, + "kuae-cloud-coding-plan": { + "id": "kuae-cloud-coding-plan", + "env": [ + "KUAE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://coding-plan-endpoint.kuaecloud.net/v1", + "name": "KUAE Cloud Coding Plan", + "doc": "https://docs.mthreads.com/kuaecloud/kuaecloud-doc-online/coding_plan/", + "models": { + "GLM-4.7": { + "id": "GLM-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 204800, + "output": 131072 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "gpt-image-2.5-flare": { - "id": "gpt-image-2.5-flare", - "name": "GPT Image 2.5 Flare", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + } + } + }, + "gitlab": { + "id": "gitlab", + "env": [ + "GITLAB_TOKEN" + ], + "npm": "gitlab-ai-provider", + "name": "GitLab Duo", + "doc": "https://docs.gitlab.com/user/duo_agent_platform/", + "models": { + "duo-chat-opus-4-5": { + "id": "duo-chat-opus-4-5", + "name": "Agentic Chat (Claude Opus 4.5)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2026-01-08", + "last_updated": "2026-01-08", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "output": 64000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 1.25 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "claude-fable-5-1": { - "id": "claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", + "duo-chat-opus-4-8": { + "id": "duo-chat-opus-4-8", + "name": "Agentic Chat (Claude Opus 4.8)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "knowledge": "2026-01-31", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -128008,33 +121117,31 @@ "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" - }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "mistral-medium-2505": { - "id": "mistral-medium-2505", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "duo-chat-opus-4-7": { + "id": "duo-chat-opus-4-7", + "name": "Agentic Chat (Claude Opus 4.7)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -128042,49 +121149,57 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.4, - "output": 2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "cohere-embed-v3-multilingual": { - "id": "cohere-embed-v3-multilingual", - "name": "Embed v3 Multilingual", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", - "attachment": false, - "reasoning": false, - "tool_call": false, + "duo-chat-gpt-5-2-codex": { + "id": "duo-chat-gpt-5-2-codex", + "name": "Agentic Chat (GPT-5.2 Codex)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2023-11-07", - "last_updated": "2023-11-07", + "knowledge": "2025-08-31", + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512, - "output": 1024 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.1, + "input": 0, "output": 0 } }, - "o1": { - "id": "o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": false, + "duo-chat-fable-5": { + "id": "duo-chat-fable-5", + "name": "Agentic Chat (Claude Fable 5)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -128092,19 +121207,22 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -128112,41 +121230,35 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1000000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "codex-mini": { - "id": "codex-mini", - "name": "Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex-mini", + "duo-chat-gpt-5-5": { + "id": "duo-chat-gpt-5-5", + "name": "Agentic Chat (GPT-5.5)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2024-04", - "release_date": "2025-05-16", - "last_updated": "2025-05-16", + "knowledge": "2025-08-31", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -128154,72 +121266,64 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1050000, + "input": 922000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 1.5, - "output": 6, - "cache_read": 0.375 + "input": 0, + "output": 0 } }, - "phi-4-reasoning-plus": { - "id": "phi-4-reasoning-plus", - "name": "Phi-4-reasoning-plus", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", - "attachment": false, + "duo-chat-opus-4-6": { + "id": "duo-chat-opus-4-6", + "name": "Agentic Chat (Claude Opus 4.6)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 4096 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.125, - "output": 0.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "claude-mythos-5": { - "id": "claude-mythos-5", - "name": "Claude Mythos 5", - "description": "Restricted Claude model for advanced cybersecurity and biology research workflows", - "family": "claude-mythos", + "duo-chat-gpt-5-4": { + "id": "duo-chat-gpt-5-4", + "name": "Agentic Chat (GPT-5.4)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -128232,42 +121336,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, - "status": "beta", - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" - }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0, + "output": 0 } }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", - "attachment": true, + "duo-chat-gpt-5-codex": { + "id": "duo-chat-gpt-5-codex", + "name": "Agentic Chat (GPT-5 Codex)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": false, "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ "text", @@ -128280,28 +121371,32 @@ "open_weights": false, "limit": { "context": 400000, - "output": 272000 + "input": 272000, + "output": 128000 }, "cost": { - "input": 15, - "output": 120 + "input": 0, + "output": 0 } }, - "gpt-3.5-turbo-instruct": { - "id": "gpt-3.5-turbo-instruct", - "name": "GPT-3.5 Turbo Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-09-21", - "last_updated": "2023-09-21", + "duo-chat-gpt-5-4-nano": { + "id": "duo-chat-gpt-5-4-nano", + "name": "Agentic Chat (GPT-5.4 Nano)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -128309,30 +121404,33 @@ }, "open_weights": false, "limit": { - "context": 4096, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 1.5, - "output": 2 + "input": 0, + "output": 0 } }, - "gpt-3.5-turbo-0125": { - "id": "gpt-3.5-turbo-0125", - "name": "GPT-3.5 Turbo 0125", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, + "duo-chat-sonnet-4-6": { + "id": "duo-chat-sonnet-4-6", + "name": "Agentic Chat (Claude Sonnet 4.6)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2021-08", - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -128340,28 +121438,34 @@ }, "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 1000000, + "output": 64000 }, - "status": "deprecated", "cost": { - "input": 0.5, - "output": 1.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "text-embedding-3-large": { - "id": "text-embedding-3-large", - "name": "text-embedding-3-large", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "duo-chat-gpt-5-mini": { + "id": "duo-chat-gpt-5-mini", + "name": "Agentic Chat (GPT-5 Mini)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -128369,33 +121473,42 @@ }, "open_weights": false, "limit": { - "context": 8191, - "output": 3072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.13, + "input": 0, "output": 0 } }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "duo-chat-sonnet-5": { + "id": "duo-chat-sonnet-5", + "name": "Agentic Chat (Claude Sonnet 5)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", @@ -128408,31 +121521,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 1000000, + "output": 64000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "cohere-embed-v-4-0": { - "id": "cohere-embed-v-4-0", - "name": "Embed v4", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", + "duo-chat-gpt-5-4-mini": { + "id": "duo-chat-gpt-5-4-mini", + "name": "Agentic Chat (GPT-5.4 Mini)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-mini", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -128442,42 +121554,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 1536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.12, + "input": 0, "output": 0 } }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", + "duo-chat-gpt-5-3-codex": { + "id": "duo-chat-gpt-5-3-codex", + "name": "Agentic Chat (GPT-5.3 Codex)", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", @@ -128490,50 +121591,33 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } + "input": 0, + "output": 0 } }, - "codestral-2501": { - "id": "codestral-2501", - "name": "Codestral 25.01", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "codestral", - "attachment": false, - "reasoning": false, + "duo-chat-haiku-4-5": { + "id": "duo-chat-haiku-4-5", + "name": "Agentic Chat (Claude Haiku 4.5)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-03", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2025-02-28", + "release_date": "2026-01-08", + "last_updated": "2026-01-08", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -128541,33 +121625,30 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": false, + "duo-chat-gpt-5-2": { + "id": "duo-chat-gpt-5-2", + "name": "Agentic Chat (GPT-5.2)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-01-23", + "last_updated": "2026-01-23", "modalities": { "input": [ "text", @@ -128577,55 +121658,73 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" + "cost": { + "input": 0, + "output": 0 + } + }, + "duo-chat-sonnet-4-5": { + "id": "duo-chat-sonnet-4-5", + "name": "Agentic Chat (Claude Sonnet 4.5)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2026-01-08", + "last_updated": "2026-01-08", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 3 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "Speech generation model for controllable voice, narration, and audio delivery", + "duo-chat-gpt-5-1": { + "id": "duo-chat-gpt-5-1", + "name": "Agentic Chat (GPT-5.1)", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": false, "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "release_date": "2026-01-22", + "last_updated": "2026-01-22", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ - "text", - "image", - "audio" + "text" ] }, "open_weights": false, @@ -128635,41 +121734,65 @@ "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0, + "output": 0 + } + } + } + }, + "kilo": { + "id": "kilo", + "env": [ + "KILO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.kilo.ai/api/gateway", + "name": "Kilo Gateway", + "doc": "https://kilo.ai", + "models": { + "inclusionai/ling-2.6-1t": { + "id": "inclusionai/ling-2.6-1t", + "name": "inclusionAI: Ling-2.6-1T", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2026-04-23", + "last_updated": "2026-05-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.06 } }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, + "inclusionai/ring-2.6-1t": { + "id": "inclusionai/ring-2.6-1t", + "name": "inclusionAI: Ring-2.6-1T", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "temperature": true, + "release_date": "2026-05-08", + "last_updated": "2026-05-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -128677,32 +121800,54 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 262144, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.075, + "output": 0.625, + "cache_read": 0.015 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "inclusionai/ling-2.6-flash": { + "id": "inclusionai/ling-2.6-flash", + "name": "inclusionAI: Ling-2.6 Flash", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2026-04-21", + "last_updated": "2026-05-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.08, + "output": 0.24, + "cache_read": 0.016 + } + }, + "ibm-granite/granite-4.0-h-micro": { + "id": "ibm-granite/granite-4.0-h-micro", + "name": "IBM: Granite 4.0 Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-10-20", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -128713,30 +121858,27 @@ }, "open_weights": true, "limit": { - "context": 128000, + "context": 131000, "output": 32768 }, "cost": { - "input": 0.71, - "output": 0.71 + "input": 0.017, + "output": 0.11 } }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, + "ibm-granite/granite-4.1-8b": { + "id": "ibm-granite/granite-4.1-8b", + "name": "IBM: Granite 4.1 8B", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-04-30", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -128744,29 +121886,81 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.05, + "output": 0.1, + "cache_read": 0.05 + } + }, + "meta-llama/llama-3.1-8b-instruct": { + "id": "meta-llama/llama-3.1-8b-instruct", + "name": "Meta: Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-07-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 16384, "output": 16384 }, - "status": "deprecated", "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.02, + "output": 0.05 } }, - "cohere-command-a": { - "id": "cohere-command-a", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", + "meta-llama/llama-3-70b-instruct": { + "id": "meta-llama/llama-3-70b-instruct", + "name": "Meta: Llama 3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8000 + }, + "cost": { + "input": 0.51, + "output": 0.74 + } + }, + "meta-llama/llama-3.1-70b-instruct": { + "id": "meta-llama/llama-3.1-70b-instruct", + "name": "Meta: Llama 3.1 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "release_date": "2024-07-16", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -128778,24 +121972,23 @@ "open_weights": true, "limit": { "context": 131072, - "output": 8192 + "output": 26215 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.4, + "output": 0.4 } }, - "cohere-embed-v3-english": { - "id": "cohere-embed-v3-english", - "name": "Embed v3 English", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", + "meta-llama/llama-3.2-1b-instruct": { + "id": "meta-llama/llama-3.2-1b-instruct", + "name": "Meta: Llama 3.2 1B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2023-11-07", - "last_updated": "2023-11-07", + "temperature": true, + "release_date": "2024-09-18", + "last_updated": "2026-01-27", "modalities": { "input": [ "text" @@ -128806,24 +121999,24 @@ }, "open_weights": true, "limit": { - "context": 512, - "output": 1024 + "context": 60000, + "output": 12000 }, "cost": { - "input": 0.1, - "output": 0 + "input": 0.027, + "output": 0.2 } }, - "model-router": { - "id": "model-router", - "name": "Model Router", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "model-router", + "meta-llama/llama-4-maverick": { + "id": "meta-llama/llama-4-maverick", + "name": "Meta: Llama 4 Maverick", + "description": "Open multimodal Llama model for strong reasoning and fast responses", "attachment": true, "reasoning": false, "tool_call": true, - "release_date": "2025-05-19", - "last_updated": "2025-11-18", + "temperature": true, + "release_date": "2025-04-05", + "last_updated": "2025-12-24", "modalities": { "input": [ "text", @@ -128833,271 +122026,244 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, + "context": 1048576, "output": 16384 }, "cost": { - "input": 0.14, - "output": 0 + "input": 0.15, + "output": 0.6 } }, - "gpt-image-2.5-sunburst": { - "id": "gpt-image-2.5-sunburst", - "name": "GPT Image 2.5 Sunburst", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "meta-llama/llama-3.2-11b-vision-instruct": { + "id": "meta-llama/llama-3.2-11b-vision-instruct", + "name": "Meta: Llama 3.2 11B Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", "attachment": true, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "temperature": true, + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text", "image" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "output": 16384 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 1.25 + "input": 0.049, + "output": 0.049 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "meta-llama/llama-3.3-70b-instruct": { + "id": "meta-llama/llama-3.3-70b-instruct", + "name": "Meta: Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "release_date": "2024-08-01", + "last_updated": "2026-02-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 16384 }, - "status": "beta", - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "cost": { + "input": 0.1, + "output": 0.32 + } + }, + "meta-llama/llama-guard-3-8b": { + "id": "meta-llama/llama-guard-3-8b", + "name": "Llama Guard 3 8B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-04-18", + "last_updated": "2026-02-04", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 26215 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.02, + "output": 0.06 } }, - "gpt-image-1": { - "id": "gpt-image-1", - "name": "GPT-Image-1", - "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", - "family": "gpt-image", + "meta-llama/llama-guard-4-12b": { + "id": "meta-llama/llama-guard-4-12b", + "name": "Meta: Llama Guard 4 12B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", "attachment": true, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-04-24", - "last_updated": "2025-04-24", + "temperature": true, + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 163840, + "output": 32768 }, "cost": { - "input": 5, - "output": 40, - "cache_read": 1.25 + "input": 0.18, + "output": 0.18 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 Mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "meta-llama/llama-3-8b-instruct": { + "id": "meta-llama/llama-3-8b-instruct", + "name": "Meta: Llama 3 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2024-04-25", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 8192, + "output": 16384 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.03, + "output": 0.04 } }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", + "meta-llama/llama-4-scout": { + "id": "meta-llama/llama-4-scout", + "name": "Meta: Llama 4 Scout", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 327680, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } + "input": 0.08, + "output": 0.3 + } + }, + "meta-llama/llama-3.2-3b-instruct": { + "id": "meta-llama/llama-3.2-3b-instruct", + "name": "Meta: Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-09-18", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 80000, + "output": 16384 + }, + "cost": { + "input": 0.051, + "output": 0.34 } }, - "gpt-5.1-codex-max": { - "id": "gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "~anthropic/claude-haiku-latest": { + "id": "~anthropic/claude-haiku-latest", + "name": "Anthropic: Claude Haiku Latest", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -129105,44 +122271,45 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "~anthropic/claude-sonnet-latest": { + "id": "~anthropic/claude-sonnet-latest", + "name": "Anthropic: Claude Sonnet Latest", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh" + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -129150,24 +122317,26 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.125 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "~anthropic/claude-opus-latest": { + "id": "~anthropic/claude-opus-latest", + "name": "Anthropic: Claude Opus Latest", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -129181,9 +122350,8 @@ ], "tool_call": true, "temperature": false, - "knowledge": "2025-12-31", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "release_date": "2026-04-16", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", @@ -129199,47 +122367,24 @@ "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" - }, "cost": { "input": 5, "output": 25, "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "cache_write": 6.25 } }, - "phi-4": { - "id": "phi-4", - "name": "Phi-4", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", + "kilo-auto/balanced": { + "id": "kilo-auto/balanced", + "name": "Kilo Auto Balanced", + "description": "Automatic model router for matching prompts to suitable backends and budgets", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2026-03-15", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -129248,46 +122393,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.125, - "output": 0.5 + "input": 0.6, + "output": 3 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "kilo-auto/small": { + "id": "kilo-auto/small", + "name": "Kilo Auto Small", + "description": "Automatic model router for matching prompts to suitable backends and budgets", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": true, + "release_date": "2026-03-15", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" @@ -129295,48 +122425,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.05, + "output": 0.4 } }, - "mistral-small-2503": { - "id": "mistral-small-2503", - "name": "Mistral Small 3.1", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, - "reasoning": false, + "kilo-auto/free": { + "id": "kilo-auto/free", + "name": "Kilo Auto Free", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2025-03-01", - "last_updated": "2025-03-01", + "release_date": "2026-03-15", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -129344,215 +122454,156 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0, + "output": 0 } }, - "phi-4-mini-reasoning": { - "id": "phi-4-mini-reasoning", - "name": "Phi-4-mini-reasoning", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, + "kilo-auto/frontier": { + "id": "kilo-auto/frontier", + "name": "Kilo Auto Frontier", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2026-03-15", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 5, + "output": 25 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, + "moonshotai/kimi-k2": { + "id": "moonshotai/kimi-k2", + "name": "MoonshotAI: Kimi K2 0711", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-07-11", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 131000, + "output": 26215 }, - "status": "deprecated", "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.55, + "output": 2.2 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "MoonshotAI: Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "release_date": "2025-11-06", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 131072, + "output": 65535 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.47, + "output": 2, + "cache_read": 0.2 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "MoonshotAI: Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-07-17", + "release_date": "2026-01-27", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 262144, + "output": 65535 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 0.45, + "output": 2.2 } }, - "gpt-4-turbo-vision": { - "id": "gpt-4-turbo-vision", - "name": "GPT-4 Turbo Vision", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "MoonshotAI: Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2026-04-20", + "last_updated": "2026-05-12", "modalities": { "input": [ "text", @@ -129562,29 +122613,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 65535 }, - "status": "deprecated", "cost": { - "input": 10, - "output": 30 + "input": 0.75, + "output": 3.5, + "cache_read": 0.375 } }, - "grok-4-20-non-reasoning": { - "id": "grok-4-20-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "MoonshotAI: Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -129593,29 +122642,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 8192 + "context": 131072, + "output": 26215 }, - "status": "beta", "cost": { - "input": 2, - "output": 6 + "input": 0.4, + "output": 2, + "cache_read": 0.15 } }, - "gpt-3.5-turbo-1106": { - "id": "gpt-3.5-turbo-1106", - "name": "GPT-3.5 Turbo 1106", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "baidu/ernie-4.5-300b-a47b": { + "id": "baidu/ernie-4.5-300b-a47b", + "name": "Baidu: ERNIE 4.5 300B A47B ", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-11-06", - "last_updated": "2023-11-06", + "release_date": "2025-06-30", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -129624,29 +122671,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "output": 16384 + "context": 123000, + "output": 12000 }, - "status": "deprecated", "cost": { - "input": 1, - "output": 2 + "input": 0.28, + "output": 1.1 } }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", + "baidu/ernie-4.5-vl-28b-a3b": { + "id": "baidu/ernie-4.5-vl-28b-a3b", + "name": "Baidu: ERNIE 4.5 VL 28B A3B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ "text", @@ -129656,98 +122701,61 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 30000, + "output": 8000 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.14, + "output": 0.56 } }, - "gpt-6-luna": { - "id": "gpt-6-luna", - "name": "GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", + "baidu/ernie-4.5-vl-424b-a47b": { + "id": "baidu/ernie-4.5-vl-424b-a47b", + "name": "Baidu: ERNIE 4.5 VL 424B A47B ", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "tool_call": false, + "temperature": true, + "release_date": "2025-06-30", + "last_updated": "2026-01", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 123000, + "output": 16000 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25 - } + "input": 0.42, + "output": 1.25 } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "baidu/ernie-4.5-21b-a3b-thinking": { + "id": "baidu/ernie-4.5-21b-a3b-thinking", + "name": "Baidu: ERNIE 4.5 21B A3B Thinking", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "release_date": "2025-09-19", + "last_updated": "2025-09-19", "modalities": { "input": [ "text" @@ -129758,29 +122766,25 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 131072, + "output": 65536 }, - "status": "deprecated", "cost": { - "input": 1.35, - "output": 5.4 + "input": 0.07, + "output": 0.28 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek-V4-Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "baidu/cobuddy:free": { + "id": "baidu/cobuddy:free", + "name": "Baidu: CoBuddy (free)", + "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "tool_call": true, + "temperature": false, + "release_date": "2026-05-06", + "last_updated": "2026-05-07", "modalities": { "input": [ "text" @@ -129789,151 +122793,114 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" + "context": 131072, + "output": 65536 }, "cost": { - "input": 1.74, - "output": 3.48 + "input": 0, + "output": 0 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "baidu/ernie-4.5-21b-a3b": { + "id": "baidu/ernie-4.5-21b-a3b", + "name": "Baidu: ERNIE 4.5 21B A3B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 120000, + "output": 8000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.07, + "output": 0.28 } }, - "llama-4-scout-17b-16e-instruct": { - "id": "llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout 17B 16E Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + "baidu/qianfan-ocr-fast": { + "id": "baidu/qianfan-ocr-fast", + "name": "Baidu: Qianfan-OCR-Fast", + "description": "OCR model for extracting structured text from documents and screenshots", "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2026-04-20", + "last_updated": "2026-05-16", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 65536, + "output": 28672 }, "cost": { - "input": 0.2, - "output": 0.78 + "input": 0.68, + "output": 2.81 } }, - "gpt-image-1.5": { - "id": "gpt-image-1.5", - "name": "GPT-Image-1.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "perceptron/perceptron-mk1": { + "id": "perceptron/perceptron-mk1", + "name": "Perceptron: Perceptron Mk1", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "temperature": false, - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "temperature": true, + "release_date": "2026-05-12", + "last_updated": "2026-05-16", "modalities": { "input": [ + "image", "text", - "image" + "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 32768, + "output": 8192 }, "cost": { - "input": 5, - "output": 32, - "cache_read": 1.25 + "input": 0.15, + "output": 1.5 } }, - "grok-4-20-reasoning": { - "id": "grok-4-20-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "alfredpros/codellama-7b-instruct-solidity": { + "id": "alfredpros/codellama-7b-instruct-solidity", + "name": "AlfredPros: CodeLLaMa 7B Instruct Solidity", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "release_date": "2025-04-14", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -129942,48 +122909,44 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 8192 + "context": 4096, + "output": 4096 }, - "status": "beta", "cost": { - "input": 2, - "output": 6 + "input": 0.8, + "output": 1.2 } }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Google: Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-05-07", + "last_updated": "2026-05-16", "modalities": { "input": [ - "text", + "audio", "image", - "pdf" + "pdf", + "text", + "video" ], "output": [ "text" @@ -129991,48 +122954,27 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 0.25, + "output": 1.5, + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.08333 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "google/gemma-3n-e4b-it": { + "id": "google/gemma-3n-e4b-it", + "name": "Google: Gemma 3n 4B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text" @@ -130043,30 +122985,38 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 128000 + "context": 32768, + "output": 6554 }, "cost": { - "input": 0.58, - "output": 1.68 + "input": 0.02, + "output": 0.04 } }, - "gpt-4-turbo": { - "id": "gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Google: Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2025-03-20", + "last_updated": "2026-03-15", "modalities": { "input": [ + "audio", + "image", + "pdf", "text", - "image" + "video" ], "output": [ "text" @@ -130074,42 +123024,44 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 65536 }, - "status": "deprecated", "cost": { - "input": 10, - "output": 30 + "input": 1.25, + "output": 10, + "reasoning": 10, + "cache_read": 0.125, + "cache_write": 0.375 } }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Google: Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "temperature": true, + "release_date": "2025-07-17", + "last_updated": "2026-03-15", "modalities": { "input": [ + "audio", + "image", + "pdf", "text", - "image" + "video" ], "output": [ "text" @@ -130117,21 +123069,21 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.3, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333 } }, - "o3": { - "id": "o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Google: Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -130140,19 +123092,22 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-05-19", + "last_updated": "2026-05-27", "modalities": { "input": [ + "audio", + "image", + "pdf", "text", - "image" + "video" ], "output": [ "text" @@ -130160,44 +123115,71 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 1.5, + "output": 9, + "reasoning": 9, + "cache_read": 0.15, + "cache_write": 0.08333 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "google/gemini-2.0-flash-lite-001": { + "id": "google/gemini-2.0-flash-lite-001", + "name": "Google: Gemini 2.0 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-12-11", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "audio", + "image", + "pdf", + "text", + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 8192 + }, + "cost": { + "input": 0.075, + "output": 0.3 + } + }, + "google/gemini-2.5-flash-lite-preview-09-2025": { + "id": "google/gemini-2.5-flash-lite-preview-09-2025", + "name": "Google: Gemini 2.5 Flash Lite Preview 09-2025", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens" } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "release_date": "2025-09-25", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", + "audio", "image", - "pdf" + "pdf", + "text", + "video" ], "output": [ "text" @@ -130205,111 +123187,120 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.1, + "output": 0.4, + "reasoning": 0.4, + "cache_read": 0.01, + "cache_write": 0.083333 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "google/gemini-2.0-flash-001": { + "id": "google/gemini-2.0-flash-001", + "name": "Google: Gemini 2.0 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2024-12-11", + "last_updated": "2026-03-15", "modalities": { "input": [ + "audio", + "image", + "pdf", "text", - "image" + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" + "context": 1048576, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025, + "cache_write": 0.083333 } }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Google: Gemma 4 31B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-11", "modalities": { "input": [ + "image", "text", - "image" + "video" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.4 + } + }, + "google/lyria-3-clip-preview": { + "id": "google/lyria-3-clip-preview", + "name": "Google: Lyria 3 Clip Preview", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-03-30", + "last_updated": "2026-04-11", + "modalities": { + "input": [ + "image", + "text" ], "output": [ + "audio", "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0, + "output": 0 } }, - "grok-4.6": { - "id": "grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Google: Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -130324,15 +123315,16 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "release_date": "2026-02-26", + "last_updated": "2026-03-15", "modalities": { "input": [ + "audio", + "image", + "pdf", "text", - "image" + "video" ], "output": [ "text" @@ -130340,254 +123332,195 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 1048576, + "output": 65536 }, - "status": "beta" + "cost": { + "input": 2, + "output": 12, + "reasoning": 12 + } }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + "google/gemini-3-pro-image-preview": { + "id": "google/gemini-3-pro-image-preview", + "name": "Google: Nano Banana Pro (Gemini 3 Pro Image Preview)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "tool_call": false, + "temperature": true, + "release_date": "2025-11-20", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 65536, + "output": 32768 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 2, + "output": 12, + "reasoning": 12 } }, - "phi-4-mini": { - "id": "phi-4-mini", - "name": "Phi-4-mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, + "google/gemini-2.5-flash-image": { + "id": "google/gemini-2.5-flash-image", + "name": "Google: Nano Banana (Gemini 2.5 Flash Image)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2025-10-08", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", "text" ], "output": [ + "image", "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 0.3, + "output": 2.5 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek-V4-Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Google: Gemini 2.5 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2025-06-17", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text" + "audio", + "image", + "pdf", + "text", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" + "context": 1048576, + "output": 65535 }, "cost": { - "input": 0.19, - "output": 0.51 + "input": 0.1, + "output": 0.4, + "reasoning": 0.4, + "cache_read": 0.01, + "cache_write": 0.083333 } }, - "gpt-6-sol": { - "id": "gpt-6-sol", - "name": "GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", + "google/gemini-3.1-flash-image-preview": { + "id": "google/gemini-3.1-flash-image-preview", + "name": "Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "tool_call": false, + "temperature": true, + "release_date": "2026-02-26", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 0.5, + "output": 3 } - } - } - }, - "freemodel": { - "id": "freemodel", - "env": [ - "FREEMODEL_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://cc.freemodel.dev/v1", - "name": "FreeModel", - "doc": "https://freemodel.dev", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + }, + "google/gemini-2.5-pro-preview-05-06": { + "id": "google/gemini-2.5-pro-preview-05-06", + "name": "Google: Gemini 2.5 Pro Preview 05-06", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2025-05-06", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", + "audio", "image", - "pdf" + "pdf", + "text", + "video" ], "output": [ "text" @@ -130595,33 +123528,27 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.freemodel.dev/v1" + "context": 1048576, + "output": 65535 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 2.5 + "input": 1.25, + "output": 10, + "reasoning": 10, + "cache_read": 0.125, + "cache_write": 0.375 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Google: Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -130630,16 +123557,16 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2026-02-19", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", + "audio", "image", - "pdf" + "pdf", + "text", + "video" ], "output": [ "text" @@ -130647,96 +123574,71 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.freemodel.dev/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175, - "cache_write": 1.75 + "input": 2, + "output": 12, + "reasoning": 12 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Google: Gemma 4 26B A4B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "release_date": "2026-04-03", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", "image", - "pdf" + "text", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.12, + "output": 0.4 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "google/gemini-2.5-pro-preview": { + "id": "google/gemini-2.5-pro-preview", + "name": "Google: Gemini 2.5 Pro Preview 06-05", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2025-06-05", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "audio", + "image", + "pdf", + "text" ], "output": [ "text" @@ -130744,26 +123646,21 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.freemodel.dev/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075, - "cache_write": 0.75 + "input": 1.25, + "output": 10, + "reasoning": 10, + "cache_read": 0.125, + "cache_write": 0.375 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Google: Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -130773,21 +123670,21 @@ "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "release_date": "2025-12-17", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", + "audio", "image", - "pdf" + "pdf", + "text", + "video" ], "output": [ "text" @@ -130795,276 +123692,174 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.5, + "output": 3, + "reasoning": 3, + "cache_read": 0.05, + "cache_write": 0.083333 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "google/gemma-3-12b-it": { + "id": "google/gemma-3-12b-it", + "name": "Google: Gemma 3 12B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-03-13", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.freemodel.dev/v1" + "context": 131072, + "output": 131072 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 5 + "input": 0.04, + "output": 0.13, + "cache_read": 0.015 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "google/gemma-3-4b-it": { + "id": "google/gemma-3-4b-it", + "name": "Google: Gemma 3 4B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2025-03-13", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 19200 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.04, + "output": 0.08 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Google: Gemma 3 27B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2025-03-12", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 65536 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.03, + "output": 0.11, + "cache_read": 0.02 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "google/lyria-3-pro-preview": { + "id": "google/lyria-3-pro-preview", + "name": "Google: Lyria 3 Pro Preview", + "description": "Speech generation model for controllable voice, narration, and audio delivery", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-03-30", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ + "audio", "text" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "google/gemma-2-27b-it": { + "id": "google/gemma-2-27b-it", + "name": "Google: Gemma 2 27B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-06-24", + "last_updated": "2024-06-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 8192, + "output": 2048 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.65, + "output": 0.65 } - } - } - }, - "azure-cognitive-services": { - "id": "azure-cognitive-services", - "env": [ - "AZURE_COGNITIVE_SERVICES_RESOURCE_NAME", - "AZURE_COGNITIVE_SERVICES_API_KEY" - ], - "npm": "@ai-sdk/azure", - "name": "Azure Cognitive Services", - "doc": "https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + }, + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Google: Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -131073,16 +123868,16 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2026-03-03", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", + "audio", "image", - "pdf" + "pdf", + "text", + "video" ], "output": [ "text" @@ -131090,56 +123885,63 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } + "input": 0.25, + "output": 1.5, + "reasoning": 1.5 + } + }, + "liquid/lfm-2-24b-a2b": { + "id": "liquid/lfm-2-24b-a2b", + "name": "LiquidAI: LFM2-24B-A2B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-02-26", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.03, + "output": 0.12 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "x-ai/grok-4.20": { + "id": "x-ai/grok-4.20", + "name": "xAI: Grok 4.20", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2026-03-31", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", "image", - "pdf" + "pdf", + "text" ], "output": [ "text" @@ -131147,47 +123949,41 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "xAI: Grok 4.3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -131195,41 +123991,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "output": 4096 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 Nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "x-ai/grok-4.20-multi-agent": { + "id": "x-ai/grok-4.20-multi-agent", + "name": "xAI: Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -131237,17 +124020,46 @@ ] } ], + "tool_call": false, + "temperature": true, + "release_date": "2026-03-31", + "last_updated": "2026-04-11", + "modalities": { + "input": [ + "image", + "pdf", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 2000000, + "output": 2000000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.2 + } + }, + "x-ai/grok-build-0.1": { + "id": "x-ai/grok-build-0.1", + "name": "xAI: Grok Build 0.1", + "description": "Grok coding model for agentic engineering, edits, and codebase workflows", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-05-20", + "last_updated": "2026-05-27", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" @@ -131255,20 +124067,19 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 1, + "output": 2, + "cache_read": 0.2 } }, - "gpt-chat-latest": { - "id": "gpt-chat-latest", - "name": "GPT Chat Latest", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "~google/gemini-pro-latest": { + "id": "~google/gemini-pro-latest", + "name": "Google: Gemini Pro Latest", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -131282,15 +124093,15 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-05-05", - "last_updated": "2026-05-28", + "temperature": true, + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -131299,47 +124110,43 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 111616, - "output": 16384 + "context": 1048576, + "output": 65536 }, - "status": "beta", "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 2, + "output": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "~google/gemini-flash-latest": { + "id": "~google/gemini-flash-latest", + "name": "Google: Gemini Flash Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-08-01", + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -131348,43 +124155,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.08333333333333334 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "microsoft/phi-4-mini-instruct": { + "id": "microsoft/phi-4-mini-instruct", + "name": "Microsoft: Phi 4 Mini Instruct", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2025-10-17", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -131392,242 +124186,146 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4 + "input": 0.08, + "output": 0.35, + "cache_read": 0.08 } }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, + "microsoft/phi-4": { + "id": "microsoft/phi-4", + "name": "Microsoft: Phi 4", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 16384, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.06, + "output": 0.14 } }, - "claude-opus-5-5": { - "id": "claude-opus-5-5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "microsoft/wizardlm-2-8x22b": { + "id": "microsoft/wizardlm-2-8x22b", + "name": "WizardLM-2 8x22B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-04-24", + "last_updated": "2024-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 65535, + "output": 8000 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 + "input": 0.62, + "output": 0.62 } }, - "claude-fable-5-1": { - "id": "claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, + "poolside/laguna-xs.2:free": { + "id": "poolside/laguna-xs.2:free", + "name": "Poolside: Laguna XS.2 (free)", + "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", + "family": "laguna", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "structured_output": false, + "temperature": true, + "release_date": "2026-04-28", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 262144, + "output": 32768 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 0, + "output": 0 } }, - "claude-mythos-5": { - "id": "claude-mythos-5", - "name": "Claude Mythos 5", - "description": "Restricted Claude model for advanced cybersecurity and biology research workflows", - "family": "claude-mythos", - "attachment": true, + "poolside/laguna-m.1:free": { + "id": "poolside/laguna-m.1:free", + "name": "Poolside: Laguna M.1 (free)", + "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", + "family": "laguna", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "structured_output": false, + "temperature": true, + "release_date": "2026-04-28", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "status": "beta", - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 262144, + "output": 32768 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0, + "output": 0 } }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, + "writer/palmyra-x5": { + "id": "writer/palmyra-x5", + "name": "Writer: Palmyra X5", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -131635,43 +124333,39 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 1040000, + "output": 8192 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.6, + "output": 6 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "Z.ai: GLM 4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-12-22", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -131679,307 +124373,253 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models", - "shape": "completions" + "context": 202752, + "output": 65535 }, "cost": { - "input": 0.6, - "output": 3 + "input": 0.38, + "output": 1.98, + "cache_read": 0.2 } }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", + "z-ai/glm-4.5v": { + "id": "z-ai/glm-4.5v", + "name": "Z.ai: GLM 4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "temperature": true, + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 65536, + "output": 16384 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.6, + "output": 1.8, + "cache_read": 0.11 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, + "z-ai/glm-4.5": { + "id": "z-ai/glm-4.5", + "name": "Z.ai: GLM 4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "release_date": "2025-07-28", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "status": "beta", - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 131072, + "output": 98304 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.6, + "output": 2.2, + "cache_read": 0.175 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 Mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "Z.ai: GLM 5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", "high", "xhigh" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 1.26, + "output": 3.96 } }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "Z.ai: GLM 4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2025-09-30", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 204800, + "output": 204800 }, "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1, - "cache_write": 1.25, - "tiers": [ - { - "input": 2, - "output": 9, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 272000 - } - } + "input": 0.39, + "output": 1.9, + "cache_read": 0.175 + } + }, + "z-ai/glm-4-32b": { + "id": "z-ai/glm-4-32b", + "name": "Z.ai: GLM 4 32B ", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-07-25", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 2, - "output": 9, - "cache_read": 0.2, - "cache_write": 2.5 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.1 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "z-ai/glm-4.6v": { + "id": "z-ai/glm-4.6v", + "name": "Z.ai: GLM 4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2025-12-31", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "release_date": "2025-09-30", + "last_updated": "2026-01-10", "modalities": { "input": [ - "text", "image", - "pdf" + "text", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 131072, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 0.3, + "output": 0.9 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "z-ai/glm-5v-turbo": { + "id": "z-ai/glm-5v-turbo", + "name": "Z.ai: GLM 5V Turbo", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -131989,267 +124629,214 @@ { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", "image", - "pdf" + "text", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 202752, + "output": 131072 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "z-ai/glm-4.5-air": { + "id": "z-ai/glm-4.5-air", + "name": "Z.ai: GLM 4.5 Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", "high", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-07-17", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 131072, + "output": 98304 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 0.13, + "output": 0.85, + "cache_read": 0.025 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, + "z-ai/glm-4.7-flash": { + "id": "z-ai/glm-4.7-flash", + "name": "Z.ai: GLM 4.7 Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", "high", - "max" + "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 202752, + "output": 40551 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.06, + "output": 0.4, + "cache_read": 0.01 } }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "Z.ai: GLM 5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25 - } + "input": 0.72, + "output": 2.3 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "Z.ai: GLM 5 Turbo", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "release_date": "2026-03-15", + "last_updated": "2026-04-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.24 + } + }, + "openai/gpt-4o-mini-2024-07-18": { + "id": "openai/gpt-4o-mini-2024-07-18", + "name": "OpenAI: GPT-4o-mini (2024-07-18)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-07-18", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "pdf", + "text" ], "output": [ "text" @@ -132257,51 +124844,37 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + "context": 128000, + "output": 16384 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.15, + "output": 0.6 } }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, + "openai/gpt-oss-safeguard-20b": { + "id": "openai/gpt-oss-safeguard-20b", + "name": "OpenAI: gpt-oss-safeguard-20b", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -132309,44 +124882,25 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.075, + "output": 0.3, + "cache_read": 0.037 } }, - "phi-4-mini": { - "id": "phi-4-mini", - "name": "Phi-4-mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", + "openai/gpt-3.5-turbo-instruct": { + "id": "openai/gpt-3.5-turbo-instruct", + "name": "OpenAI: GPT-3.5 Turbo Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2023-03-01", + "last_updated": "2023-09-21", "modalities": { "input": [ "text" @@ -132355,43 +124909,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 4095, "output": 4096 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 1.5, + "output": 2 } }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "openai/gpt-5.2-chat": { + "id": "openai/gpt-5.2-chat", + "name": "OpenAI: GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2025-12-11", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -132399,21 +124941,19 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "o3": { - "id": "o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "openai/o3": { + "id": "openai/o3", + "name": "OpenAI: o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -132428,13 +124968,13 @@ ], "tool_call": true, "temperature": false, - "knowledge": "2024-05", "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -132451,33 +124991,34 @@ "cache_read": 0.5 } }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "openai/o4-mini-high": { + "id": "openai/o4-mini-high", + "name": "OpenAI: o4 Mini High", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "release_date": "2025-04-17", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -132485,123 +125026,94 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 1.1, + "output": 4.4 } }, - "gpt-4-turbo": { - "id": "gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "openai/gpt-audio": { + "id": "openai/gpt-audio", + "name": "OpenAI: GPT Audio", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2026-01-20", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "audio", + "text" ], "output": [ + "audio", "text" ] }, "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 16384 }, - "status": "deprecated", "cost": { - "input": 10, - "output": 30 + "input": 2.5, + "output": 10 } }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "OpenAI: GPT-5.2 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "temperature": false, + "release_date": "2025-12-11", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 400000, "output": 128000 }, "cost": { - "input": 0.58, - "output": 1.68 + "input": 21, + "output": 168 } }, - "llama-4-scout-17b-16e-instruct": { - "id": "llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout 17B 16E Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.2, - "output": 0.78 - } - }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "openai/gpt-4o-mini-search-preview": { + "id": "openai/gpt-4o-mini-search-preview", + "name": "OpenAI: GPT-4o-mini Search Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "temperature": false, + "release_date": "2025-01", + "last_updated": "2025-01", "modalities": { "input": [ "text" @@ -132610,33 +125122,42 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 128000, + "output": 16384 }, - "status": "deprecated", "cost": { - "input": 1.35, - "output": 5.4 + "input": 0.15, + "output": 0.6 } }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "OpenAI: GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "temperature": false, + "release_date": "2025-08-07", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -132644,30 +125165,29 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 400000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "gpt-3.5-turbo-1106": { - "id": "gpt-3.5-turbo-1106", - "name": "GPT-3.5 Turbo 1106", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, + "openai/gpt-5-chat": { + "id": "openai/gpt-5-chat", + "name": "OpenAI: GPT-5 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-11-06", - "last_updated": "2023-11-06", + "temperature": false, + "release_date": "2025-08-07", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ @@ -132676,31 +125196,28 @@ }, "open_weights": false, "limit": { - "context": 16384, + "context": 128000, "output": 16384 }, - "status": "deprecated", "cost": { - "input": 1, - "output": 2 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "gpt-4-turbo-vision": { - "id": "gpt-4-turbo-vision", - "name": "GPT-4 Turbo Vision", + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "OpenAI: GPT-3.5 Turbo", "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -132708,31 +125225,37 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 16385, "output": 4096 }, - "status": "deprecated", "cost": { - "input": 10, - "output": 30 + "input": 0.5, + "output": 1.5 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "OpenAI: GPT-5 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "temperature": false, + "release_date": "2025-10-06", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -132740,63 +125263,58 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 400000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 15, + "output": 120 } }, - "phi-4-mini-reasoning": { - "id": "phi-4-mini-reasoning", - "name": "Phi-4-mini-reasoning", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "OpenAI: GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2024-05-13", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 16384 }, "cost": { - "input": 0.075, - "output": 0.3 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "mistral-small-2503": { - "id": "mistral-small-2503", - "name": "Mistral Small 3.1", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, + "openai/gpt-4": { + "id": "openai/gpt-4", + "name": "OpenAI: GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2025-03-01", - "last_updated": "2025-03-01", + "release_date": "2023-03-14", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -132804,26 +125322,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 8191, + "output": 4096 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 30, + "output": 60 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "OpenAI: o4 Mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -132832,16 +125351,14 @@ } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2025-04-16", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "pdf", + "text" ], "output": [ "text" @@ -132849,44 +125366,25 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "phi-4": { - "id": "phi-4", - "name": "Phi-4", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "phi", + "openai/gpt-3.5-turbo-16k": { + "id": "openai/gpt-3.5-turbo-16k", + "name": "OpenAI: GPT-3.5 Turbo 16k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2023-08-28", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -132895,76 +125393,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 16385, "output": 4096 }, "cost": { - "input": 0.125, - "output": 0.5 + "input": 3, + "output": 4 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "openai/o3-pro": { + "id": "openai/o3-pro", + "name": "OpenAI: o3 Pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2025-04-16", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" - ], - "output": [ + "image", + "pdf", "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.125 - } - }, - "model-router": { - "id": "model-router", - "name": "Model Router", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "model-router", - "attachment": true, - "reasoning": false, - "tool_call": true, - "release_date": "2025-05-19", - "last_updated": "2025-11-18", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" @@ -132973,89 +125436,59 @@ "open_weights": false, "limit": { "context": 200000, - "output": 16384 + "output": 100000 }, "cost": { - "input": 0.14, - "output": 0 + "input": 20, + "output": 80 } }, - "cohere-embed-v3-english": { - "id": "cohere-embed-v3-english", - "name": "Embed v3 English", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", - "attachment": false, + "openai/gpt-5.1-chat": { + "id": "openai/gpt-5.1-chat", + "name": "OpenAI: GPT-5.1 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2023-11-07", - "last_updated": "2023-11-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 512, - "output": 1024 - }, - "cost": { - "input": 0.1, - "output": 0 - } - }, - "cohere-command-a": { - "id": "cohere-command-a", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", - "attachment": false, - "reasoning": true, - "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "temperature": false, + "release_date": "2025-11-13", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 128000, + "output": 16384 }, "cost": { - "input": 2.5, - "output": 10 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", + "openai/gpt-4o-2024-05-13": { + "id": "openai/gpt-4o-2024-05-13", + "name": "OpenAI: GPT-4o (2024-05-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2024-05-13", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -133064,27 +125497,23 @@ "open_weights": false, "limit": { "context": 128000, - "output": 16384 + "output": 4096 }, - "status": "deprecated", "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 5, + "output": 15 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "openai/gpt-4-0314": { + "id": "openai/gpt-4-0314", + "name": "OpenAI: GPT-4 (older v0314)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "2023-05-28", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -133093,21 +125522,20 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 8191, + "output": 4096 }, "cost": { - "input": 0.71, - "output": 0.71 + "input": 30, + "output": 60 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "OpenAI: GPT-5.4 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -133117,54 +125545,49 @@ "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "release_date": "2026-03-17", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", "image", - "audio" + "pdf", + "text" ], "output": [ - "text", - "image", - "audio" + "text" ] }, "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "codestral-2501": { - "id": "codestral-2501", - "name": "Codestral 25.01", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "codestral", - "attachment": false, + "openai/gpt-5.3-chat": { + "id": "openai/gpt-5.3-chat", + "name": "OpenAI: GPT-5.3 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-03", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-03-04", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ @@ -133172,55 +125595,25 @@ ] }, "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.3, - "output": 0.9 - } - }, - "cohere-embed-v-4-0": { - "id": "cohere-embed-v-4-0", - "name": "Embed v4", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, "limit": { "context": 128000, - "output": 1536 + "output": 16384 }, "cost": { - "input": 0.12, - "output": 0 + "input": 1.75, + "output": 14 } }, - "text-embedding-3-large": { - "id": "text-embedding-3-large", - "name": "text-embedding-3-large", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "openai/gpt-3.5-turbo-0613": { + "id": "openai/gpt-3.5-turbo-0613", + "name": "OpenAI: GPT-3.5 Turbo (older v0613)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, - "tool_call": false, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "tool_call": true, + "temperature": true, + "release_date": "2023-06-13", + "last_updated": "2023-06-13", "modalities": { "input": [ "text" @@ -133231,60 +125624,70 @@ }, "open_weights": false, "limit": { - "context": 8191, - "output": 3072 + "context": 4095, + "output": 4096 }, "cost": { - "input": 0.13, - "output": 0 + "input": 1, + "output": 2 } }, - "gpt-3.5-turbo-0125": { - "id": "gpt-3.5-turbo-0125", - "name": "GPT-3.5 Turbo 0125", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, + "openai/gpt-5-image-mini": { + "id": "openai/gpt-5-image-mini", + "name": "OpenAI: GPT-5 Image Mini", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "knowledge": "2021-08", - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "release_date": "2025-10-16", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 400000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 0.5, - "output": 1.5 + "input": 2.5, + "output": 2 } }, - "gpt-3.5-turbo-instruct": { - "id": "gpt-3.5-turbo-instruct", - "name": "GPT-3.5 Turbo Instruct", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2021-08", - "release_date": "2023-09-21", - "last_updated": "2023-09-21", + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "OpenAI: GPT-5.1-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -133292,36 +125695,36 @@ }, "open_weights": false, "limit": { - "context": 4096, - "output": 4096 + "context": 400000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 1.5, - "output": 2 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "OpenAI: GPT-5.1-Codex-Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high" + "low", + "medium", + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -133334,68 +125737,58 @@ "open_weights": false, "limit": { "context": 400000, - "output": 272000 + "output": 128000 }, "cost": { - "input": 15, - "output": 120 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "phi-4-reasoning-plus": { - "id": "phi-4-reasoning-plus", - "name": "Phi-4-reasoning-plus", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "openai/gpt-4o-2024-08-06": { + "id": "openai/gpt-4o-2024-08-06", + "name": "OpenAI: GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, + "reasoning": false, + "tool_call": true, "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "release_date": "2024-08-06", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 4096 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.125, - "output": 0.5 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "codex-mini": { - "id": "codex-mini", - "name": "Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex-mini", + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "OpenAI: o3 Mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "temperature": false, - "knowledge": "2024-04", - "release_date": "2025-05-16", - "last_updated": "2025-05-16", + "release_date": "2024-12-20", + "last_updated": "2026-03-15", "modalities": { "input": [ + "pdf", "text" ], "output": [ @@ -133407,39 +125800,39 @@ "context": 200000, "output": 100000 }, - "status": "deprecated", "cost": { - "input": 1.5, - "output": 6, - "cache_read": 0.375 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "o1": { - "id": "o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": false, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "OpenAI: GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "release_date": "2025-12-11", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -133447,136 +125840,153 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 400000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "cohere-embed-v3-multilingual": { - "id": "cohere-embed-v3-multilingual", - "name": "Embed v3 Multilingual", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2023-11-07", - "last_updated": "2023-11-07", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "OpenAI: GPT-5.3-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "release_date": "2026-02-25", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512, - "output": 1024 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0 + "input": 1.75, + "output": 14 } }, - "mistral-medium-2505": { - "id": "mistral-medium-2505", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, + "openai/gpt-audio-mini": { + "id": "openai/gpt-audio-mini", + "name": "OpenAI: GPT Audio Mini", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", + "release_date": "2026-01-20", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "audio", + "text" ], "output": [ + "audio", "text" ] }, "open_weights": false, "limit": { "context": 128000, - "output": 128000 + "output": 16384 }, "cost": { - "input": 0.4, - "output": 2 + "input": 0.6, + "output": 2.4 } }, - "phi-4-multimodal": { - "id": "phi-4-multimodal", - "name": "Phi-4-multimodal", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "phi", + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "OpenAI: GPT-5.1-Codex-Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", "image", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 400000, + "output": 100000 }, "cost": { - "input": 0.08, - "output": 0.32, - "input_audio": 4 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "openai/o4-mini-deep-research": { + "id": "openai/o4-mini-deep-research", + "name": "OpenAI: o4 Mini Deep Research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2024-06-26", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -133584,32 +125994,30 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "OpenAI: GPT-4.1 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-04", "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -133620,71 +126028,64 @@ "context": 1047576, "output": 32768 }, - "status": "deprecated", "cost": { "input": 0.1, "output": 0.4, "cache_read": 0.025 } }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "OpenAI: gpt-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 26215 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.039, + "output": 0.19 } }, - "text-embedding-ada-002": { - "id": "text-embedding-ada-002", - "name": "text-embedding-ada-002", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, + "openai/gpt-4o-2024-11-20": { + "id": "openai/gpt-4o-2024-11-20", + "name": "OpenAI: GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, "reasoning": false, - "tool_call": false, - "release_date": "2022-12-15", - "last_updated": "2022-12-15", + "tool_call": true, + "temperature": true, + "release_date": "2024-11-20", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ @@ -133693,20 +126094,51 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": false, + "openai/o1": { + "id": "openai/o1", + "name": "OpenAI: o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": false, + "release_date": "2024-12-05", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "image", + "pdf", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 15, + "output": 60, + "cache_read": 7.5 + } + }, + "openai/o1-pro": { + "id": "openai/o1-pro", + "name": "OpenAI: o1-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -133718,13 +126150,14 @@ ] } ], - "tool_call": true, + "tool_call": false, "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "release_date": "2025-03-19", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ @@ -133736,70 +126169,102 @@ "context": 200000, "output": 100000 }, - "status": "deprecated", "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 150, + "output": 600 } }, - "llama-4-maverick-17b-128e-instruct-fp8": { - "id": "llama-4-maverick-17b-128e-instruct-fp8", - "name": "Llama 4 Maverick 17B 128E Instruct FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", + "openai/gpt-chat-latest": { + "id": "openai/gpt-chat-latest", + "name": "OpenAI: GPT Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": false, + "release_date": "2026-05-05", + "last_updated": "2026-05-07", + "modalities": { + "input": [ + "image", + "pdf", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + } + }, + "openai/gpt-5-image": { + "id": "openai/gpt-5-image", + "name": "OpenAI: GPT-5 Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "release_date": "2025-10-14", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ + "image", "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 16384 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1 + "input": 10, + "output": 10 } }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "OpenAI: GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "release_date": "2026-03-06", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -133807,43 +126272,41 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1050000, + "output": 128000 }, - "status": "deprecated", "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 2.5, + "output": 15 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "OpenAI: GPT-5.4 Mini", "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2026-03-17", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -133852,26 +126315,153 @@ "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.03 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "OpenAI: GPT-4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-04-14", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "image", + "pdf", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 + } + }, + "openai/gpt-4o-audio-preview": { + "id": "openai/gpt-4o-audio-preview", + "name": "OpenAI: GPT-4o Audio", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-08-15", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "audio", + "text" + ], + "output": [ + "audio", + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "openai/o3-deep-research": { + "id": "openai/o3-deep-research", + "name": "OpenAI: o3 Deep Research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2024-06-26", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "image", + "pdf", + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 10, + "output": 40, + "cache_read": 2.5 + } + }, + "openai/gpt-4-turbo-preview": { + "id": "openai/gpt-4-turbo-preview", + "name": "OpenAI: GPT-4 Turbo Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-01-25", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 10, + "output": 30 + } + }, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "OpenAI: GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -133880,13 +126470,13 @@ ], "tool_call": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "release_date": "2025-08-07", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -133895,57 +126485,55 @@ "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "ministral-3b": { - "id": "ministral-3b", - "name": "Ministral 3B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "OpenAI: GPT-4.1 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", + "release_date": "2025-04-14", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 0.04, - "output": 0.04 + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "OpenAI: GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "release_date": "2023-09-13", + "last_updated": "2024-04-09", "modalities": { "input": [ "text", @@ -133958,26 +126546,24 @@ "open_weights": false, "limit": { "context": 128000, - "output": 16384 + "output": 4096 }, - "status": "deprecated", "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 10, + "output": 30 } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt-codex", - "attachment": false, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "OpenAI: GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -133985,47 +126571,40 @@ } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "release_date": "2025-08-07", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "audio" + "pdf", + "text" ], "output": [ - "text", - "image", - "audio" + "text" ] }, "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "OpenAI: GPT-5.4 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", "high", "xhigh" @@ -134033,15 +126612,13 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-14", - "last_updated": "2026-01-14", + "release_date": "2026-03-06", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -134049,28 +126626,27 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, "output": 128000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 30, + "output": 180 } }, - "text-embedding-3-small": { - "id": "text-embedding-3-small", - "name": "text-embedding-3-small", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, + "openai/o3-mini-high": { + "id": "openai/o3-mini-high", + "name": "OpenAI: o3 Mini High", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "attachment": true, "reasoning": false, - "tool_call": false, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "tool_call": true, + "temperature": false, + "release_date": "2025-01-31", + "last_updated": "2026-03-15", "modalities": { "input": [ + "pdf", "text" ], "output": [ @@ -134079,58 +126655,57 @@ }, "open_weights": false, "limit": { - "context": 8191, - "output": 1536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.02, - "output": 0 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "phi-4-reasoning": { - "id": "phi-4-reasoning", - "name": "Phi-4-reasoning", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "phi", - "attachment": false, + "openai/gpt-5.4-image-2": { + "id": "openai/gpt-5.4-image-2", + "name": "OpenAI: GPT-5.4 Image 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": false, - "temperature": true, - "knowledge": "2023-10", - "release_date": "2024-12-11", - "last_updated": "2024-12-11", + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ + "image", "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 4096 + "context": 272000, + "output": 128000 }, "cost": { - "input": 0.125, - "output": 0.5 + "input": 8, + "output": 15, + "cache_read": 2 } }, - "deepseek-v3.2-speciale": { - "id": "deepseek-v3.2-speciale", - "name": "DeepSeek-V3.2-Speciale", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "openai/gpt-4o-search-preview": { + "id": "openai/gpt-4o-search-preview", + "name": "OpenAI: GPT-4o Search Preview", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2025-03-13", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -134139,93 +126714,91 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 128000 + "output": 16384 }, "cost": { - "input": 0.58, - "output": 1.68 + "input": 2.5, + "output": 10 } - } - } - }, - "pendra": { - "id": "pendra", - "env": [ - "PENDRA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.pendra.ai/api/v1", - "name": "Pendra", - "doc": "https://pendra.ai/docs/integrations/opencode", - "models": { - "llama3.3:70b": { - "id": "llama3.3:70b", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, + }, + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "OpenAI: GPT-5.5 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "temperature": false, + "release_date": "2026-04-24", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 30, + "output": 180 } }, - "qwen3-coder:30b": { - "id": "qwen3-coder:30b", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", - "attachment": false, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "OpenAI: GPT-4o-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2024-07-18", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "gpt-oss:120b": { - "id": "gpt-oss:120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "OpenAI: gpt-oss-20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -134239,7 +126812,6 @@ } ], "tool_call": true, - "structured_output": true, "temperature": true, "release_date": "2025-08-05", "last_updated": "2025-08-05", @@ -134254,218 +126826,217 @@ "open_weights": true, "limit": { "context": 131072, - "output": 32768 + "output": 26215 }, "cost": { - "input": 0, - "output": 0 + "input": 0.03, + "output": 0.14 } }, - "qwen3.6:27b": { - "id": "qwen3.6:27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4-1106-preview": { + "id": "openai/gpt-4-1106-preview", + "name": "OpenAI: GPT-4 Turbo (older v1106)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2023-11-06", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0, - "output": 0 + "input": 10, + "output": 30 } }, - "glm-4.7-flash": { - "id": "glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "OpenAI: GPT-5 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "temperature": false, + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "OpenAI: GPT-5.2-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } - } - } - }, - "moark": { - "id": "moark", - "env": [ - "MOARK_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://moark.com/v1", - "name": "Moark", - "doc": "https://moark.com/docs/openapi/v1#tag/%E6%96%87%E6%9C%AC%E7%94%9F%E6%88%90", - "models": { - "GLM-4.7": { - "id": "GLM-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + }, + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "OpenAI: GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "temperature": false, + "release_date": "2025-11-13", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 3.5, - "output": 14 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "OpenAI: GPT-5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "temperature": false, + "release_date": "2026-04-24", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 2.1, - "output": 8.4, - "cache_read": 2.1, - "cache_write": 8.4 + "input": 5, + "output": 30, + "cache_read": 0.5 } - } - } - }, - "atomic-chat": { - "id": "atomic-chat", - "env": [ - "ATOMIC_CHAT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "http://127.0.0.1:1337/v1", - "name": "Atomic Chat", - "doc": "https://atomic.chat", - "models": { - "gemma-4-E4B-it-MLX-4bit": { - "id": "gemma-4-E4B-it-MLX-4bit", - "name": "Gemma 4 E4B Instruct (MLX 4-bit)", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + }, + "thedrummer/cydonia-24b-v4.1": { + "id": "thedrummer/cydonia-24b-v4.1", + "name": "TheDrummer: Cydonia 24B V4.1", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-09-27", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -134476,25 +127047,24 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 8192 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 0.5 } }, - "Meta-Llama-3_1-8B-Instruct-GGUF": { - "id": "Meta-Llama-3_1-8B-Instruct-GGUF", - "name": "Meta Llama 3.1 8B Instruct (GGUF)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "thedrummer/skyfall-36b-v2": { + "id": "thedrummer/skyfall-36b-v2", + "name": "TheDrummer: Skyfall 36B V2", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2025-03-11", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -134505,29 +127075,27 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 4096 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.55, + "output": 0.8 } }, - "Qwen3_5-9B-Q4_K_M": { - "id": "Qwen3_5-9B-Q4_K_M", - "name": "Qwen 3.5 9B (Q4_K_M)", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "thedrummer/unslopnemo-12b": { + "id": "thedrummer/unslopnemo-12b", + "name": "TheDrummer: UnslopNemo 12B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-05", - "last_updated": "2026-04-04", + "release_date": "2024-11-09", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -134536,24 +127104,23 @@ "open_weights": true, "limit": { "context": 32768, - "output": 8192 + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 0.4 } }, - "gemma-4-E4B-it-IQ4_XS": { - "id": "gemma-4-E4B-it-IQ4_XS", - "name": "Gemma 4 E4B Instruct (IQ4_XS)", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "thedrummer/rocinante-12b": { + "id": "thedrummer/rocinante-12b", + "name": "TheDrummer: Rocinante 12B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2024-09-30", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -134565,255 +127132,207 @@ "open_weights": true, "limit": { "context": 32768, - "output": 8192 + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.17, + "output": 0.43 } }, - "Qwen3_5-9B-MLX-4bit": { - "id": "Qwen3_5-9B-MLX-4bit", - "name": "Qwen 3.5 9B (MLX 4-bit)", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "bytedance/ui-tars-1.5-7b": { + "id": "bytedance/ui-tars-1.5-7b", + "name": "ByteDance: UI-TARS 7B ", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2026-03-05", - "last_updated": "2026-04-04", + "release_date": "2025-07-23", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 128000, + "output": 2048 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1, + "output": 0.2 } - } - } - }, - "qihang-ai": { - "id": "qihang-ai", - "env": [ - "QIHANG_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.qhaigc.net/v1", - "name": "QiHang", - "doc": "https://www.qhaigc.net/docs", - "models": { - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + }, + "rekaai/reka-flash-3": { + "id": "rekaai/reka-flash-3", + "name": "Reka Flash 3", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "tool_call": false, + "temperature": true, + "release_date": "2025-03-12", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 1.14 + "input": 0.1, + "output": 0.2 } }, - "gemini-3-pro-preview": { - "id": "gemini-3-pro-preview", - "name": "Gemini 3 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "rekaai/reka-edge": { + "id": "rekaai/reka-edge", + "name": "Reka Edge", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "release_date": "2026-03-20", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", "image", - "audio", + "text", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65000 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0.57, - "output": 3.43 + "input": 0.1, + "output": 0.1 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5-Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistralai/mistral-large-2407": { + "id": "mistralai/mistral-large-2407", + "name": "Mistral Large 2407", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "release_date": "2024-11-19", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.04, - "output": 0.29 + "input": 2, + "output": 6 } }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "mistralai/mistral-small-3.2-24b-instruct": { + "id": "mistralai/mistral-small-3.2-24b-instruct", + "name": "Mistral: Mistral Small 3.2 24B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.71, - "output": 3.57 + "input": 0.06, + "output": 0.18, + "cache_read": 0.03 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistralai/mistral-nemo": { + "id": "mistralai/mistral-nemo", + "name": "Mistral: Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2024-07-01", + "last_updated": "2024-07-30", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.09, - "output": 0.71, - "tiers": [ - { - "input": 0.09, - "output": 0.71, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 0.09, - "output": 0.71 - } + "input": 0.02, + "output": 0.04 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "mistralai/mistral-medium-3-5": { + "id": "mistralai/mistral-medium-3-5", + "name": "Mistral: Mistral Medium 3.5", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-04-30", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" @@ -134821,269 +127340,171 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.25, - "output": 2 + "input": 1.5, + "output": 7.5 } }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "mistralai/ministral-8b-2512": { + "id": "mistralai/ministral-8b-2512", + "name": "Mistral: Ministral 3 8B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2025-12-02", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.43, - "output": 2.14 + "input": 0.15, + "output": 0.15 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistralai/devstral-small": { + "id": "mistralai/devstral-small", + "name": "Mistral: Devstral Small 1.1", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2025-05-07", + "last_updated": "2025-07-10", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 26215 }, "cost": { - "input": 0.07, - "output": 0.43, - "tiers": [ - { - "input": 0.07, - "output": 0.43, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 0.07, - "output": 0.43 - } + "input": 0.1, + "output": 0.3 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "mistralai/mistral-small-3.1-24b-instruct": { + "id": "mistralai/mistral-small-3.1-24b-instruct", + "name": "Mistral: Mistral Small 3.1 24B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-10-01", - "last_updated": "2025-10-01", + "release_date": "2025-03-17", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.71 + "input": 0.35, + "output": 0.56, + "cache_read": 0.015 } - } - } - }, - "ai-router": { - "id": "ai-router", - "env": [ - "AI_ROUTER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.ai-router.dev/v1", - "name": "AI-ROUTER", - "doc": "https://ai-router.dev/openai-compatible-api-gateway/", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + }, + "mistralai/mistral-saba": { + "id": "mistralai/mistral-saba", + "name": "Mistral: Saba", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2025-02-17", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 0.2, + "output": 0.6 } }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "mistralai/mistral-large": { + "id": "mistralai/mistral-large", + "name": "Mistral Large", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2024-07-24", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 25600 }, "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 2, + "output": 6 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "mistralai/mistral-medium-3.1": { + "id": "mistralai/mistral-medium-3.1", + "name": "Mistral: Mistral Medium 3.1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2025-08-12", + "last_updated": "2025-08-12", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -135091,144 +127512,84 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 26215 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.4, + "output": 2 } }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", + "mistralai/pixtral-large-2411": { + "id": "mistralai/pixtral-large-2411", + "name": "Mistral: Pixtral Large 2411", + "description": "Mistral vision-language model for image understanding and multimodal chat", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2024-11-19", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125 + "input": 2, + "output": 6 } }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "mistralai/devstral-medium": { + "id": "mistralai/devstral-medium", + "name": "Mistral: Devstral Medium", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2025-07-10", + "last_updated": "2025-07-10", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 26215 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.4, + "output": 2 } - } - } - }, - "llmtr": { - "id": "llmtr", - "env": [ - "LLMTR_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://llmtr.com/v1", - "name": "LLMTR", - "doc": "https://llmtr.com/docs", - "models": { - "muse-glimmer-30b-tr": { - "id": "muse-glimmer-30b-tr", - "name": "Muse Glimmer 30B (TR)", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + }, + "mistralai/mistral-small-24b-instruct-2501": { + "id": "mistralai/mistral-small-24b-instruct-2501", + "name": "Mistral: Mistral Small 3", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "release_date": "2025-12-29", + "last_updated": "2026-01-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -135236,64 +127597,58 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 32768, + "output": 16384 }, "cost": { - "input": 2, - "output": 5, - "cache_read": 0.5 + "input": 0.05, + "output": 0.08 } }, - "magibu-11b-v8": { - "id": "magibu-11b-v8", - "name": "Magibu 11B v8", - "description": "Turkish-language chat model for instruction following and assistant flows", - "attachment": false, + "mistralai/ministral-3b-2512": { + "id": "mistralai/ministral-3b-2512", + "name": "Mistral: Ministral 3 3B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-06-05", - "last_updated": "2026-08-16", + "release_date": "2025-12-02", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 131072, + "output": 32768 }, "cost": { "input": 0.1, - "output": 0.5 + "output": 0.1 } }, - "gemma-4": { - "id": "gemma-4", - "name": "Gemma 4", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "mistralai/mistral-small-2603": { + "id": "mistralai/mistral-small-2603", + "name": "Mistral: Mistral Small 4", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-03-16", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" @@ -135301,30 +127656,28 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 2, - "output": 5, - "cache_read": 0.5 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "trendyol-asure-12b": { - "id": "trendyol-asure-12b", - "name": "Trendyol Asure 12B", - "description": "Turkish-language multimodal instruct model built on Gemma 3 12B for e-commerce text, chat, and image-text tasks", - "family": "gemma", - "attachment": true, + "mistralai/mistral-large-2411": { + "id": "mistralai/mistral-large-2411", + "name": "Mistral Large 2411", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-02-19", - "last_updated": "2026-02-20", + "release_date": "2024-07-24", + "last_updated": "2024-11-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -135332,100 +127685,81 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 40960 + "context": 131072, + "output": 26215 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.025 + "input": 2, + "output": 6 } }, - "medgemma-4b": { - "id": "medgemma-4b", - "name": "MedGemma 4B", - "description": "Multimodal medical-domain Gemma variant for text and image analysis", - "attachment": true, + "mistralai/mistral-7b-instruct-v0.1": { + "id": "mistralai/mistral-7b-instruct-v0.1", + "name": "Mistral: Mistral 7B Instruct v0.1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-04-26", - "last_updated": "2026-08-16", + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 2824, + "output": 565 }, - "status": "deprecated", "cost": { - "input": 3, - "output": 5 + "input": 0.11, + "output": 0.19 } }, - "qwen3-6-35b": { - "id": "qwen3-6-35b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "mistralai/ministral-14b-2512": { + "id": "mistralai/ministral-14b-2512", + "name": "Mistral: Ministral 3 14B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": false, - "structured_output": true, + "reasoning": false, + "tool_call": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2025-12-16", + "last_updated": "2025-12-16", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 16384, - "output": 16384 + "context": 262144, + "output": 52429 }, "cost": { - "input": 5, - "output": 10 + "input": 0.2, + "output": 0.2 } }, - "poolside/laguna-xs-2.1": { - "id": "poolside/laguna-xs-2.1", - "name": "Laguna XS 2.1", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", + "mistralai/devstral-2512": { + "id": "mistralai/devstral-2512", + "name": "Mistral: Devstral 2 2512", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-07-02", - "last_updated": "2026-07-02", + "release_date": "2025-09-12", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -135437,24 +127771,24 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 2, + "cache_read": 0.025 } }, - "publicai/apertus-70b-instruct": { - "id": "publicai/apertus-70b-instruct", - "name": "Apertus 70B Instruct", - "description": "Fully open 70B multilingual LLM supporting 1800+ languages with 65K context. Trained on 15T tokens of compliant open data. Apache 2.0, EU AI Act compliant.", + "mistralai/mixtral-8x22b-instruct": { + "id": "mistralai/mixtral-8x22b-instruct", + "name": "Mistral: Mixtral 8x22B Instruct", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-02", - "last_updated": "2025-09-02", + "release_date": "2024-04-17", + "last_updated": "2024-04-17", "modalities": { "input": [ "text" @@ -135466,27 +127800,56 @@ "open_weights": true, "limit": { "context": 65536, - "output": 8192 + "output": 13108 }, "cost": { - "input": 0.82, - "output": 2.92 + "input": 2, + "output": 6 } }, - "publicai/apertus-8b-instruct": { - "id": "publicai/apertus-8b-instruct", - "name": "Apertus 8B Instruct", - "description": "Fully open 8B multilingual LLM supporting 1800+ languages with 65K context. Trained on compliant open data. Apache 2.0, EU AI Act compliant.", - "attachment": false, + "mistralai/mistral-medium-3": { + "id": "mistralai/mistral-medium-3", + "name": "Mistral: Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-02", - "last_updated": "2025-09-02", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "output": 26215 + }, + "cost": { + "input": 0.4, + "output": 2 + } + }, + "mistralai/voxtral-small-24b-2507": { + "id": "mistralai/voxtral-small-24b-2507", + "name": "Mistral: Voxtral Small 24B 2507", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-07-01", + "last_updated": "2025-07-01", + "modalities": { + "input": [ + "text", + "audio" ], "output": [ "text" @@ -135494,43 +127857,28 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 8192 + "context": 32000, + "output": 6400 }, "cost": { "input": 0.1, - "output": 0.2 + "output": 0.3 } }, - "thinkingmachines/inkling-small": { - "id": "thinkingmachines/inkling-small", - "name": "Inkling Small", - "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", - "family": "ling", + "mistralai/mistral-large-2512": { + "id": "mistralai/mistral-large-2512", + "name": "Mistral: Mistral Large 3 2512", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "release_date": "2024-11-01", + "last_updated": "2025-12-16", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" @@ -135539,42 +127887,26 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "output": 52429 }, "cost": { - "input": 0.58, - "output": 1.44 + "input": 0.5, + "output": 1.5 } }, - "thinkingmachines/inkling": { - "id": "thinkingmachines/inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "mistralai/codestral-2508": { + "id": "mistralai/codestral-2508", + "name": "Mistral: Codestral 2508", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "release_date": "2025-08-01", + "last_updated": "2025-08-01", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ "text" @@ -135582,27 +127914,24 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 51200 }, "cost": { - "input": 1.87, - "output": 4.68 + "input": 0.3, + "output": 0.9 } }, - "perplexity/sonar-deep-research": { - "id": "perplexity/sonar-deep-research", - "name": "Sonar Deep Research", - "description": "Sonar search model for autonomous research and citation-backed long-form reports", - "family": "sonar", + "morph/morph-v3-fast": { + "id": "morph/morph-v3-fast", + "name": "Morph: Morph V3 Fast", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-02-01", - "last_updated": "2025-09-01", + "temperature": true, + "release_date": "2024-08-15", + "last_updated": "2024-08-15", "modalities": { "input": [ "text" @@ -135613,91 +127942,115 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 81920, + "output": 38000 }, "cost": { - "input": 2, - "output": 8, - "reasoning": 3 + "input": 0.8, + "output": 1.2 } }, - "meta/muse-spark-1.2-contributor": { - "id": "meta/muse-spark-1.2-contributor", - "name": "Muse Spark 1.2 Contributor", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", + "morph/morph-v3-large": { + "id": "morph/morph-v3-large", + "name": "Morph: Morph V3 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-08-15", + "last_updated": "2024-08-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.9, + "output": 1.9 + } + }, + "bytedance-seed/seed-1.6-flash": { + "id": "bytedance-seed/seed-1.6-flash", + "name": "ByteDance Seed: Seed 1.6 Flash", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "release_date": "2025-12-23", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "video", - "pdf", - "audio" + "text", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.2 + "input": 0.075, + "output": 0.3 } }, - "upstage/solar-pro4": { - "id": "upstage/solar-pro4", - "name": "Solar Pro 4", - "description": "Upstage's flagship model, specialized for agentic use", - "family": "solar-pro", - "attachment": false, + "bytedance-seed/seed-1.6": { + "id": "bytedance-seed/seed-1.6", + "name": "ByteDance Seed: Seed 1.6", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-08-06", - "last_updated": "2026-08-06", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" @@ -135705,25 +128058,28 @@ }, "open_weights": false, "limit": { - "context": 524288, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.03, - "output": 0.12 + "input": 0.25, + "output": 2 } }, - "upstage/solar-pro3": { - "id": "upstage/solar-pro3", - "name": "Solar Pro 3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "solar-pro", - "attachment": false, + "bytedance-seed/seed-2.0-mini": { + "id": "bytedance-seed/seed-2.0-mini", + "name": "ByteDance Seed: Seed-2.0-Mini", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -135732,149 +128088,156 @@ ], "tool_call": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-02-27", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.1, + "output": 0.4 } }, - "upstage/solar-pro2": { - "id": "upstage/solar-pro2", - "name": "Solar Pro 2", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "solar-pro", - "attachment": false, + "bytedance-seed/seed-2.0-lite": { + "id": "bytedance-seed/seed-2.0-lite", + "name": "ByteDance Seed: Seed-2.0-Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "minimal", + "low", + "medium", "high" ] } ], "tool_call": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "release_date": "2026-03-10", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 8192 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.25, + "output": 2 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "anthracite-org/magnum-v4-72b": { + "id": "anthracite-org/magnum-v4-72b", + "name": "Magnum v4 72B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2024-10-22", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 16384, + "output": 2048 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 3, + "output": 5 } }, - "mimo/mimo-v2.5": { - "id": "mimo/mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free": { + "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free", + "name": "NVIDIA: Nemotron 3 Nano Omni (free)", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-04-28", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", - "image", "audio", + "image", "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28 + "input": 0, + "output": 0 } }, - "mimo/mimo-v2.5-pro": { - "id": "mimo/mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "nvidia/nemotron-3-nano-30b-a3b": { + "id": "nvidia/nemotron-3-nano-30b-a3b", + "name": "NVIDIA: Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2024-12", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -135885,39 +128248,33 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 52429 }, "cost": { - "input": 0.435, - "output": 0.87 + "input": 0.05, + "output": 0.2 } }, - "sakana/fugu-ultra": { - "id": "sakana/fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, + "nvidia/llama-3.3-nemotron-super-49b-v1.5": { + "id": "nvidia/llama-3.3-nemotron-super-49b-v1.5", + "name": "NVIDIA: Llama 3.3 Nemotron Super 49B V1.5", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "temperature": true, + "release_date": "2025-03-16", + "last_updated": "2025-03-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -135925,29 +128282,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 131072, + "output": 26215 }, "cost": { - "input": 5, - "output": 30 + "input": 0.1, + "output": 0.4 } }, - "mistral/voxtral-small-latest": { - "id": "mistral/voxtral-small-latest", - "name": "Voxtral Small (latest)", - "description": "Instruct model with native audio input for speech understanding and tool use", - "family": "voxtral", - "attachment": true, - "reasoning": false, + "nvidia/nemotron-nano-9b-v2": { + "id": "nvidia/nemotron-nano-9b-v2", + "name": "NVIDIA: Nemotron Nano 9B V2", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", + "release_date": "2025-08-18", + "last_updated": "2025-08-18", "modalities": { "input": [ - "text", - "audio" + "text" ], "output": [ "text" @@ -135955,34 +128312,30 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 32000 + "context": 131072, + "output": 26215 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.04, + "output": 0.16 } }, - "qwen/qwen-flash": { - "id": "qwen/qwen-flash", - "name": "Qwen Flash", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", + "nvidia/nemotron-3-super-120b-a12b:free": { + "id": "nvidia/nemotron-3-super-120b-a12b:free", + "name": "NVIDIA: Nemotron 3 Super (free)", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-03-12", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -135991,118 +128344,125 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.05, - "output": 0.4 + "input": 0, + "output": 0 } }, - "qwen/qwen3-vl-plus": { - "id": "qwen/qwen3-vl-plus", - "name": "Qwen3-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "NVIDIA: Nemotron 3 Super", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-03-11", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.2, - "output": 1.6 + "input": 0.1, + "output": 0.5, + "cache_read": 0.1 } }, - "qwen/qwen3.5-plus": { - "id": "qwen/qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "Xiaomi: MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { "input": 0.4, - "output": 2.4 + "output": 2, + "cache_read": 0.08, + "tiers": [ + { + "input": 0.8, + "output": 4, + "cache_read": 0.16, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.8, + "output": 4, + "cache_read": 0.16 + } } }, - "qwen/qwen-plus": { - "id": "qwen/qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "xiaomi/mimo-v2-omni": { + "id": "xiaomi/mimo-v2-omni", + "name": "Xiaomi: MiMo-V2-Omni", + "description": "MiMo omni model for text, image, video, audio, and agents", + "family": "mimo", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -136110,34 +128470,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 262144, + "output": 65536 }, "cost": { "input": 0.4, - "output": 1.2 + "output": 2, + "cache_read": 0.08 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", + "xiaomi/mimo-v2-flash": { + "id": "xiaomi/mimo-v2-flash", + "name": "Xiaomi: MiMo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2024-12-01", + "release_date": "2025-12-16", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -136146,41 +128500,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, + "context": 262144, "output": 65536 }, "cost": { - "input": 1.2, - "output": 6 + "input": 0.09, + "output": 0.29, + "cache_read": 0.045 } }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, + "xiaomi/mimo-v2-pro": { + "id": "xiaomi/mimo-v2-pro", + "name": "Xiaomi: MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -136188,26 +128534,47 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 1.5 + "input": 1, + "output": 3, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "qwen/qwen3-coder-flash": { - "id": "qwen/qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "Xiaomi: MiMo V2.5 Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -136216,28 +128583,56 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 1, + "output": 3, + "cache_read": 0.2, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", + "inception/mercury-2": { + "id": "inception/mercury-2", + "name": "Inception: Mercury 2", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text" @@ -136248,60 +128643,50 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "output": 50000 }, "cost": { - "input": 1, - "output": 5 + "input": 0.25, + "output": 0.75, + "cache_read": 0.025 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "anthropic/claude-3.5-haiku": { + "id": "anthropic/claude-3.5-haiku", + "name": "Anthropic: Claude 3.5 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 65536 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 } }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Anthropic: Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -136309,19 +128694,28 @@ "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-09-29", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "video" + "pdf", + "text" ], "output": [ "text" @@ -136330,18 +128724,19 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 65536 + "output": 64000 }, "cost": { - "input": 0.5, - "output": 3 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Anthropic: Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -136349,19 +128744,28 @@ "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "release_date": "2025-05-22", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "video" + "pdf", + "text" ], "output": [ "text" @@ -136369,48 +128773,49 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 200000, "output": 64000 }, "cost": { - "input": 0.4, - "output": 1.6 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } - } - } - }, - "alibaba": { - "id": "alibaba", - "env": [ - "DASHSCOPE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1", - "name": "Alibaba", - "doc": "https://www.alibabacloud.com/help/en/model-studio/models", - "models": { - "qwen-flash": { - "id": "qwen-flash", - "name": "Qwen Flash", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, + }, + "anthropic/claude-opus-4.6-fast": { + "id": "anthropic/claude-opus-4.6-fast", + "name": "Anthropic: Claude Opus 4.6 (Fast)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-05-31", + "release_date": "2026-04-07", + "last_updated": "2026-04-11", "modalities": { "input": [ + "image", "text" ], "output": [ @@ -136420,34 +128825,46 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 32768 + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.4 + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Anthropic: Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ + "image", "text" ], "output": [ @@ -136456,61 +128873,45 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 - } - }, - "qwen2-5-32b-instruct": { - "id": "qwen2-5-32b-instruct", - "name": "Qwen2.5 32B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.7, - "output": 2.8 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "qwq-plus": { - "id": "qwq-plus", - "name": "QwQ Plus", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "anthropic/claude-opus-4.7-fast": { + "id": "anthropic/claude-opus-4.7-fast", + "name": "Anthropic: Claude Opus 4.7 (Fast)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-03-05", - "last_updated": "2025-03-05", + "temperature": false, + "release_date": "2026-05-12", + "last_updated": "2026-05-16", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ @@ -136519,69 +128920,97 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.8, - "output": 2.4 + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 } }, - "qwen2-5-vl-72b-instruct": { - "id": "qwen2-5-vl-72b-instruct", - "name": "Qwen2.5-VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Anthropic: Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "temperature": false, + "release_date": "2026-04-16", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 2.8, - "output": 8.4 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen3-vl-plus": { - "id": "qwen3-vl-plus", - "name": "Qwen3-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Anthropic: Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2025-08-05", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "image", + "pdf", + "text" ], "output": [ "text" @@ -136589,20 +129018,20 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.2, - "output": 1.6, - "reasoning": 4.8 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "qwen3.5-27b": { - "id": "qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Anthropic: Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -136610,41 +129039,51 @@ "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2025-11-24", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "video", - "audio" + "pdf", + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 2.4 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Anthropic: Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -136653,54 +129092,55 @@ { "type": "effort", "values": [ + "low", + "medium", "high", "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.4, - "cache_read": 0.04 + "input": 3, + "output": 15 } }, - "qwen-max": { - "id": "qwen-max", - "name": "Qwen Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "anthropic/claude-3-haiku": { + "id": "anthropic/claude-3-haiku", + "name": "Anthropic: Claude 3 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", + "release_date": "2024-03-07", + "last_updated": "2024-03-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -136708,54 +129148,71 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 200000, + "output": 4096 }, "cost": { - "input": 1.6, - "output": 6.4 + "input": 0.25, + "output": 1.25, + "cache_read": 0.03, + "cache_write": 0.3 } }, - "qwen3-next-80b-a3b-thinking": { - "id": "qwen3-next-80b-a3b-thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", - "attachment": false, + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Anthropic: Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2025-05-22", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", + "pdf", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.5, - "output": 6 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Anthropic: Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -136767,29 +129224,25 @@ "values": [ "low", "medium", - "xhigh" + "high", + "xhigh", + "max" ] }, { "type": "budget_tokens", - "min": 0, - "max": 262144 + "min": 1024 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" @@ -136798,115 +129251,67 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 128000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "tencent/hunyuan-a13b-instruct": { + "id": "tencent/hunyuan-a13b-instruct", + "name": "Tencent: Hunyuan A13B Instruct", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "release_date": "2025-06-30", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 131072, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "qwen-plus-character-ja": { - "id": "qwen-plus-character-ja", - "name": "Qwen Plus Character (Japanese)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01", - "last_updated": "2024-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 512 - }, - "cost": { - "input": 0.5, - "output": 1.4 + "input": 0.14, + "output": 0.57 } }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "tencent/hy3-preview": { + "id": "tencent/hy3-preview", + "name": "Tencent: Hy3 Preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "none", + "low", + "high" + ] } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "release_date": "2026-04-22", + "last_updated": "2026-05-16", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -136914,63 +129319,54 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.4, - "output": 2.4, - "reasoning": 2.4 + "input": 0.066, + "output": 0.26, + "cache_read": 0.029 } }, - "qwen3-livetranslate-flash-realtime": { - "id": "qwen3-livetranslate-flash-realtime", - "name": "Qwen3-LiveTranslate Flash Realtime", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "qwen", + "deepcogito/cogito-v2.1-671b": { + "id": "deepcogito/cogito-v2.1-671b", + "name": "Deep Cogito: Cogito v2.1 671B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "release_date": "2025-11-14", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 53248, - "output": 4096 + "context": 128000, + "output": 32768 }, "cost": { - "input": 10, - "output": 10, - "input_audio": 10, - "output_audio": 38 + "input": 1.25, + "output": 1.25 } }, - "qwen2-5-72b-instruct": { - "id": "qwen2-5-72b-instruct", - "name": "Qwen2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "cohere/command-a": { + "id": "cohere/command-a", + "name": "Cohere: Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ "text" @@ -136981,34 +129377,24 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 256000, "output": 8192 }, "cost": { - "input": 1.4, - "output": 5.6 + "input": 2.5, + "output": 10 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", + "cohere/command-r-08-2024": { + "id": "cohere/command-r-08-2024", + "name": "Cohere: Command R (08-2024)", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -137019,35 +129405,24 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.7, - "output": 2.8, - "reasoning": 8.4 + "input": 0.15, + "output": 0.6 } }, - "qwen-plus": { - "id": "qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "cohere/command-r7b-12-2024": { + "id": "cohere/command-r7b-12-2024", + "name": "Cohere: Command R7B (12-2024)", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", + "release_date": "2024-12-02", + "last_updated": "2024-12-02", "modalities": { "input": [ "text" @@ -137056,29 +129431,26 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.4, - "output": 1.2, - "reasoning": 4 + "input": 0.0375, + "output": 0.15 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", + "cohere/command-r-plus-08-2024": { + "id": "cohere/command-r-plus-08-2024", + "name": "Cohere: Command R+ (08-2024)", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -137087,63 +129459,26 @@ "text" ] }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 1.2, - "output": 6 - } - }, - "qwen-omni-turbo-realtime": { - "id": "qwen-omni-turbo-realtime", - "name": "Qwen-Omni Turbo Realtime", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-05-08", - "last_updated": "2025-05-08", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 2048 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.27, - "output": 1.07, - "input_audio": 4.44, - "output_audio": 8.89 + "input": 2.5, + "output": 10 } }, - "qwen3-coder-480b-a35b-instruct": { - "id": "qwen3-coder-480b-a35b-instruct", - "name": "Qwen3-Coder 480B-A35B Instruct", - "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", - "family": "qwen", + "gryphe/mythomax-l2-13b": { + "id": "gryphe/mythomax-l2-13b", + "name": "MythoMax 13B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2024-04-25", + "last_updated": "2024-04-25", "modalities": { "input": [ "text" @@ -137154,52 +129489,34 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 4096, + "output": 4096 }, "cost": { - "input": 1.5, - "output": 7.5, - "tiers": [ - { - "input": 2.7, - "output": 13.5, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 4.5, - "output": 22.5, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.06, + "output": 0.06 } }, - "qwen-turbo": { - "id": "qwen-turbo", - "name": "Qwen Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", + "stepfun/step-3.5-flash": { + "id": "stepfun/step-3.5-flash", + "name": "StepFun: Step 3.5 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11-01", - "last_updated": "2025-04-28", + "release_date": "2026-01-29", + "last_updated": "2026-01-29", "modalities": { "input": [ "text" @@ -137208,29 +129525,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 16384 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.05, - "output": 0.2, - "reasoning": 0.5 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", + "prime-intellect/intellect-3": { + "id": "prime-intellect/intellect-3", + "name": "Prime Intellect: INTELLECT-3", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-11-26", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -137241,49 +129557,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.45, - "output": 2.25, - "tiers": [ - { - "input": 0.75, - "output": 3.75, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.2, - "output": 6, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.2, + "output": 1.1 } }, - "qvq-max": { - "id": "qvq-max", - "name": "QVQ Max", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qvq", + "nex-agi/deepseek-v3.1-nex-n1": { + "id": "nex-agi/deepseek-v3.1-nex-n1", + "name": "Nex AGI: DeepSeek V3.1 Nex N1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "release_date": "2025-01-01", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -137292,25 +129586,23 @@ "open_weights": false, "limit": { "context": 131072, - "output": 8192 + "output": 163840 }, "cost": { - "input": 1.2, - "output": 4.8 + "input": 0.27, + "output": 1 } }, - "qwen2-5-7b-instruct": { - "id": "qwen2-5-7b-instruct", - "name": "Qwen2.5 7B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "undi95/remm-slerp-l2-13b": { + "id": "undi95/remm-slerp-l2-13b", + "name": "ReMM SLERP 13B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "release_date": "2023-07-22", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -137321,71 +129613,43 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 6144, + "output": 4096 }, "cost": { - "input": 0.175, - "output": 0.7 + "input": 0.45, + "output": 0.65 } }, - "qwen3-omni-flash": { - "id": "qwen3-omni-flash", - "name": "Qwen3-Omni Flash", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, + "~openai/gpt-mini-latest": { + "id": "~openai/gpt-mini-latest", + "name": "OpenAI: GPT Mini Latest", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "temperature": false, + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", "image", - "audio", - "video" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 16384 - }, - "cost": { - "input": 0.43, - "output": 1.66, - "input_audio": 3.81, - "output_audio": 15.11 - } - }, - "qwen-vl-max": { - "id": "qwen-vl-max", - "name": "Qwen-VL Max", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-08", - "last_updated": "2025-08-13", - "modalities": { - "input": [ - "text", - "image" + "pdf" ], "output": [ "text" @@ -137393,70 +129657,75 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.8, - "output": 3.2 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "qwen3-235b-a22b": { - "id": "qwen3-235b-a22b", - "name": "Qwen3 235B-A22B", - "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", - "family": "qwen", - "attachment": false, + "~openai/gpt-latest": { + "id": "~openai/gpt-latest", + "name": "OpenAI: GPT Latest", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "budget_tokens" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "temperature": false, + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1050000, + "output": 128000 }, "cost": { - "input": 0.7, - "output": 2.8, - "reasoning": 8.4 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "qwen3-vl-30b-a3b": { - "id": "qwen3-vl-30b-a3b", - "name": "Qwen3-VL 30B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "~moonshotai/kimi-latest": { + "id": "~moonshotai/kimi-latest", + "name": "MoonshotAI: Kimi Latest", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "toggle" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", @@ -137466,111 +129735,82 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 262142, + "output": 262142 }, "cost": { - "input": 0.2, - "output": 0.8, - "reasoning": 2.4 + "input": 0.74, + "output": 3.49, + "cache_read": 0.14 } }, - "qwen3.5-122b-a10b": { - "id": "qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "relace/relace-search": { + "id": "relace/relace-search", + "name": "Relace: Relace Search", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2025-12-09", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 3.2 + "input": 1, + "output": 3 } }, - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "relace/relace-apply-3": { + "id": "relace/relace-apply-3", + "name": "Relace: Relace Apply 3", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2025-09-26", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.248, - "output": 1.485 + "input": 0.85, + "output": 1.25 } }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "ai21/jamba-large-1.7": { + "id": "ai21/jamba-large-1.7", + "name": "AI21: Jamba Large 1.7", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2025-08-09", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -137579,165 +129819,140 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 256000, + "output": 4096 }, "cost": { - "input": 0.5, - "output": 2 + "input": 2, + "output": 8 } }, - "qwen-vl-plus": { - "id": "qwen-vl-plus", - "name": "Qwen-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "arcee-ai/coder-large": { + "id": "arcee-ai/coder-large", + "name": "Arcee AI: Coder Large", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-08-15", + "release_date": "2025-05-06", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.21, - "output": 0.63 + "input": 0.5, + "output": 0.8 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "arcee-ai/virtuoso-large": { + "id": "arcee-ai/virtuoso-large", + "name": "Arcee AI: Virtuoso Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "release_date": "2025-05-06", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 131072, + "output": 64000 }, "cost": { - "input": 0.1875, - "output": 1.125, - "cache_write": 0.234375 + "input": 0.75, + "output": 1.2 } }, - "qwen3-coder-flash": { - "id": "qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "arcee-ai/spotlight": { + "id": "arcee-ai/spotlight", + "name": "Arcee AI: Spotlight", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2025-05-06", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 131072, + "output": 65537 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 0.18, + "output": 0.18 } }, - "qwen2-5-omni-7b": { - "id": "qwen2-5-omni-7b", - "name": "Qwen2.5-Omni 7B", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", + "arcee-ai/trinity-large-thinking": { + "id": "arcee-ai/trinity-large-thinking", + "name": "Arcee AI: Trinity Large Thinking", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-12", - "last_updated": "2024-12", + "release_date": "2026-04-01", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": true, "limit": { - "context": 32768, - "output": 2048 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.1, - "output": 0.4, - "input_audio": 6.76 + "input": 0.22, + "output": 0.85 } }, - "qwen-mt-plus": { - "id": "qwen-mt-plus", - "name": "Qwen-MT Plus", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "family": "qwen", + "arcee-ai/maestro-reasoning": { + "id": "arcee-ai/maestro-reasoning", + "name": "Arcee AI: Maestro Reasoning", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01", - "last_updated": "2025-01", + "release_date": "2025-05-06", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -137746,37 +129961,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "output": 8192 + "context": 131072, + "output": 32000 }, "cost": { - "input": 2.46, - "output": 7.37 + "input": 0.9, + "output": 3.3 } }, - "qwen3-vl-235b-a22b": { - "id": "qwen3-vl-235b-a22b", - "name": "Qwen3-VL 235B-A22B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "arcee-ai/trinity-mini": { + "id": "arcee-ai/trinity-mini", + "name": "Arcee AI: Trinity Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-12", + "last_updated": "2026-01-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -137785,26 +129993,23 @@ "open_weights": true, "limit": { "context": 131072, - "output": 32768 + "output": 131072 }, "cost": { - "input": 0.7, - "output": 2.8, - "reasoning": 8.4 + "input": 0.045, + "output": 0.15 } }, - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", + "mancer/weaver": { + "id": "mancer/weaver", + "name": "Mancer: Weaver (alpha)", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "release_date": "2023-08-02", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -137813,82 +130018,60 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 8000, + "output": 2000 }, "cost": { - "input": 1, - "output": 5 + "input": 0.75, + "output": 1 } }, - "qwen2-5-14b-instruct": { - "id": "qwen2-5-14b-instruct", - "name": "Qwen2.5 14B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, + "perplexity/sonar-reasoning-pro": { + "id": "perplexity/sonar-reasoning-pro", + "name": "Perplexity: Sonar Reasoning Pro", + "description": "Web-grounded reasoning model for multi-step research and cited answers", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 128000, + "output": 25600 }, "cost": { - "input": 0.35, - "output": 1.4 + "input": 2, + "output": 8 } }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "perplexity/sonar": { + "id": "perplexity/sonar", + "name": "Perplexity: Sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -137896,28 +130079,24 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 127072, + "output": 25415 }, "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016, - "cache_write": 0.2 + "input": 1, + "output": 1 } }, - "qwen2-5-vl-7b-instruct": { - "id": "qwen2-5-vl-7b-instruct", - "name": "Qwen2.5-VL 7B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "perplexity/sonar-pro": { + "id": "perplexity/sonar-pro", + "name": "Perplexity: Sonar Pro", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text", @@ -137927,77 +130106,57 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 200000, + "output": 8000 }, "cost": { - "input": 0.35, - "output": 1.05 + "input": 3, + "output": 15 } }, - "qwen3.5-35b-a3b": { - "id": "qwen3.5-35b-a3b", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "perplexity/sonar-pro-search": { + "id": "perplexity/sonar-pro-search", + "name": "Perplexity: Sonar Pro Search", + "description": "Advanced Sonar search model for deeper research and cited synthesis", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2025-10-31", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 8000 }, "cost": { - "input": 0.25, - "output": 2 + "input": 3, + "output": 15 } }, - "qwen3.6-max-preview": { - "id": "qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "perplexity/sonar-deep-research": { + "id": "perplexity/sonar-deep-research", + "name": "Perplexity: Sonar Deep Research", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", "modalities": { "input": [ "text" @@ -138008,28 +130167,25 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 25600 }, "cost": { - "input": 1.3, - "output": 7.8, - "cache_read": 0.13, - "cache_write": 1.625 + "input": 2, + "output": 8 } }, - "qwen-mt-turbo": { - "id": "qwen-mt-turbo", - "name": "Qwen-MT Turbo", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "family": "qwen", + "switchpoint/router": { + "id": "switchpoint/router", + "name": "Switchpoint Router", + "description": "Automatic model router for matching prompts to suitable backends and budgets", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01", - "last_updated": "2025-01", + "release_date": "2025-07-12", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -138040,123 +130196,84 @@ }, "open_weights": false, "limit": { - "context": 16384, - "output": 8192 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.16, - "output": 0.49 + "input": 0.85, + "output": 3.4 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "openrouter/bodybuilder": { + "id": "openrouter/bodybuilder", + "name": "Body Builder (beta)", + "description": "Preview model for early access evaluation, prototyping, and compatibility testing", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-03-15", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 32768 }, + "status": "beta", "cost": { - "input": 0.6, - "output": 3.6 + "input": 0, + "output": 0 } }, - "qwen3-8b": { - "id": "qwen3-8b", - "name": "Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "openrouter/free": { + "id": "openrouter/free", + "name": "Free Models Router", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-02-01", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 200000, + "output": 32768 }, "cost": { - "input": 0.18, - "output": 0.7, - "reasoning": 2.1 + "input": 0, + "output": 0 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "openrouter/owl-alpha": { + "id": "openrouter/owl-alpha", + "name": "Owl Alpha", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-04-28", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" @@ -138165,125 +130282,83 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048756, + "output": 262144 }, + "status": "alpha", "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.28, - "cache_write": 0 + "input": 0, + "output": 0 } }, - "qwen-omni-turbo": { - "id": "qwen-omni-turbo", - "name": "Qwen-Omni Turbo", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", + "openrouter/pareto-code": { + "id": "openrouter/pareto-code", + "name": "Pareto Code Router", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01-19", - "last_updated": "2025-03-26", + "tool_call": false, + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 2048 + "context": 200000, + "output": 65536 }, "cost": { - "input": 0.07, - "output": 0.27, - "input_audio": 4.44, - "output_audio": 8.89 + "input": 0, + "output": 0 } }, - "qwen3-omni-flash-realtime": { - "id": "qwen3-omni-flash-realtime", - "name": "Qwen3-Omni Flash Realtime", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openrouter/auto": { + "id": "openrouter/auto", + "name": "Auto Router", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "release_date": "2026-03-15", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", "audio", - "video" - ], - "output": [ + "image", + "pdf", "text", - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 16384 - }, - "cost": { - "input": 0.52, - "output": 1.99, - "input_audio": 4.57, - "output_audio": 18.13 - } - }, - "qwen3-asr-flash": { - "id": "qwen3-asr-flash", - "name": "Qwen3-ASR Flash", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2024-04", - "release_date": "2025-09-08", - "last_updated": "2025-09-08", - "modalities": { - "input": [ - "audio" + "video" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 53248, - "output": 4096 + "context": 2000000, + "output": 32768 }, "cost": { - "input": 0.035, - "output": 0.035 + "input": 0, + "output": 0 } }, - "qwen3.6-27b": { - "id": "qwen3.6-27b", - "name": "Qwen3.6 27B", + "qwen/qwen3.5-plus-20260420": { + "id": "qwen/qwen3.5-plus-20260420", + "name": "Qwen: Qwen3.5 Plus 2026-04-20", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -138291,117 +130366,79 @@ "type": "toggle" }, { - "type": "budget_tokens" + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", "image", - "video", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.4, + "output": 2.4 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": false, + "qwen/qwen3-vl-235b-a22b-thinking": { + "id": "qwen/qwen3-vl-235b-a22b-thinking", + "name": "Qwen: Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-09-24", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 0.26, + "output": 2.6 } }, - "qwen3-14b": { - "id": "qwen3-14b", - "name": "Qwen3 14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "qwen/qwen3-vl-30b-a3b-thinking": { + "id": "qwen/qwen3-vl-30b-a3b-thinking", + "name": "Qwen: Qwen3 VL 30B A3B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-10-11", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", "text" ], "output": [ @@ -138411,90 +130448,55 @@ "open_weights": true, "limit": { "context": 131072, - "output": 8192 + "output": 32768 }, "cost": { - "input": 0.35, - "output": 1.4, - "reasoning": 4.2 + "input": 0.13, + "output": 1.56 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen: Qwen3 Coder Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-04", + "release_date": "2025-07-01", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, "output": 65536 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 0.65, + "output": 3.25, + "cache_read": 0.2 } }, - "qwen-vl-ocr": { - "id": "qwen-vl-ocr", - "name": "Qwen-VL OCR", - "description": "OCR model for extracting structured text from documents and screenshots", - "family": "qwen", + "qwen/qwen-plus": { + "id": "qwen/qwen-plus", + "name": "Qwen: Qwen-Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-10-28", - "last_updated": "2025-04-13", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -138502,90 +130504,58 @@ }, "open_weights": false, "limit": { - "context": 34096, - "output": 4096 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 0.4, + "output": 1.2, + "cache_read": 0.08 } - } - } - }, - "auriko": { - "id": "auriko", - "env": [ - "AURIKO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.auriko.ai/v1", - "name": "Auriko", - "doc": "https://docs.auriko.ai", - "models": { - "qwen-3.6-plus": { - "id": "qwen-3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + }, + "qwen/qwen3-coder-30b-a3b-instruct": { + "id": "qwen/qwen3-coder-30b-a3b-instruct", + "name": "Qwen: Qwen3 Coder 30B A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-07-31", + "last_updated": "2025-07-31", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 160000, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 0.07, + "output": 0.27 } }, - "minimax-m2-7-highspeed": { - "id": "minimax-m2-7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", + "qwen/qwen3-32b": { + "id": "qwen/qwen3-32b", + "name": "Qwen: Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2024-12-01", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -138596,51 +130566,28 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 40960, + "output": 40960 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_write": 0.375 + "input": 0.08, + "output": 0.24, + "cache_read": 0.04 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen: Qwen3 Next 80B A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2025-09-11", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -138648,107 +130595,63 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 52429 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.09, + "output": 1.1 } }, - "grok-4.3": { - "id": "grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "qwen/qwen3-vl-8b-instruct": { + "id": "qwen/qwen3-vl-8b-instruct", + "name": "Qwen: Qwen3 VL 8B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2025-10-15", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 30000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.08, + "output": 0.5 } }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "qwen/qwen3.6-35b-a3b": { + "id": "qwen/qwen3.6-35b-a3b", + "name": "Qwen: Qwen3.6 35B A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" @@ -138756,114 +130659,68 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 262144, "output": 65536 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.1612, + "output": 0.96525, + "cache_read": 0.1612 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen: Qwen3.7 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens", + "min": 1, + "max": 262144 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": true, + "release_date": "2025-08-26", + "last_updated": "2026-05-27", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 2.8 + "input": 1.625, + "output": 4.875, + "cache_read": 0.1625, + "cache_write": 2.03125 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen: Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2025-09-05", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -138871,94 +130728,59 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 1.2, + "output": 6, + "cache_read": 0.24 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, + "qwen/qwen3-8b": { + "id": "qwen/qwen3-8b", + "name": "Qwen: Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2025-04", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 40960, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 0.05, + "output": 0.4, + "cache_read": 0.05 } }, - "minimax-m2-7": { - "id": "minimax-m2-7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "qwen/qwen-plus-2025-07-28": { + "id": "qwen/qwen-plus-2025-07-28", + "name": "Qwen: Qwen Plus 0728", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2025-09-09", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -138969,20 +130791,18 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_write": 0.375 + "input": 0.26, + "output": 0.78 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "qwen/qwen3.5-flash-02-23": { + "id": "qwen/qwen3.5-flash-02-23", + "name": "Qwen: Qwen3.5-Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -138990,58 +130810,45 @@ "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-02-26", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 128000 + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.1, + "output": 0.4 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "qwen/qwen3-30b-a3b-instruct-2507": { + "id": "qwen/qwen3-30b-a3b-instruct-2507", + "name": "Qwen: Qwen3 30B A3B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-07-29", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -139052,46 +130859,25 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.09, + "output": 0.3, + "cache_read": 0.04 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "qwen/qwen-2.5-coder-32b-instruct": { + "id": "qwen/qwen-2.5-coder-32b-instruct", + "name": "Qwen2.5 Coder 32B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2024-11-11", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -139102,95 +130888,87 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0.2, + "output": 0.2, + "cache_read": 0.015 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen: Qwen3 Next 80B A3B Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "release_date": "2025-09-11", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.0975, + "output": 0.78 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "qwen/qwen3-235b-a22b-thinking-2507": { + "id": "qwen/qwen3-235b-a22b-thinking-2507", + "name": "Qwen: Qwen3 235B A22B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "release_date": "2025-07-25", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.11, + "output": 0.6 + } + }, + "qwen/qwen3-vl-32b-instruct": { + "id": "qwen/qwen3-vl-32b-instruct", + "name": "Qwen: Qwen3 VL 32B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-10-21", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -139198,124 +130976,109 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.104, + "output": 0.416 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "qwen/qwen3-coder": { + "id": "qwen/qwen3-coder", + "name": "Qwen: Qwen3 Coder 480B A35B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2025-07-23", + "last_updated": "2025-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 52429 + }, + "cost": { + "input": 0.22, + "output": 1, + "cache_read": 0.022 + } + }, + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen: Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.25, + "output": 1.5, + "cache_write": 0.3125 } - } - } - }, - "zenmux": { - "id": "zenmux", - "env": [ - "ZENMUX_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://zenmux.ai/api/v1", - "name": "ZenMux", - "doc": "https://docs.zenmux.ai", - "models": { - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + }, + "qwen/qwen3.5-plus-02-15": { + "id": "qwen/qwen3.5-plus-02-15", + "name": "Qwen: Qwen3.5 Plus 2026-02-15", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-02-15", + "last_updated": "2026-03-15", "modalities": { "input": [ "image", "text", - "pdf" + "video" ], "output": [ "text" @@ -139323,44 +131086,55 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 65536 }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "cost": { + "input": 0.26, + "output": 1.56 + } + }, + "qwen/qwen-2.5-7b-instruct": { + "id": "qwen/qwen-2.5-7b-instruct", + "name": "Qwen: Qwen2.5 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-09", + "last_updated": "2025-04-16", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 6554 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.04, + "output": 0.1 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "qwen/qwen3-vl-8b-thinking": { + "id": "qwen/qwen3-vl-8b-thinking", + "name": "Qwen: Qwen3 VL 8B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2025-10-15", + "last_updated": "2025-11-25", "modalities": { "input": [ - "pdf", "image", "text" ], @@ -139370,44 +131144,33 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.117, + "output": 1.365 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "attachment": true, + "qwen/qwen3-max-thinking": { + "id": "qwen/qwen3-max-thinking", + "name": "Qwen: Qwen3 Max Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "release_date": "2026-01-23", + "last_updated": "2026-03-15", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -139416,357 +131179,259 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 262144, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.78, + "output": 3.9 } }, - "anthropic/claude-3.5-haiku": { - "id": "anthropic/claude-3.5-haiku", - "name": "Claude 3.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "attachment": true, - "reasoning": false, + "qwen/qwen3-30b-a3b-thinking-2507": { + "id": "qwen/qwen3-30b-a3b-thinking-2507", + "name": "Qwen: Qwen3 30B A3B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2024-11-04", - "last_updated": "2024-11-04", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 32768, + "output": 6554 }, "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "input": 0.051, + "output": 0.34 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "qwen/qwen2.5-vl-72b-instruct": { + "id": "qwen/qwen2.5-vl-72b-instruct", + "name": "Qwen: Qwen2.5 VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-02-01", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 32768, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.8, + "output": 0.8, + "cache_read": 0.075 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen: Qwen3.5-27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "release_date": "2026-02-26", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 262144, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.195, + "output": 1.56 } }, - "anthropic/claude-sonnet-5-free": { - "id": "anthropic/claude-sonnet-5-free", - "name": "Claude Sonnet 5 (Free)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "qwen/qwen3-235b-a22b": { + "id": "qwen/qwen3-235b-a22b", + "name": "Qwen: Qwen3 235B A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 38912 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "release_date": "2024-12-01", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 131072, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.455, + "output": 1.82, + "cache_read": 0.15 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "attachment": true, + "qwen/qwen-2.5-72b-instruct": { + "id": "qwen/qwen-2.5-72b-instruct", + "name": "Qwen2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2024-09", + "last_updated": "2026-01-10", "modalities": { "input": [ - "image", "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 32768, + "output": 16384 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.12, + "output": 0.39 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "attachment": true, + "qwen/qwen-plus-2025-07-28:thinking": { + "id": "qwen/qwen-plus-2025-07-28:thinking", + "name": "Qwen: Qwen Plus 0728 (thinking)", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2025-09-09", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.26, + "output": 0.78 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "Qwen: Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "release_date": "2026-02-02", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 262144, + "output": 65536 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.12, + "output": 0.75, + "cache_read": 0.035 } }, - "anthropic/claude-3.7-sonnet": { - "id": "anthropic/claude-3.7-sonnet", - "name": "Claude 3.7 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "qwen/qwen3.6-27b": { + "id": "qwen/qwen3.6-27b", + "name": "Qwen: Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-02-24", - "last_updated": "2025-02-24", + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -139774,141 +131439,104 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 256000, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.325, + "output": 3.25 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen: Qwen3.5-35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-02-26", + "last_updated": "2026-03-15", "modalities": { "input": [ "image", "text", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 262144, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.1625, + "output": 1.3 } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "qwen/qwen3.5-9b": { + "id": "qwen/qwen3.5-9b", + "name": "Qwen: Qwen3.5-9B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "release_date": "2026-03-10", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "pdf" + "text", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 256000, + "output": 32768 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 4 + "input": 0.05, + "output": 0.15 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen: Qwen3.5 397B A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-18", - "last_updated": "2026-02-18", + "release_date": "2026-02-15", + "last_updated": "2026-03-15", "modalities": { "input": [ + "image", "text", - "image" + "video" ], "output": [ "text" @@ -139916,78 +131544,53 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 262144, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.39, + "output": 2.34 } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "qwen/qwen3-vl-30b-a3b-instruct": { + "id": "qwen/qwen3-vl-30b-a3b-instruct", + "name": "Qwen: Qwen3 VL 30B A3B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2025-10-05", + "last_updated": "2025-11-25", "modalities": { "input": [ - "image", "text", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 131072, + "output": 32768 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 0.13, + "output": 0.52 } }, - "deepseek/deepseek-chat": { - "id": "deepseek/deepseek-chat", - "name": "DeepSeek-V3.2 (Non-thinking Mode)", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "qwen/qwen3-235b-a22b-2507": { + "id": "qwen/qwen3-235b-a22b-2507", + "name": "Qwen: Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2025-04", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -139996,34 +131599,26 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 64000 + "context": 262144, + "output": 52429 }, - "status": "deprecated", "cost": { - "input": 0.28, - "output": 0.42, - "cache_read": 0.03 + "input": 0.071, + "output": 0.1 } }, - "deepseek/deepseek-v3.2-exp": { - "id": "deepseek/deepseek-v3.2-exp", - "name": "DeepSeek-V3.2-Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "qwen/qwen3-coder-flash": { + "id": "qwen/qwen3-coder-flash", + "name": "Qwen: Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2025-07-23", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -140034,43 +131629,30 @@ }, "open_weights": false, "limit": { - "context": 163000, - "output": 64000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.22, - "output": 0.33 + "input": 0.195, + "output": 0.975, + "cache_read": 0.06 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "qwen/qwen3-14b": { + "id": "qwen/qwen3-14b", + "name": "Qwen: Qwen3 14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2025-04", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -140081,78 +131663,60 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 40960, + "output": 40960 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0.06, + "output": 0.24, + "cache_read": 0.025 } }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen/qwen3-vl-235b-a22b-instruct": { + "id": "qwen/qwen3-vl-235b-a22b-instruct", + "name": "Qwen: Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-05", - "last_updated": "2025-12-05", + "release_date": "2025-09-23", + "last_updated": "2026-01-10", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 64000 + "context": 262144, + "output": 52429 }, "cost": { - "input": 0.28, - "output": 0.43 + "input": 0.2, + "output": 0.88, + "cache_read": 0.11 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "qwen/qwen3-30b-a3b": { + "id": "qwen/qwen3-30b-a3b", + "name": "Qwen: Qwen3 30B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2025-04", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -140163,36 +131727,35 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 40960, + "output": 40960 }, "cost": { - "input": 0.14, + "input": 0.08, "output": 0.28, - "cache_read": 0.0028 + "cache_read": 0.03 } }, - "tencent/hy3-preview": { - "id": "tencent/hy3-preview", - "name": "Hy3 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + "qwen/qwen3.6-max-preview": { + "id": "qwen/qwen3.6-max-preview", + "name": "Qwen: Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 131072 } ], "tool_call": true, "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "release_date": "2026-04-27", + "last_updated": "2026-05-01", "modalities": { "input": [ "text" @@ -140201,23 +131764,21 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.172, - "output": 0.572, - "cache_read": 0.058, - "cache_write": 0 + "input": 1.04, + "output": 6.24, + "cache_write": 1.3 } }, - "z-ai/glm-4.6v": { - "id": "z-ai/glm-4.6v", - "name": "GLM 4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen: Qwen3.5-122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -140227,13 +131788,12 @@ ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2026-02-26", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", + "text", "video" ], "output": [ @@ -140242,99 +131802,69 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.1456, - "output": 0.4367, - "cache_read": 0.0291, - "tiers": [ - { - "input": 0.2911, - "output": 0.8734, - "cache_read": 0.0582, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.26, + "output": 2.08 } }, - "z-ai/glm-4.5": { - "id": "z-ai/glm-4.5", - "name": "GLM 4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen: Qwen3.6 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2025-08-26", + "last_updated": "2026-04-11", "modalities": { "input": [ + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 96000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.2911, - "output": 1.1645, - "cache_read": 0.0582, - "tiers": [ - { - "input": 0.5823, - "output": 2.3291, - "cache_read": 0.1165, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.325, + "output": 1.95, + "cache_read": 0.0325, + "cache_write": 0.40625 } }, - "z-ai/glm-5v-turbo": { - "id": "z-ai/glm-5v-turbo", - "name": "GLM 5V Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", + "amazon/nova-lite-v1": { + "id": "amazon/nova-lite-v1", + "name": "Amazon: Nova Lite 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "release_date": "2024-12-06", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -140342,141 +131872,82 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 300000, + "output": 5120 }, "cost": { - "input": 0.726, - "output": 3.1946, - "cache_read": 0.1743, - "tiers": [ - { - "input": 1.0165, - "output": 3.7754, - "cache_read": 0.2614, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.06, + "output": 0.24 } }, - "z-ai/glm-5.3-flash": { - "id": "z-ai/glm-5.3-flash", - "name": "GLM 5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "amazon/nova-premier-v1": { + "id": "amazon/nova-premier-v1", + "name": "Amazon: Nova Premier 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2025-11-01", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 32000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 2.5, + "output": 12.5 } }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "GLM 4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "amazon/nova-pro-v1": { + "id": "amazon/nova-pro-v1", + "name": "Amazon: Nova Pro 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 300000, + "output": 5120 }, "cost": { - "input": 0.2911, - "output": 1.1645, - "cache_read": 0.0582, - "tiers": [ - { - "input": 0.5823, - "output": 2.3291, - "cache_read": 0.1165, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.8, + "output": 3.2 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM 5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "amazon/nova-micro-v1": { + "id": "amazon/nova-micro-v1", + "name": "Amazon: Nova Micro 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2024-12-06", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -140485,144 +131956,116 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 128000, + "output": 5120 }, "cost": { - "input": 0.58, - "output": 2.6, - "cache_read": 0.14, - "tiers": [ - { - "input": 0.87, - "output": 3.18, - "cache_read": 0.22, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.035, + "output": 0.14 } }, - "z-ai/glm-4.6v-flash": { - "id": "z-ai/glm-4.6v-flash", - "name": "GLM 4.6V FlashX", - "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", + "amazon/nova-2-lite-v1": { + "id": "amazon/nova-2-lite-v1", + "name": "Amazon: Nova 2 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2024-12-01", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", "image", + "pdf", + "text", "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 1000000, + "output": 65535 }, "cost": { - "input": 0.0218, - "output": 0.2184, - "cache_read": 0.0044, - "tiers": [ - { - "input": 0.0437, - "output": 0.4367, - "cache_read": 0.0044, - "tier": { - "type": "context", - "size": 32000 - } - } + "input": 0.3, + "output": 2.5 + } + }, + "aion-labs/aion-rp-llama-3.1-8b": { + "id": "aion-labs/aion-rp-llama-3.1-8b", + "name": "AionLabs: Aion-RP 1.0 (8B)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-02-05", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" ] + }, + "open_weights": false, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.8, + "output": 1.6 } }, - "z-ai/glm-5.3-flashx": { - "id": "z-ai/glm-5.3-flashx", - "name": "GLM 5.3 FlashX", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + "aion-labs/aion-1.0-mini": { + "id": "aion-labs/aion-1.0-mini", + "name": "AionLabs: Aion-1.0-Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2025-02-05", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.375, - "output": 1.25, - "cache_read": 0.075 + "input": 0.7, + "output": 1.4 } }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "GLM 4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", + "aion-labs/aion-2.0": { + "id": "aion-labs/aion-2.0", + "name": "AionLabs: Aion-2.0", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-02-24", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -140631,48 +132074,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.2911, - "output": 1.1645, - "cache_read": 0.0582, - "tiers": [ - { - "input": 0.5823, - "output": 2.3291, - "cache_read": 0.1165, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.8, + "output": 1.6 } }, - "z-ai/glm-4.7-flash-free": { - "id": "z-ai/glm-4.7-flash-free", - "name": "GLM 4.7 Flash (Free)", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", + "aion-labs/aion-1.0": { + "id": "aion-labs/aion-1.0", + "name": "AionLabs: Aion-1.0", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-02-05", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -140681,41 +132103,54 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 4, + "output": 8 } }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "inflection/inflection-3-pi": { + "id": "inflection/inflection-3-pi", + "name": "Inflection: Inflection 3 Pi", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2024-10-11", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "structured_output": true, + "open_weights": false, + "limit": { + "context": 8000, + "output": 1024 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "inflection/inflection-3-productivity": { + "id": "inflection/inflection-3-productivity", + "name": "Inflection: Inflection 3 Productivity", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2024-10-11", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -140724,37 +132159,26 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 8000, + "output": 1024 }, "cost": { - "input": 0.98, - "output": 3.08, - "cache_read": 0.182 + "input": 2.5, + "output": 10 } }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "sao10k/l3.1-euryale-70b": { + "id": "sao10k/l3.1-euryale-70b", + "name": "Sao10K: Llama 3.1 Euryale 70B v2.2", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2024-08-28", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -140765,46 +132189,24 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.8781, - "output": 3.5126, - "cache_read": 0.1903, - "tiers": [ - { - "input": 1.1709, - "output": 4.098, - "cache_read": 0.2927, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.85, + "output": 0.85 } }, - "z-ai/glm-4.7-flashx": { - "id": "z-ai/glm-4.7-flashx", - "name": "GLM 4.7 FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "sao10k/l3.3-euryale-70b": { + "id": "sao10k/l3.3-euryale-70b", + "name": "Sao10K: Llama 3.3 Euryale 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2024-12-18", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -140815,35 +132217,24 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.0728, - "output": 0.4367, - "cache_read": 0.0146 + "input": 0.65, + "output": 0.75 } }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "GLM 5 Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", + "sao10k/l3-lunaris-8b": { + "id": "sao10k/l3-lunaris-8b", + "name": "Sao10K: Llama 3 8B Lunaris", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "release_date": "2024-08-13", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -140852,50 +132243,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.73, - "output": 3.19, - "cache_read": 0.174, - "tiers": [ - { - "input": 1.02, - "output": 3.77, - "cache_read": 0.261, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.04, + "output": 0.05 } }, - "z-ai/glm-4.6v-flash-free": { - "id": "z-ai/glm-4.6v-flash-free", - "name": "GLM 4.6V Flash (Free)", - "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "sao10k/l3-euryale-70b": { + "id": "sao10k/l3-euryale-70b", + "name": "Sao10k: Llama 3 Euryale 70B v2.1", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2024-06-18", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -140903,67 +132273,53 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "tiers": [ - { - "input": 0, - "output": 0, - "cache_read": 0, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 1.48, + "output": 1.48 } }, - "z-ai/glm-image": { - "id": "z-ai/glm-image", - "name": "GLM-Image", - "description": "GLM-Image is an image generation model adopts a hybrid autoregressive + diffusion decoder architecture. In general image generation quality, GLM‑Image aligns with mainstream latent diffusion approaches, but it shows significant advantages in text-rendering and knowledge‑intensive generation scenarios. It performs especially well in tasks requiring precise semantic understanding and complex information expression, while maintaining strong capabilities in high‑fidelity and fine‑grained detail generation. In addition to text‑to‑image generation, GLM‑Image also supports a rich set of image‑to‑image tasks including image editing, style transfer, identity‑preserving generation, and multi‑subject consistency.", + "sao10k/l3.1-70b-hanami-x1": { + "id": "sao10k/l3.1-70b-hanami-x1", + "name": "Sao10K: Llama 3.1 70B Hanami x1", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "temperature": true, + "release_date": "2025-01-08", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": true, "limit": { - "context": 10240, - "output": 0 + "context": 16000, + "output": 16000 + }, + "cost": { + "input": 3, + "output": 3 } }, - "z-ai/glm-4.5-air": { - "id": "z-ai/glm-4.5-air", - "name": "GLM 4.5 Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "upstage/solar-pro-3": { + "id": "upstage/solar-pro-3", + "name": "Upstage: Solar Pro 3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2026-01-27", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -140972,53 +132328,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 96000 + "output": 32768 }, "cost": { - "input": 0.1165, - "output": 0.2911, - "cache_read": 0.0233, - "tiers": [ - { - "input": 0.1747, - "output": 1.1645, - "cache_read": 0.0349, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0.15, + "output": 0.6 } }, - "z-ai/glm-5.3": { - "id": "z-ai/glm-5.3", - "name": "GLM 5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "allenai/olmo-3-32b-think": { + "id": "allenai/olmo-3-32b-think", + "name": "AllenAI: Olmo 3 32B Think", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2025-11-22", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -141029,25 +132359,24 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.15, + "output": 0.5 } }, - "kuaishou/kat-coder-pro-v2": { - "id": "kuaishou/kat-coder-pro-v2", - "name": "KAT-Coder-Pro-V2", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "essentialai/rnj-1-instruct": { + "id": "essentialai/rnj-1-instruct", + "name": "EssentialAI: Rnj 1 Instruct", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", + "release_date": "2025-12-05", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -141056,77 +132385,71 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 80000 + "context": 32768, + "output": 6554 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.15, + "output": 0.15 } }, - "baidu/ernie-5.0-thinking-preview": { - "id": "baidu/ernie-5.0-thinking-preview", - "name": "ERNIE 5.0", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "deepseek/deepseek-r1-0528": { + "id": "deepseek/deepseek-r1-0528", + "name": "DeepSeek: R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-01-22", - "last_updated": "2026-01-22", + "release_date": "2025-05-28", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 64000 + "context": 163840, + "output": 65536 }, "cost": { - "input": 0.84, - "output": 3.37 + "input": 0.45, + "output": 2.15, + "cache_read": 0.2 } }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek: DeepSeek V4 Flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-04-24", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -141134,449 +132457,440 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "x-ai/grok-4.5": { - "id": "x-ai/grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + "deepseek/deepseek-v3.1-terminus": { + "id": "deepseek/deepseek-v3.1-terminus", + "name": "DeepSeek: DeepSeek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 163840, + "output": 32768 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "input": 0.21, + "output": 0.79, + "cache_read": 0.13 } }, - "x-ai/grok-voice-tts-1.0": { - "id": "x-ai/grok-voice-tts-1.0", - "name": "Grok Voice TTS 1.0", - "description": "Convert text into spoken audio with a single API call. The API supports a rich set of expressive voices, inline speech tags for fine-grained delivery control, and output formats from high-fidelity MP3 to telephony-optimized μ-law.", + "deepseek/deepseek-r1-distill-llama-70b": { + "id": "deepseek/deepseek-r1-distill-llama-70b", + "name": "DeepSeek: R1 Distill Llama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": false, - "temperature": false, - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "temperature": true, + "release_date": "2025-01-23", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 15000, - "output": 15000 + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0.7, + "output": 0.8, + "cache_read": 0.015 } }, - "x-ai/grok-imagine-image-2.0": { - "id": "x-ai/grok-imagine-image-2.0", - "name": "Grok Imagine Image 2.0", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-08-07", - "last_updated": "2026-08-07", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek: DeepSeek V4 Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-04-24", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 66000, - "output": 0 + "context": 1048576, + "output": 384000 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "x-ai/grok-voice-stt-1.0": { - "id": "x-ai/grok-voice-stt-1.0", - "name": "Grok Voice STT 1.0", - "description": "Grok Voice STT 1.0 is xAI's speech-to-text model. It supports transcription with word-level timestamps, optional speaker diarization, and multichannel audio.", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-08-04", - "last_updated": "2026-08-04", + "deepseek/deepseek-r1": { + "id": "deepseek/deepseek-r1", + "name": "DeepSeek: R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 15000, - "output": 15000 + "context": 64000, + "output": 16000 + }, + "cost": { + "input": 0.7, + "output": 2.5 } }, - "x-ai/grok-build-0.1": { - "id": "x-ai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, + "deepseek/deepseek-v3.2-speciale": { + "id": "deepseek/deepseek-v3.2-speciale", + "name": "DeepSeek: DeepSeek V3.2 Speciale", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "release_date": "2025-12-01", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 163840, + "output": 163840 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 + "input": 0.4, + "output": 1.2, + "cache_read": 0.135 } }, - "x-ai/grok-4.2-fast": { - "id": "x-ai/grok-4.2-fast", - "name": "Grok 4.2 Fast", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", - "attachment": true, - "reasoning": false, + "deepseek/deepseek-v3.2-exp": { + "id": "deepseek/deepseek-v3.2-exp", + "name": "DeepSeek: DeepSeek V3.2 Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "release_date": "2025-01-01", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 163840, + "output": 65536 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 0.2, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.27, + "output": 0.41 } }, - "x-ai/grok-4.2-fast-non-reasoning": { - "id": "x-ai/grok-4.2-fast-non-reasoning", - "name": "Grok 4.2 Fast Non Reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "deepseek/deepseek-chat-v3-0324": { + "id": "deepseek/deepseek-chat-v3-0324", + "name": "DeepSeek: DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "release_date": "2025-03-24", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 30000 + "context": 163840, + "output": 65536 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 0.2, - "tier": { - "type": "context", - "size": 128000 - } - } + "input": 0.2, + "output": 0.77, + "cache_read": 0.095 + } + }, + "deepseek/deepseek-r1-distill-qwen-32b": { + "id": "deepseek/deepseek-r1-distill-qwen-32b", + "name": "DeepSeek: R1 Distill Qwen 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "release_date": "2025-01-01", + "last_updated": "2025-11-25", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0.29, + "output": 0.29 } }, - "x-ai/grok-4.6": { - "id": "x-ai/grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, + "deepseek/deepseek-chat": { + "id": "deepseek/deepseek-chat", + "name": "DeepSeek: DeepSeek V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-12-01", + "last_updated": "2026-03-15", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 163840, + "output": 163840 + }, + "cost": { + "input": 0.32, + "output": 0.89, + "cache_read": 0.15 + } + }, + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek: DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "release_date": "2025-12-01", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 163840, + "output": 65536 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "input": 0.26, + "output": 0.38, + "cache_read": 0.125 } }, - "volcengine/doubao-seed-2.0-mini": { - "id": "volcengine/doubao-seed-2.0-mini", - "name": "Doubao-Seed-2.0-mini", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "deepseek/deepseek-chat-v3.1": { + "id": "deepseek/deepseek-chat-v3.1", + "name": "DeepSeek: DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "temperature": true, - "knowledge": "2026-02-14", - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 32768, + "output": 7168 }, "cost": { - "input": 0.03, - "output": 0.28, - "cache_read": 0.01, - "cache_write": 0.0024 + "input": 0.15, + "output": 0.75 } }, - "volcengine/doubao-seed-2.0-lite": { - "id": "volcengine/doubao-seed-2.0-lite", - "name": "Doubao-Seed-2.0-lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax: MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2026-02-14", - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-02-12", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 196608, + "output": 196608 }, "cost": { - "input": 0.09, - "output": 0.51, - "cache_read": 0.02, - "cache_write": 0.0024 + "input": 0.25, + "output": 1.2, + "cache_read": 0.029 } }, - "volcengine/doubao-seed-code": { - "id": "volcengine/doubao-seed-code", - "name": "Doubao-Seed-Code", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": true, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax: MiniMax M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-11", - "last_updated": "2025-11-11", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 196608, + "output": 39322 + }, + "cost": { + "input": 0.27, + "output": 0.95, + "cache_read": 0.03 + } + }, + "minimax/minimax-01": { + "id": "minimax/minimax-01", + "name": "MiniMax: MiniMax-01", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2025-01-15", + "last_updated": "2025-01-15", "modalities": { "input": [ "text", @@ -141586,39 +132900,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 1000192, + "output": 1000192 }, - "status": "deprecated", "cost": { - "input": 0.17, - "output": 1.12, - "cache_read": 0.03 + "input": 0.2, + "output": 1.1 } }, - "volcengine/doubao-seed-2.0-pro": { - "id": "volcengine/doubao-seed-2.0-pro", - "name": "Doubao-Seed-2.0-pro", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax: MiniMax M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "temperature": true, - "knowledge": "2026-02-14", - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-06-01", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", @@ -141629,72 +132936,57 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 1048576, + "output": 512000 }, "cost": { - "input": 0.45, - "output": 2.24, - "cache_read": 0.09, - "cache_write": 0.0024 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "volcengine/doubao-seed-1.8": { - "id": "volcengine/doubao-seed-1.8", - "name": "Doubao-Seed-1.8", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax: MiniMax M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "release_date": "2025-10-23", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 196608, + "output": 196608 }, "cost": { - "input": 0.11, - "output": 0.28, - "cache_read": 0.02, - "cache_write": 0.0024 + "input": 0.255, + "output": 1, + "cache_read": 0.03 } }, - "volcengine/doubao-seed-2.0-code": { - "id": "volcengine/doubao-seed-2.0-code", - "name": "Doubao Seed 2.0 Code", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": true, + "minimax/minimax-m2-her": { + "id": "minimax/minimax-m2-her", + "name": "MiniMax: MiniMax M2-her", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "release_date": "2026-01-23", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -141703,135 +132995,106 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 65536, + "output": 2048 }, "cost": { - "input": 0.9, - "output": 4.48 + "input": 0.3, + "output": 1.2 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "attachment": true, + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax: MiniMax M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax-m2.7", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2026-02-19", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048000, - "output": 64000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 4.5 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "minimax/minimax-m1": { + "id": "minimax/minimax-m1", + "name": "MiniMax: MiniMax M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 40000 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 + "input": 0.4, + "output": 2.2 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "stealth/claude-opus-4.7": { + "id": "stealth/claude-opus-4.7", + "name": "Stealth: Claude Opus 4.7 (20% off)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "temperature": false, + "release_date": "2026-04-16", + "last_updated": "2026-05-27", "modalities": { "input": [ - "pdf", "image", - "text", - "audio", - "video" + "pdf", + "text" ], "output": [ "text" @@ -141839,43 +133102,45 @@ }, "open_weights": false, "limit": { - "context": 1048000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.31, - "cache_write": 4.5 + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "stealth/claude-sonnet-4.6": { + "id": "stealth/claude-sonnet-4.6", + "name": "Stealth: Claude Sonnet 4.6 (20% off)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "max" ] } ], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-02-17", + "last_updated": "2026-05-27", "modalities": { "input": [ - "pdf", "image", - "text", - "audio" + "pdf", + "text" ], "output": [ "text" @@ -141883,43 +133148,46 @@ }, "open_weights": false, "limit": { - "context": 1048000, + "context": 1000000, "output": 64000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.07, - "cache_write": 1 + "input": 2.4, + "output": 12, + "cache_read": 0.24, + "cache_write": 3 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "stealth/claude-opus-4.6": { + "id": "stealth/claude-opus-4.6", + "name": "Stealth: Claude Opus 4.6 (20% off)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-02-05", + "last_updated": "2026-05-27", "modalities": { "input": [ - "text", "image", "pdf", - "audio" + "text" ], "output": [ "text" @@ -141927,167 +133195,144 @@ }, "open_weights": false, "limit": { - "context": 1048000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 1 + "input": 4, + "output": 20, + "cache_read": 0.4, + "cache_write": 5 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "attachment": true, + "kwaipilot/kat-coder-pro-v2": { + "id": "kwaipilot/kat-coder-pro-v2", + "name": "Kwaipilot: KAT-Coder-Pro V2", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-07-22", - "last_updated": "2025-07-22", + "release_date": "2026-03-27", + "last_updated": "2026-04-11", "modalities": { "input": [ - "pdf", - "image", - "text", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048000, - "output": 64000 + "context": 256000, + "output": 80000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.03, - "cache_write": 1 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "attachment": true, + "nousresearch/hermes-2-pro-llama-3-8b": { + "id": "nousresearch/hermes-2-pro-llama-3-8b", + "name": "NousResearch: Hermes 2 Pro - Llama-3 8B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2025-03-20", - "last_updated": "2025-03-20", + "release_date": "2024-05-27", + "last_updated": "2024-06-27", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "output": 65530 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.25, - "output": 1.5 + "input": 0.14, + "output": 0.14 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "nousresearch/hermes-4-405b": { + "id": "nousresearch/hermes-4-405b", + "name": "Nous: Hermes 4 405B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2025-08-25", + "last_updated": "2025-08-25", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 26215 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025 + "input": 1, + "output": 3 } }, - "sapiens-ai/agnes-1.5-lite": { - "id": "sapiens-ai/agnes-1.5-lite", - "name": "Agnes 1.5 Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": true, + "nousresearch/hermes-3-llama-3.1-70b": { + "id": "nousresearch/hermes-3-llama-3.1-70b", + "name": "Nous: Hermes 3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2026-03-26", - "last_updated": "2026-03-26", + "release_date": "2024-08-18", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.12, - "output": 0.6 + "input": 0.3, + "output": 0.3 } }, - "sapiens-ai/agnes-1.5-pro": { - "id": "sapiens-ai/agnes-1.5-pro", - "name": "Agnes 1.5 Pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "nousresearch/hermes-3-llama-3.1-405b": { + "id": "nousresearch/hermes-3-llama-3.1-405b", + "name": "Nous: Hermes 3 405B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-03-21", - "last_updated": "2026-03-21", + "release_date": "2024-08-16", + "last_updated": "2024-08-16", "modalities": { "input": [ "text" @@ -142096,28 +133341,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.16, - "output": 0.8 + "input": 1, + "output": 1 } }, - "inclusionai/ring-2.6-1t": { - "id": "inclusionai/ring-2.6-1t", - "name": "inclusionAI: Ring-2.6-1T", + "nousresearch/hermes-4-70b": { + "id": "nousresearch/hermes-4-70b", + "name": "Nous: Hermes 4 70B", "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": true, + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, "temperature": true, - "knowledge": "2025-12-31", - "release_date": "2026-05-07", - "last_updated": "2026-05-14", + "release_date": "2025-08-25", + "last_updated": "2026-03-15", "modalities": { "input": [ "text" @@ -142128,27 +133376,40 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 65000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.06 + "input": 0.13, + "output": 0.4, + "cache_read": 0.055 } - }, - "inclusionai/ring-1t": { - "id": "inclusionai/ring-1t", - "name": "Ring-1T", + } + } + }, + "lucidquery": { + "id": "lucidquery", + "env": [ + "LUCIDQUERY_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.lucidquery.com/v1", + "name": "LucidQuery", + "doc": "https://lucidquery.com/docs", + "models": { + "lucidnova-rf1-100b": { + "id": "lucidnova-rf1-100b", + "name": "LucidNova RF1 100B", "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "family": "nova", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-10-12", - "last_updated": "2025-10-12", + "temperature": false, + "knowledge": "2025-09-16", + "release_date": "2024-12-28", + "last_updated": "2025-09-10", "modalities": { "input": [ "text" @@ -142159,27 +133420,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 64000 + "context": 120000, + "output": 8000 }, - "status": "deprecated", "cost": { - "input": 0.56, - "output": 2.24, - "cache_read": 0.11 + "input": 2, + "output": 5 } }, - "inclusionai/ling-1t": { - "id": "inclusionai/ling-1t", - "name": "Ling-1T", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, - "reasoning": false, + "lucidquery-nexus-coder": { + "id": "lucidquery-nexus-coder", + "name": "LucidQuery Nexus Coder", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "lucid", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-10-09", - "last_updated": "2025-10-09", + "temperature": false, + "knowledge": "2025-08-01", + "release_date": "2025-09-01", + "last_updated": "2025-09-01", "modalities": { "input": [ "text" @@ -142190,87 +133451,104 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 64000 + "context": 250000, + "output": 60000 }, - "status": "deprecated", "cost": { - "input": 0.56, - "output": 2.24, - "cache_read": 0.11 + "input": 2, + "output": 5 } }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "lucidquery-agi-01-swift": { + "id": "lucidquery-agi-01-swift", + "name": "AGI-01 Swift", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "agi", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2026-06-05", + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 300000, + "output": 120000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 2.5, + "output": 15 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "lucidquery-agi-01-frontier": { + "id": "lucidquery-agi-01-frontier", + "name": "AGI-01 Frontier", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "agi", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2025-01-01", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "temperature": true, + "knowledge": "2026-06-05", + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ "text", - "image", - "video" + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 300000, + "output": 120000 + }, + "cost": { + "input": 4.5, + "output": 22 + } + } + } + }, + "meganova": { + "id": "meganova", + "env": [ + "MEGANOVA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.meganova.ai/v1", + "name": "Meganova", + "doc": "https://docs.meganova.ai", + "models": { + "meta-llama/Llama-3.3-70B-Instruct": { + "id": "meta-llama/Llama-3.3-70B-Instruct", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": true, + "temperature": true, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -142278,25 +133556,28 @@ }, "open_weights": true, "limit": { - "context": 262140, - "output": 262140 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.1, + "output": 0.3 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", + "moonshotai/Kimi-K2-Thinking": { + "id": "moonshotai/Kimi-K2-Thinking", "name": "Kimi K2 Thinking", "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01-01", + "knowledge": "2024-08", "release_date": "2025-11-06", "last_updated": "2025-11-06", "modalities": { @@ -142307,89 +133588,67 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 64000 + "context": 262144, + "output": 262144 }, - "status": "deprecated", "cost": { "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "output": 2.6 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", "name": "Kimi K2.5", "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "attachment": true, + "family": "kimi-k2", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "temperature": false, - "knowledge": "2025-01-01", + "temperature": true, + "knowledge": "2026-01", "release_date": "2026-01-27", "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.58, - "output": 3.02, - "cache_read": 0.1 + "input": 0.45, + "output": 2.8 } }, - "moonshotai/kimi-k3-free": { - "id": "moonshotai/kimi-k3-free", - "name": "Kimi K3 (Free)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "Qwen/Qwen2.5-VL-32B-Instruct": { + "id": "Qwen/Qwen2.5-VL-32B-Instruct", + "name": "Qwen2.5 VL 32B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -142397,29 +133656,32 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.2, + "output": 0.6 } }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "Qwen/Qwen3.5-Plus": { + "id": "Qwen/Qwen3.5-Plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-09-04", - "last_updated": "2025-09-04", + "knowledge": "2025-04", + "release_date": "2026-02", + "last_updated": "2026-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -142427,35 +133689,30 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 64000 + "context": 1000000, + "output": 65536 }, - "status": "deprecated", "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 0.4, + "output": 2.4, + "reasoning": 2.4 } }, - "moonshotai/kimi-k2.7-code-free": { - "id": "moonshotai/kimi-k2.7-code-free", - "name": "Kimi K2.7 Code (Free)", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -142463,27 +133720,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.09, + "output": 0.6 } }, - "moonshotai/kimi-k2-thinking-turbo": { - "id": "moonshotai/kimi-k2-thinking-turbo", - "name": "Kimi K2 Thinking Turbo", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "XiaomiMiMo/MiMo-V2-Flash": { + "id": "XiaomiMiMo/MiMo-V2-Flash", + "name": "MiMo V2 Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "knowledge": "2024-12-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text" @@ -142492,37 +133749,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 64000 + "context": 262144, + "output": 32000 }, - "status": "deprecated", "cost": { - "input": 1.15, - "output": 8, - "cache_read": 0.15 + "input": 0.1, + "output": 0.3 } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "mistralai/Mistral-Small-3.2-24B-Instruct-2506": { + "id": "mistralai/Mistral-Small-3.2-24B-Instruct-2506", + "name": "Mistral Small 3.2 24B Instruct", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -142530,32 +133783,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0, + "output": 0 } }, - "xiaomi/mimo-v2-flash": { - "id": "xiaomi/mimo-v2-flash", - "name": "MiMo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", + "mistralai/Mistral-Nemo-Instruct-2407": { + "id": "mistralai/Mistral-Nemo-Instruct-2407", + "name": "Mistral Nemo Instruct 2407", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12-01", - "release_date": "2025-12-16", - "last_updated": "2026-02-04", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text" @@ -142566,41 +133813,33 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 131072, "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01 + "input": 0.02, + "output": 0.04 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, + "zai-org/GLM-4.6": { + "id": "zai-org/GLM-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -142608,37 +133847,20 @@ }, "open_weights": true, "limit": { - "context": 1048576, + "context": 202752, "output": 131072 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08, - "tiers": [ - { - "input": 0.8, - "output": 4, - "cache_read": 0.16, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.8, - "output": 4, - "cache_read": 0.16 - } + "input": 0.45, + "output": 1.9 } }, - "xiaomi/mimo-v2-omni": { - "id": "xiaomi/mimo-v2-omni", - "name": "MiMo V2 Omni", - "description": "MiMo omni model for text, image, video, audio, and agents", - "family": "mimo", - "attachment": true, + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -142646,50 +133868,46 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 265000, - "output": 265000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08 + "input": 0.8, + "output": 2.56 } }, - "xiaomi/mimo-v2-pro": { - "id": "xiaomi/mimo-v2-pro", - "name": "MiMo V2 Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", + "zai-org/GLM-4.7": { + "id": "zai-org/GLM-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -142698,53 +133916,27 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 256000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 0.2, + "output": 0.8 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "deepseek-ai/DeepSeek-V3-0324": { + "id": "deepseek-ai/DeepSeek-V3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ "text" @@ -142755,43 +133947,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 163840, + "output": 163840 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } + "input": 0.25, + "output": 0.88 } }, - "minimax/minimax-m2.5-lightning": { - "id": "minimax/minimax-m2.5-lightning", - "name": "MiniMax M2.5 highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "knowledge": "2024-07", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -142800,34 +133979,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 163840, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 4.8, - "cache_read": 0.06, - "cache_write": 0.75 + "input": 0.5, + "output": 2.15 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "release_date": "2025-08-25", + "last_updated": "2025-08-25", "modalities": { "input": [ "text" @@ -142836,34 +134009,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.27, + "output": 1 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "deepseek-ai/DeepSeek-V3.2-Exp": { + "id": "deepseek-ai/DeepSeek-V3.2-Exp", + "name": "DeepSeek V3.2 Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "release_date": "2025-10-10", + "last_updated": "2025-10-10", "modalities": { "input": [ "text" @@ -142872,41 +134039,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 131070 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.3055, - "output": 1.2219 + "input": 0.27, + "output": 0.4 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "deepseek-ai/DeepSeek-V3.2": { + "id": "deepseek-ai/DeepSeek-V3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2025-12-03", + "last_updated": "2025-12-03", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -142914,30 +134071,29 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 512000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 164000, + "output": 164000 }, "cost": { - "input": 0.6, - "output": 2.4 + "input": 0.26, + "output": 0.38 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax M2.7 highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "attachment": true, + "MiniMaxAI/MiniMax-M2.1": { + "id": "MiniMaxAI/MiniMax-M2.1", + "name": "MiniMax M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -142946,32 +134102,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 131070 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 196608, + "output": 131072 }, "cost": { - "input": 0.611, - "output": 2.4439 + "input": 0.28, + "output": 1.2 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax M2.1", + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax M2.5", "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -142980,37 +134135,54 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 204800, + "output": 131072 }, "cost": { "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.38 + "output": 1.2 } - }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "attachment": false, + } + } + }, + "perplexity": { + "id": "perplexity", + "env": [ + "PERPLEXITY_API_KEY" + ], + "npm": "@ai-sdk/perplexity", + "name": "Perplexity", + "doc": "https://docs.perplexity.ai", + "models": { + "sonar-reasoning-pro": { + "id": "sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", + "family": "sonar-reasoning", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -143018,35 +134190,28 @@ }, "open_weights": false, "limit": { - "context": 204000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://zenmux.ai/api/anthropic/v1" + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.38 + "input": 2, + "output": 8 } }, - "stepfun/step-3": { - "id": "stepfun/step-3", - "name": "Step-3", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "sonar": { + "id": "sonar", + "name": "Sonar", + "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", + "family": "sonar", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-07-31", - "last_updated": "2025-07-31", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "image", "text" ], "output": [ @@ -143055,29 +134220,30 @@ }, "open_weights": false, "limit": { - "context": 65536, - "output": 64000 + "context": 128000, + "output": 4096 }, - "status": "deprecated", "cost": { - "input": 0.21, - "output": 0.57 + "input": 1, + "output": 1 } }, - "stepfun/step-3.5-flash": { - "id": "stepfun/step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, + "sonar-pro": { + "id": "sonar-pro", + "name": "Sonar Pro", + "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", + "family": "sonar-pro", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-02-02", - "last_updated": "2026-02-02", + "knowledge": "2025-09-01", + "release_date": "2024-01-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -143085,113 +134251,134 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 200000, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 3, + "output": 15 } }, - "stepfun/step-3.7-flash": { - "id": "stepfun/step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, + "sonar-deep-research": { + "id": "sonar-deep-research", + "name": "Perplexity Sonar Deep Research", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" ] } ], - "tool_call": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-02-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 1.15 + "input": 2, + "output": 8, + "reasoning": 3 } - }, - "stepfun/step-3.7-flash-free": { - "id": "stepfun/step-3.7-flash-free", - "name": "Step 3.7 Flash (Free)", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + } + } + }, + "amazon-bedrock": { + "id": "amazon-bedrock", + "env": [ + "AWS_ACCESS_KEY_ID", + "AWS_SECRET_ACCESS_KEY", + "AWS_REGION", + "AWS_BEARER_TOKEN_BEDROCK" + ], + "npm": "@ai-sdk/amazon-bedrock", + "name": "Amazon Bedrock", + "doc": "https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html", + "models": { + "global.anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "global.anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5 (Global)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "qwen/qwen3.5-flash": { - "id": "qwen/qwen3.5-flash", - "name": "Qwen3.5 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "global.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "global.anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5 (Global)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -143199,33 +134386,40 @@ }, "open_weights": false, "limit": { - "context": 1020000, - "output": 1020000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "jp.anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "jp.anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5 (JP)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -143233,37 +134427,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "qwen/qwen3.5-plus": { - "id": "qwen/qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "us.meta.llama4-scout-17b-instruct-v1:0": { + "id": "us.meta.llama4-scout-17b-instruct-v1:0", + "name": "Llama 4 Scout 17B Instruct (US)", + "description": "Open Llama with long-context vision for efficient multimodal agents", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -143273,28 +134458,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 3500000, + "output": 16384 }, "cost": { - "input": 0.8, - "output": 4.8 + "input": 0.17, + "output": 0.66 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3-Max-Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "minimax.minimax-m2": { + "id": "minimax.minimax-m2", + "name": "MiniMax M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-01-23", - "last_updated": "2026-01-23", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ "text" @@ -143303,30 +134489,45 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 204608, + "output": 128000 }, "cost": { - "input": 1.2, - "output": 6 + "input": 0.3, + "output": 1.2 } }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen3-Coder-Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "attachment": false, - "reasoning": false, + "anthropic.claude-opus-4-7": { + "id": "anthropic.claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -143335,33 +134536,48 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6-Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "eu.anthropic.claude-sonnet-4-6": { + "id": "eu.anthropic.claude-sonnet-4-6", + "name": "Claude Sonnet 4.6 (EU)", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -143373,92 +134589,80 @@ "output": 64000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 } }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "mistral.voxtral-small-24b-2507": { + "id": "mistral.voxtral-small-24b-2507", + "name": "Voxtral Small 24B 2507", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ "text", - "image", - "video" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 32000, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "cache_write": 0.5, - "tiers": [ - { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24, - "cache_write": 1.5, - "tier": { - "type": "context", - "size": 256000 - } - } + "input": 0.15, + "output": 0.35 + } + }, + "mistral.ministral-3-3b-instruct": { + "id": "mistral.ministral-3-3b-instruct", + "name": "Ministral 3 3B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" ], - "context_over_200k": { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24, - "cache_write": 1.5 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 0.1, + "output": 0.1 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "attachment": true, + "openai.gpt-oss-20b": { + "id": "openai.gpt-oss-20b", + "name": "gpt-oss-20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -143471,14 +134675,13 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -143486,43 +134689,52 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "output": 128000 + "context": 128000, + "output": 16384 }, "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/v1", + "shape": "responses" }, "cost": { - "input": 3.75, - "output": 18.75 + "input": 0.07, + "output": 0.3 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "anthropic.claude-opus-4-6-v1": { + "id": "anthropic.claude-opus-4-6-v1", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -143530,29 +134742,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, + "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" - }, "cost": { - "input": 45, - "output": 225 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/gpt-5.3-chat": { - "id": "openai/gpt-5.3-chat", - "name": "GPT-5.3 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "attachment": true, + "openai.gpt-oss-safeguard-20b": { + "id": "openai.gpt-oss-safeguard-20b", + "name": "GPT OSS Safeguard 20B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "release_date": "2025-10-29", + "last_updated": "2025-10-29", "modalities": { "input": [ "text" @@ -143564,40 +134775,40 @@ "open_weights": false, "limit": { "context": 128000, - "output": 16380 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.07, + "output": 0.2 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "anthropic.claude-opus-4-5-20251101-v1:0": { + "id": "anthropic.claude-opus-4-5-20251101-v1:0", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", - "high", - "xhigh" + "high" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-08-01", "modalities": { "input": [ "text", @@ -143610,43 +134821,45 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": false, - "reasoning": false, + "global.anthropic.claude-fable-5": { + "id": "global.anthropic.claude-fable-5", + "name": "Claude Fable 5 (Global)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -143654,23 +134867,22 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" - }, "cost": { - "input": 0.2, - "output": 1.25 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, + "openai.gpt-oss-120b-1:0": { + "id": "openai.gpt-oss-120b-1:0", + "name": "gpt-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -143683,15 +134895,13 @@ } ], "tool_call": true, - "temperature": false, - "knowledge": "2025-01-01", - "release_date": "2026-01-15", - "last_updated": "2026-01-15", + "structured_output": true, + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -143699,44 +134909,38 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.17 + "input": 0.15, + "output": 0.6 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -143744,44 +134948,33 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 200000, "output": 64000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" - }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "GPT-5 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "amazon.nova-pro-v1:0": { + "id": "amazon.nova-pro-v1:0", + "name": "Nova Pro", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -143789,40 +134982,70 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 64000 + "context": 300000, + "output": 8192 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "cost": { + "input": 0.8, + "output": 3.2, + "cache_read": 0.2 + } + }, + "qwen.qwen3-coder-next": { + "id": "qwen.qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-06", + "last_updated": "2026-02-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12 + "input": 0.22, + "output": 1.8 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2-Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "us.anthropic.claude-opus-4-7": { + "id": "us.anthropic.claude-opus-4-7", + "name": "Claude Opus 4.7 (US)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -143835,39 +135058,28 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" - }, "cost": { - "input": 21, - "output": 168 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "nvidia.nemotron-nano-9b-v2": { + "id": "nvidia.nemotron-nano-9b-v2", + "name": "NVIDIA Nemotron Nano 9B v2", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ "text" @@ -143878,66 +135090,51 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.06, + "output": 0.23 } }, - "openai/gpt-5.5-instant": { - "id": "openai/gpt-5.5-instant", - "name": "GPT-5.5 Instant", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, + "qwen.qwen3-32b-v1:0": { + "id": "qwen.qwen3-32b-v1:0", + "name": "Qwen3 32B (dense)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-12-01", - "release_date": "2026-05-05", - "last_updated": "2026-05-28", + "knowledge": "2024-04", + "release_date": "2025-09-18", + "last_updated": "2025-09-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 400000, - "output": 128000 + "context": 16384, + "output": 16384 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.15, + "output": 0.6 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "jp.anthropic.claude-sonnet-4-6": { + "id": "jp.anthropic.claude-sonnet-4-6", + "name": "Claude Sonnet 4.6 (JP)", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -143946,19 +135143,25 @@ "values": [ "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "image", "text", + "image", "pdf" ], "output": [ @@ -143967,30 +135170,29 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 64000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" - }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "attachment": true, - "reasoning": false, + "deepseek.r1-v1:0": { + "id": "deepseek.r1-v1:0", + "name": "DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-20", - "last_updated": "2026-03-20", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" @@ -144001,89 +135203,82 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 128000, + "output": 32768 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "cost": { + "input": 1.35, + "output": 5.4 + } + }, + "mistral.mistral-large-3-675b-instruct": { + "id": "mistral.mistral-large-3-675b-instruct", + "name": "Mistral Large 3", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 256000, + "output": 8192 }, "cost": { - "input": 0.75, - "output": 4.5 + "input": 0.5, + "output": 1.5 } }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", + "google.gemma-3-27b-it": { + "id": "google.gemma-3-27b-it", + "name": "Google Gemma 3 27B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-07", + "release_date": "2025-07-27", + "last_updated": "2025-07-27", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 202752, + "output": 8192 }, "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1, - "cache_write": 1.25, - "tiers": [ - { - "input": 2, - "output": 9, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 9, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 0.12, + "output": 0.2 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "anthropic.claude-sonnet-4-6": { + "id": "anthropic.claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -144092,19 +135287,25 @@ "values": [ "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "temperature": false, - "knowledge": "2025-01-01", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "structured_output": true, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "image", "text", + "image", "pdf" ], "output": [ @@ -144113,27 +135314,27 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 64000 }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" - }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.17 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "amazon.nova-2-lite-v1:0": { + "id": "amazon.nova-2-lite-v1:0", + "name": "Nova 2 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -144144,16 +135345,14 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -144161,63 +135360,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 128000, + "output": 4096 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.33, + "output": 2.75 } }, - "openai/gpt-5.1-chat": { - "id": "openai/gpt-5.1-chat", - "name": "GPT-5.1 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "attachment": true, + "openai.gpt-oss-safeguard-120b": { + "id": "openai.gpt-oss-safeguard-120b", + "name": "GPT OSS Safeguard 120B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2025-10-29", + "last_updated": "2025-10-29", "modalities": { "input": [ - "pdf", - "image", "text" ], "output": [ @@ -144227,50 +135391,28 @@ "open_weights": false, "limit": { "context": 128000, - "output": 64000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12 + "input": 0.15, + "output": 0.6 } }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "mistral.ministral-3-8b-instruct": { + "id": "mistral.ministral-3-8b-instruct", + "name": "Ministral 3 8B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -144278,39 +135420,19 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25 - } + "input": 0.15, + "output": 0.15 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1-Codex-Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "eu.anthropic.claude-opus-4-6-v1": { + "id": "eu.anthropic.claude-opus-4-6-v1", + "name": "Claude Opus 4.6 (EU)", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -144319,19 +135441,26 @@ "values": [ "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ + "text", "image", - "text" + "pdf" ], "output": [ "text" @@ -144339,23 +135468,21 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.03 + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "au.anthropic.claude-opus-4-6-v1": { + "id": "au.anthropic.claude-opus-4-6-v1", + "name": "AU Anthropic Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -144364,15 +135491,21 @@ "values": [ "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2025-05", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", @@ -144385,31 +135518,30 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://zenmux.ai/api/v1" + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12 + "input": 16.5, + "output": 82.5, + "cache_read": 1.65, + "cache_write": 20.625 } }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + "jp.anthropic.claude-sonnet-5": { + "id": "jp.anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5 (JP)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", "medium", "high", @@ -144421,9 +135553,9 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", @@ -144436,66 +135568,41 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } - } - } - }, - "unorouter": { - "id": "unorouter", - "env": [ - "UNOROUTER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.unorouter.com/v1", - "name": "UnoRouter", - "doc": "https://unorouter.com/models", - "models": { - "gpt-5.5:free": { - "id": "gpt-5.5:free", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + }, + "openai.gpt-oss-120b": { + "id": "openai.gpt-oss-120b", + "name": "gpt-oss-120b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -144503,29 +135610,47 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 16384 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/v1", + "shape": "responses" }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.6 } }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "global.anthropic.claude-sonnet-5": { + "id": "global.anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5 (Global)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", @@ -144538,31 +135663,45 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 1.8, - "output": 10.8 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "glm-4.5-flash:free": { - "id": "glm-4.5-flash:free", - "name": "GLM-4.5-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, + "global.anthropic.claude-opus-4-8": { + "id": "global.anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8 (Global)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -144570,27 +135709,28 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemma-4-31b-it:free": { - "id": "gemma-4-31b-it:free", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "us.meta.llama4-maverick-17b-instruct-v1:0": { + "id": "us.meta.llama4-maverick-17b-instruct-v1:0", + "name": "Llama 4 Maverick 17B Instruct (US)", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -144602,92 +135742,93 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.24, + "output": 0.97 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "openai.gpt-5.4": { + "id": "openai.gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.2675, - "output": 5.3368 - } - }, - "nemotron-3-ultra-550b-a55b:free": { - "id": "nemotron-3-ultra-550b-a55b:free", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", - "modalities": { - "input": [ - "text" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, + "context": 272000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" + }, "cost": { - "input": 0, - "output": 0 + "input": 2.75, + "output": 16.5, + "cache_read": 0.275 } }, - "gpt-5.4:free": { - "id": "gpt-5.4:free", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "anthropic.claude-sonnet-5": { + "id": "anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", @@ -144700,27 +135841,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "mistral.devstral-2-123b": { + "id": "mistral.devstral-2-123b", + "name": "Devstral 2 123B", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ "text" @@ -144731,68 +135873,78 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 256000, + "output": 8192 }, "cost": { - "input": 0.819, - "output": 3.276 + "input": 0.4, + "output": 2 } }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "zai.glm-4.7": { + "id": "zai.glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.1857, - "output": 1.1142 + "input": 0.6, + "output": 2.2 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "anthropic.claude-fable-5": { + "id": "anthropic.claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -144800,33 +135952,31 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 1.05, - "output": 8.4 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "writer.palmyra-x4-v1:0": { + "id": "writer.palmyra-x4-v1:0", + "name": "Palmyra X4", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "palmyra", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -144834,95 +135984,89 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 122880, + "output": 8192 }, "cost": { - "input": 0.425, - "output": 2.125 + "input": 2.5, + "output": 10 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "mistral.magistral-small-2509": { + "id": "mistral.magistral-small-2509", + "name": "Magistral Small 1.2", + "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", + "family": "magistral", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 40000 }, "cost": { - "input": 0.1875, - "output": 1.125 + "input": 0.5, + "output": 1.5 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen.qwen3-coder-480b-a35b-v1:0": { + "id": "qwen.qwen3-coder-480b-a35b-v1:0", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "structured_output": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-09-18", + "last_updated": "2025-09-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 1.44, - "output": 7.2 + "input": 0.22, + "output": 1.8 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "amazon.nova-micro-v1:0": { + "id": "amazon.nova-micro-v1:0", + "name": "Nova Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text" @@ -144931,94 +136075,121 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 8192 }, "cost": { - "input": 1.6001, - "output": 5.0288 + "input": 0.035, + "output": 0.14, + "cache_read": 0.00875 } }, - "step-3.7-flash:free": { - "id": "step-3.7-flash:free", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistral.pixtral-large-2502-v1:0": { + "id": "mistral.pixtral-large-2502-v1:0", + "name": "Pixtral Large (25.02)", + "description": "Mistral vision-language model for image understanding and multimodal chat", + "family": "mistral", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2025-04-08", + "last_updated": "2025-04-08", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 6 } }, - "deepseek-v4-flash:free": { - "id": "deepseek-v4-flash:free", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "us.anthropic.claude-opus-4-6-v1": { + "id": "us.anthropic.claude-opus-4-6-v1", + "name": "Claude Opus 4.6 (US)", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "jp.anthropic.claude-opus-4-7": { + "id": "jp.anthropic.claude-opus-4-7", + "name": "Claude Opus 4.7 (JP)", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -145031,59 +136202,71 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1.2, - "output": 6 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-v4-pro:free": { - "id": "deepseek-v4-pro:free", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "au.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "au.anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5 (AU)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "glm-5.2:free": { - "id": "glm-5.2:free", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "deepseek.v3-v1:0": { + "id": "deepseek.v3-v1:0", + "name": "DeepSeek-V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-07", + "release_date": "2025-09-18", + "last_updated": "2025-09-18", "modalities": { "input": [ "text" @@ -145094,170 +136277,178 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 163840, + "output": 81920 }, "cost": { - "input": 0, - "output": 0 + "input": 0.58, + "output": 1.68 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "anthropic.claude-opus-4-1-20250805-v1:0": { + "id": "anthropic.claude-opus-4-1-20250805-v1:0", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.8999, - "output": 1.7999 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "qwen3.5-397b-a17b:free": { - "id": "qwen3.5-397b-a17b:free", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "jp.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "jp.anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5 (JP)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "minimax-m2.7:free": { - "id": "minimax-m2.7:free", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "google.gemma-3-4b-it": { + "id": "google.gemma-3-4b-it", + "name": "Gemma 3 4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0, - "output": 0 + "input": 0.04, + "output": 0.08 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "eu.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "eu.anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5 (EU)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.0625, - "output": 0.125 + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 } - } - } - }, - "salad-cloud": { - "id": "salad-cloud", - "env": [ - "SALAD_CLOUD_API_KEY" - ], - "npm": "@saladtechnologies-oss/ai-sdk-provider", - "name": "SaladCloud AI Gateway", - "doc": "https://docs.salad.com/ai-gateway/explanation/overview", - "models": { - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Qwen MoE for agentic tasks, complex reasoning, code generation, and instruction following", + }, + "qwen.qwen3-vl-235b-a22b": { + "id": "qwen.qwen3-vl-235b-a22b", + "name": "Qwen/Qwen3-VL-235B-A22B-Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2025-10-04", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -145267,48 +136458,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 262144 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.09, - "output": 0.6 + "input": 0.3, + "output": 1.5 } - } - } - }, - "vispark": { - "id": "vispark", - "env": [ - "VISPARK_LAB_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.lab.vispark.in/v1", - "name": "Vispark", - "doc": "https://lab.vispark.in/#vision", - "models": { - "vispark/vision-large": { - "id": "vispark/vision-large", - "name": "Vision Large", - "description": "Most capable Vision model for complex reasoning, detailed media analysis, and structured output over a 1M-token context window.", - "attachment": true, + }, + "writer.palmyra-x5-v1:0": { + "id": "writer.palmyra-x5-v1:0", + "name": "Palmyra X5", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "palmyra", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-05-15", - "last_updated": "2026-09", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -145316,32 +136490,46 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 1040000, + "output": 8192 }, "cost": { - "input": 7.37, - "output": 22.11 + "input": 0.6, + "output": 6 } }, - "vispark/vision-small": { - "id": "vispark/vision-small", - "name": "Vision Small", - "description": "Fast, low-cost multimodal model for understanding text, images, audio, video, and PDFs, with tool calling and a 1M-token context window.", + "us.anthropic.claude-sonnet-4-6": { + "id": "us.anthropic.claude-sonnet-4-6", + "name": "Claude Sonnet 4.6 (US)", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-05-15", - "last_updated": "2026-09", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -145351,31 +136539,38 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 65536 + "output": 64000 }, "cost": { - "input": 1.05, - "output": 3.16 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "vispark/vision-medium": { - "id": "vispark/vision-medium", - "name": "Vision Medium", - "description": "Balanced multimodal model pairing a 1M-token context window with deeper reasoning for analysis, content creation, and tool use across text, image, audio, video, and PDF inputs.", + "au.anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "au.anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5 (AU)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-05-15", - "last_updated": "2026-09", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -145384,38 +136579,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 4.21, - "output": 12.63 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } - } - } - }, - "siliconflow-cn": { - "id": "siliconflow-cn", - "env": [ - "SILICONFLOW_CN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.siliconflow.cn/v1", - "name": "SiliconFlow (China)", - "doc": "https://cloud.siliconflow.com/models", - "models": { - "ByteDance-Seed/Seed-OSS-36B-Instruct": { - "id": "ByteDance-Seed/Seed-OSS-36B-Instruct", - "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "seed", + }, + "meta.llama3-3-70b-instruct-v1:0": { + "id": "meta.llama3-3-70b-instruct-v1:0", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-09-04", - "last_updated": "2025-11-25", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -145424,36 +136609,79 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.21, - "output": 0.57 + "input": 0.72, + "output": 0.72 } }, - "tencent/Hunyuan-A13B-Instruct": { - "id": "tencent/Hunyuan-A13B-Instruct", - "name": "tencent/Hunyuan-A13B-Instruct", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", - "attachment": false, + "zai.glm-5": { + "id": "zai.glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 101376 + }, + "cost": { + "input": 1, + "output": 3.2 + } + }, + "us.anthropic.claude-opus-4-8": { + "id": "us.anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8 (US)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -145461,100 +136689,120 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "Qwen/Qwen3.5-4B": { - "id": "Qwen/Qwen3.5-4B", - "name": "Qwen/Qwen3.5-4B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "global.anthropic.claude-opus-4-7": { + "id": "global.anthropic.claude-opus-4-7", + "name": "Claude Opus 4.7 (Global)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "Qwen/Qwen3.5-27B": { - "id": "Qwen/Qwen3.5-27B", - "name": "Qwen/Qwen3.5-27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "us.anthropic.claude-fable-5": { + "id": "us.anthropic.claude-fable-5", + "name": "Claude Fable 5 (US)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-25", - "last_updated": "2026-02-25", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.26, - "output": 2.09 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen/Qwen3.6-35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "amazon.nova-lite-v1:0": { + "id": "amazon.nova-lite-v1:0", + "name": "Nova Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2024-10", + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text", @@ -145565,38 +136813,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 300000, + "output": 8192 }, "cost": { - "input": 0.23, - "output": 1.86 + "input": 0.06, + "output": 0.24, + "cache_read": 0.015 } }, - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "us.anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "us.anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5 (US)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "budget_tokens", - "min": 128, - "max": 32768 + "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-11-25", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -145604,39 +136855,43 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.13, - "output": 0.6 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "Qwen/Qwen3-14B": { - "id": "Qwen/Qwen3-14B", - "name": "Qwen/Qwen3-14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "xai.grok-4.3": { + "id": "xai.grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", + "release_date": "2026-04-17", + "last_updated": "2026-06-28", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -145644,38 +136899,35 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" }, "cost": { - "input": 0.07, - "output": 0.28 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "Qwen/Qwen3-8B": { - "id": "Qwen/Qwen3-8B", - "name": "Qwen/Qwen3-8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "mistral.voxtral-mini-3b-2507": { + "id": "mistral.voxtral-mini-3b-2507", + "name": "Voxtral Mini 3B 2507", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ + "audio", "text" ], "output": [ @@ -145684,36 +136936,28 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.06, - "output": 0.06 + "input": 0.04, + "output": 0.04 } }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen/Qwen3-32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "moonshot.kimi-k2-thinking": { + "id": "moonshot.kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2025-11-25", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text" @@ -145722,38 +136966,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 262143, + "output": 16000 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 0.6, + "output": 2.5 } }, - "Qwen/Qwen3.5-35B-A3B": { - "id": "Qwen/Qwen3.5-35B-A3B", - "name": "Qwen/Qwen3.5-35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "meta.llama3-1-70b-instruct-v1:0": { + "id": "meta.llama3-1-70b-instruct-v1:0", + "name": "Llama 3.1 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-25", - "last_updated": "2026-02-25", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -145761,36 +136998,30 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.23, - "output": 1.86 + "input": 0.72, + "output": 0.72 } }, - "Qwen/Qwen3.5-122B-A10B": { - "id": "Qwen/Qwen3.5-122B-A10B", - "name": "Qwen/Qwen3.5-122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "us.deepseek.r1-v1:0": { + "id": "us.deepseek.r1-v1:0", + "name": "DeepSeek-R1 (US)", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -145798,134 +137029,167 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.29, - "output": 2.32 + "input": 1.35, + "output": 5.4 } }, - "Qwen/Qwen3.5-9B": { - "id": "Qwen/Qwen3.5-9B", - "name": "Qwen/Qwen3.5-9B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "global.anthropic.claude-sonnet-4-6": { + "id": "global.anthropic.claude-sonnet-4-6", + "name": "Claude Sonnet 4.6 (Global)", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.22, - "output": 1.74 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen/Qwen3.5-397B-A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.29, - "output": 1.74 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "name": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "moonshotai.kimi-k2.5": { + "id": "moonshotai.kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-31", - "last_updated": "2025-11-25", + "release_date": "2026-02-06", + "last_updated": "2026-02-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 262143, + "output": 16000 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.6, + "output": 3 } }, - "Qwen/Qwen3-VL-32B-Instruct": { - "id": "Qwen/Qwen3-VL-32B-Instruct", - "name": "Qwen/Qwen3-VL-32B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "au.anthropic.claude-opus-4-8": { + "id": "au.anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8 (AU)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-21", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -145933,29 +137197,32 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "Qwen/Qwen2.5-7B-Instruct": { - "id": "Qwen/Qwen2.5-7B-Instruct", - "name": "Qwen/Qwen2.5-7B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "nvidia.nemotron-nano-12b-v2": { + "id": "nvidia.nemotron-nano-12b-v2", + "name": "NVIDIA Nemotron Nano 12B v2 VL BF16", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2025-11-25", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -145963,26 +137230,28 @@ }, "open_weights": false, "limit": { - "context": 33000, - "output": 4000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.05, - "output": 0.05 + "input": 0.2, + "output": 0.6 } }, - "Qwen/Qwen3-Coder-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "name": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "zai.glm-4.7-flash": { + "id": "zai.glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-01", - "last_updated": "2025-11-25", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -145991,63 +137260,80 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 200000, + "output": 131072 }, "cost": { "input": 0.07, - "output": 0.28 + "output": 0.4 } }, - "Qwen/Qwen2.5-72B-Instruct": { - "id": "Qwen/Qwen2.5-72B-Instruct", - "name": "Qwen/Qwen2.5-72B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "meta.llama4-scout-17b-instruct-v1:0": { + "id": "meta.llama4-scout-17b-instruct-v1:0", + "name": "Llama 4 Scout 17B Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2025-11-25", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 33000, - "output": 4000 + "context": 3500000, + "output": 16384 }, "cost": { - "input": 0.59, - "output": 0.59 + "input": 0.17, + "output": 0.66 } }, - "Qwen/Qwen3-VL-30B-A3B-Thinking": { - "id": "Qwen/Qwen3-VL-30B-A3B-Thinking", - "name": "Qwen/Qwen3-VL-30B-A3B-Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "au.anthropic.claude-sonnet-5": { + "id": "au.anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5 (AU)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-10-11", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -146055,17 +137341,19 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.29, - "output": 1 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "Qwen/Qwen3-30B-A3B-Instruct-2507": { - "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen/Qwen3-30B-A3B-Instruct-2507", + "qwen.qwen3-235b-a22b-2507-v1:0": { + "id": "qwen.qwen3-235b-a22b-2507-v1:0", + "name": "Qwen3 235B A22B 2507", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, @@ -146073,8 +137361,9 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-30", - "last_updated": "2025-11-25", + "knowledge": "2024-04", + "release_date": "2025-09-18", + "last_updated": "2025-09-18", "modalities": { "input": [ "text" @@ -146083,95 +137372,40 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.09, - "output": 0.3 + "input": 0.22, + "output": 0.88 } }, - "Qwen/Qwen3-VL-32B-Thinking": { - "id": "Qwen/Qwen3-VL-32B-Thinking", - "name": "Qwen/Qwen3-VL-32B-Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "eu.anthropic.claude-haiku-4-5-20251001-v1:0": { + "id": "eu.anthropic.claude-haiku-4-5-20251001-v1:0", + "name": "Claude Haiku 4.5 (EU)", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-21", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.2, - "output": 1.5 - } - }, - "Qwen/Qwen3-VL-8B-Instruct": { - "id": "Qwen/Qwen3-VL-8B-Instruct", - "name": "Qwen/Qwen3-VL-8B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, + "knowledge": "2025-02-28", "release_date": "2025-10-15", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.18, - "output": 0.68 - } - }, - "Qwen/Qwen3-VL-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-VL-30B-A3B-Instruct", - "name": "Qwen/Qwen3-VL-30B-A3B-Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-05", - "last_updated": "2025-11-25", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -146179,33 +137413,38 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.29, - "output": 1 + "input": 1.1, + "output": 5.5, + "cache_read": 0.11, + "cache_write": 1.375 } }, - "deepseek-ai/DeepSeek-R1": { - "id": "deepseek-ai/DeepSeek-R1", - "name": "deepseek-ai/DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "openai.gpt-oss-20b-1:0": { + "id": "openai.gpt-oss-20b-1:0", + "name": "gpt-oss-20b", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-11-25", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -146216,98 +137455,119 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.5, - "output": 2.18 + "input": 0.07, + "output": 0.3 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "jp.anthropic.claude-opus-4-8": { + "id": "jp.anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8 (JP)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.003 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-ai/DeepSeek-OCR": { - "id": "deepseek-ai/DeepSeek-OCR", - "name": "deepseek-ai/DeepSeek-OCR", - "description": "OCR model for extracting structured text from documents and screenshots", + "anthropic.claude-opus-4-8": { + "id": "anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-10-20", - "last_updated": "2025-10-20", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 8192, - "output": 8192 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "deepseek-ai/DeepSeek-V4-Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", + "qwen.qwen3-coder-30b-a3b-v1:0": { + "id": "qwen.qwen3-coder-30b-a3b-v1:0", + "name": "Qwen3 Coder 30B A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-04", + "release_date": "2025-09-18", + "last_updated": "2025-09-18", "modalities": { "input": [ "text" @@ -146316,34 +137576,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1049000, - "output": 393000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.145 + "input": 0.15, + "output": 0.6 } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "deepseek-ai/DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "qwen.qwen3-next-80b-a3b": { + "id": "qwen.qwen3-next-80b-a3b", + "name": "Qwen/Qwen3-Next-80B-A3B-Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", + "release_date": "2025-09-18", + "last_updated": "2025-11-25", "modalities": { "input": [ "text" @@ -146354,34 +137608,44 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.27, - "output": 0.42 + "input": 0.14, + "output": 1.4 } }, - "deepseek-ai/DeepSeek-V3.1-Terminus": { - "id": "deepseek-ai/DeepSeek-V3.1-Terminus", - "name": "deepseek-ai/DeepSeek-V3.1-Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "openai.gpt-5.5": { + "id": "openai.gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -146389,60 +137653,53 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 - }, - "cost": { - "input": 0.27, - "output": 1 - } - }, - "deepseek-ai/DeepSeek-V3": { - "id": "deepseek-ai/DeepSeek-V3", - "name": "deepseek-ai/DeepSeek-V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 272000, + "output": 128000 }, - "open_weights": false, - "limit": { - "context": 164000, - "output": 164000 + "provider": { + "npm": "@ai-sdk/amazon-bedrock/mantle", + "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", + "shape": "responses" }, "cost": { - "input": 0.25, - "output": 1 + "input": 5.5, + "output": 33, + "cache_read": 0.55 } }, - "stepfun-ai/Step-3.5-Flash": { - "id": "stepfun-ai/Step-3.5-Flash", - "name": "stepfun-ai/Step-3.5-Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "family": "step", - "attachment": false, + "au.anthropic.claude-sonnet-4-6": { + "id": "au.anthropic.claude-sonnet-4-6", + "name": "AU Anthropic Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "knowledge": "2025-08", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -146450,29 +137707,49 @@ }, "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 3.3, + "output": 16.5, + "cache_read": 0.33, + "cache_write": 4.125 } }, - "inclusionAI/Ling-flash-2.0": { - "id": "inclusionAI/Ling-flash-2.0", - "name": "inclusionAI/Ling-flash-2.0", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", - "attachment": false, - "reasoning": false, + "us.anthropic.claude-sonnet-5": { + "id": "us.anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5 (US)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-09-18", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -146480,70 +137757,78 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 131000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.57 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "us.anthropic.claude-opus-4-5-20251101-v1:0": { + "id": "us.anthropic.claude-opus-4-5-20251101-v1:0", + "name": "Claude Opus 4.5 (US)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", - "max" + "low", + "medium", + "high" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-08-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1049000, - "output": 262000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "zai-org/GLM-4.5-Air": { - "id": "zai-org/GLM-4.5-Air", - "name": "zai-org/GLM-4.5-Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", + "minimax.minimax-m2.5": { + "id": "minimax.minimax-m2.5", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-11-25", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -146552,38 +137837,45 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131000, - "output": 131000 + "context": 196608, + "output": 98304 }, "cost": { - "input": 0.14, - "output": 0.86 + "input": 0.3, + "output": 1.2 } }, - "Pro/deepseek-ai/DeepSeek-R1": { - "id": "Pro/deepseek-ai/DeepSeek-R1", - "name": "Pro/deepseek-ai/DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "eu.anthropic.claude-opus-4-8": { + "id": "eu.anthropic.claude-opus-4-8", + "name": "Claude Opus 4.8 (EU)", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-11-25", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -146591,34 +137883,45 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 2.18 + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 } }, - "Pro/deepseek-ai/DeepSeek-V3.2": { - "id": "Pro/deepseek-ai/DeepSeek-V3.2", - "name": "Pro/deepseek-ai/DeepSeek-V3.2", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", - "attachment": false, + "eu.anthropic.claude-opus-4-7": { + "id": "eu.anthropic.claude-opus-4-7", + "name": "Claude Opus 4.7 (EU)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -146626,26 +137929,28 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 0.42 + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 } }, - "Pro/deepseek-ai/DeepSeek-V3": { - "id": "Pro/deepseek-ai/DeepSeek-V3", - "name": "Pro/deepseek-ai/DeepSeek-V3", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", + "meta.llama3-1-8b-instruct-v1:0": { + "id": "meta.llama3-1-8b-instruct-v1:0", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2025-11-25", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -146654,36 +137959,39 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 164000, - "output": 164000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.22, + "output": 0.22 } }, - "Pro/deepseek-ai/DeepSeek-V3.1-Terminus": { - "id": "Pro/deepseek-ai/DeepSeek-V3.1-Terminus", - "name": "Pro/deepseek-ai/DeepSeek-V3.1-Terminus", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", - "attachment": false, + "us.anthropic.claude-opus-4-1-20250805-v1:0": { + "id": "us.anthropic.claude-opus-4-1-20250805-v1:0", + "name": "Claude Opus 4.1 (US)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-11-25", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -146691,63 +137999,28 @@ }, "open_weights": false, "limit": { - "context": 164000, - "output": 164000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.27, - "output": 1 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "Pro/MiniMaxAI/MiniMax-M2.5": { - "id": "Pro/MiniMaxAI/MiniMax-M2.5", - "name": "Pro/MiniMaxAI/MiniMax-M2.5", - "description": "Frontier MiniMax model for engineering, office tasks, and agentic reasoning", - "family": "minimax", - "attachment": false, + "meta.llama4-maverick-17b-instruct-v1:0": { + "id": "meta.llama4-maverick-17b-instruct-v1:0", + "name": "Llama 4 Maverick 17B Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 192000, - "output": 131000 - }, - "cost": { - "input": 0.3, - "output": 1.22 - } - }, - "Pro/moonshotai/Kimi-K2.6": { - "id": "Pro/moonshotai/Kimi-K2.6", - "name": "Pro/moonshotai/Kimi-K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -146759,67 +138032,76 @@ }, "open_weights": true, "limit": { - "context": 262000, - "output": 262000 + "context": 1000000, + "output": 16384 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.24, + "output": 0.97 } }, - "Pro/moonshotai/Kimi-K2.5": { - "id": "Pro/moonshotai/Kimi-K2.5", - "name": "Pro/moonshotai/Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi", - "attachment": false, + "global.anthropic.claude-opus-4-5-20251101-v1:0": { + "id": "global.anthropic.claude-opus-4-5-20251101-v1:0", + "name": "Claude Opus 4.5 (Global)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-08-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.45, - "output": 2.25, - "cache_read": 0.07 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "Pro/zai-org/GLM-5.1": { - "id": "Pro/zai-org/GLM-5.1", - "name": "Pro/zai-org/GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "nvidia.nemotron-super-3-120b": { + "id": "nvidia.nemotron-super-3-120b", + "name": "NVIDIA Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-04-08", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text" @@ -146830,68 +138112,80 @@ }, "open_weights": true, "limit": { - "context": 205000, - "output": 205000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "input": 0.15, + "output": 0.65 } }, - "Pro/zai-org/GLM-5": { - "id": "Pro/zai-org/GLM-5", - "name": "Pro/zai-org/GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "eu.anthropic.claude-sonnet-5": { + "id": "eu.anthropic.claude-sonnet-5", + "name": "Claude Sonnet 5 (EU)", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 205000, - "output": 205000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 3.2 + "input": 2.2, + "output": 11, + "cache_read": 0.22, + "cache_write": 2.75 } }, - "PaddlePaddle/PaddleOCR-VL-1.5": { - "id": "PaddlePaddle/PaddleOCR-VL-1.5", - "name": "PaddlePaddle/PaddleOCR-VL-1.5", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, - "reasoning": false, - "tool_call": false, + "nvidia.nemotron-nano-3-30b": { + "id": "nvidia.nemotron-nano-3-30b", + "name": "NVIDIA Nemotron Nano 3 30B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-01-29", - "last_updated": "2026-01-29", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -146899,78 +138193,96 @@ }, "open_weights": true, "limit": { - "context": 16384, - "output": 16384 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0, - "output": 0 + "input": 0.06, + "output": 0.24 } - } - } - }, - "regolo-ai": { - "id": "regolo-ai", - "env": [ - "REGOLO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.regolo.ai/v1", - "name": "Regolo AI", - "doc": "https://docs.regolo.ai/", - "models": { - "qwen3.8-27b": { - "id": "qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", + }, + "eu.anthropic.claude-opus-4-5-20251101-v1:0": { + "id": "eu.anthropic.claude-opus-4-5-20251101-v1:0", + "name": "Claude Opus 4.5 (EU)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-08-01", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 120000, - "output": 120000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.58, - "output": 2.42 + "input": 5.5, + "output": 27.5, + "cache_read": 0.55, + "cache_write": 6.875 } }, - "brick-v1-beta": { - "id": "brick-v1-beta", - "name": "Brick v1 Beta", - "description": "Semantic router by Regolo.ai that directs each request to the most suitable model, optimizing costs and performance", - "family": "model-router", + "global.anthropic.claude-opus-4-6-v1": { + "id": "global.anthropic.claude-opus-4-6-v1", + "name": "Claude Opus 4.6 (Global)", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -146978,87 +138290,102 @@ }, "open_weights": false, "limit": { - "context": 100000, - "output": 15000 + "context": 1000000, + "output": 128000 }, - "status": "beta", "cost": { - "input": 0, - "output": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen3.5-122b": { - "id": "qwen3.5-122b", - "name": "Qwen3.5-122B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "us.anthropic.claude-sonnet-4-5-20250929-v1:0": { + "id": "us.anthropic.claude-sonnet-4-5-20250929-v1:0", + "name": "Claude Sonnet 4.5 (US)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.9, - "output": 3.6 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "Qwen3-Coder-Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "google.gemma-3-12b-it": { + "id": "google.gemma-3-12b-it", + "name": "Google Gemma 3 12B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "knowledge": "2024-12", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.049999999999999996, + "output": 0.09999999999999999 } }, - "apertus-70b": { - "id": "apertus-70b", - "name": "Apertus 70B", - "description": "Fully open 70B multilingual LLM supporting 1800+ languages with 65K context. Trained on 15T tokens of compliant open data. Apache 2.0, EU AI Act compliant.", + "minimax.minimax-m2.1": { + "id": "minimax.minimax-m2.1", + "name": "MiniMax M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-02", - "last_updated": "2025-09-02", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -147069,64 +138396,74 @@ }, "open_weights": true, "limit": { - "context": 30000, - "output": 30000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.46, - "output": 2.42 + "input": 0.3, + "output": 1.2 } }, - "glm5.2": { - "id": "glm5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "eu.anthropic.claude-fable-5": { + "id": "eu.anthropic.claude-fable-5", + "name": "Claude Fable 5 (EU)", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 96000, - "output": 96000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 2.31, - "output": 6 + "input": 11, + "output": 55, + "cache_read": 1.1, + "cache_write": 13.75 } }, - "qwen3-reranker-4b": { - "id": "qwen3-reranker-4b", - "name": "Qwen3-Reranker-4B", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "qwen", + "deepseek.v3.2": { + "id": "deepseek.v3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2026-02-06", + "last_updated": "2026-02-06", "modalities": { "input": [ "text" @@ -147137,77 +138474,100 @@ }, "open_weights": true, "limit": { - "context": 32768, - "output": 8192 + "context": 163840, + "output": 81920 }, "cost": { - "input": 0.12, - "output": 0.12 + "input": 0.62, + "output": 1.85 } }, - "deepseek-ocr-2": { - "id": "deepseek-ocr-2", - "name": "DeepSeek OCR 2", - "description": "High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes", - "attachment": true, + "mistral.ministral-3-14b-instruct": { + "id": "mistral.ministral-3-14b-instruct", + "name": "Ministral 14B 3.0", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, "reasoning": false, - "tool_call": false, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 4000, - "output": 4000 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 0.2 } - }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + } + } + }, + "umans-ai": { + "id": "umans-ai", + "env": [ + "UMANS_AI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.code.umans.ai/v1", + "name": "Umans AI", + "doc": "https://app.umans.ai/offers/code/docs/orgs", + "models": { + "umans-kimi-k2.7": { + "id": "umans-kimi-k2.7", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.7 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "gemma4-31b": { - "id": "gemma4-31b", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, + "umans-glm-5.1": { + "id": "umans-glm-5.1", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -147215,10 +138575,13 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text", @@ -147230,24 +138593,32 @@ }, "open_weights": true, "limit": { - "context": 100000, - "output": 100000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.46, - "output": 2.42 + "input": 1.4, + "output": 4.4, + "cache_read": 0.29 } }, - "brick-complexity-pro": { - "id": "brick-complexity-pro", - "name": "Brick Complexity Pro", - "description": "Complexity classifier that powers the Brick semantic router by extracting query difficulty", + "umans-coder": { + "id": "umans-coder", + "name": "Umans Coder", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", @@ -147257,24 +138628,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 100000, - "output": 15000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.12, - "output": 0.46 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT-OSS-20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "umans-flash": { + "id": "umans-flash", + "name": "Umans Flash", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -147285,12 +138660,17 @@ } ], "tool_call": true, - "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -147298,27 +138678,43 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.4, - "output": 1.8 + "input": 0.15, + "output": 1, + "cache_read": 0.05 } }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5-9B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "umans-glm-5.2": { + "id": "umans-glm-5.2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", - "modalities": { + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { "input": [ "text", "image" @@ -147329,25 +138725,37 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 8192 + "context": 405504, + "output": 131072 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } - }, - "qwen3-embedding-8b": { - "id": "qwen3-embedding-8b", - "name": "Qwen3-Embedding-8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", + } + } + }, + "togetherai": { + "id": "togetherai", + "env": [ + "TOGETHER_API_KEY" + ], + "npm": "@ai-sdk/togetherai", + "name": "Together AI", + "doc": "https://docs.together.ai/docs/serverless-models", + "models": { + "LiquidAI/LFM2-24B-A2B": { + "id": "LiquidAI/LFM2-24B-A2B", + "name": "LFM2-24B-A2B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "liquid", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "temperature": true, + "release_date": "2026-02-25", + "last_updated": "2026-02-25", "modalities": { "input": [ "text" @@ -147359,94 +138767,54 @@ "open_weights": true, "limit": { "context": 32768, - "output": 8192 - }, - "cost": { - "input": 0.1, - "output": 0.1 - } - }, - "mistral-small-4-119b": { - "id": "mistral-small-4-119b", - "name": "Mistral Small 4 119B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-15", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 16384 + "output": 32768 }, "cost": { - "input": 0.75, - "output": 3 + "input": 0.03, + "output": 0.12 } }, - "qwen-image": { - "id": "qwen-image", - "name": "Qwen-Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", + "meta-llama/Meta-Llama-3-8B-Instruct-Lite": { + "id": "meta-llama/Meta-Llama-3-8B-Instruct-Lite", + "name": "Meta Llama 3 8B Instruct Lite", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "release_date": "2024-04-18", + "last_updated": "2024-04-18", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 8192, - "output": 4096 + "output": 8192 }, "cost": { - "input": 0.5, - "output": 2 + "input": 0.14, + "output": 0.14 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT-OSS-120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "meta-llama/Llama-3.3-70B-Instruct-Turbo": { + "id": "meta-llama/Llama-3.3-70B-Instruct-Turbo", + "name": "Llama 3.3 70B", + "description": "Compact Llama instruction model for fast chat and local deployment", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" @@ -147455,61 +138823,21 @@ "text" ] }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1, - "output": 4.2 - } - }, - "faster-whisper-large-v3": { - "id": "faster-whisper-large-v3", - "name": "Faster Whisper Large v3", - "description": "Open Whisper checkpoint for robust multilingual transcription and captioning", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, "open_weights": true, "limit": { - "context": 448, - "output": 4096 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 1.04, + "output": 1.04 } - } - } - }, - "xiaomi-token-plan-ams": { - "id": "xiaomi-token-plan-ams", - "env": [ - "XIAOMI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://token-plan-ams.xiaomimimo.com/v1", - "name": "Xiaomi Token Plan (Europe)", - "doc": "https://platform.xiaomimimo.com/#/docs", - "models": { - "mimo-v2.6-pro": { - "id": "mimo-v2.6-pro", - "name": "MiMo-V2.6-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + }, + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -147518,17 +138846,15 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ @@ -147537,21 +138863,21 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 131000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.2, + "output": 4.5, + "cache_read": 0.2 } }, - "mimo-v2.6-flash": { - "id": "mimo-v2.6-flash", - "name": "MiMo-V2.6-Flash", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi K2.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "kimi-k2", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -147559,18 +138885,15 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "interleaved": true, "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "knowledge": "2026-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" @@ -147578,66 +138901,65 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.5, + "output": 2.8 } }, - "mimo-v2.5-tts-voiceclone": { - "id": "mimo-v2.5-tts-voiceclone", - "name": "MiMo-V2.5-TTS-VoiceClone", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Kimi coding model for software agents, refactors, and repository reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-14", + "last_updated": "2026-06-14", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "google/gemma-4-31B-it": { + "id": "google/gemma-4-31B-it", + "name": "Gemma 4 31B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -147645,63 +138967,55 @@ }, "open_weights": true, "limit": { - "context": 1048576, + "context": 262144, "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.39, + "output": 0.97 } }, - "mimo-v2-tts": { - "id": "mimo-v2-tts", - "name": "MiMo-V2-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", + "google/gemma-3n-E4B-it": { + "id": "google/gemma-3n-E4B-it", + "name": "Gemma 3N E4B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": true, + "temperature": true, + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.06, + "output": 0.12 } }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", + "Qwen/Qwen3.7-Max": { + "id": "Qwen/Qwen3.7-Max", + "name": "Qwen3.7 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-05-21", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" @@ -147712,98 +139026,98 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 500000 }, - "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.25, + "output": 3.75 } }, - "mimo-v2.5-tts": { - "id": "mimo-v2.5-tts", - "name": "MiMo-V2.5-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", + "Qwen/Qwen3.6-Plus": { + "id": "Qwen/Qwen3.6-Plus", + "name": "Qwen3.6 Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 1000000, + "output": 500000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.5, + "output": 3 } }, - "mimo-v2.5-tts-voicedesign": { - "id": "mimo-v2.5-tts-voicedesign", - "name": "MiMo-V2.5-TTS-VoiceDesign", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", + "Qwen/Qwen3.5-397B-A17B": { + "id": "Qwen/Qwen3.5-397B-A17B", + "name": "Qwen3.5 397B A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "release_date": "2026-02-16", + "last_updated": "2026-06-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "audio" + "text" ] }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 262144, + "output": 130000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.6, + "output": 3.6 } }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "Qwen/Qwen3-Coder-Next-FP8": { + "id": "Qwen/Qwen3-Coder-Next-FP8", + "name": "Qwen3 Coder Next FP8", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2026-02-03", + "release_date": "2026-02-03", + "last_updated": "2026-02-03", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -147811,52 +139125,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.5, + "output": 1.2 } - } - } - }, - "inceptron": { - "id": "inceptron", - "env": [ - "INCEPTRON_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.inceptron.io/v1", - "name": "Inceptron", - "doc": "https://docs.inceptron.io", - "models": { - "deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + }, + "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8": { + "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Legacy model retained for compatibility with older integrations", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -147867,37 +139156,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 262144, + "output": 262144 }, + "status": "deprecated", "cost": { - "input": 0.13, - "output": 0.28, - "cache_read": 0.03, - "cache_write": 0 + "input": 2, + "output": 2 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen3-235B-A22B-Instruct-2507-tput": { + "id": "Qwen/Qwen3-235B-A22B-Instruct-2507-tput", + "name": "Qwen3 235B A22B Instruct 2507 FP8", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-07", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -147909,33 +139191,25 @@ "output": 262144 }, "cost": { - "input": 0.53, - "output": 3.39, - "cache_read": 0.17, - "cache_write": 0 + "input": 0.2, + "output": 0.6 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "Qwen/Qwen2.5-7B-Instruct-Turbo": { + "id": "Qwen/Qwen2.5-7B-Instruct-Turbo", + "name": "Qwen 2.5 7B Instruct Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -147943,43 +139217,35 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.66, - "output": 3.4, - "cache_read": 0.18, - "cache_write": 0 + "input": 0.3, + "output": 0.3 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM 5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "Qwen/Qwen3.5-9B": { + "id": "Qwen/Qwen3.5-9B", + "name": "Qwen3.5 9B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -147987,55 +139253,36 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.71, - "output": 2.35, - "cache_read": 0.12, - "cache_write": 0 + "input": 0.17, + "output": 0.25 } - } - } - }, - "upstage": { - "id": "upstage", - "env": [ - "UPSTAGE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.upstage.ai/v1/solar", - "name": "Upstage", - "doc": "https://developers.upstage.ai/docs/apis/chat", - "models": { - "solar-pro4": { - "id": "solar-pro4", - "name": "Solar Pro 4", - "description": "Upstage's flagship model, specialized for agentic use", - "family": "solar-pro", + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-08-06", - "last_updated": "2026-08-06", + "knowledge": "2025-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -148044,22 +139291,21 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 524288, + "context": 131072, "output": 131072 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.15, + "output": 0.6 } }, - "solar-pro3": { - "id": "solar-pro3", - "name": "solar-pro3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "solar-pro", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -148073,10 +139319,10 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -148085,28 +139331,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, - "output": 8192 + "output": 131072 }, "cost": { - "input": 0.25, - "output": 0.25 + "input": 0.05, + "output": 0.2 } }, - "solar-mini": { - "id": "solar-mini", - "name": "solar-mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "solar-mini", + "pearl-ai/gemma-4-31b-it": { + "id": "pearl-ai/gemma-4-31b-it", + "name": "Pearl AI Gemma 4 31B Instruct", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-06-12", - "last_updated": "2025-04-22", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -148117,35 +139363,31 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 4096 + "context": 32000, + "output": 32000 }, "cost": { - "input": 0.15, - "output": 0.15 + "input": 0.28, + "output": 0.86 } }, - "solar-pro2": { - "id": "solar-pro2", - "name": "solar-pro2", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "solar-pro", + "nvidia/nemotron-3-ultra-550b-a55b": { + "id": "nvidia/nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -148154,41 +139396,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 8192 + "context": 512300, + "output": 512300 }, "cost": { - "input": 0.25, - "output": 0.25 + "input": 0.6, + "output": 3.6, + "cache_read": 0.2 } - } - } - }, - "vultr": { - "id": "vultr", - "env": [ - "VULTR_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.vultrinference.com/v1", - "name": "Vultr", - "doc": "https://api.vultrinference.com/", - "models": { - "XiaomiMiMo/MiMo-V2.5-Pro": { - "id": "XiaomiMiMo/MiMo-V2.5-Pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + }, + "deepcogito/cogito-v2-1-671b": { + "id": "deepcogito/cogito-v2-1-671b", + "name": "Cogito v2.1 671B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "cogito", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text" @@ -148197,35 +139431,39 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0.55, - "output": 1.65 + "input": 1.25, + "output": 1.25 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -148233,33 +139471,44 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 2 + "input": 1, + "output": 3.2 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-06-16", + "last_updated": "2026-06-16", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -148268,27 +139517,32 @@ "open_weights": true, "limit": { "context": 262144, - "output": 65536 + "output": 164000 }, "cost": { - "input": 0.3, - "output": 2 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-11", + "release_date": "2026-04-07", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" @@ -148299,26 +139553,27 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0.3, - "output": 1 + "input": 1.4, + "output": 4.4 } }, - "MiniMaxAI/MiniMax-M2.7": { - "id": "MiniMaxAI/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "deepseek-ai/DeepSeek-R1": { + "id": "deepseek-ai/DeepSeek-R1", + "name": "DeepSeek-R1", + "description": "Legacy model retained for compatibility with older integrations", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-03-24", "modalities": { "input": [ "text" @@ -148329,31 +139584,32 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 163839, + "output": 163839 }, + "status": "deprecated", "cost": { - "input": 0.3, - "output": 1.2 + "input": 3, + "output": 7 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "deepseek-ai/DeepSeek-V3-1": { + "id": "deepseek-ai/DeepSeek-V3-1", + "name": "DeepSeek V3.1", + "description": "Legacy model retained for compatibility with older integrations", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-08", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ "text" @@ -148364,30 +139620,42 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 131072, "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.6, + "output": 1.7 } }, - "zai-org/GLM-5.2-FP8": { - "id": "zai-org/GLM-5.2-FP8", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "DeepSeek V4 Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -148398,27 +139666,27 @@ }, "open_weights": true, "limit": { - "context": 393216, - "output": 131072 + "context": 512000, + "output": 384000 }, "cost": { - "input": 0.85, - "output": 3.1 + "input": 1.74, + "output": 3.48, + "cache_read": 0.2 } }, - "nvidia/DeepSeek-V3.2-NVFP4": { - "id": "nvidia/DeepSeek-V3.2-NVFP4", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "deepseek-ai/DeepSeek-V3": { + "id": "deepseek-ai/DeepSeek-V3", + "name": "DeepSeek-V3", + "description": "Legacy model retained for compatibility with older integrations", "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2024-12-26", + "last_updated": "2025-05-29", "modalities": { "input": [ "text" @@ -148432,24 +139700,24 @@ "context": 131072, "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.55, - "output": 1.65 + "input": 1.25, + "output": 1.25 } }, - "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16": { - "id": "nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16", - "name": "NVIDIA Nemotron 3 Nano Omni", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -148460,30 +139728,33 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 204800, "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.13, - "output": 0.38 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "nvidia/Nemotron-Cascade-2-30B-A3B": { - "id": "nvidia/Nemotron-Cascade-2-30B-A3B", - "name": "NVIDIA Nemotron Cascade 2", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, + "MiniMaxAI/MiniMax-M3": { + "id": "MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -148491,48 +139762,28 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 524288, + "output": 250000 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } - } - } - }, - "huggingface": { - "id": "huggingface", - "env": [ - "HF_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://router.huggingface.co/v1", - "name": "Hugging Face", - "doc": "https://huggingface.co/docs/inference-providers", - "models": { - "tencent/Hy3": { - "id": "tencent/Hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + }, + "MiniMaxAI/MiniMax-M2.7": { + "id": "MiniMaxAI/MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -148543,36 +139794,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "input": 192000, - "output": 128000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.58 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "tencent/Hy4-preview": { - "id": "tencent/Hy4-preview", - "name": "Hy4 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + "essentialai/Rnj-1-Instruct": { + "id": "essentialai/Rnj-1-Instruct", + "name": "Rnj-1 Instruct", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "rnj", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", + "knowledge": "2024-10", + "release_date": "2025-12-05", + "last_updated": "2025-12-05", "modalities": { "input": [ "text" @@ -148583,27 +139825,38 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.834, - "output": 2.501 + "input": 0.15, + "output": 0.15 } - }, - "meta-llama/Llama-3.1-8B-Instruct": { - "id": "meta-llama/Llama-3.1-8B-Instruct", - "name": "Llama-3.1-8B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, + } + } + }, + "frogbot": { + "id": "frogbot", + "env": [ + "FROGBOT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://app.frogbot.ai/api/v1", + "name": "FrogBot", + "doc": "https://docs.frogbot.ai", + "models": { + "minimax-m2-5": { + "id": "minimax-m2-5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2024-09", + "release_date": "2025-01-15", + "last_updated": "2025-02-22", "modalities": { "input": [ "text" @@ -148612,29 +139865,61 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 4096 + "context": 192000, + "output": 8192 }, "cost": { - "input": 0.06, - "output": 0.06 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "kimi-k2-6": { + "id": "kimi-k2-6", + "name": "Kimi-K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.95, + "output": 4, + "cache_read": 0.16 + } + }, + "zai-glm-5-1": { + "id": "zai-glm-5-1", + "name": "Z.AI GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-01-20", + "last_updated": "2025-02-22", "modalities": { "input": [ "text" @@ -148645,27 +139930,28 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 4096 + "context": 198000, + "output": 8192 }, "cost": { - "input": 0.59, - "output": 0.79 + "input": 1.4, + "output": 4.4, + "cache_read": 0.26 } }, - "XiaomiMiMo/MiMo-V2-Flash": { - "id": "XiaomiMiMo/MiMo-V2-Flash", - "name": "MiMo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", + "grok-code-fast-1": { + "id": "grok-code-fast-1", + "name": "Grok 4.1 Fast (Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "knowledge": "2023-10", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ "text" @@ -148674,144 +139960,172 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 4096 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.2, + "output": 1.5, + "cache_read": 0.02 } }, - "XiaomiMiMo/MiMo-V2.5-Pro": { - "id": "XiaomiMiMo/MiMo-V2.5-Pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 + "output": 65536 }, "cost": { - "input": 1, - "output": 3 + "input": 1.25, + "output": 10, + "cache_read": 0.31 } }, - "XiaomiMiMo/MiMo-V2.5": { - "id": "XiaomiMiMo/MiMo-V2.5", - "name": "MiMo-V2.5", - "description": "MiMo model for long-context reasoning, perception, and agentic tasks", - "family": "mimo", - "attachment": false, + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 0, + "max": 24576 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-01", + "release_date": "2025-07-17", + "last_updated": "2025-07-17", "modalities": { "input": [ + "text", + "image", + "audio", + "video", + "pdf" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.075 + } + }, + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.4, - "output": 2 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "thinkingmachines/Inkling-Small": { - "id": "thinkingmachines/Inkling-Small", - "name": "Inkling Small", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", + "qwen-3-6-plus": { + "id": "qwen-3-6-plus", + "name": "Qwen 3.6 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "release_date": "2026-04-02", + "last_updated": "2026-04-03", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 1048576 + "context": 1000000, + "output": 64000 }, "cost": { "input": 0.5, - "output": 1.2 + "output": 3, + "cache_read": 0.1 } }, - "thinkingmachines/Inkling": { - "id": "thinkingmachines/Inkling", - "name": "Inkling", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "ling", + "grok-4-3": { + "id": "grok-4-3", + "name": "Grok 4.3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -148827,104 +140141,117 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "knowledge": "2024-11", + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 4.05 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma 3 27B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek v4 Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "knowledge": "2026-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.08, - "output": 0.16 + "input": 1.74, + "output": 3.48, + "cache_read": 0.14 } }, - "google/gemma-3-12b-it": { - "id": "google/gemma-3-12b-it", - "name": "Gemma 3 12B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.15 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "google/gemma-3-4b-it": { - "id": "google/gemma-3-4b-it", - "name": "Gemma 3 4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "Grok 4.1 Fast (Non-Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "knowledge": "2025-11", + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ "text", @@ -148934,61 +140261,70 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 2000000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.1 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "minimax-m2-7": { + "id": "minimax-m2-7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-09", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 192000, + "output": 8192 }, "cost": { - "input": 0.14, - "output": 0.4 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "google/gemma-4-26B-A4B-it": { - "id": "google/gemma-4-26B-A4B-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "gpt-5-3-codex": { + "id": "gpt-5-3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", @@ -148998,29 +140334,40 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.13, - "output": 0.4 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "Qwen/Qwen3-VL-235B-A22B-Instruct": { - "id": "Qwen/Qwen3-VL-235B-A22B-Instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-5-4-nano": { + "id": "gpt-5-4-nano", + "name": "GPT-5.4 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -149030,21 +140377,22 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "Qwen/Qwen3.8-27B": { - "id": "Qwen/Qwen3.8-27B", - "name": "Qwen3.8 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-5-4-mini": { + "id": "gpt-5-4-mini", + "name": "GPT-5.4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -149053,15 +140401,16 @@ "values": [ "low", "medium", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -149071,58 +140420,71 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 3 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "Qwen/Qwen3-Next-80B-A3B-Instruct": { - "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", - "name": "Qwen3-Next-80B-A3B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi-K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-11", - "last_updated": "2025-09-11", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 66536 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "Qwen/Qwen2.5-Coder-32B-Instruct": { - "id": "Qwen/Qwen2.5-Coder-32B-Instruct", - "name": "Qwen2.5-Coder-32B-Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-11-12", - "last_updated": "2024-11-12", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ "text" @@ -149134,48 +140496,18 @@ "open_weights": true, "limit": { "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.06, - "output": 0.2 - } - }, - "Qwen/Qwen3-Coder-Next": { - "id": "Qwen/Qwen3-Coder-Next", - "name": "Qwen3-Coder-Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 + "output": 32768 }, "cost": { - "input": 0.2, - "output": 1.5 + "input": 0.15, + "output": 0.6 } }, - "Qwen/Qwen3-VL-235B-A22B-Thinking": { - "id": "Qwen/Qwen3-VL-235B-A22B-Thinking", - "name": "Qwen3 VL 235B A22B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gpt-5-5": { + "id": "gpt-5-5", + "name": "GPT-5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -149189,11 +140521,10 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -149203,217 +140534,300 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 272000, + "output": 128000 }, "cost": { - "input": 0.98, - "output": 3.95 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "Qwen/Qwen3.5-27B": { - "id": "Qwen/Qwen3.5-27B", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 2.4 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "Qwen/Qwen3-Next-80B-A3B-Thinking": { - "id": "Qwen/Qwen3-Next-80B-A3B-Thinking", - "name": "Qwen3-Next-80B-A3B-Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": false, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-11", - "last_updated": "2025-09-11", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 2 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", + "gemini-3-1-pro-preview": { + "id": "gemini-3-1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2026-01", + "release_date": "2026-02-18", + "last_updated": "2026-02-18", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.95 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen3-235B-A22B-Thinking-2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "Grok 4.1 Fast (Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "knowledge": "2025-11", + "release_date": "2025-11-25", + "last_updated": "2025-11-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 2000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 3 + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B-A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.855, - "output": 2.565 - } + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 + } }, - "Qwen/Qwen3-235B-A22B": { - "id": "Qwen/Qwen3-235B-A22B", - "name": "Qwen3 235B-A22B", - "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", - "family": "qwen", - "attachment": false, + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0.5, + "output": 3, + "cache_read": 0.05 } }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "1970-01-01", + "last_updated": "1970-01-01", "modalities": { "input": [ "text" @@ -149425,26 +140839,37 @@ "open_weights": true, "limit": { "context": 131072, - "output": 16384 + "output": 32768 }, "cost": { - "input": 0.29, - "output": 0.59 + "input": 0.07, + "output": 0.2 } - }, - "Qwen/Qwen3-Coder-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", + } + } + }, + "openrouter": { + "id": "openrouter", + "env": [ + "OPENROUTER_API_KEY" + ], + "npm": "@openrouter/ai-sdk-provider", + "api": "https://openrouter.ai/api/v1", + "name": "OpenRouter", + "doc": "https://openrouter.ai/models", + "models": { + "inclusionai/ling-2.6-1t": { + "id": "inclusionai/ling-2.6-1t", + "name": "Ling-2.6-1T", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "family": "ling", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text" @@ -149453,172 +140878,103 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 65536 + "output": 32768 }, "cost": { - "input": 0.07, - "output": 0.26 + "input": 0.075, + "output": 0.625, + "cache_read": 0.015 } }, - "Qwen/Qwen3-Embedding-8B": { - "id": "Qwen/Qwen3-Embedding-8B", - "name": "Qwen 3 Embedding 8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", + "inclusionai/ring-2.6-1t": { + "id": "inclusionai/ring-2.6-1t", + "name": "Ring-2.6-1T", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "ring", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32000, - "output": 4096 - }, - "cost": { - "input": 0.01, - "output": 0 - } - }, - "Qwen/Qwen3.5-35B-A3B": { - "id": "Qwen/Qwen3.5-35B-A3B", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-05-08", + "last_updated": "2026-05-08", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 2 - } - }, - "Qwen/Qwen3.5-122B-A10B": { - "id": "Qwen/Qwen3.5-122B-A10B", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, "output": 65536 }, "cost": { - "input": 0.4, - "output": 3.2 + "input": 0.075, + "output": 0.625, + "cache_read": 0.015 } }, - "Qwen/Qwen3.5-9B": { - "id": "Qwen/Qwen3.5-9B", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "inclusionai/ling-2.6-flash": { + "id": "inclusionai/ling-2.6-flash", + "name": "Ling-2.6-flash", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "ling", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 65536 + "output": 32768 }, "cost": { - "input": 0.17, - "output": 0.25 + "input": 0.01, + "output": 0.03, + "cache_read": 0.002 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen3.5-397B-A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "ibm-granite/granite-4.0-h-micro": { + "id": "ibm-granite/granite-4.0-h-micro", + "name": "Granite 4.0 Micro", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "granite", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-01", - "last_updated": "2026-02-01", + "release_date": "2025-10-20", + "last_updated": "2025-10-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -149626,26 +140982,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 131000, + "output": 131000 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.017, + "output": 0.112 } }, - "Qwen/Qwen3-Embedding-4B": { - "id": "Qwen/Qwen3-Embedding-4B", - "name": "Qwen 3 Embedding 4B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", + "ibm-granite/granite-4.1-8b": { + "id": "ibm-granite/granite-4.1-8b", + "name": "Granite 4.1 8B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "granite", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2024-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text" @@ -149656,36 +141012,28 @@ }, "open_weights": true, "limit": { - "context": 32000, - "output": 2048 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.01, - "output": 0 + "input": 0.05, + "output": 0.1, + "cache_read": 0.05 } }, - "Qwen/Qwen3.8-2.4T-A95B": { - "id": "Qwen/Qwen3.8-2.4T-A95B", - "name": "Qwen3.8 2.4T A95B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "meta-llama/llama-3.1-8b-instruct": { + "id": "meta-llama/llama-3.1-8b-instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2023-12-31", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -149696,27 +141044,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 131072 + "context": 131072, + "output": 16384 }, "cost": { - "input": 2.5, - "output": 6.25 + "input": 0.02, + "output": 0.03 } }, - "Qwen/Qwen3-30B-A3B": { - "id": "Qwen/Qwen3-30B-A3B", - "name": "Qwen3 30B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "meta-llama/llama-3.1-70b-instruct": { + "id": "meta-llama/llama-3.1-70b-instruct", + "name": "Llama 3.1 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", + "knowledge": "2023-12-31", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -149727,26 +141075,27 @@ }, "open_weights": true, "limit": { - "context": 40960, + "context": 131072, "output": 16384 }, "cost": { - "input": 0.12, - "output": 0.5 + "input": 0.4, + "output": 0.4 } }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct", - "name": "Qwen3-Coder-480B-A35B-Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "meta-llama/llama-3.2-1b-instruct": { + "id": "meta-llama/llama-3.2-1b-instruct", + "name": "Llama 3.2 1B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "knowledge": "2023-12-31", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text" @@ -149757,27 +141106,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 66536 + "context": 131072, + "output": 60000 }, "cost": { - "input": 2, - "output": 2 + "input": 0.027, + "output": 0.201 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "meta-llama/llama-4-maverick": { + "id": "meta-llama/llama-4-maverick", + "name": "Llama 4 Maverick", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2024-08-31", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", @@ -149789,30 +141138,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 1048576, + "output": 16384 }, "cost": { - "input": 0.47, - "output": 3.19 + "input": 0.15, + "output": 0.6 } }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek-V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "meta-llama/llama-3.2-11b-vision-instruct": { + "id": "meta-llama/llama-3.2-11b-vision-instruct", + "name": "Llama 3.2 11B Vision Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "knowledge": "2023-12-31", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -149821,38 +141171,26 @@ "open_weights": true, "limit": { "context": 131072, - "output": 8192 + "output": 16384 }, "cost": { - "input": 0.27, - "output": 1 + "input": 0.345, + "output": 0.345 } }, - "deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + "meta-llama/llama-3.3-70b-instruct:free": { + "id": "meta-llama/llama-3.3-70b-instruct:free", + "name": "Llama 3.3 70B Instruct (free)", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -149863,28 +141201,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 384000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.28 + "input": 0, + "output": 0 } }, - "deepseek-ai/DeepSeek-R1": { - "id": "deepseek-ai/DeepSeek-R1", - "name": "DeepSeek-R1", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", + "meta-llama/llama-3.3-70b-instruct": { + "id": "meta-llama/llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -149895,36 +141232,27 @@ }, "open_weights": true, "limit": { - "context": 64000, - "output": 32768 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.7, - "output": 2.5 + "input": 0.1, + "output": 0.32 } }, - "deepseek-ai/DeepSeek-V4-Pro-0813": { - "id": "deepseek-ai/DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", + "meta-llama/llama-3.2-3b-instruct:free": { + "id": "meta-llama/llama-3.2-3b-instruct:free", + "name": "Llama 3.2 3B Instruct (free)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2023-12-31", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text" @@ -149935,43 +141263,31 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.32, - "output": 3.96 + "input": 0, + "output": 0 } }, - "deepseek-ai/DeepSeek-V4.1-Flash": { - "id": "deepseek-ai/DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", + "meta-llama/llama-guard-4-12b": { + "id": "meta-llama/llama-guard-4-12b", + "name": "Llama Guard 4 12B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2024-08-31", + "release_date": "2025-04-30", + "last_updated": "2025-04-30", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ "text" @@ -149979,28 +141295,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 384000 + "context": 163840, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.18, + "output": 0.18 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "meta-llama/llama-3-8b-instruct": { + "id": "meta-llama/llama-3-8b-instruct", + "name": "Llama 3 8B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-12-31", + "release_date": "2024-04-18", + "last_updated": "2024-04-18", "modalities": { "input": [ "text" @@ -150011,41 +141326,31 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 384000 + "context": 8192, + "output": 8192 }, "cost": { "input": 0.14, - "output": 0.28 + "output": 0.14 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], + "meta-llama/llama-4-scout": { + "id": "meta-llama/llama-4-scout", + "name": "Llama 4 Scout", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-08-31", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -150053,28 +141358,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 393216 + "context": 10000000, + "output": 16384 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0.1, + "output": 0.3 } }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "meta-llama/llama-3.2-3b-instruct": { + "id": "meta-llama/llama-3.2-3b-instruct", + "name": "Llama 3.2 3B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "knowledge": "2023-12-31", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text" @@ -150085,199 +141389,219 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 65536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.28, - "output": 0.4 + "input": 0.05, + "output": 0.33 } }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", - "name": "DeepSeek-R1-0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "~anthropic/claude-haiku-latest": { + "id": "~anthropic/claude-haiku-latest", + "name": "Anthropic Claude Haiku Latest", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 163840 - }, - "cost": { - "input": 3, - "output": 5 - } - }, - "deepseek-ai/DeepSeek-V3": { - "id": "deepseek-ai/DeepSeek-V3", - "name": "DeepSeek-V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2024-12-26", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 64000, - "output": 8192 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.4, - "output": 1.3 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "deepseek-ai/DeepSeek-V3-0324": { - "id": "deepseek-ai/DeepSeek-V3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "~anthropic/claude-fable-latest": { + "id": "~anthropic/claude-fable-latest", + "name": "Claude Fable Latest", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "temperature": false, + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 1.12 + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp": { - "id": "deepseek-ai/DeepSeek-V4-Flash-Vision-Exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", + "~anthropic/claude-sonnet-latest": { + "id": "~anthropic/claude-sonnet-latest", + "name": "Anthropic Claude Sonnet Latest", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", + "medium", "high", "max" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", + "temperature": false, + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.44, - "output": 1.32 + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "stepfun-ai/Step-3.7-Flash": { - "id": "stepfun-ai/Step-3.7-Flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "~anthropic/claude-opus-latest": { + "id": "~anthropic/claude-opus-latest", + "name": "Claude Opus Latest", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 256000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.15 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "stepfun-ai/Step-3.5-Flash": { - "id": "stepfun-ai/Step-3.5-Flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "moonshotai/kimi-k2": { + "id": "moonshotai/kimi-k2", + "name": "Kimi K2 0711", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", + "knowledge": "2024-12-31", + "release_date": "2025-07-11", + "last_updated": "2025-07-11", "modalities": { "input": [ "text" @@ -150288,27 +141612,28 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 256000 + "context": 131072, + "output": 100352 }, "cost": { - "input": 0.1, - "output": 0.3 + "input": 0.57, + "output": 2.3 } }, - "MiniMaxAI/MiniMax-M3": { - "id": "MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", @@ -150320,30 +141645,32 @@ }, "open_weights": true, "limit": { - "context": 524288, - "output": 512000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.72, + "output": 3.49, + "cache_read": 0.159 } }, - "MiniMaxAI/MiniMax-M2.1": { - "id": "MiniMaxAI/MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", + "family": "kimi-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "interleaved": { - "field": "reasoning_content" + "field": "reasoning_details" }, + "structured_output": true, "temperature": true, - "knowledge": "2025-10", - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ "text" @@ -150354,32 +141681,36 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 100352 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "interleaved": { - "field": "reasoning_content" + "field": "reasoning_details" }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -150387,34 +141718,36 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 256000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.375, + "output": 2.025, + "cache_read": 0.203 } }, - "MiniMaxAI/MiniMax-M2.7": { - "id": "MiniMaxAI/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "interleaved": { - "field": "reasoning_content" + "field": "reasoning_details" }, "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -150422,27 +141755,28 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.66, + "output": 3.41, + "cache_read": 0.15 } }, - "MiniMaxAI/MiniMax-M2": { - "id": "MiniMaxAI/MiniMax-M2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "knowledge": "2024-12-31", + "release_date": "2025-09-04", + "last_updated": "2025-09-04", "modalities": { "input": [ "text" @@ -150453,106 +141787,31 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 100352 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.6, + "output": 2.5 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi-K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "baidu/ernie-4.5-vl-424b-a47b": { + "id": "baidu/ernie-4.5-vl-424b-a47b", + "name": "ERNIE 4.5 VL 424B A47B ", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ernie", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "knowledge": "2025-03-31", + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ - "text", "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 - } - }, - "moonshotai/Kimi-K3": { - "id": "moonshotai/Kimi-K3", - "name": "Kimi K3", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15 - } - }, - "moonshotai/Kimi-K2-Instruct": { - "id": "moonshotai/Kimi-K2-Instruct", - "name": "Kimi-K2-Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-07-14", - "last_updated": "2025-07-14", - "modalities": { - "input": [ "text" ], "output": [ @@ -150562,130 +141821,148 @@ "open_weights": true, "limit": { "context": 131072, - "output": 16384 + "output": 16000 }, "cost": { - "input": 1, - "output": 3 + "input": 0.42, + "output": 1.25 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "perceptron/perceptron-mk1": { + "id": "perceptron/perceptron-mk1", + "name": "Perceptron Mk1", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 32768, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 4 + "input": 0.15, + "output": 1.5 } }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, "knowledge": "2025-01", - "release_date": "2026-01-01", - "last_updated": "2026-01-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 0.25, + "output": 1.5, + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.083333 } }, - "moonshotai/Kimi-K2-Thinking": { - "id": "moonshotai/Kimi-K2-Thinking", - "name": "Kimi-K2-Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, + "google/gemini-3.1-flash-image": { + "id": "google/gemini-3.1-flash-image", + "name": "Nano Banana 2 (Gemini 3.1 Flash Image)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "minimal" + ] + } + ], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "release_date": "2026-06-18", + "last_updated": "2026-06-18", "modalities": { "input": [ + "image", "text" ], "output": [ + "image", "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.15 + "input": 0.5, + "output": 3 } }, - "moonshotai/Kimi-K2-Instruct-0905": { - "id": "moonshotai/Kimi-K2-Instruct-0905", - "name": "Kimi-K2-Instruct-0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "google/gemma-3n-e4b-it": { + "id": "google/gemma-3n-e4b-it", + "name": "Gemma 3n 4B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-04", - "last_updated": "2025-09-04", + "knowledge": "2024-08-31", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text" @@ -150696,19 +141973,19 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 16384 + "context": 32768, + "output": 32768 }, "cost": { - "input": 1, - "output": 3 + "input": 0.06, + "output": 0.12 } }, - "zai-org/GLM-4.6V-Flash": { - "id": "zai-org/GLM-4.6V-Flash", - "name": "GLM-4.6V-Flash", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", + "google/gemma-4-26b-a4b-it:free": { + "id": "google/gemma-4-26b-a4b-it:free", + "name": "Gemma 4 26B A4B (free)", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -150717,13 +141994,15 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ + "image", "text", - "image" + "video" ], "output": [ "text" @@ -150731,129 +142010,178 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 262144, "output": 32768 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0, + "output": 0 } }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-04-03", - "last_updated": "2026-04-03", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 1.25, + "output": 10, + "reasoning": 10, + "cache_read": 0.125, + "cache_write": 0.375 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 1.4, - "output": 4.4 + "input": 0.3, + "output": 2.5, + "reasoning": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333 } }, - "zai-org/GLM-4.5-Air": { - "id": "zai-org/GLM-4.5-Air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", - "attachment": false, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.13, - "output": 0.85 + "input": 1.5, + "output": 9, + "reasoning": 9, + "cache_read": 0.15, + "cache_write": 0.083333 } }, - "zai-org/GLM-4.7": { - "id": "zai-org/GLM-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" @@ -150861,54 +142189,52 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0.12, + "output": 0.35, + "cache_read": 0.09 } }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "google/lyria-3-clip-preview": { + "id": "google/lyria-3-clip-preview", + "name": "Lyria 3 Clip Preview", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "lyria", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "release_date": "2026-03-30", + "last_updated": "2026-03-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0, + "output": 0 } }, - "zai-org/GLM-5.3-Flash": { - "id": "zai-org/GLM-5.3-Flash", - "name": "GLM-5.3-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -150916,300 +142242,315 @@ "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_details" + }, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1048756, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.5 + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "zai-org/GLM-4.5V": { - "id": "zai-org/GLM-4.5V", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "google/gemini-3.1-flash-lite-image": { + "id": "google/gemini-3.1-flash-lite-image", + "name": "Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "high" + ] + } + ], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "knowledge": "2025-01-01", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text", - "image" + "image", + "text" ], "output": [ + "image", "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 65536, - "output": 16384 + "output": 66000 }, "cost": { - "input": 0.6, - "output": 1.8 + "input": 0.25, + "output": 1.5 } }, - "zai-org/GLM-4.6": { - "id": "zai-org/GLM-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, + "google/gemini-3-pro-image-preview": { + "id": "google/gemini-3-pro-image-preview", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2025-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 65536, + "output": 32768 }, "cost": { - "input": 0.55, - "output": 2.2 + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "zai-org/GLM-4.7-Flash": { - "id": "zai-org/GLM-4.7-Flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "google/gemini-2.5-flash-image": { + "id": "google/gemini-2.5-flash-image", + "name": "Nano Banana", + "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", + "family": "gemini-flash", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-08", - "last_updated": "2025-08-08", + "knowledge": "2025-06", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "cache_write": 0.083333 } }, - "zai-org/GLM-5.3": { - "id": "zai-org/GLM-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "google/gemini-2.5-flash-lite": { + "id": "google/gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 512, + "max": 24576 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4 - } - }, - "zai-org/GLM-4.5": { - "id": "zai-org/GLM-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 98304 + "output": 65535 }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 0.1, + "output": 0.4, + "reasoning": 0.4, + "cache_read": 0.01, + "cache_write": 0.083333 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "google/gemini-3.1-flash-image-preview": { + "id": "google/gemini-3.1-flash-image-preview", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", + "minimal", "high" ] } ], - "tool_call": true, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ + "image", "text" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 131072, "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.5 + "input": 0.5, + "output": 3 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "google/gemini-2.5-pro-preview-05-06": { + "id": "google/gemini-2.5-pro-preview-05-06", + "name": "Gemini 2.5 Pro Preview 05-06", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01-31", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ - "text" + "text", + "image", + "pdf", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 0.25, - "output": 0.69 + "input": 1.25, + "output": 10, + "reasoning": 10, + "cache_read": 0.125, + "cache_write": 0.375 } - } - } - }, - "volcengine": { - "id": "volcengine", - "env": [ - "ARK_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://ark.cn-beijing.volces.com/api/v3", - "name": "Volcengine Ark", - "doc": "https://www.volcengine.com/docs/82379/1330310", - "models": { - "glm-5-3-flash-260828": { - "id": "glm-5-3-flash-260828", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + }, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -151217,143 +142558,124 @@ "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, "interleaved": { - "field": "reasoning_content" + "field": "reasoning_details" }, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", "video", + "audio", "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.11875, - "output": 0.41563, - "cache_read": 0.03414 + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "doubao-seed-1-6-251015": { - "id": "doubao-seed-1-6-251015", - "name": "Seed 1.6", - "description": "ByteDance Seed model for long-context reasoning, instruction following, and tool-assisted tasks", - "family": "seed", - "attachment": false, + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.11875, - "output": 1.18747, - "cache_read": 0.02375, - "tiers": [ - { - "input": 0.17812, - "output": 2.37494, - "cache_read": 0.02375, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.35624, - "output": 3.56241, - "cache_read": 0.02375, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.06, + "output": 0.33 } }, - "doubao-seed-2-0-pro-260215": { - "id": "doubao-seed-2-0-pro-260215", - "name": "Seed 2.0 Pro", - "description": "Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows", - "family": "seed", + "google/gemini-2.5-pro-preview": { + "id": "google/gemini-2.5-pro-preview", + "name": "Gemini 2.5 Pro Preview 06-05", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-01-31", + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text", + "pdf", "image", - "video" + "text", + "audio" ], "output": [ "text" @@ -151361,69 +142683,51 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.47499, - "output": 2.37494, - "cache_read": 0.095, - "tiers": [ - { - "input": 0.71248, - "output": 3.56241, - "cache_read": 0.1425, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.42496, - "output": 7.12482, - "cache_read": 0.28499, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 1.25, + "output": 10, + "reasoning": 10, + "cache_read": 0.125, + "cache_write": 0.375 } }, - "doubao-seed-2-0-lite-260428": { - "id": "doubao-seed-2-0-lite-260428", - "name": "Seed 2.0 Lite", - "description": "Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation", - "family": "seed", + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "interleaved": { - "field": "reasoning_content" + "field": "reasoning_details" }, "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -151431,105 +142735,94 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 131072 - }, + "context": 1048576, + "output": 65535 + }, "cost": { - "input": 0.08906, - "output": 0.53436, - "cache_read": 0.01781, - "tiers": [ - { - "input": 0.13359, - "output": 0.80154, - "cache_read": 0.02672, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.26718, - "output": 1.60308, - "cache_read": 0.05344, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.5, + "output": 3, + "reasoning": 3, + "cache_read": 0.05, + "cache_write": 0.083333 } }, - "doubao-seed-2-1-pro-260628": { - "id": "doubao-seed-2-1-pro-260628", - "name": "Seed 2.1 Pro", - "description": "Flagship ByteDance Seed 2.1 model for complex multimodal reasoning, coding, and agents", - "family": "seed", + "google/gemma-3-12b-it": { + "id": "google/gemma-3-12b-it", + "name": "Gemma 3 12B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "structured_output": true, + "temperature": true, + "knowledge": "2024-08-31", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 16384 }, + "cost": { + "input": 0.05, + "output": 0.15 + } + }, + "google/gemma-3-4b-it": { + "id": "google/gemma-3-4b-it", + "name": "Gemma 3 4B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": true, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", + "knowledge": "2024-08-31", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.8906, - "output": 4.45301, - "cache_read": 0.17812 + "input": 0.05, + "output": 0.1 } }, - "doubao-seed-1-6-vision-250815": { - "id": "doubao-seed-1-6-vision-250815", - "name": "Seed 1.6 Vision", - "description": "ByteDance Seed multimodal model for image understanding, visual reasoning, and tool-assisted tasks", - "family": "seed", + "google/gemma-3-27b-it": { + "id": "google/gemma-3-27b-it", + "name": "Gemma 3 27B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2025-08-15", - "last_updated": "2025-08-15", + "knowledge": "2024-08-31", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ "text", @@ -151539,211 +142832,183 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.11875, - "output": 1.18747, - "cache_read": 0.02375, - "tiers": [ - { - "input": 0.17812, - "output": 2.37494, - "cache_read": 0.02375, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.35624, - "output": 3.56241, - "cache_read": 0.02375, - "tier": { - "type": "context", - "size": 128000 - } - } + "input": 0.08, + "output": 0.16 + } + }, + "google/lyria-3-pro-preview": { + "id": "google/lyria-3-pro-preview", + "name": "Lyria 3 Pro Preview", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "lyria", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-03-30", + "last_updated": "2026-03-30", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "audio" ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 } }, - "doubao-seed-2-0-mini-260428": { - "id": "doubao-seed-2-0-mini-260428", - "name": "Seed 2.0 Mini", - "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", - "family": "seed", + "google/gemma-4-31b-it:free": { + "id": "google/gemma-4-31b-it:free", + "name": "Gemma 4 31B (free)", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text", "image", + "text", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 131072 + "context": 262144, + "output": 8192 }, "cost": { - "input": 0.02969, - "output": 0.29687, - "cache_read": 0.00594, - "tiers": [ - { - "input": 0.05937, - "output": 0.59374, - "cache_read": 0.01187, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.11875, - "output": 1.18747, - "cache_read": 0.02375, - "tier": { - "type": "context", - "size": 128000 - } - } + "input": 0, + "output": 0 + } + }, + "google/gemma-2-27b-it": { + "id": "google/gemma-2-27b-it", + "name": "Gemma 2 27B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2024-06-30", + "release_date": "2024-07-13", + "last_updated": "2024-07-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 2048 + }, + "cost": { + "input": 0.65, + "output": 0.65 } }, - "doubao-seed-2-0-code-preview-260215": { - "id": "doubao-seed-2-0-code-preview-260215", - "name": "Seed 2.0 Code", - "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", - "family": "seed", + "google/gemini-3-pro-image": { + "id": "google/gemini-3-pro-image", + "name": "Nano Banana Pro (Gemini 3 Pro Image)", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-06-18", + "last_updated": "2026-06-18", "modalities": { "input": [ - "text", "image", - "video" + "text" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 65536, + "output": 32768 }, "cost": { - "input": 0.47499, - "output": 2.37494, - "cache_read": 0.095, - "tiers": [ - { - "input": 0.71248, - "output": 3.56241, - "cache_read": 0.1425, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.42496, - "output": 7.12482, - "cache_read": 0.28499, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "doubao-seed-2-1-turbo-260628": { - "id": "doubao-seed-2-1-turbo-260628", - "name": "Seed 2.1 Turbo", - "description": "Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows", - "family": "seed", + "google/gemini-3.1-flash-lite-preview": { + "id": "google/gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -151751,44 +143016,31 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.4453, - "output": 2.22651, - "cache_read": 0.08906 + "input": 0.25, + "output": 1.5, + "reasoning": 1.5, + "cache_read": 0.025, + "cache_write": 0.083333 } }, - "glm-5-2-260617": { - "id": "glm-5-2-260617", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "liquid/lfm-2.5-1.2b-thinking:free": { + "id": "liquid/lfm-2.5-1.2b-thinking:free", + "name": "LFM2.5-1.2B-Thinking (free)", + "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", + "family": "liquid", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-06", + "release_date": "2026-01-20", + "last_updated": "2026-01-20", "modalities": { "input": [ "text" @@ -151799,118 +143051,64 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 32768, + "output": 32768 }, "cost": { - "input": 1.18747, - "output": 4.15615, - "cache_read": 0.29687 + "input": 0, + "output": 0 } }, - "doubao-seed-1-8-251228": { - "id": "doubao-seed-1-8-251228", - "name": "Seed 1.8", - "description": "ByteDance Seed model for multimodal reasoning, long-context analysis, and agent workflows", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "liquid/lfm-2.5-1.2b-instruct:free": { + "id": "liquid/lfm-2.5-1.2b-instruct:free", + "name": "LFM2.5-1.2B-Instruct (free)", + "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", + "family": "liquid", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2025-12-28", - "last_updated": "2025-12-28", + "knowledge": "2025-06", + "release_date": "2026-01-20", + "last_updated": "2026-01-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 64000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.11875, - "output": 1.18747, - "cache_read": 0.02375, - "tiers": [ - { - "input": 0.17812, - "output": 2.37494, - "cache_read": 0.02375, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.35624, - "output": 3.56241, - "cache_read": 0.02375, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0, + "output": 0 } }, - "doubao-seed-evolving": { - "id": "doubao-seed-evolving", - "name": "Seed Evolving", - "description": "Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities", - "family": "seed", + "x-ai/grok-4.20": { + "id": "x-ai/grok-4.20", + "name": "Grok 4.20", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", + "knowledge": "2025-09-01", + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -151918,39 +143116,43 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.8906, - "output": 4.45301, - "cache_read": 0.17812 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "doubao-seed-1-6-flash-250828": { - "id": "doubao-seed-1-6-flash-250828", - "name": "Seed 1.6 Flash", - "description": "Low-latency ByteDance Seed model for high-throughput chat, extraction, and lightweight tool use", - "family": "seed", + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -151958,68 +143160,58 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 1000000, + "output": 1000000 }, "cost": { - "input": 0.02227, - "output": 0.22265, - "cache_read": 0.00445, + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, "tiers": [ { - "input": 0.04453, - "output": 0.4453, - "cache_read": 0.00445, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.08906, - "output": 0.8906, - "cache_read": 0.00445, + "input": 2.5, + "output": 5, + "cache_read": 0.4, "tier": { "type": "context", - "size": 128000 + "size": 200000 } } - ] + ], + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "doubao-seed-character-260628": { - "id": "doubao-seed-character-260628", - "name": "Seed Character", - "description": "ByteDance Seed model optimized for character-driven dialogue and consistent conversational behavior", - "family": "seed", + "x-ai/grok-4.5": { + "id": "x-ai/grok-4.5", + "name": "Grok 4.5", + "description": "xAI's latest Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", "low", + "medium", "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -152027,141 +143219,114 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 500000, + "output": 500000 }, "cost": { - "input": 0.11875, - "output": 0.29687, - "cache_read": 0.02375, + "input": 2, + "output": 6, + "cache_read": 0.5, "tiers": [ { - "input": 0.17812, - "output": 0.8906, - "cache_read": 0.02375, + "input": 4, + "output": 12, + "cache_read": 1, "tier": { "type": "context", - "size": 32000 + "size": 200000 } } - ] + ], + "context_over_200k": { + "input": 4, + "output": 12, + "cache_read": 1 + } } }, - "deepseek-v4-pro-ga-260813": { - "id": "deepseek-v4-pro-ga-260813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, + "x-ai/grok-4.20-multi-agent": { + "id": "x-ai/grok-4.20-multi-agent", + "name": "Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "minimal", "low", - "medium", "high" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2025-09-01", + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 1.3359, - "output": 4.00771, - "cache_read": 0.04453 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "deepseek-v4-flash-ga-260731": { - "id": "deepseek-v4-flash-ga-260731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "x-ai/grok-build-0.1": { + "id": "x-ai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.4453, - "output": 1.3359, - "cache_read": 0.01484 + "input": 1, + "output": 2, + "cache_read": 0.2 } - } - } - }, - "impossibl": { - "id": "impossibl", - "env": [ - "IMPOSSIBL_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.impossibl.com/v1", - "name": "Impossibl", - "doc": "https://impossibl.com/docs/models", - "models": { - "cerebras/gpt-oss-120b": { - "id": "cerebras/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + }, + "~google/gemini-pro-latest": { + "id": "~google/gemini-pro-latest", + "name": "Google Gemini Pro Latest", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -152176,37 +143341,45 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "audio", + "pdf", + "image", + "text", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.35, - "output": 0.75 + "input": 2, + "output": 12, + "reasoning": 12, + "cache_read": 0.2, + "cache_write": 0.375 } }, - "groq/gpt-oss-20b": { - "id": "groq/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "~google/gemini-flash-latest": { + "id": "~google/gemini-flash-latest", + "name": "Google Gemini Flash Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -152216,49 +143389,47 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01-01", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.075, - "output": 0.3, - "cache_read": 0.0375 + "input": 1.5, + "output": 9, + "reasoning": 9, + "cache_read": 0.15, + "cache_write": 0.083333 } }, - "groq/gpt-oss-120b": { - "id": "groq/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "microsoft/phi-4": { + "id": "microsoft/phi-4", + "name": "Phi 4", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-06-30", + "release_date": "2025-01-10", + "last_updated": "2025-01-10", "modalities": { "input": [ "text" @@ -152269,32 +143440,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 0.07, + "output": 0.14 } }, - "zai/glm-4.5": { - "id": "zai/glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", + "microsoft/wizardlm-2-8x22b": { + "id": "microsoft/wizardlm-2-8x22b", + "name": "WizardLM-2 8x22B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2024-04-30", + "release_date": "2024-04-16", + "last_updated": "2024-04-16", "modalities": { "input": [ "text" @@ -152305,32 +143470,27 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 65536, + "output": 8000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0.62, + "output": 0.62 } }, - "zai/glm-4.6": { - "id": "zai/glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", + "poolside/laguna-xs-2.1:free": { + "id": "poolside/laguna-xs-2.1:free", + "name": "Laguna XS 2.1 (free)", + "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", + "family": "laguna", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2026-07-02", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" @@ -152341,31 +143501,27 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0, + "output": 0 } }, - "zai/glm-5": { - "id": "zai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "poolside/laguna-m.1": { + "id": "poolside/laguna-m.1", + "name": "Laguna M.1", + "description": "Poolside's flagship agentic coding model for long-horizon work", + "family": "laguna", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-04-28", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -152376,32 +143532,28 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.2, + "output": 0.4, + "cache_read": 0.1 } }, - "zai/glm-4.7": { - "id": "zai/glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", + "poolside/laguna-xs-2.1": { + "id": "poolside/laguna-xs-2.1", + "name": "Laguna XS 2.1", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "release_date": "2026-07-02", + "last_updated": "2026-07-02", "modalities": { "input": [ "text" @@ -152412,27 +143564,30 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0.06, + "output": 0.12, + "cache_read": 0.03 } }, - "zai/glm-5.2": { - "id": "zai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "poolside/laguna-m.1:free": { + "id": "poolside/laguna-m.1:free", + "name": "Laguna M.1 (free)", + "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", + "family": "laguna", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "release_date": "2026-06-13", + "release_date": "2026-04-28", "last_updated": "2026-06-13", "modalities": { "input": [ @@ -152444,32 +143599,61 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0, + "output": 0 } }, - "zai/glm-5.1": { - "id": "zai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "writer/palmyra-x5": { + "id": "writer/palmyra-x5", + "name": "Palmyra X5", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "palmyra", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-01-21", + "last_updated": "2026-01-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1040000, + "output": 8192 + }, + "cost": { + "input": 0.6, + "output": 6 + } + }, + "z-ai/glm-4.7": { + "id": "z-ai/glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_details" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -152480,21 +143664,21 @@ }, "open_weights": true, "limit": { - "context": 200000, + "context": 202752, "output": 131072 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.4, + "output": 1.75, + "cache_read": 0.08 } }, - "zai/glm-5-turbo": { - "id": "zai/glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", + "z-ai/glm-4.5v": { + "id": "z-ai/glm-4.5v", + "name": "GLM-4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "family": "glm", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -152502,34 +143686,36 @@ } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "knowledge": "2025-04", + "release_date": "2025-08-11", + "last_updated": "2025-08-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 65536, + "output": 16384 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 0.6, + "output": 1.8, + "cache_read": 0.11 } }, - "zai/glm-4.5-air": { - "id": "zai/glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "z-ai/glm-4.5": { + "id": "z-ai/glm-4.5", + "name": "GLM-4.5", + "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -152538,6 +143724,7 @@ } ], "tool_call": true, + "structured_output": false, "temperature": true, "knowledge": "2025-04", "release_date": "2025-07-28", @@ -152556,180 +143743,175 @@ "output": 98304 }, "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 202752, + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.966, + "output": 3.036, + "cache_read": 0.1794 } }, - "anthropic/claude-opus-4-5": { - "id": "anthropic/claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "GLM-4.6", + "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-09-30", + "last_updated": "2025-09-30", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 131072 + }, + "cost": { + "input": 0.43, + "output": 1.74, + "cache_read": 0.08 + } + }, + "z-ai/glm-5.2": { + "id": "z-ai/glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - }, { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 101376 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.54, + "output": 1.76, + "cache_read": 0.1 } }, - "anthropic/claude-sonnet-4-5": { - "id": "anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "z-ai/glm-4.6v": { + "id": "z-ai/glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.3, + "output": 0.9, + "cache_read": 0.055 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "z-ai/glm-5v-turbo": { + "id": "z-ai/glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", + "family": "glm", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "structured_output": false, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ - "text", "image", - "pdf" + "text", + "video" ], "output": [ "text" @@ -152737,149 +143919,142 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 202752, + "output": 131072 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "anthropic/claude-opus-4-8": { - "id": "anthropic/claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "z-ai/glm-4.5-air": { + "id": "z-ai/glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", + "family": "glm-air", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 98304 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.13, + "output": 0.85, + "cache_read": 0.025 } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, + "z-ai/glm-4.7-flash": { + "id": "z-ai/glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "interleaved": { + "field": "reasoning_details" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 202752, + "output": 16384 + }, + "cost": { + "input": 0.06, + "output": 0.4, + "cache_read": 0.01 + } + }, + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, "output": 128000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.6, + "output": 1.92, + "cache_read": 0.12 } }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Faster GLM-5 lane for coding agents that need lower latency", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": false, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -152887,46 +144062,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "openai/gpt-4o-mini-2024-07-18": { + "id": "openai/gpt-4o-mini-2024-07-18", + "name": "GPT-4o-mini (2024-07-18)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "o-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2023-10-31", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", @@ -152939,79 +144096,69 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "openai/gpt-oss-safeguard-20b": { + "id": "openai/gpt-oss-safeguard-20b", + "name": "gpt-oss-safeguard-20b", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "temperature": true, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.075, + "output": 0.3, + "cache_read": 0.0375 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "openai/gpt-3.5-turbo-instruct": { + "id": "openai/gpt-3.5-turbo-instruct", + "name": "GPT-3.5 Turbo Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2021-09-30", + "release_date": "2023-09-28", + "last_updated": "2023-09-28", "modalities": { "input": [ "text" @@ -153020,55 +144167,55 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 4095, + "output": 4096 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.145 + "input": 1.5, + "output": 2 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-5.2-chat": { + "id": "openai/gpt-5.2-chat", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-10", + "last_updated": "2025-12-10", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.19, - "output": 0.51, - "cache_read": 0.028 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "thinkingmachines/inkling": { - "id": "thinkingmachines/inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", + "openai/gpt-5.6-luna-pro": { + "id": "openai/gpt-5.6-luna-pro", + "name": "GPT-5.6 Luna Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -153076,50 +144223,54 @@ "type": "effort", "values": [ "none", - "minimal", "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.87, - "output": 4.68, - "cache_read": 0.374 + "input": 1, + "output": 6, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "google/gemini-3.6-flash": { - "id": "google/gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "openai/o3": { + "id": "openai/o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -153128,16 +144279,14 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -153146,45 +144295,45 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "google/gemini-3.5-flash-lite": { - "id": "google/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", + "openai/gpt-5.6-sol-pro": { + "id": "openai/gpt-5.6-sol-pro", + "name": "GPT-5.6 Sol Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -153193,44 +144342,42 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "openai/o4-mini-high": { + "id": "openai/o4-mini-high", + "name": "o4 Mini High", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "temperature": false, + "knowledge": "2024-06-30", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text", "image", - "video", - "audio", + "text", "pdf" ], "output": [ @@ -153239,104 +144386,74 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 100000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "openai/gpt-audio": { + "id": "openai/gpt-audio", + "name": "GPT Audio", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 + "input": 2.5, + "output": 10 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT-5.2 Pro", + "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 128, - "max": 32768 + "type": "effort", + "values": [ + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", "image", - "audio", - "video", + "text", "pdf" ], "output": [ @@ -153345,61 +144462,31 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 21, + "output": 168 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], - "tool_call": true, + "openai/gpt-4o-mini-search-preview": { + "id": "openai/gpt-4o-mini-search-preview", + "name": "GPT-4o-mini Search Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "o-mini", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "temperature": false, + "knowledge": "2023-10-31", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -153407,44 +144494,44 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 0.15, + "output": 0.6 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", + "openai/gpt-5.6-terra-pro": { + "id": "openai/gpt-5.6-terra-pro", + "name": "GPT-5.6 Terra Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", "image", - "audio", - "video", "pdf" ], "output": [ @@ -153453,20 +144540,22 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -153482,16 +144571,14 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", "image", - "video", - "audio", "pdf" ], "output": [ @@ -153500,33 +144587,34 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "xai/grok-4.3": { - "id": "xai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "openai/gpt-5-chat": { + "id": "openai/gpt-5-chat", + "name": "GPT-5 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text", + "pdf", "image", - "pdf" + "text" ], "output": [ "text" @@ -153534,33 +144622,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 128000, + "output": 16384 }, "cost": { "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "output": 10, + "cache_read": 0.125 } }, - "xai/grok-4.20-0309-reasoning": { - "id": "xai/grok-4.20-0309-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "knowledge": "2021-09-01", + "release_date": "2023-03-01", + "last_updated": "2023-11-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -153568,32 +144654,40 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 16385, + "output": 4096 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.5, + "output": 1.5 } }, - "xai/grok-4.5": { - "id": "xai/grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ "text" @@ -153601,27 +144695,28 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3 + "input": 15, + "output": 120 } }, - "xai/grok-4.20-0309-non-reasoning": { - "id": "xai/grok-4.20-0309-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", @@ -153634,33 +144729,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 2.5, + "output": 10 } }, - "xai/grok-build-0.1": { - "id": "xai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4": { + "id": "openai/gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -153668,21 +144760,20 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 8191, + "output": 4096 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 + "input": 30, + "output": 60 } }, - "fireworks/gpt-oss-20b": { - "id": "fireworks/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -153696,41 +144787,44 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.07, - "output": 0.3, - "cache_read": 0.035 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "fireworks/glm-5.2": { - "id": "fireworks/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "openai/gpt-3.5-turbo-16k": { + "id": "openai/gpt-3.5-turbo-16k", + "name": "GPT-3.5 Turbo 16k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2021-09-30", + "release_date": "2023-08-28", + "last_updated": "2023-08-28", "modalities": { "input": [ "text" @@ -153739,23 +144833,22 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 16385, + "output": 4096 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.14 + "input": 3, + "output": 4 } }, - "fireworks/gpt-oss-120b": { - "id": "fireworks/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "openai/o3-pro": { + "id": "openai/o3-pro", + "name": "o3-pro", + "description": "High-effort o3 tier for difficult technical reasoning and careful answers", + "family": "o-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -153769,115 +144862,126 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", "modalities": { "input": [ - "text" + "text", + "pdf", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 20, + "output": 80 } }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "openai/gpt-5.1-chat": { + "id": "openai/gpt-5.1-chat", + "name": "GPT-5.1 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text", + "pdf", "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 32000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "openai/gpt-4o-2024-05-13": { + "id": "openai/gpt-4o-2024-05-13", + "name": "GPT-4o (2024-05-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.003 + "input": 5, + "output": 15 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ @@ -153886,41 +144990,33 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "qwen/qwen3.8-max-preview": { - "id": "qwen/qwen3.8-max-preview", - "name": "Qwen3.8 Max Preview", - "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", - "family": "qwen", + "openai/gpt-5.3-chat": { + "id": "openai/gpt-5.3-chat", + "name": "GPT-5.3 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-07-19", - "last_updated": "2026-07-19", + "structured_output": true, + "temperature": false, + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -153928,36 +145024,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 2.5, - "output": 7.5 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "openai/gpt-3.5-turbo-0613": { + "id": "openai/gpt-3.5-turbo-0613", + "name": "GPT-3.5 Turbo (older v0613)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2021-09-30", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -153965,115 +145056,76 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 4095, + "output": 4096 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.05, - "tiers": [ - { - "input": 1, - "output": 4, - "cache_read": 0.2, - "tier": { - "type": "context", - "size": 262144 - } - } - ], - "context_over_200k": { - "input": 1, - "output": 4, - "cache_read": 0.2 - } + "input": 1, + "output": 2 } }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "openai/gpt-5-image-mini": { + "id": "openai/gpt-5-image-mini", + "name": "GPT-5 Image Mini", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "release_date": "2025-10-16", + "last_updated": "2025-10-16", "modalities": { "input": [ - "text", + "pdf", "image", - "video" + "text" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "tiers": [ - { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24, - "tier": { - "type": "context", - "size": 262144 - } - } - ], - "context_over_200k": { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24 - } + "input": 2.5, + "output": 2, + "cache_read": 0.25 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Codex GPT for repository edits, code review, and practical software agents", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -154081,43 +145133,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "low", "medium", "high", "xhigh" @@ -154125,11 +145162,11 @@ } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -154141,45 +145178,38 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 30, - "output": 180, - "cache_read": 3, - "tiers": [ - { - "input": 60, - "output": 270, - "cache_read": 3, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270, - "cache_read": 3 - } + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "openai/gpt-oss-120b:free": { + "id": "openai/gpt-oss-120b:free", + "name": "gpt-oss-120b (free)", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "structured_output": false, "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -154188,39 +145218,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16385, - "output": 4096 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 0, + "output": 0 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "openai/gpt-4o-2024-08-06": { + "id": "openai/gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-08-06", + "last_updated": "2024-08-06", "modalities": { "input": [ "text", @@ -154233,61 +145253,42 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 30, - "output": 180, - "cache_read": 3, - "tiers": [ - { - "input": 60, - "output": 270, - "cache_read": 3, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270, - "cache_read": 3 - } + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" @@ -154295,27 +145296,27 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -154331,9 +145332,9 @@ "last_updated": "2025-12-11", "modalities": { "input": [ - "text", + "pdf", "image", - "pdf" + "text" ], "output": [ "text" @@ -154351,10 +145352,10 @@ "cache_read": 0.175 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "family": "gpt-codex", "attachment": true, "reasoning": true, @@ -154364,20 +145365,22 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -154390,32 +145393,31 @@ "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "openai/gpt-audio-mini": { + "id": "openai/gpt-audio-mini", + "name": "GPT Audio Mini", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "o-mini", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text", - "image", - "pdf" + "audio" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, @@ -154424,38 +145426,41 @@ "output": 16384 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.6, + "output": 2.4 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, + "openai/gpt-5.6-luna": { + "id": "openai/gpt-5.6-luna", + "name": "GPT-5.6 Luna", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -154463,28 +145468,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1, + "output": 6, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -154494,9 +145499,9 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -154510,7 +145515,7 @@ "limit": { "context": 400000, "input": 272000, - "output": 128000 + "output": 100000 }, "cost": { "input": 0.25, @@ -154518,33 +145523,37 @@ "cache_read": 0.025 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "openai/gpt-5.6-terra": { + "id": "openai/gpt-5.6-terra", + "name": "GPT-5.6 Terra", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -154552,40 +145561,35 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "cache_write": 3.125 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "openai/o4-mini-deep-research": { + "id": "openai/o4-mini-deep-research", + "name": "o4-mini-deep-research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", "family": "o-mini", - "attachment": false, + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "release_date": "2024-06-26", + "last_updated": "2024-06-26", "modalities": { "input": [ + "pdf", + "image", "text" ], "output": [ @@ -154598,56 +145602,9 @@ "output": 100000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 - } - }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, "openai/gpt-4.1-nano": { @@ -154665,8 +145622,9 @@ "last_updated": "2025-04-14", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ "text" @@ -154683,18 +145641,17 @@ "cache_read": 0.025 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" @@ -154703,14 +145660,45 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.036, + "output": 0.18 + } + }, + "openai/gpt-4o-2024-11-20": { + "id": "openai/gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-11-20", + "last_updated": "2024-11-20", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -154718,14 +145706,13 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, "openai/o1": { @@ -154772,31 +145759,25 @@ "cache_read": 7.5 } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", + "openai/o1-pro": { + "id": "openai/o1-pro", + "name": "o1-pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "knowledge": "2023-09", + "release_date": "2025-03-19", + "last_updated": "2025-03-19", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -154804,43 +145785,31 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 272000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 15, - "output": 120 + "input": 150, + "output": 600 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "openai/gpt-chat-latest": { + "id": "openai/gpt-chat-latest", + "name": "GPT Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "temperature": false, + "release_date": "2026-05-05", + "last_updated": "2026-05-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -154849,54 +145818,55 @@ "open_weights": false, "limit": { "context": 400000, - "input": 272000, "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", + "openai/gpt-5-image": { + "id": "openai/gpt-5-image", + "name": "GPT-5 Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", "attachment": true, - "reasoning": false, - "tool_call": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "knowledge": "2024-10-01", + "release_date": "2025-10-14", + "last_updated": "2025-10-14", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 10, + "output": 10, + "cache_read": 1.25 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -154913,14 +145883,15 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, + "temperature": false, "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -154928,21 +145899,21 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -154960,14 +145931,14 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text", + "pdf", "image", - "pdf" + "text" ], "output": [ "text" @@ -154975,64 +145946,34 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -155040,45 +145981,33 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "openai/o3-deep-research": { + "id": "openai/o3-deep-research", + "name": "o3-deep-research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "family": "o", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "knowledge": "2024-05", + "release_date": "2024-06-26", + "last_updated": "2024-06-26", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ @@ -155087,50 +146016,31 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 10, + "output": 40, + "cache_read": 2.5 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "openai/gpt-4-turbo-preview": { + "id": "openai/gpt-4-turbo-preview", + "name": "GPT-4 Turbo Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "family": "gpt", - "attachment": true, + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2023-12-31", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -155138,28 +146048,38 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 128000, + "output": 4096 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 10, + "output": 30 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", "family": "gpt-mini", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -155172,40 +146092,29 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT-4.1 mini", + "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", @@ -155218,33 +146127,13 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1047576, + "output": 32768 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, "openai/gpt-4-turbo": { @@ -155255,7 +146144,7 @@ "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, "knowledge": "2023-12", "release_date": "2023-11-06", @@ -155279,17 +146168,18 @@ "output": 30 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -155299,13 +146189,14 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -155318,34 +146209,34 @@ "output": 128000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.05, + "output": 0.4, + "cache_read": 0.01 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", @@ -155358,29 +146249,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 30, + "output": 180 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", + "openai/o3-mini-high": { + "id": "openai/o3-mini-high", + "name": "o3 Mini High", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", "high" ] } @@ -155388,13 +146276,13 @@ "tool_call": true, "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "knowledge": "2023-10-31", + "release_date": "2025-02-12", + "last_updated": "2025-02-12", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" @@ -155402,21 +146290,20 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + "openai/gpt-5.4-image-2": { + "id": "openai/gpt-5.4-image-2", + "name": "GPT-5.4 Image 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -155431,97 +146318,49 @@ ] } ], - "tool_call": true, + "tool_call": false, "structured_output": true, "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text", "image", + "text", "pdf" ], "output": [ + "image", "text" ] }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 272000, "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 8, + "output": 15, + "cache_read": 2 } - } - } - }, - "xpersona": { - "id": "xpersona", - "env": [ - "XPERSONA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://www.xpersona.co/v1", - "name": "Xpersona", - "doc": "https://www.xpersona.co/docs", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + }, + "openai/gpt-4o-search-preview": { + "id": "openai/gpt-4o-search-preview", + "name": "GPT-4o Search Preview", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": false, + "knowledge": "2023-10-31", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -155529,47 +146368,42 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "output": 128000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.75, - "output": 6, - "reasoning": 6, - "cache_read": 0.075 + "input": 2.5, + "output": 10 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -155577,44 +146411,33 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 0.6, - "output": 3.7, - "reasoning": 3.7, - "cache_read": 0.06 + "input": 30, + "output": 180 } }, - "xpersona-frieren-coder": { - "id": "xpersona-frieren-coder", - "name": "Xpersona Frieren 1", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-30", - "release_date": "2026-05-01", - "last_updated": "2026-05-25", + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -155622,22 +146445,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 128000, + "output": 16384 }, "cost": { - "input": 1.5, - "output": 6, - "reasoning": 6, - "cache_read": 0.15 + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "gpt-5.6": { - "id": "gpt-5.6", - "name": "GPT-5.6", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -155645,53 +146467,45 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 372000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 12, - "reasoning": 12, - "cache_read": 0.15 + "input": 0.029, + "output": 0.14 } }, - "xpersona-gpt-5.5": { - "id": "xpersona-gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "openai/gpt-5.6-sol": { + "id": "openai/gpt-5.6-sol", + "name": "GPT-5.6 Sol", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", "family": "gpt", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -155701,18 +146515,16 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "knowledge": "2025-12-30", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "knowledge": "2026-02-16", + "release_date": "2026-07-09", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -155720,21 +146532,22 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 3, - "output": 18, - "reasoning": 18, - "cache_read": 0.3 + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -155743,21 +146556,16 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", @@ -155769,21 +146577,21 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 1.55, - "output": 12.2, - "reasoning": 12.2, - "cache_read": 0.155 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -155793,19 +146601,16 @@ "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -155817,22 +146622,22 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 3, - "output": 18.5, - "reasoning": 18.5, - "cache_read": 0.3 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, + "openai/gpt-oss-20b:free": { + "id": "openai/gpt-oss-20b:free", + "name": "gpt-oss-20b (free)", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -155840,73 +146645,62 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 272000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.375, - "output": 4, - "reasoning": 4, - "cache_read": 0.0375 + "input": 0, + "output": 0 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ "text" @@ -155914,18 +146708,18 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 1.5, - "output": 9.25, - "reasoning": 9.25, - "cache_read": 0.15 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "gpt-5.5": { - "id": "gpt-5.5", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", "name": "GPT-5.5", "description": "Default frontier GPT for coding, computer use, research, and knowledge work", "family": "gpt", @@ -155935,18 +146729,15 @@ { "type": "effort", "values": [ + "none", "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, "knowledge": "2025-12-01", @@ -155955,7 +146746,8 @@ "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -155964,189 +146756,119 @@ "open_weights": false, "limit": { "context": 1050000, + "input": 922000, "output": 128000 }, "cost": { - "input": 1.5, - "output": 12, - "reasoning": 12, - "cache_read": 0.15 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "thedrummer/cydonia-24b-v4.1": { + "id": "thedrummer/cydonia-24b-v4.1", + "name": "Cydonia 24B V4.1", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2024-04-30", + "release_date": "2025-09-27", + "last_updated": "2025-09-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.9, - "output": 5.55, - "reasoning": 5.55, - "cache_read": 0.09 + "input": 0.3, + "output": 0.5, + "cache_read": 0.15 } }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "thedrummer/skyfall-36b-v2": { + "id": "thedrummer/skyfall-36b-v2", + "name": "Skyfall 36B V2", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2024-06-30", + "release_date": "2025-03-10", + "last_updated": "2025-03-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 372000, - "output": 128000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 2, - "reasoning": 2, - "cache_read": 0.15 + "input": 0.55, + "output": 0.8, + "cache_read": 0.25 } }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "thedrummer/unslopnemo-12b": { + "id": "thedrummer/unslopnemo-12b", + "name": "UnslopNemo 12B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2024-04-30", + "release_date": "2024-11-08", + "last_updated": "2024-11-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 372000, - "output": 128000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 12, - "reasoning": 12, - "cache_read": 0.15 + "input": 0.4, + "output": 0.4 } - } - } - }, - "qiniu-ai": { - "id": "qiniu-ai", - "env": [ - "QINIU_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.qnaigc.com/v1", - "name": "Qiniu", - "doc": "https://developer.qiniu.com/aitokenapi", - "models": { - "glm-4.5": { - "id": "glm-4.5", - "name": "GLM 4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + }, + "thedrummer/rocinante-12b": { + "id": "thedrummer/rocinante-12b", + "name": "Rocinante 12B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-04-30", + "release_date": "2024-09-30", + "last_updated": "2024-09-30", "modalities": { "input": [ "text" @@ -156155,90 +146877,97 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 65536, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 0.5 } }, - "gemini-3.0-flash-preview": { - "id": "gemini-3.0-flash-preview", - "name": "Gemini 3.0 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "bytedance/ui-tars-1.5-7b": { + "id": "bytedance/ui-tars-1.5-7b", + "name": "UI-TARS 7B ", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "knowledge": "2025-01-31", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ - "text", "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "output": 2048 + }, + "cost": { + "input": 0.1, + "output": 0.2, + "cache_read": 0.1 } }, - "doubao-seed-2.0-mini": { - "id": "doubao-seed-2.0-mini", - "name": "Doubao Seed 2.0 Mini", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "rekaai/reka-flash-3": { + "id": "rekaai/reka-flash-3", + "name": "Reka Flash 3", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "reka", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-01-31", + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 65536, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.2 } }, - "mimo-v2-flash": { - "id": "mimo-v2-flash", - "name": "Mimo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "rekaai/reka-edge": { + "id": "rekaai/reka-edge", + "name": "Reka Edge", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "reka", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-12-01", - "release_date": "2025-12-16", - "last_updated": "2026-02-04", + "release_date": "2026-03-20", + "last_updated": "2026-03-20", "modalities": { "input": [ - "text" + "image", + "text", + "video" ], "output": [ "text" @@ -156246,36 +146975,31 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 16384, + "output": 16384 }, "cost": { "input": 0.1, - "output": 0.3, - "cache_read": 0.01 + "output": 0.1 } }, - "doubao-seed-2.0-lite": { - "id": "doubao-seed-2.0-lite", - "name": "Doubao Seed 2.0 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "mistralai/mistral-large-2407": { + "id": "mistralai/mistral-large-2407", + "name": "Mistral Large 2407", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2024-03-31", + "release_date": "2024-11-19", + "last_updated": "2024-11-19", "modalities": { "input": [ "text", - "image", - "video" + "pdf" ], "output": [ "text" @@ -156283,102 +147007,131 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "doubao-seed-1.6-thinking": { - "id": "doubao-seed-1.6-thinking", - "name": "Doubao-Seed 1.6 Thinking", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "mistralai/mistral-small-3.2-24b-instruct": { + "id": "mistralai/mistral-small-3.2-24b-instruct", + "name": "Mistral Small 3.2 24B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-15", - "last_updated": "2025-08-15", + "knowledge": "2023-10-31", + "release_date": "2025-06-20", + "last_updated": "2025-06-20", "modalities": { "input": [ "image", - "text", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.075, + "output": 0.2 } }, - "gemini-2.5-flash-image": { - "id": "gemini-2.5-flash-image", - "name": "Gemini 2.5 Flash Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, + "mistralai/mistral-nemo": { + "id": "mistralai/mistral-nemo", + "name": "Mistral Nemo", + "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", + "family": "mistral-nemo", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-10-22", - "last_updated": "2025-10-22", + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 8192 + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.02, + "output": 0.03 } }, - "gemini-3.0-pro-image-preview": { - "id": "gemini-3.0-pro-image-preview", - "name": "Gemini 3.0 Pro Image Preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "mistralai/mistral-medium-3-5": { + "id": "mistralai/mistral-medium-3-5", + "name": "Mistral Medium 3.5", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 1.5, + "output": 7.5 } }, - "claude-3.5-haiku": { - "id": "claude-3.5-haiku", - "name": "Claude 3.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "mistralai/ministral-8b-2512": { + "id": "mistralai/ministral-8b-2512", + "name": "Ministral 3 8B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", @@ -156388,56 +147141,66 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.15, + "output": 0.15, + "cache_read": 0.015 } }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235b A22B Instruct 2507", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, + "mistralai/mistral-small-3.1-24b-instruct": { + "id": "mistralai/mistral-small-3.1-24b-instruct", + "name": "Mistral Small 3.1 24B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": false, "temperature": true, - "release_date": "2025-08-12", - "last_updated": "2025-08-12", + "knowledge": "2023-10-31", + "release_date": "2025-03-17", + "last_updated": "2025-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 64000 + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0.351, + "output": 0.555 } }, - "qwen3-next-80b-a3b-thinking": { - "id": "qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B Thinking", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "mistralai/mistral-saba": { + "id": "mistralai/mistral-saba", + "name": "Saba", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-09-12", - "last_updated": "2025-09-12", + "knowledge": "2024-09-30", + "release_date": "2025-02-17", + "last_updated": "2025-02-17", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -156445,24 +147208,32 @@ }, "open_weights": false, "limit": { - "context": 131072, + "context": 32768, "output": 32768 + }, + "cost": { + "input": 0.2, + "output": 0.6, + "cache_read": 0.02 } }, - "qwen-max-2025-01-25": { - "id": "qwen-max-2025-01-25", - "name": "Qwen2.5-Max-2025-01-25", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "attachment": false, + "mistralai/mistral-large": { + "id": "mistralai/mistral-large", + "name": "Mistral Large", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-11-30", + "release_date": "2024-02-26", + "last_updated": "2024-02-26", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -156471,23 +147242,32 @@ "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "output": 128000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "deepseek-v3-0324": { - "id": "deepseek-v3-0324", - "name": "DeepSeek-V3-0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "mistralai/mistral-medium-3.1": { + "id": "mistralai/mistral-medium-3.1", + "name": "Mistral Medium 3.1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-06-30", + "release_date": "2025-08-13", + "last_updated": "2025-08-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -156495,26 +147275,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16000 + "context": 131072, + "output": 262144 + }, + "cost": { + "input": 0.4, + "output": 2, + "cache_read": 0.04 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "mistralai/mistral-small-24b-instruct-2501": { + "id": "mistralai/mistral-small-24b-instruct-2501", + "name": "Mistral Small 3", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2023-10-31", + "release_date": "2025-01-30", + "last_updated": "2025-01-30", "modalities": { "input": [ "text" @@ -156523,49 +147305,70 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 40000, - "output": 4096 + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 0.05, + "output": 0.08 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, + "mistralai/ministral-3b-2512": { + "id": "mistralai/ministral-3b-2512", + "name": "Ministral 3 3B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.1, + "cache_read": 0.01 } }, - "claude-4.5-sonnet": { - "id": "claude-4.5-sonnet", - "name": "Claude 4.5 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "mistralai/mistral-small-2603": { + "id": "mistralai/mistral-small-2603", + "name": "Mistral Small 4", + "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", + "family": "mistral-small", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2025-06", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text", @@ -156575,109 +147378,134 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "deepseek-v3": { - "id": "deepseek-v3", - "name": "DeepSeek-V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "mistralai/ministral-14b-2512": { + "id": "mistralai/ministral-14b-2512", + "name": "Ministral 3 14B 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-13", - "last_updated": "2025-08-13", + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16000 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.2, + "output": 0.2, + "cache_read": 0.02 } }, - "qwen3-coder-480b-a35b-instruct": { - "id": "qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, + "mistralai/devstral-2512": { + "id": "mistralai/devstral-2512", + "name": "Devstral 2", + "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", + "family": "devstral", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-14", - "last_updated": "2025-08-14", + "knowledge": "2025-12", + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262000, - "output": 4096 + "context": 262144, + "output": 262144 + }, + "status": "deprecated", + "cost": { + "input": 0.4, + "output": 2, + "cache_read": 0.04 } }, - "qwen-turbo": { - "id": "qwen-turbo", - "name": "Qwen-Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "mistralai/mixtral-8x22b-instruct": { + "id": "mistralai/mixtral-8x22b-instruct", + "name": "Mixtral 8x22B Instruct", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": true, + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-01-31", + "release_date": "2024-04-17", + "last_updated": "2024-04-17", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 4096 + "context": 65536, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.2 } }, - "gemini-2.0-flash": { - "id": "gemini-2.0-flash", - "name": "Gemini 2.0 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "mistralai/mistral-medium-3": { + "id": "mistralai/mistral-medium-3", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-03-31", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -156685,83 +147513,99 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 8192 + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.4, + "output": 2, + "cache_read": 0.04 } }, - "gemini-2.0-flash-lite": { - "id": "gemini-2.0-flash-lite", - "name": "Gemini 2.0 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "mistralai/voxtral-small-24b-2507": { + "id": "mistralai/voxtral-small-24b-2507", + "name": "Voxtral Small 24B 2507", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-10-30", + "last_updated": "2025-10-30", "modalities": { "input": [ "text", - "image", "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 8192 + "context": 32000, + "output": 32000 + }, + "cost": { + "input": 0.1, + "output": 0.3, + "cache_read": 0.01 } }, - "claude-4.0-sonnet": { - "id": "claude-4.0-sonnet", - "name": "Claude 4.0 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "mistralai/mistral-large-2512": { + "id": "mistralai/mistral-large-2512", + "name": "Mistral Large 3", + "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", + "family": "mistral-large", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-11", + "release_date": "2024-11-01", + "last_updated": "2025-12-02", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.5, + "output": 1.5, + "cache_read": 0.05 } }, - "qwen2.5-vl-7b-instruct": { - "id": "qwen2.5-vl-7b-instruct", - "name": "Qwen 2.5 VL 7B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "mistralai/codestral-2508": { + "id": "mistralai/codestral-2508", + "name": "Codestral 2508", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "codestral", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-03-31", + "release_date": "2025-08-01", + "last_updated": "2025-08-01", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -156769,26 +147613,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.3, + "output": 0.9, + "cache_read": 0.03 } }, - "qwen3-235b-a22b-thinking-2507": { - "id": "qwen3-235b-a22b-thinking-2507", - "name": "Qwen3 235B A22B Thinking 2507", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "morph/morph-v3-fast": { + "id": "morph/morph-v3-fast", + "name": "Morph V3 Fast", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "morph", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": false, "temperature": true, - "release_date": "2025-08-12", - "last_updated": "2025-08-12", + "release_date": "2025-07-07", + "last_updated": "2025-07-07", "modalities": { "input": [ "text" @@ -156799,21 +147644,26 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 4096 + "context": 81920, + "output": 38000 + }, + "cost": { + "input": 0.8, + "output": 1.2 } }, - "qwen3-235b-a22b": { - "id": "qwen3-235b-a22b", - "name": "Qwen 3 235B A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "morph/morph-v3-large": { + "id": "morph/morph-v3-large", + "name": "Morph V3 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "morph", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-07-07", + "last_updated": "2025-07-07", "modalities": { "input": [ "text" @@ -156824,30 +147674,31 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.9, + "output": 1.9 } }, - "doubao-seed-2.0-pro": { - "id": "doubao-seed-2.0-pro", - "name": "Doubao Seed 2.0 Pro", + "bytedance-seed/seed-1.6-flash": { + "id": "bytedance-seed/seed-1.6-flash", + "name": "Seed 1.6 Flash", "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", "image", + "text", "video" ], "output": [ @@ -156856,30 +147707,31 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.075, + "output": 0.3 } }, - "doubao-seed-1.6-flash": { - "id": "doubao-seed-1.6-flash", - "name": "Doubao-Seed 1.6 Flash", + "bytedance-seed/seed-1.6": { + "id": "bytedance-seed/seed-1.6", + "name": "Seed 1.6", "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-15", - "last_updated": "2025-08-15", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", "image", + "text", "video" ], "output": [ @@ -156888,24 +147740,42 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.25, + "output": 2 } }, - "qwen3-max-preview": { - "id": "qwen3-max-preview", - "name": "Qwen3 Max Preview", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, - "reasoning": false, + "bytedance-seed/seed-2.0-mini": { + "id": "bytedance-seed/seed-2.0-mini", + "name": "Seed-2.0-Mini", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-09-06", - "last_updated": "2025-09-06", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -156913,24 +147783,42 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.4 } }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, - "reasoning": false, + "bytedance-seed/seed-2.0-lite": { + "id": "bytedance-seed/seed-2.0-lite", + "name": "Seed-2.0-Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-09-12", - "last_updated": "2025-09-12", + "release_date": "2026-03-10", + "last_updated": "2026-03-10", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -156938,26 +147826,40 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.25, + "output": 2 } }, - "qwen3-vl-30b-a3b-thinking": { - "id": "qwen3-vl-30b-a3b-thinking", - "name": "Qwen3-Vl 30b A3b Thinking", - "description": "Multimodal model for analyzing text, images, documents, and rich media", + "sakana/fugu-ultra": { + "id": "sakana/fugu-ultra", + "name": "Fugu Ultra", + "description": "Quality-first multi-agent model for hard research, analysis, and competitions", + "family": "fugu", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "max", + "xhigh", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-09", - "last_updated": "2026-02-09", + "structured_output": true, + "temperature": false, + "release_date": "2026-06-24", + "last_updated": "2026-06-24", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -156965,21 +147867,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "doubao-1.5-pro-32k": { - "id": "doubao-1.5-pro-32k", - "name": "Doubao 1.5 Pro 32k", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "anthracite-org/magnum-v4-72b": { + "id": "anthracite-org/magnum-v4-72b", + "name": "Magnum v4 72B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-06-30", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ "text" @@ -156988,51 +147896,60 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 12000 + "context": 32768, + "output": 2048 + }, + "cost": { + "input": 3, + "output": 5 } }, - "claude-3.5-sonnet": { - "id": "claude-3.5-sonnet", - "name": "Claude 3.5 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "attachment": true, + "nvidia/nemotron-3-nano-30b-a3b:free": { + "id": "nvidia/nemotron-3-nano-30b-a3b:free", + "name": "Nemotron 3 Nano 30B A3B (free)", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-09-09", - "last_updated": "2025-09-09", + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8200 + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "gpt-oss-20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "nvidia/nemotron-nano-9b-v2:free": { + "id": "nvidia/nemotron-nano-9b-v2:free", + "name": "Nemotron Nano 9B V2 (free)", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "release_date": "2025-08-18", + "last_updated": "2025-08-18", "modalities": { "input": [ "text" @@ -157041,110 +147958,153 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 128000, - "output": 4096 + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "nvidia/nemotron-nano-12b-v2-vl:free": { + "id": "nvidia/nemotron-nano-12b-v2-vl:free", + "name": "Nemotron Nano 12B 2 VL (free)", + "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", + "family": "nemotron", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-10-28", + "last_updated": "2025-10-28", "modalities": { "input": [ "text", "image", - "video", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 - } - }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free": { + "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free", + "name": "Nemotron 3 Nano Omni (free)", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ "text", "image", - "audio", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 64000 + "context": 256000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 } }, - "qwen-vl-max-2025-01-25": { - "id": "qwen-vl-max-2025-01-25", - "name": "Qwen VL-MAX-2025-01-25", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "attachment": true, - "reasoning": false, + "nvidia/nemotron-3-ultra-550b-a55b:free": { + "id": "nvidia/nemotron-3-ultra-550b-a55b:free", + "name": "Nemotron 3 Ultra (free)", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0, + "output": 0 } }, - "doubao-1.5-thinking-pro": { - "id": "doubao-1.5-thinking-pro", - "name": "Doubao 1.5 Thinking Pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "nvidia/nemotron-3-ultra-550b-a55b": { + "id": "nvidia/nemotron-3-ultra-550b-a55b", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium", + "high" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text" @@ -157153,24 +148113,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16000 + "context": 1000000, + "output": 16384 + }, + "cost": { + "input": 0.5, + "output": 2.2, + "cache_read": 0.1 } }, - "claude-4.5-haiku": { - "id": "claude-4.5-haiku", - "name": "Claude 4.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "nvidia/nemotron-3.5-content-safety:free": { + "id": "nvidia/nemotron-3.5-content-safety:free", + "name": "Nemotron 3.5 Content Safety (free)", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "nemotron", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, "structured_output": false, "temperature": true, - "release_date": "2025-10-16", - "last_updated": "2025-10-16", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ "text", @@ -157180,109 +148146,144 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 } }, - "kling-v2-6": { - "id": "kling-v2-6", - "name": "Kling-V2 6", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "nvidia/nemotron-3-nano-30b-a3b": { + "id": "nvidia/nemotron-3-nano-30b-a3b", + "name": "Nemotron 3 Nano 30B A3B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-01-13", - "last_updated": "2026-01-13", + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 99999999, - "output": 99999999 + "context": 262144, + "output": 228000 + }, + "cost": { + "input": 0.05, + "output": 0.2 } }, - "claude-4.1-opus": { - "id": "claude-4.1-opus", - "name": "Claude 4.1 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "attachment": true, + "nvidia/llama-3.3-nemotron-super-49b-v1.5": { + "id": "nvidia/llama-3.3-nemotron-super-49b-v1.5", + "name": "Llama 3.3 Nemotron Super 49B v1.5", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0.4, + "output": 0.4 } }, - "claude-3.7-sonnet": { - "id": "claude-3.7-sonnet", - "name": "Claude 3.7 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "attachment": true, + "nvidia/nemotron-3-super-120b-a12b:free": { + "id": "nvidia/nemotron-3-super-120b-a12b:free", + "name": "Nemotron 3 Super (free)", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium" + ] + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 1000000, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 } }, - "MiniMax-M1": { - "id": "MiniMax-M1", - "name": "MiniMax M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium" + ] + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text" @@ -157291,16 +148292,52 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 80000 + "output": 16384 + }, + "cost": { + "input": 0.08, + "output": 0.45 } }, - "doubao-seed-1.6": { - "id": "doubao-seed-1.6", - "name": "Doubao-Seed 1.6", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "cognitivecomputations/dolphin-mistral-24b-venice-edition:free": { + "id": "cognitivecomputations/dolphin-mistral-24b-venice-edition:free", + "name": "Uncensored (free)", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2024-04-30", + "release_date": "2025-07-09", + "last_updated": "2025-07-09", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32768, + "output": 32768 + }, + "cost": { + "input": 0, + "output": 0 + } + }, + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -157309,30 +148346,41 @@ } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_details" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-08-15", - "last_updated": "2025-08-15", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.105, + "output": 0.28, + "cache_read": 0.028 } }, - "qwen3-30b-a3b": { - "id": "qwen3-30b-a3b", - "name": "Qwen3 30B A3B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -157341,10 +148389,14 @@ } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -157353,32 +148405,43 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 40000, - "output": 4096 + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "inception/mercury-2": { + "id": "inception/mercury-2", + "name": "Mercury 2", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "mercury", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-22", - "last_updated": "2026-02-22", + "release_date": "2026-03-04", + "last_updated": "2026-03-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -157386,31 +148449,43 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 128000, + "output": 50000 + }, + "cost": { + "input": 0.25, + "output": 0.75, + "cache_read": 0.025 } }, - "doubao-seed-2.0-code": { - "id": "doubao-seed-2.0-code", - "name": "Doubao Seed 2.0 Code", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "anthropic/claude-sonnet-4.5": { + "id": "anthropic/claude-sonnet-4.5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -157418,26 +148493,62 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "claude-4.0-opus": { - "id": "claude-4.0-opus", - "name": "Claude 4.0 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ + "image", "text", - "image" + "pdf" ], "output": [ "text" @@ -157445,24 +148556,62 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "kimi-k2": { - "id": "kimi-k2", - "name": "Kimi K2", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "attachment": false, - "reasoning": false, + "anthropic/claude-haiku-4.5": { + "id": "anthropic/claude-haiku-4.5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -157470,24 +148619,48 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "qwen3-30b-a3b-instruct-2507": { - "id": "qwen3-30b-a3b-instruct-2507", - "name": "Qwen3 30b A3b Instruct 2507", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, - "reasoning": false, + "anthropic/claude-opus-4.7-fast": { + "id": "anthropic/claude-opus-4.7-fast", + "name": "Claude Opus 4.7 (Fast)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-04", - "last_updated": "2026-02-04", + "structured_output": true, + "temperature": false, + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -157495,27 +148668,46 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 } }, - "qwen2.5-vl-72b-instruct": { - "id": "qwen2.5-vl-72b-instruct", - "name": "Qwen 2.5 VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "anthropic/claude-sonnet-5": { + "id": "anthropic/claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -157523,26 +148715,49 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "doubao-1.5-vision-pro": { - "id": "doubao-1.5-vision-pro", - "name": "Doubao 1.5 Vision Pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -157550,26 +148765,67 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16000 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "claude-4.5-opus": { - "id": "claude-4.5-opus", - "name": "Claude 4.5 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -157577,25 +148833,46 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 200000 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-v3.1": { - "id": "deepseek-v3.1", - "name": "DeepSeek-V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "attachment": false, + "anthropic/claude-fable-5": { + "id": "anthropic/claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-19", - "last_updated": "2025-08-19", + "structured_output": true, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -157603,25 +148880,44 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek-R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "attachment": false, + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, + "knowledge": "2025-03-31", "release_date": "2025-08-05", "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -157629,29 +148925,52 @@ }, "open_weights": false, "limit": { - "context": 128000, + "context": 200000, "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "qwen3-30b-a3b-thinking-2507": { - "id": "qwen3-30b-a3b-thinking-2507", - "name": "Qwen3 30b A3b Thinking 2507", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, + "anthropic/claude-opus-4.5": { + "id": "anthropic/claude-opus-4.5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-04", - "last_updated": "2026-02-04", + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -157659,27 +148978,53 @@ }, "open_weights": false, "limit": { - "context": 126000, - "output": 32000 + "context": 200000, + "output": 64000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "audio", - "video" + "pdf" ], "output": [ "text" @@ -157687,25 +149032,51 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 64000 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "deepseek-r1-0528": { - "id": "deepseek-r1-0528", - "name": "DeepSeek-R1-0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "anthropic/claude-3-haiku": { + "id": "anthropic/claude-3-haiku", + "name": "Claude 3 Haiku", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2023-08-31", + "release_date": "2024-03-13", + "last_updated": "2024-03-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -157713,29 +149084,44 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 200000, + "output": 4096 + }, + "cost": { + "input": 0.25, + "output": 1.25, + "cache_read": 0.03, + "cache_write": 0.3 } }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "GLM 4.5 Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "attachment": false, + "anthropic/claude-opus-4": { + "id": "anthropic/claude-opus-4", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 } ], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-01-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "image", + "text", + "pdf" ], "output": [ "text" @@ -157743,25 +149129,48 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 4096 - } - }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "attachment": false, + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 + } + }, + "anthropic/claude-opus-4.8-fast": { + "id": "anthropic/claude-opus-4.8-fast", + "name": "Claude Opus 4.8 (Fast)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-06", - "last_updated": "2025-08-06", + "structured_output": true, + "temperature": false, + "release_date": "2026-05-27", + "last_updated": "2026-05-27", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -157769,29 +149178,53 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "gemini-3.0-pro-preview": { - "id": "gemini-3.0-pro-preview", - "name": "Gemini 3.0 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", "image", - "video", - "pdf", - "audio" + "pdf" ], "output": [ "text" @@ -157800,21 +149233,55 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "output": 128000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "deepseek/deepseek-v3.2-exp-thinking": { - "id": "deepseek/deepseek-v3.2-exp-thinking", - "name": "DeepSeek/DeepSeek-V3.2-Exp-Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "tencent/hy3:free": { + "id": "tencent/hy3:free", + "name": "Hy3 (free)", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hy3", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ "text" @@ -157823,24 +149290,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 } }, - "deepseek/deepseek-math-v2": { - "id": "deepseek/deepseek-math-v2", - "name": "Deepseek/Deepseek-Math-V2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "tencent/hy3": { + "id": "tencent/hy3", + "name": "Hy3", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hy3", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-04", - "last_updated": "2025-12-04", + "release_date": "2026-07-06", + "last_updated": "2026-07-06", "modalities": { "input": [ "text" @@ -157849,23 +149330,31 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 160000, - "output": 160000 + "context": 262144, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.58, + "cache_read": 0.035 } }, - "deepseek/deepseek-v3.1-terminus": { - "id": "deepseek/deepseek-v3.1-terminus", - "name": "DeepSeek/DeepSeek-V3.1-Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "tencent/hunyuan-a13b-instruct": { + "id": "tencent/hunyuan-a13b-instruct", + "name": "Hunyuan A13B Instruct", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "hunyuan", "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "knowledge": "2025-03-31", + "release_date": "2025-07-08", + "last_updated": "2025-07-08", "modalities": { "input": [ "text" @@ -157874,24 +149363,38 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.14, + "output": 0.57 } }, - "deepseek/deepseek-v3.2-251201": { - "id": "deepseek/deepseek-v3.2-251201", - "name": "Deepseek/DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "tencent/hy3-preview": { + "id": "tencent/hy3-preview", + "name": "Hy3 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", + "family": "Hy", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "high" + ] + } + ], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text" @@ -157900,24 +149403,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.063, + "output": 0.21, + "cache_read": 0.021 } }, - "deepseek/deepseek-v3.1-terminus-thinking": { - "id": "deepseek/deepseek-v3.1-terminus-thinking", - "name": "DeepSeek/DeepSeek-V3.1-Terminus-Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "deepcogito/cogito-v2.1-671b": { + "id": "deepcogito/cogito-v2.1-671b", + "name": "Cogito v2.1 671B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "cogito", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": false, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text" @@ -157929,20 +149438,26 @@ "open_weights": false, "limit": { "context": 128000, - "output": 32000 + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 1.25 } }, - "deepseek/deepseek-v3.2-exp": { - "id": "deepseek/deepseek-v3.2-exp", - "name": "DeepSeek/DeepSeek-V3.2-Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", + "cohere/command-a": { + "id": "cohere/command-a", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2024-08-31", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ "text" @@ -157951,23 +149466,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 32000 + "context": 256000, + "output": 8192 + }, + "cost": { + "input": 2.5, + "output": 10 } }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "Z-AI/GLM 4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "cohere/command-r-08-2024": { + "id": "cohere/command-r-08-2024", + "name": "Command R", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-10-11", - "last_updated": "2025-10-11", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -157976,28 +149497,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 200000 + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.15, + "output": 0.6 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "Z-Ai/GLM 5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "cohere/command-r7b-12-2024": { + "id": "cohere/command-r7b-12-2024", + "name": "Command R7B", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2024-06-01", + "release_date": "2024-12-02", + "last_updated": "2024-12-02", "modalities": { "input": [ "text" @@ -158006,28 +149528,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 0.0375, + "output": 0.15 } }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "Z-Ai/GLM 4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "cohere/command-r-plus-08-2024": { + "id": "cohere/command-r-plus-08-2024", + "name": "Command R+", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -158036,52 +149559,75 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 200000 + "context": 128000, + "output": 4000 + }, + "cost": { + "input": 2.5, + "output": 10 } }, - "z-ai/autoglm-phone-9b": { - "id": "z-ai/autoglm-phone-9b", - "name": "Z-Ai/Autoglm Phone 9b", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "attachment": true, - "reasoning": false, + "cohere/north-mini-code:free": { + "id": "cohere/north-mini-code:free", + "name": "North Mini Code (free)", + "description": "Cohere coding model for practical software engineering and agentic edits", + "family": "north", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2026-06-17", + "last_updated": "2026-06-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 12800, - "output": 4096 + "context": 256000, + "output": 64000 + }, + "cost": { + "input": 0, + "output": 0 } }, - "meituan/longcat-flash-chat": { - "id": "meituan/longcat-flash-chat", - "name": "Meituan/Longcat-Flash-Chat", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, + "~x-ai/grok-latest": { + "id": "~x-ai/grok-latest", + "name": "Grok Latest", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-11-05", - "last_updated": "2025-11-05", + "release_date": "2026-07-08", + "last_updated": "2026-07-08", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -158089,21 +149635,27 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 500000, + "output": 1000000 + }, + "cost": { + "input": 2, + "output": 6, + "cache_read": 0.5 } }, - "meituan/longcat-flash-lite": { - "id": "meituan/longcat-flash-lite", - "name": "Meituan/Longcat-Flash-Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "gryphe/mythomax-l2-13b": { + "id": "gryphe/mythomax-l2-13b", + "name": "MythoMax 13B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-02-06", - "last_updated": "2026-02-06", + "knowledge": "2023-06-30", + "release_date": "2023-07-02", + "last_updated": "2023-07-02", "modalities": { "input": [ "text" @@ -158112,136 +149664,169 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 320000 + "context": 4096, + "output": 4096 + }, + "cost": { + "input": 0.06, + "output": 0.06 } }, - "x-ai/grok-4-fast-reasoning": { - "id": "x-ai/grok-4-fast-reasoning", - "name": "X-Ai/Grok-4-Fast-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "stepfun/step-3.7-flash": { + "id": "stepfun/step-3.7-flash", + "name": "Step 3.7 Flash", + "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "knowledge": "2026-01-01", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 256000, + "input": 256000, + "output": 256000 + }, + "cost": { + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 } }, - "x-ai/grok-4.1-fast-reasoning": { - "id": "x-ai/grok-4.1-fast-reasoning", - "name": "X-Ai/Grok 4.1 Fast Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": true, + "stepfun/step-3.5-flash": { + "id": "stepfun/step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-12-19", - "last_updated": "2025-12-19", + "knowledge": "2025-01", + "release_date": "2026-01-29", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 20000000, - "output": 2000000 + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.1, + "output": 0.3 } }, - "x-ai/grok-4-fast-non-reasoning": { - "id": "x-ai/grok-4-fast-non-reasoning", - "name": "X-Ai/Grok-4-Fast-Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "nex-agi/nex-n2-mini": { + "id": "nex-agi/nex-n2-mini", + "name": "Nex-N2-Mini", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "agi", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-18", + "release_date": "2026-06-24", + "last_updated": "2026-06-24", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.025, + "output": 0.1, + "cache_read": 0.0025 } }, - "x-ai/grok-code-fast-1": { - "id": "x-ai/grok-code-fast-1", - "name": "x-AI/Grok-Code-Fast 1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "attachment": false, + "nex-agi/nex-n2-pro": { + "id": "nex-agi/nex-n2-pro", + "name": "Nex-N2-Pro", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "agi", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-09-02", - "last_updated": "2025-09-02", + "release_date": "2026-06-08", + "last_updated": "2026-06-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 10000 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.25, + "output": 1, + "cache_read": 0.025 } }, - "x-ai/grok-4.1-fast": { - "id": "x-ai/grok-4.1-fast", - "name": "x-AI/Grok-4.1-Fast", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "undi95/remm-slerp-l2-13b": { + "id": "undi95/remm-slerp-l2-13b", + "name": "ReMM SLERP 13B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "knowledge": "2023-06-30", + "release_date": "2023-07-22", + "last_updated": "2023-07-22", "modalities": { "input": [ "text" @@ -158250,29 +149835,46 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 6144, + "output": 4096 + }, + "cost": { + "input": 0.45, + "output": 0.65 } }, - "x-ai/grok-4.1-fast-non-reasoning": { - "id": "x-ai/grok-4.1-fast-non-reasoning", - "name": "X-Ai/Grok 4.1 Fast Non Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "~openai/gpt-mini-latest": { + "id": "~openai/gpt-mini-latest", + "name": "OpenAI GPT Mini Latest", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-19", - "last_updated": "2025-12-19", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text", + "pdf", "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -158280,28 +149882,45 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "x-ai/grok-4-fast": { - "id": "x-ai/grok-4-fast", - "name": "x-AI/Grok-4-Fast", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "~openai/gpt-latest": { + "id": "~openai/gpt-latest", + "name": "OpenAI GPT Latest", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-20", - "last_updated": "2025-09-20", + "structured_output": true, + "temperature": false, + "knowledge": "2026-02-16", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text", + "pdf", "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -158309,21 +149928,29 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 1050000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "stepfun-ai/gelab-zero-4b-preview": { - "id": "stepfun-ai/gelab-zero-4b-preview", - "name": "Stepfun-Ai/Gelab Zero 4b Preview", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "~moonshotai/kimi-latest": { + "id": "~moonshotai/kimi-latest", + "name": "MoonshotAI Kimi Latest", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", @@ -158335,21 +149962,26 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 4096 + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.66, + "output": 3.41, + "cache_read": 0.15 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "relace/relace-search": { + "id": "relace/relace-search", + "name": "Relace Search", + "description": "Tool-capable chat model for instruction following and agentic application workflows", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-11-07", - "last_updated": "2025-11-07", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text" @@ -158361,25 +149993,27 @@ "open_weights": false, "limit": { "context": 256000, - "output": 100000 + "output": 128000 + }, + "cost": { + "input": 1, + "output": 3 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Moonshotai/Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "attachment": true, + "relace/relace-apply-3": { + "id": "relace/relace-apply-3", + "name": "Relace Apply 3", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": false, - "temperature": true, - "release_date": "2026-01-28", - "last_updated": "2026-01-28", + "temperature": false, + "release_date": "2025-09-26", + "last_updated": "2025-09-26", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -158388,20 +150022,26 @@ "open_weights": false, "limit": { "context": 256000, - "output": 256000 + "output": 128000 + }, + "cost": { + "input": 0.85, + "output": 1.25 } }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "ai21/jamba-large-1.7": { + "id": "ai21/jamba-large-1.7", + "name": "Jamba Large 1.7", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "jamba", "attachment": false, "reasoning": false, "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2025-09-08", - "last_updated": "2025-09-08", + "knowledge": "2024-08-31", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ "text" @@ -158410,25 +150050,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 256000, - "output": 100000 + "output": 4096 + }, + "cost": { + "input": 2, + "output": 8 } }, - "xiaomi/mimo-v2-flash": { - "id": "xiaomi/mimo-v2-flash", - "name": "Xiaomi/Mimo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", + "arcee-ai/coder-large": { + "id": "arcee-ai/coder-large", + "name": "Coder Large", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-12-01", - "release_date": "2025-12-16", - "last_updated": "2026-02-04", + "knowledge": "2025-03-31", + "release_date": "2025-05-05", + "last_updated": "2025-05-05", "modalities": { "input": [ "text" @@ -158437,33 +150080,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01 + "input": 0.5, + "output": 0.8 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "Minimax/Minimax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "arcee-ai/virtuoso-large": { + "id": "arcee-ai/virtuoso-large", + "name": "Virtuoso Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-03-31", + "release_date": "2025-05-05", + "last_updated": "2025-05-05", "modalities": { "input": [ "text" @@ -158474,26 +150112,27 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 128000 + "context": 131072, + "output": 64000 + }, + "cost": { + "input": 0.75, + "output": 1.2 } }, - "minimax/minimax-m2.5-highspeed": { - "id": "minimax/minimax-m2.5-highspeed", - "name": "Minimax/Minimax-M2.5 Highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "arcee-ai/trinity-large-thinking": { + "id": "arcee-ai/trinity-large-thinking", + "name": "Trinity Large Thinking", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "trinity", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text" @@ -158502,28 +150141,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 128000 + "context": 262144, + "output": 80000 + }, + "cost": { + "input": 0.25, + "output": 0.8, + "cache_read": 0.06 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "Minimax/Minimax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "arcee-ai/trinity-mini": { + "id": "arcee-ai/trinity-mini", + "name": "Trinity Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "trinity-mini", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text" @@ -158532,28 +150173,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 204800, - "output": 128000 + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.045, + "output": 0.15 } }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "Minimax/Minimax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "mancer/weaver": { + "id": "mancer/weaver", + "name": "Weaver (alpha)", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "alpha", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", + "knowledge": "2023-06-30", + "release_date": "2023-08-02", + "last_updated": "2023-08-02", "modalities": { "input": [ "text" @@ -158564,21 +150206,27 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 8000, + "output": 2000 + }, + "cost": { + "input": 0.75, + "output": 1 } }, - "stepfun/step-3.5-flash": { - "id": "stepfun/step-3.5-flash", - "name": "Stepfun/Step-3.5 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "perplexity/sonar-reasoning-pro": { + "id": "perplexity/sonar-reasoning-pro", + "name": "Sonar Reasoning Pro", + "description": "Web-grounded reasoning model for multi-step research and cited answers", + "family": "sonar-reasoning", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, "structured_output": false, "temperature": true, - "release_date": "2026-02-02", - "last_updated": "2026-02-02", + "release_date": "2025-03-07", + "last_updated": "2025-03-07", "modalities": { "input": [ "text", @@ -158590,22 +150238,26 @@ }, "open_weights": false, "limit": { - "context": 64000, - "output": 4096 + "context": 128000, + "output": 128000 + }, + "cost": { + "input": 2, + "output": 8 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "OpenAI/GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "perplexity/sonar": { + "id": "perplexity/sonar", + "name": "Sonar", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": false, "temperature": true, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2025-01-27", + "last_updated": "2025-01-27", "modalities": { "input": [ "text", @@ -158617,24 +150269,30 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 127072, + "output": 127072 + }, + "cost": { + "input": 1, + "output": 1 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "OpenAI/GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "attachment": false, + "perplexity/sonar-pro": { + "id": "perplexity/sonar-pro", + "name": "Sonar Pro", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": false, "temperature": true, - "release_date": "2025-09-19", - "last_updated": "2025-09-19", + "release_date": "2025-03-07", + "last_updated": "2025-03-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -158642,74 +150300,59 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 200000, + "output": 8000 + }, + "cost": { + "input": 3, + "output": 15 } - } - } - }, - "modelscope": { - "id": "modelscope", - "env": [ - "MODELSCOPE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api-inference.modelscope.cn/v1", - "name": "ModelScope", - "doc": "https://modelscope.cn/docs/model-service/API-Inference/intro", - "models": { - "ZhipuAI/GLM-4.6": { - "id": "ZhipuAI/GLM-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + }, + "perplexity/sonar-pro-search": { + "id": "perplexity/sonar-pro-search", + "name": "Sonar Pro Search", + "description": "Advanced Sonar search model for deeper research and cited synthesis", + "family": "sonar-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "release_date": "2025-10-30", + "last_updated": "2025-10-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 98304 + "context": 200000, + "output": 8000 }, "cost": { - "input": 0, - "output": 0 + "input": 3, + "output": 15 } }, - "ZhipuAI/GLM-4.5": { - "id": "ZhipuAI/GLM-4.5", - "name": "GLM-4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "perplexity/sonar-deep-research": { + "id": "perplexity/sonar-deep-research", + "name": "Sonar Deep Research", + "description": "Sonar search model for current answers, retrieval, and citation-backed chat", + "family": "sonar-deep-research", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "release_date": "2025-03-07", + "last_updated": "2025-03-07", "modalities": { "input": [ "text" @@ -158718,28 +150361,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 8, + "reasoning": 3 } }, - "Qwen/Qwen3-30B-A3B-Instruct-2507": { - "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen3 30B A3B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "openrouter/bodybuilder": { + "id": "openrouter/bodybuilder", + "name": "Body Builder (beta)", + "description": "Preview model for early access evaluation, prototyping, and compatibility testing", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-30", - "last_updated": "2025-07-30", + "tool_call": false, + "structured_output": false, + "temperature": false, + "release_date": "2025-12-05", + "last_updated": "2025-12-05", "modalities": { "input": [ "text" @@ -158748,59 +150391,55 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 + "context": 128000, + "output": 128000 } }, - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen3-235B-A22B-Thinking-2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "openrouter/free": { + "id": "openrouter/free", + "name": "Free Models Router", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", + "release_date": "2026-02-01", + "last_updated": "2026-02-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 8000 }, "cost": { "input": 0, "output": 0 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "openrouter/fusion": { + "id": "openrouter/fusion", + "name": "Fusion", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-07-21", + "tool_call": false, + "structured_output": false, + "temperature": false, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -158809,29 +150448,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 + "context": 1000000, + "output": 128000 } }, - "Qwen/Qwen3-30B-A3B-Thinking-2507": { - "id": "Qwen/Qwen3-30B-A3B-Thinking-2507", - "name": "Qwen3 30B A3B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "openrouter/pareto-code": { + "id": "openrouter/pareto-code", + "name": "Pareto Code Router", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-30", - "last_updated": "2025-07-30", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text" @@ -158840,89 +150473,71 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 + "context": 2000000, + "output": 200000 } }, - "Qwen/Qwen3-Coder-30B-A3B-Instruct": { - "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct", - "name": "Qwen3 Coder 30B A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openrouter/auto": { + "id": "openrouter/auto", + "name": "Auto Router", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "auto", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-31", - "last_updated": "2025-07-31", + "release_date": "2023-11-08", + "last_updated": "2023-11-08", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "pdf", + "video" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 + "context": 2000000, + "output": 2000000 } - } - } - }, - "google": { - "id": "google", - "env": [ - "GOOGLE_API_KEY", - "GOOGLE_GENERATIVE_AI_API_KEY", - "GEMINI_API_KEY" - ], - "npm": "@ai-sdk/google", - "name": "Google", - "doc": "https://ai.google.dev/gemini-api/docs/models", - "models": { - "gemini-flash-latest": { - "id": "gemini-flash-latest", - "name": "Gemini Flash Latest", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", + }, + "qwen/qwen3.5-plus-20260420": { + "id": "qwen/qwen3.5-plus-20260420", + "name": "Qwen3.5 Plus 2026-04-20", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.5", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" @@ -158930,37 +150545,31 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "input_audio": 0.75 + "input": 0.3, + "output": 1.8, + "cache_write": 0.375 } }, - "gemma-4-31b-it": { - "id": "gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen/qwen3-next-80b-a3b-instruct:free": { + "id": "qwen/qwen3-next-80b-a3b-instruct:free", + "name": "Qwen3 Next 80B A3B Instruct (free)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -158969,319 +150578,301 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 262144 + }, + "cost": { + "input": 0, + "output": 0 } }, - "gemini-omni-flash-preview": { - "id": "gemini-omni-flash-preview", - "name": "Gemini Omni Flash Preview", - "description": "Video generation and editing model for fast, conversational text- and image-to-video workflows", - "family": "gemini", + "qwen/qwen3-vl-235b-a22b-thinking": { + "id": "qwen/qwen3-vl-235b-a22b-thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, - "output": 65536 + "output": 32768 }, "cost": { - "input": 1.5, - "output": 17.5 + "input": 0.26, + "output": 2.6 } }, - "gemini-3.1-flash-image-preview": { - "id": "gemini-3.1-flash-image-preview", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", + "qwen/qwen3-vl-30b-a3b-thinking": { + "id": "qwen/qwen3-vl-30b-a3b-thinking", + "name": "Qwen3 VL 30B A3B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "knowledge": "2025-03-31", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 60 + "input": 0.13, + "output": 1.56 } }, - "veo-3.1-lite-generate-preview": { - "id": "veo-3.1-lite-generate-preview", - "name": "Veo 3.1 lite", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-03-31", - "last_updated": "2026-03-31", + "tool_call": true, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 480, - "output": 8192 + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.65, + "output": 3.25, + "cache_read": 0.13, + "cache_write": 0.8125 } }, - "lyria-3-pro-preview": { - "id": "lyria-3-pro-preview", - "name": "Lyria 3 Pro Preview", - "description": "Music generation model for full-length songs from text or images with vocals and structure", - "family": "lyria", - "attachment": true, + "qwen/qwen-plus": { + "id": "qwen/qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-25", - "last_updated": "2026-03-25", + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.26, + "output": 0.78, + "cache_read": 0.052, + "cache_write": 0.325 } }, - "gemini-3.1-flash-tts-preview": { - "id": "gemini-3.1-flash-tts-preview", - "name": "Gemini 3.1 Flash TTS Preview", - "description": "Low-latency speech generation with steerable prompts and expressive audio tags", - "family": "gemini-flash", + "qwen/qwen3-coder-30b-a3b-instruct": { + "id": "qwen/qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-15", - "last_updated": "2026-04-15", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 16384 + "context": 160000, + "output": 32768 }, "cost": { - "input": 1, - "output": 20 + "input": 0.07, + "output": 0.27 } }, - "gemini-2.5-flash-image": { - "id": "gemini-2.5-flash-image", - "name": "Nano Banana", - "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", - "family": "gemini-flash", - "attachment": true, + "qwen/qwen3-32b": { + "id": "qwen/qwen3-32b", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 30, - "cache_read": 0.075 + "input": 0.08, + "output": 0.28 } }, - "gemini-flash-lite-latest": { - "id": "gemini-flash-lite-latest", - "name": "Gemini Flash-Lite Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 0.09, + "output": 1.1 } }, - "gemini-embedding-2": { - "id": "gemini-embedding-2", - "name": "Gemini Embedding 2", - "description": "Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space", - "family": "gemini", + "qwen/qwen3-vl-8b-instruct": { + "id": "qwen/qwen3-vl-8b-instruct", + "name": "Qwen3 VL 8B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-11", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-14", + "last_updated": "2025-10-14", "modalities": { "input": [ - "text", "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1 + "context": 256000, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 0, - "input_audio": 6.5 + "input": 0.117, + "output": 0.455 } }, - "gemini-3.6-flash": { - "id": "gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "qwen/qwen3.7-plus": { + "id": "qwen/qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -159289,21 +150880,21 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "input_audio": 0.75 + "input": 0.32, + "output": 1.28, + "cache_read": 0.064, + "cache_write": 0.4 } }, - "gemma-4-26b-a4b-it": { - "id": "gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "qwen/qwen3.6-35b-a3b": { + "id": "qwen/qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -159314,12 +150905,13 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -159328,103 +150920,140 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 262144 + }, + "cost": { + "input": 0.14, + "output": 1 } }, - "gemini-3-pro-image-preview": { - "id": "gemini-3-pro-image-preview", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", - "attachment": true, + "qwen/qwen3.7-max": { + "id": "qwen/qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 262144 + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 2, - "output": 120 + "input": 1.25, + "output": 3.75, + "cache_read": 0.25, + "cache_write": 1.5625 } }, - "gemini-2.5-pro-preview-tts": { - "id": "gemini-2.5-pro-preview-tts", - "name": "Gemini 2.5 Pro Preview TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gemini-flash", + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-05-01", - "last_updated": "2025-05-01", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 8192, - "output": 16384 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1, - "output": 20 + "input": 0.78, + "output": 3.9, + "cache_read": 0.156, + "cache_write": 0.975 } }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, + "qwen/qwen3-8b": { + "id": "qwen/qwen3-8b", + "name": "Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.117, + "output": 0.455 + } + }, + "qwen/qwen-plus-2025-07-28": { + "id": "qwen/qwen-plus-2025-07-28", + "name": "Qwen Plus 0728", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2025-03-31", + "release_date": "2025-09-08", + "last_updated": "2025-09-08", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -159432,628 +151061,446 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 0.26, + "output": 0.78 } }, - "gemini-3.1-flash-image": { - "id": "gemini-3.1-flash-image", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", + "qwen/qwen3.5-flash-02-23": { + "id": "qwen/qwen3.5-flash-02-23", + "name": "Qwen3.5-Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "release_date": "2026-02-25", + "last_updated": "2026-02-25", "modalities": { "input": [ "text", "image", - "video", - "pdf" + "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.5, - "output": 60 + "input": 0.065, + "output": 0.26 } }, - "lyria-3-clip-preview": { - "id": "lyria-3-clip-preview", - "name": "Lyria 3 Clip Preview", - "description": "Music generation model for short 30-second clips, loops, and previews from text or image prompts", - "family": "lyria", - "attachment": true, + "qwen/qwen3-coder:free": { + "id": "qwen/qwen3-coder:free", + "name": "Qwen3 Coder 480B A35B (free)", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-03-25", - "last_updated": "2026-03-25", + "knowledge": "2025-06-30", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, - "output": 65536 + "output": 262000 }, "cost": { "input": 0, "output": 0 } }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "qwen/qwen3-30b-a3b-instruct-2507": { + "id": "qwen/qwen3-30b-a3b-instruct-2507", + "name": "Qwen3 30B A3B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "knowledge": "2025-06-30", + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 32000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.04815, + "output": 0.19305 } }, - "gemini-2.5-computer-use-preview-10-2025": { - "id": "gemini-2.5-computer-use-preview-10-2025", - "name": "Gemini 2.5 Computer Use Preview 10-2025", - "description": "Specialized Gemini 2.5 model for browser-control agents that automate UI tasks", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "qwen/qwen-2.5-coder-32b-instruct": { + "id": "qwen/qwen-2.5-coder-32b-instruct", + "name": "Qwen2.5 Coder 32B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-10-07", - "last_updated": "2025-10-07", + "knowledge": "2024-06-30", + "release_date": "2024-11-11", + "last_updated": "2024-11-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 128000, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "tiers": [ - { - "input": 2.5, - "output": 15, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15 - } + "input": 0.66, + "output": 1 } }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 + "input": 0.0975, + "output": 0.78 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "qwen/qwen3-235b-a22b-thinking-2507": { + "id": "qwen/qwen3-235b-a22b-thinking-2507", + "name": "Qwen3 235B A22B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2025-06-30", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.1495, + "output": 1.495 } }, - "gemini-3-pro-image": { - "id": "gemini-3-pro-image", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", + "qwen/qwen3-vl-32b-instruct": { + "id": "qwen/qwen3-vl-32b-instruct", + "name": "Qwen3 VL 32B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], - "tool_call": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "release_date": "2025-10-23", + "last_updated": "2025-10-23", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, + "context": 262144, "output": 32768 }, "cost": { - "input": 2, - "output": 120 + "input": 0.104, + "output": 0.416 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], + "qwen/qwen3-coder": { + "id": "qwen/qwen3-coder", + "name": "Qwen3 Coder 480B A35B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2025-06-30", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1048576, "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 - } - }, - "gemini-2.5-flash-preview-tts": { - "id": "gemini-2.5-flash-preview-tts", - "name": "Gemini 2.5 Flash Preview TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gemini-flash", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-05-01", - "last_updated": "2025-05-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 16384 - }, - "cost": { - "input": 0.5, - "output": 10 + "input": 0.22, + "output": 1.8 } }, - "gemini-3.1-flash-lite-image": { - "id": "gemini-3.1-flash-lite-image", - "name": "Nano Banana 2 Lite", - "description": "Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing", - "family": "gemini-flash-lite", + "qwen/qwen3.6-flash": { + "id": "qwen/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, + "context": 1000000, "output": 65536 }, "cost": { - "input": 0.25, - "output": 30 + "input": 0.1875, + "output": 1.125, + "cache_write": 0.234375 } }, - "deep-research-preview-04-2026": { - "id": "deep-research-preview-04-2026", - "name": "Deep Research Preview (Apr-21-2026)", - "description": "Agentic model for autonomous multi-step research, synthesis, and cited reports", - "family": "gemini-pro", + "qwen/qwen3.5-plus-02-15": { + "id": "qwen/qwen3.5-plus-02-15", + "name": "Qwen3.5 Plus 2026-02-15", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 81920 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 131072, + "context": 1000000, "output": 65536 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.26, + "output": 1.56 } }, - "deep-research-max-preview-04-2026": { - "id": "deep-research-max-preview-04-2026", - "name": "Deep Research Max Preview (Apr-21-2026)", - "description": "Maximum-comprehensiveness agentic researcher for multi-step investigation, synthesis, and cited reports", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen/qwen-2.5-7b-instruct": { + "id": "qwen/qwen-2.5-7b-instruct", + "name": "Qwen2.5 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2024-06-30", + "release_date": "2024-10-16", + "last_updated": "2024-10-16", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, - "output": 65536 + "output": 32768 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.04, + "output": 0.1 } }, - "veo-3.1-fast-generate-preview": { - "id": "veo-3.1-fast-generate-preview", - "name": "Veo 3.1 fast", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", + "qwen/qwen3-vl-8b-thinking": { + "id": "qwen/qwen3-vl-8b-thinking", + "name": "Qwen3 VL 8B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-10-15", - "last_updated": "2026-01-01", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2025-10-14", + "last_updated": "2025-10-14", "modalities": { "input": [ - "text", "image", - "video" + "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 8192 + "context": 256000, + "output": 32768 + }, + "cost": { + "input": 0.117, + "output": 1.365 } }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, + "qwen/qwen3-max-thinking": { + "id": "qwen/qwen3-max-thinking", + "name": "Qwen3 Max Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "release_date": "2026-02-09", + "last_updated": "2026-02-09", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -160061,262 +151508,211 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "input_audio": 0.75 + "input": 0.78, + "output": 3.9 } }, - "veo-3.1-generate-preview": { - "id": "veo-3.1-generate-preview", - "name": "Veo 3.1", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-10-15", - "last_updated": "2026-01", + "qwen/qwen3-30b-a3b-thinking-2507": { + "id": "qwen/qwen3-30b-a3b-thinking-2507", + "name": "Qwen3 30B A3B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "temperature": true, + "knowledge": "2025-06-30", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 480, - "output": 8192 + "context": 131072, + "output": 32768 }, - "status": "beta" + "cost": { + "input": 0.13, + "output": 1.56 + } }, - "gemini-3.1-pro-preview-customtools": { - "id": "gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "qwen/qwen2.5-vl-72b-instruct": { + "id": "qwen/qwen2.5-vl-72b-instruct", + "name": "Qwen2.5 VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "knowledge": "2024-06-30", + "release_date": "2025-02-01", + "last_updated": "2025-02-01", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 128000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.8, + "output": 1, + "cache_read": 0.4 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", + "qwen/qwen3.5-27b": { + "id": "qwen/qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 262144, "output": 65536 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 + "input": 0.195, + "output": 1.56 } }, - "gemini-3.8-flash": { - "id": "gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, + "qwen/qwen3-235b-a22b": { + "id": "qwen/qwen3-235b-a22b", + "name": "Qwen3 235B-A22B", + "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 38912 } ], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "input_audio": 0.75 + "input": 0.455, + "output": 1.82 } }, - "gemini-3.5-live-translate-preview": { - "id": "gemini-3.5-live-translate-preview", - "name": "Gemini 3.5 Live Translate Preview", - "description": "Low-latency audio-to-audio model for real-time speech translation across 70+ languages", - "family": "gemini-pro", + "qwen/qwen-2.5-72b-instruct": { + "id": "qwen/qwen-2.5-72b-instruct", + "name": "Qwen2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "knowledge": "2024-06-30", + "release_date": "2024-09-19", + "last_updated": "2024-09-19", "modalities": { "input": [ - "audio" + "text" ], "output": [ - "audio", "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "output": 32768 + "context": 131072, + "output": 16384 }, "cost": { - "input": 3.5, - "output": 21, - "input_audio": 3.5, - "output_audio": 21 + "input": 0.36, + "output": 0.4 } }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, + "qwen/qwen-plus-2025-07-28:thinking": { + "id": "qwen/qwen-plus-2025-07-28:thinking", + "name": "Qwen Plus 0728 (thinking)", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "budget_tokens", - "min": 512, - "max": 24576 + "min": 1, + "max": 81920 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2025-03-31", + "release_date": "2025-09-08", + "last_updated": "2025-09-08", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -160324,28 +151720,27 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 + "input": 0.26, + "output": 0.78, + "cache_write": 0.325 } }, - "gemini-embedding-001": { - "id": "gemini-embedding-001", - "name": "Gemini Embedding 001", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "gemini", + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-05", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-02-04", + "last_updated": "2026-02-04", "modalities": { "input": [ "text" @@ -160354,178 +151749,135 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2048, - "output": 1 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.15, - "output": 0 + "input": 0.11, + "output": 0.8, + "cache_read": 0.07 } }, - "gemini-3.1-flash-lite-preview": { - "id": "gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gemini-flash-lite", + "qwen/qwen3.6-27b": { + "id": "qwen/qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262140 }, - "status": "deprecated", "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 0.285, + "output": 2.4, + "cache_read": 0.15 } }, - "gemini-3.1-flash-live-preview": { - "id": "gemini-3.1-flash-live-preview", - "name": "Gemini 3.1 Flash Live Preview", - "description": "High-quality, low-latency Live API model for real-time dialogue and voice-first AI applications", - "family": "gemini-flash", + "qwen/qwen3.5-35b-a3b": { + "id": "qwen/qwen3.5-35b-a3b", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-26", - "last_updated": "2026-03-26", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "video", - "audio" + "video" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 262144, + "output": 81920 }, "cost": { - "input": 0.75, - "output": 4.5, - "input_audio": 3, - "output_audio": 12 + "input": 0.14, + "output": 1, + "cache_read": 0.05 } }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "qwen/qwen3.5-9b": { + "id": "qwen/qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 0.1, + "output": 0.15 } - } - } - }, - "vancine": { - "id": "vancine", - "env": [ - "VANCINE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://vancine.com/v1", - "name": "Vancine", - "doc": "https://vancine.com/docs", - "models": { - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + }, + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -160534,12 +151886,10 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", @@ -160552,43 +151902,32 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 512000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.24, - "output": 0.96, - "cache_read": 0.048 + "input": 0.385, + "output": 2.45, + "cache_read": 0.111 } }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "qwen/qwen3-vl-30b-a3b-instruct": { + "id": "qwen/qwen3-vl-30b-a3b-instruct", + "name": "Qwen3 VL 30B A3B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-03-31", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" @@ -160596,151 +151935,99 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.12, - "output": 0.4, - "cache_read": 0.024 + "input": 0.13, + "output": 0.52 } }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", + "qwen/qwen3-235b-a22b-2507": { + "id": "qwen/qwen3-235b-a22b-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 - } - ], + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "knowledge": "2025-06-30", + "release_date": "2025-07-21", + "last_updated": "2025-07-21", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 16384 }, "cost": { - "input": 1.6, - "output": 4.8, - "cache_read": 0.2 + "input": 0.09, + "output": 0.1 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "qwen/qwen3-coder-flash": { + "id": "qwen/qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "structured_output": false, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 2.4, - "output": 12, - "cache_read": 0.24 + "input": 0.195, + "output": 0.975, + "cache_read": 0.039, + "cache_write": 0.24375 } }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "qwen/qwen3-14b": { + "id": "qwen/qwen3-14b", + "name": "Qwen3 14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-03-31", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -160748,79 +152035,65 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 131702, + "output": 40960 }, "cost": { - "input": 0.24, - "output": 0.96, - "cache_read": 0.0048 + "input": 0.1, + "output": 0.24 } }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "qwen/qwen3-vl-235b-a22b-instruct": { + "id": "qwen/qwen3-vl-235b-a22b-instruct", + "name": "Qwen3 VL 235B A22B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 16384 }, "cost": { - "input": 0.12, - "output": 0.38, - "cache_read": 0.013 + "input": 0.2, + "output": 0.88, + "cache_read": 0.11 } }, - "hy4-preview": { - "id": "hy4-preview", - "name": "Hy4 preview", - "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", - "family": "Hy", + "qwen/qwen3-30b-a3b": { + "id": "qwen/qwen3-30b-a3b", + "name": "Qwen3 30B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", + "knowledge": "2025-03-31", + "release_date": "2025-04-28", + "last_updated": "2025-04-28", "modalities": { "input": [ "text" @@ -160831,37 +152104,37 @@ }, "open_weights": true, "limit": { - "context": 1024000, - "output": 64000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.67, - "output": 2, - "cache_read": 0.034 + "input": 0.12, + "output": 0.5 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "qwen/qwen3.6-max-preview": { + "id": "qwen/qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 131072 } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text" @@ -160870,60 +152143,39 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.12, - "output": 3.52, - "cache_read": 0.208 + "input": 1.04, + "output": 6.24, + "cache_write": 1.3 } - } - } - }, - "zhipuai-coding-plan": { - "id": "zhipuai-coding-plan", - "env": [ - "ZHIPU_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://open.bigmodel.cn/api/coding/paas/v4", - "name": "Zhipu AI Coding Plan", - "doc": "https://docs.bigmodel.cn/cn/coding-plan/overview", - "models": { - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + }, + "qwen/qwen3.5-122b-a10b": { + "id": "qwen/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "video", - "pdf" + "video" ], "output": [ "text" @@ -160931,168 +152183,135 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.26, + "output": 2.08 } }, - "glm-5.3-highspeed": { - "id": "glm-5.3-highspeed", - "name": "GLM-5.3 Highspeed", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "qwen/qwen3.6-plus": { + "id": "qwen/qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1, + "max": 81920 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.325, + "output": 1.95, + "cache_write": 0.40625 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "amazon/nova-lite-v1": { + "id": "amazon/nova-lite-v1", + "name": "Nova Lite 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2024-10-31", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 300000, + "output": 5120 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.06, + "output": 0.24 } }, - "glm-4.6v": { - "id": "glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", + "amazon/nova-premier-v1": { + "id": "amazon/nova-premier-v1", + "name": "Nova Premier 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", + "release_date": "2025-10-31", + "last_updated": "2025-10-31", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 2.5, + "output": 12.5, + "cache_read": 0.625 } - } - } - }, - "lucidquery": { - "id": "lucidquery", - "env": [ - "LUCIDQUERY_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.lucidquery.com/v1", - "name": "LucidQuery", - "doc": "https://lucidquery.com/docs", - "models": { - "lucidquery-agi-01-frontier": { - "id": "lucidquery-agi-01-frontier", - "name": "AGI-01 Frontier", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "agi", + }, + "amazon/nova-pro-v1": { + "id": "amazon/nova-pro-v1", + "name": "Nova Pro 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2026-06-05", - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "knowledge": "2024-10-31", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text", @@ -161105,26 +152324,26 @@ "open_weights": false, "limit": { "context": 300000, - "output": 120000 + "output": 5120 }, "cost": { - "input": 4.5, - "output": 22 + "input": 0.8, + "output": 3.2 } }, - "lucidquery-nexus-coder": { - "id": "lucidquery-nexus-coder", - "name": "LucidQuery Nexus Coder", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "lucid", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "amazon/nova-micro-v1": { + "id": "amazon/nova-micro-v1", + "name": "Nova Micro 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2025-08-01", - "release_date": "2025-09-01", - "last_updated": "2025-09-01", + "structured_output": false, + "temperature": true, + "knowledge": "2024-10-31", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text" @@ -161135,30 +152354,33 @@ }, "open_weights": false, "limit": { - "context": 250000, - "output": 60000 + "context": 128000, + "output": 5120 }, "cost": { - "input": 2, - "output": 5 + "input": 0.035, + "output": 0.14 } }, - "lucidnova-rf1-100b": { - "id": "lucidnova-rf1-100b", - "name": "LucidNova RF1 100B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "amazon/nova-2-lite-v1": { + "id": "amazon/nova-2-lite-v1", + "name": "Nova 2 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "family": "nova", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2025-09-16", - "release_date": "2024-12-28", - "last_updated": "2025-09-10", + "structured_output": false, + "temperature": true, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "pdf" ], "output": [ "text" @@ -161166,31 +152388,29 @@ }, "open_weights": false, "limit": { - "context": 120000, - "output": 8000 + "context": 1000000, + "output": 65535 }, "cost": { - "input": 2, - "output": 5 + "input": 0.3, + "output": 2.5 } }, - "lucidquery-agi-01-swift": { - "id": "lucidquery-agi-01-swift", - "name": "AGI-01 Swift", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "agi", - "attachment": true, + "aion-labs/aion-3.0-mini": { + "id": "aion-labs/aion-3.0-mini", + "name": "Aion-3.0-Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2026-06-05", - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "release_date": "2026-07-07", + "last_updated": "2026-07-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -161198,45 +152418,28 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 120000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 15 + "input": 0.7, + "output": 1.4, + "cache_read": 0.18 } - } - } - }, - "gmicloud": { - "id": "gmicloud", - "env": [ - "GMICLOUD_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.gmi-serving.com/v1", - "name": "GMI Cloud", - "doc": "https://docs.gmicloud.ai/inference-engine/api-reference/llm-api-reference", - "models": { - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 - } - ], - "tool_call": true, + }, + "aion-labs/aion-rp-llama-3.1-8b": { + "id": "aion-labs/aion-rp-llama-3.1-8b", + "name": "Aion-RP 1.0 (8B)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2023-12-31", + "release_date": "2025-02-04", + "last_updated": "2025-02-04", "modalities": { "input": [ "text" @@ -161247,39 +152450,26 @@ }, "open_weights": false, "limit": { - "context": 409600, - "output": 128000 + "context": 32768, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.8, + "output": 1.6 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, + "aion-labs/aion-3.0": { + "id": "aion-labs/aion-3.0", + "name": "Aion-3.0", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": false, + "temperature": true, + "release_date": "2026-07-07", + "last_updated": "2026-07-07", "modalities": { "input": [ "text" @@ -161290,64 +152480,30 @@ }, "open_weights": false, "limit": { - "context": 409600, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "context": 131072, + "output": 32768 }, "cost": { - "input": 4.5, - "output": 22.5, - "cache_read": 0.45 + "input": 3, + "output": 6, + "cache_read": 0.75 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "aion-labs/aion-2.0": { + "id": "aion-labs/aion-2.0", + "name": "Aion-2.0", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": false, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -161355,34 +152511,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 + "input": 0.8, + "output": 1.6, + "cache_read": 0.2 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, + "inflection/inflection-3-pi": { + "id": "inflection/inflection-3-pi", + "name": "Inflection 3 Pi", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2024-10-31", + "release_date": "2024-10-11", + "last_updated": "2024-10-11", "modalities": { "input": [ "text" @@ -161393,27 +152542,26 @@ }, "open_weights": false, "limit": { - "context": 409600, - "output": 64000 + "context": 8000, + "output": 1024 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 2.5, + "output": 10 } }, - "Qwen/Qwen3.7-Max": { - "id": "Qwen/Qwen3.7-Max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "inflection/inflection-3-productivity": { + "id": "inflection/inflection-3-productivity", + "name": "Inflection 3 Productivity", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2024-10-31", + "release_date": "2024-10-11", + "last_updated": "2024-10-11", "modalities": { "input": [ "text" @@ -161424,33 +152572,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 8000, + "output": 1024 }, "cost": { "input": 2.5, - "output": 7.5, - "cache_read": 0.25, - "cache_write": 3.125 + "output": 10 } }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "sao10k/l3.1-euryale-70b": { + "id": "sao10k/l3.1-euryale-70b", + "name": "Llama 3.1 Euryale 70B v2.2", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-12-31", + "release_date": "2024-08-28", + "last_updated": "2024-08-28", "modalities": { "input": [ "text" @@ -161461,32 +152603,27 @@ }, "open_weights": true, "limit": { - "context": 1048575, - "output": 384000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.112, - "output": 0.224, - "cache_read": 0.022 + "input": 0.85, + "output": 0.85 } }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "sao10k/l3.3-euryale-70b": { + "id": "sao10k/l3.3-euryale-70b", + "name": "Llama 3.3 Euryale 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "reasoning": false, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2023-12-31", + "release_date": "2024-12-18", + "last_updated": "2024-12-18", "modalities": { "input": [ "text" @@ -161497,36 +152634,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 384000 + "context": 131072, + "output": 16384 }, "cost": { - "input": 1.392, - "output": 2.784, - "cache_read": 0.116 + "input": 0.65, + "output": 0.75 } }, - "MiniMaxAI/MiniMax-M3": { - "id": "MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "sao10k/l3-lunaris-8b": { + "id": "sao10k/l3-lunaris-8b", + "name": "Llama 3 8B Lunaris", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "knowledge": "2023-12-31", + "release_date": "2024-08-13", + "last_updated": "2024-08-13", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -161534,43 +152665,58 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 512000 + "context": 8192, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12, - "tiers": [ - { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24, - "tier": { - "type": "context", - "size": 512000 - } - } + "input": 0.04, + "output": 0.05 + } + }, + "sao10k/l3.1-70b-hanami-x1": { + "id": "sao10k/l3.1-70b-hanami-x1", + "name": "Llama 3.1 70B Hanami x1", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": true, + "temperature": true, + "knowledge": "2023-12-31", + "release_date": "2025-01-08", + "last_updated": "2025-01-08", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24 - } + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 16000, + "output": 16000 + }, + "cost": { + "input": 3, + "output": 3 } }, - "MiniMaxAI/MiniMax-M2.7": { - "id": "MiniMaxAI/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "upstage/solar-pro-3": { + "id": "upstage/solar-pro-3", + "name": "Solar Pro 3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "solar-pro", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text" @@ -161579,34 +152725,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 131072 + "context": 128000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "allenai/olmo-3-32b-think": { + "id": "allenai/olmo-3-32b-think", + "name": "Olmo 3 32B Think", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "allenai", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2025-11-21", + "last_updated": "2025-11-21", "modalities": { "input": [ "text" @@ -161621,28 +152763,24 @@ "output": 65536 }, "cost": { - "input": 0.855, - "output": 3.6, - "cache_read": 0.144 + "input": 0.15, + "output": 0.5 } }, - "moonshotai/kimi-k2.7-code-highspeed": { - "id": "moonshotai/kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": true, + "deepseek/deepseek-r1-0528": { + "id": "deepseek/deepseek-r1-0528", + "name": "R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -161653,28 +152791,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 163840, + "output": 32768 }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 + "input": 0.5, + "output": 2.15, + "cache_read": 0.35 } }, - "zai-org/GLM-5.2-FP8": { - "id": "zai-org/GLM-5.2-FP8", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "high", - "max" + "xhigh" ] } ], @@ -161684,8 +152825,9 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -161696,30 +152838,33 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.979, - "output": 3.08, - "cache_read": 0.182 + "input": 0.09, + "output": 0.18, + "cache_read": 0.018 } }, - "zai-org/GLM-5-FP8": { - "id": "zai-org/GLM-5-FP8", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "deepseek/deepseek-v3.1-terminus": { + "id": "deepseek/deepseek-v3.1-terminus", + "name": "DeepSeek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2025-03-31", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ "text" @@ -161730,31 +152875,29 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 163840, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 1.92, - "cache_read": 0.12 + "input": 0.27, + "output": 0.95, + "cache_read": 0.13 } }, - "zai-org/GLM-5.1-FP8": { - "id": "zai-org/GLM-5.1-FP8", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "deepseek/deepseek-r1-distill-llama-70b": { + "id": "deepseek/deepseek-r1-distill-llama-70b", + "name": "R1 Distill Llama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2024-07-31", + "release_date": "2025-01-23", + "last_updated": "2025-01-23", "modalities": { "input": [ "text" @@ -161765,193 +152908,142 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.98, - "output": 3.08, - "cache_read": 0.182 + "input": 0.8, + "output": 0.8 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", "high", "xhigh" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 384000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } - } - } - }, - "oci": { - "id": "oci", - "env": [ - "OCI_GENAI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.generativeai.us-chicago-1.oci.oraclecloud.com/openai/v1", - "name": "OCI Generative AI", - "doc": "https://docs.oracle.com/en-us/iaas/Content/generative-ai/pretrained-models.htm", - "models": { - "xai.grok-4.3": { - "id": "xai.grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + }, + "deepseek/deepseek-r1": { + "id": "deepseek/deepseek-r1", + "name": "DeepSeek-R1", + "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-05-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 30000 + "context": 163840, + "output": 16000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.7, + "output": 2.5 } }, - "xai.grok-4.20-non-reasoning": { - "id": "xai.grok-4.20-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, + "deepseek/deepseek-v3.2-exp": { + "id": "deepseek/deepseek-v3.2-exp", + "name": "DeepSeek V3.2 Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 30000 + "context": 163840, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.27, + "output": 0.41 } }, - "meta.llama-3.3-70b-instruct": { - "id": "meta.llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "deepseek/deepseek-chat-v3-0324": { + "id": "deepseek/deepseek-chat-v3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2024-07-31", + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ "text" @@ -161962,30 +153054,31 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 163840, + "output": 16384 }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 0.24, + "output": 0.9, + "cache_read": 0.135 } }, - "meta.llama-4-scout-17b-16e-instruct": { - "id": "meta.llama-4-scout-17b-16e-instruct", - "name": "Llama 4 Scout 17B Instruct", - "description": "Open Llama with long-context vision for efficient multimodal agents", - "family": "llama", - "attachment": true, + "deepseek/deepseek-chat": { + "id": "deepseek/deepseek-chat", + "name": "DeepSeek Chat", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -161993,90 +153086,72 @@ }, "open_weights": true, "limit": { - "context": 192000, - "output": 16384 + "context": 131072, + "output": 16000 }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 0.2002, + "output": 0.8001 } }, - "xai.grok-4.6": { - "id": "xai.grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 131072, + "output": 64000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "input": 0.2288, + "output": 0.3432, + "cache_read": 0.02288 } }, - "meta.llama-4-maverick-17b-128e-instruct-fp8": { - "id": "meta.llama-4-maverick-17b-128e-instruct-fp8", - "name": "Llama 4 Maverick 17B Instruct", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, + "deepseek/deepseek-chat-v3.1": { + "id": "deepseek/deepseek-chat-v3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-03-31", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -162084,39 +153159,31 @@ }, "open_weights": true, "limit": { - "context": 512000, - "output": 16384 + "context": 163840, + "output": 32768 }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 0.21, + "output": 0.79, + "cache_read": 0.13 } }, - "openai.gpt-oss-120b": { - "id": "openai.gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { - "field": "reasoning_content" + "field": "reasoning_details" }, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -162127,39 +153194,31 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 196608 }, "cost": { "input": 0.15, - "output": 0.6 + "output": 0.9, + "cache_read": 0.05 } }, - "openai.gpt-oss-20b": { - "id": "openai.gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax-M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { - "field": "reasoning_content" + "field": "reasoning_details" }, - "structured_output": true, + "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -162170,27 +153229,28 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.07, - "output": 0.3 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "xai.grok-4.20-reasoning": { - "id": "xai.grok-4.20-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", + "minimax/minimax-01": { + "id": "minimax/minimax-01", + "name": "MiniMax-01", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "knowledge": "2024-03-31", + "release_date": "2025-01-15", + "last_updated": "2025-01-15", "modalities": { "input": [ "text", @@ -162200,181 +153260,99 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 30000 + "context": 1000192, + "output": 1000192 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.2, + "output": 1.1 } - } - } - }, - "cloudflare-ai-gateway": { - "id": "cloudflare-ai-gateway", - "env": [ - "CLOUDFLARE_API_TOKEN", - "CLOUDFLARE_ACCOUNT_ID", - "CLOUDFLARE_GATEWAY_ID" - ], - "npm": "ai-gateway-provider", - "name": "Cloudflare AI Gateway", - "doc": "https://developers.cloudflare.com/ai-gateway/", - "models": { - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + }, + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 1048576, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "minimax/minimax-m2": { + "id": "minimax/minimax-m2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_details" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 204800, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.255, + "output": 1.02 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "minimax/minimax-m2-her": { + "id": "minimax/minimax-m2-her", + "name": "MiniMax M2-her", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2026-01-23", + "last_updated": "2026-01-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -162382,93 +153360,63 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 65536, + "output": 2048 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 204800, + "output": 196608 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.24, + "output": 0.96 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, + "minimax/minimax-m1": { + "id": "minimax/minimax-m1", + "name": "MiniMax M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2024-06-30", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -162476,49 +153424,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 1000000, + "output": 40000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.4, + "output": 2.2 } }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "kwaipilot/kat-coder-pro-v2": { + "id": "kwaipilot/kat-coder-pro-v2", + "name": "KAT-Coder-Pro V2", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "kat-coder", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "temperature": true, + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -162526,324 +153454,234 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 256000, + "output": 80000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "anthropic/claude-fable-5.1": { - "id": "anthropic/claude-fable-5.1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "nousresearch/hermes-3-llama-3.1-405b:free": { + "id": "nousresearch/hermes-3-llama-3.1-405b:free", + "name": "Hermes 3 405B Instruct (free)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "hermes", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "knowledge": "2023-12-31", + "release_date": "2024-08-16", + "last_updated": "2024-08-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 131072, + "output": 131072 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 0, + "output": 0 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "nousresearch/hermes-4-405b": { + "id": "nousresearch/hermes-4-405b", + "name": "Hermes 4 405B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "hermes", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2024-08-31", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 131072, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 1, + "output": 3 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, + "nousresearch/hermes-3-llama-3.1-70b": { + "id": "nousresearch/hermes-3-llama-3.1-70b", + "name": "Hermes 3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "nousresearch", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "temperature": true, + "knowledge": "2023-12-31", + "release_date": "2024-08-18", + "last_updated": "2024-08-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 131072, + "output": 16384 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0.7, + "output": 0.7 } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, + "nousresearch/hermes-3-llama-3.1-405b": { + "id": "nousresearch/hermes-3-llama-3.1-405b", + "name": "Hermes 3 405B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "nousresearch", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "knowledge": "2023-12-31", + "release_date": "2024-08-16", + "last_updated": "2024-08-16", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 131072, + "output": 16384 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 1, + "output": 1 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, + "nousresearch/hermes-4-70b": { + "id": "nousresearch/hermes-4-70b", + "name": "Hermes 4 70B", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "hermes", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2024-08-31", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" + "context": 131072, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.13, + "output": 0.4 } - }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + } + } + }, + "jiekou": { + "id": "jiekou", + "env": [ + "JIEKOU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.jiekou.ai/openai", + "name": "Jiekou.AI", + "doc": "https://docs.jiekou.ai/docs/support/quickstart?utm_source=github_models.dev", + "models": { + "o3": { + "id": "o3", + "name": "o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "attachment": true, + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 131072, - "output": 384000 + "output": 131072 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.145 + "input": 10, + "output": 40 } }, - "unbiased/pareto": { - "id": "unbiased/pareto", - "name": "Pareto", - "description": "Blended multimodal model that runs multiple language models in parallel and returns a synthesized answer", + "grok-code-fast-1": { + "id": "grok-code-fast-1", + "name": "grok-code-fast-1", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -162855,37 +153693,31 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 256000, + "output": 256000 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.25 + "input": 0.18, + "output": 1.35 } }, - "alibaba/qwen3.7-max": { - "id": "alibaba/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, + "gpt-5.2-pro": { + "id": "gpt-5.2-pro", + "name": "gpt-5.2-pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -162893,53 +153725,43 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 3.75, - "cache_read": 0.25 + "input": 18.9, + "output": 151.2 } }, - "alibaba/qwen3.8-max": { - "id": "alibaba/qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "gemini-2.5-pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "xhigh" + "high" ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", "video", - "pdf" + "audio" ], "output": [ "text" @@ -162947,30 +153769,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 1.125, + "output": 9 } }, - "alibaba/qwen3-max": { - "id": "alibaba/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "claude-haiku-4-5-20251001", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -162978,80 +153800,75 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 20000, + "output": 64000 }, "cost": { - "input": 1.2, - "output": 6 + "input": 0.9, + "output": 4.5 } }, - "alibaba/qwen3.5-397b-a17b": { - "id": "alibaba/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "gpt-5-pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "output": 272000 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 13.5, + "output": 108 } }, - "alibaba/qwen3.7-plus": { - "id": "alibaba/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "gemini-2.5-flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" @@ -163059,44 +153876,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 0.32, - "output": 1.28, - "cache_read": 0.064 + "input": 0.27, + "output": 2.25 } }, - "xai/grok-4.7": { - "id": "xai/grok-4.7", - "name": "Grok 4.7", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -163104,36 +153907,19 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "input": 1.1, + "output": 4.4 } }, - "xai/grok-4.3": { - "id": "xai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "gemini-2.5-flash-lite-preview-09-2025": { + "id": "gemini-2.5-flash-lite-preview-09-2025", + "name": "gemini-2.5-flash-lite-preview-09-2025", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -163150,13 +153936,14 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -163164,33 +153951,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0.09, + "output": 0.36 } }, - "xai/grok-4.20-0309-reasoning": { - "id": "xai/grok-4.20-0309-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "claude-opus-4-1-20250805", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -163198,37 +153982,26 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 13.5, + "output": 67.5 } }, - "xai/grok-4.5": { - "id": "xai/grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", + "grok-4-1-fast-non-reasoning": { + "id": "grok-4-1-fast-non-reasoning", + "name": "grok-4-1-fast-non-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -163240,48 +154013,30 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 0.6 - } + "input": 0.18, + "output": 0.45 } }, - "xai/grok-4.20-0309-non-reasoning": { - "id": "xai/grok-4.20-0309-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "gpt-5-chat-latest": { + "id": "gpt-5-chat-latest", + "name": "gpt-5-chat-latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -163289,38 +154044,26 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 30000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 1.125, + "output": 9 } }, - "xai/grok-4.6": { - "id": "xai/grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "claude-opus-4-5-20251101", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -163332,107 +154075,63 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 200000, + "output": 65536 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "input": 4.5, + "output": 22.5 } }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "gpt-5.1-codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 1.125, + "output": 9 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "gpt-5.1-codex-max": { + "id": "gpt-5.1-codex-max", + "name": "gpt-5.1-codex-max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -163440,42 +154139,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, + "context": 400000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai" - }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 1.125, + "output": 9 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "claude-opus-4-20250514": { + "id": "claude-opus-4-20250514", + "name": "claude-opus-4-20250514", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": true, + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -163487,46 +154170,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 200000, + "output": 32000 }, "cost": { - "input": 30, - "output": 180 + "input": 13.5, + "output": 67.5 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -163534,43 +154201,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 131072, + "output": 131072 }, "cost": { - "input": 30, - "output": 180 + "input": 1.1, + "output": 4.4 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "gpt-5.2": { + "id": "gpt-5.2", + "name": "gpt-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -163582,37 +154232,30 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 272000, + "context": 400000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai" - }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 1.575, + "output": 12.6 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "grok-4-0709": { + "id": "grok-4-0709", + "name": "grok-4-0709", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -163620,30 +154263,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 256000, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 5, - "cache_read": 0.625 + "input": 2.7, + "output": 13.5 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "gemini-2.5-flash-lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", "high" @@ -163652,14 +154291,15 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" @@ -163667,42 +154307,26 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 1048576, + "output": 65535 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.09, + "output": 0.36 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "claude-sonnet-4-20250514": { + "id": "claude-sonnet-4-20250514", + "name": "claude-sonnet-4-20250514", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -163715,43 +154339,30 @@ "open_weights": false, "limit": { "context": 200000, - "output": 100000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "output": 64000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 2.7, + "output": 13.5 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": false, + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "gpt-5.1-codex-mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -163759,31 +154370,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 400000, + "output": 128000 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 0.225, + "output": 1.8 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", + "grok-4-fast-reasoning": { + "id": "grok-4-fast-reasoning", + "name": "grok-4-fast-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -163795,42 +154401,43 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "input": 0.18, + "output": 0.45 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "claude-opus-4-6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02", + "last_updated": "2026-02", "modalities": { "input": [ "text", @@ -163842,47 +154449,30 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 272000, + "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai" - }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 5, + "output": 25 } }, - "openai/gpt-6-astra": { - "id": "openai/gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "claude-sonnet-4-5-20250929", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -163890,62 +154480,26 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 200000, + "output": 64000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } + "input": 2.7, + "output": 13.5 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", + "grok-4-1-fast-reasoning": { + "id": "grok-4-1-fast-reasoning", + "name": "grok-4-1-fast-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -163957,37 +154511,32 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.18, + "output": 0.45 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", + "gemini-3-pro-preview": { + "id": "gemini-3-pro-preview", + "name": "gemini-3-pro-preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -163995,43 +154544,27 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.075, - "output": 0.3, - "cache_read": 0.0375 + "input": 1.8, + "output": 10.8 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "gpt-5-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -164043,50 +154576,30 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 272000, + "context": 400000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai" - }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.225, + "output": 1.8 } }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "gpt-5-nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -164094,50 +154607,30 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai" - }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 + "input": 0.045, + "output": 0.36 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "grok-4-fast-non-reasoning": { + "id": "grok-4-fast-non-reasoning", + "name": "grok-4-fast-non-reasoning", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", + "family": "grok", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -164145,73 +154638,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 2000000, + "output": 2000000 }, "cost": { - "input": 5, - "output": 30 + "input": 0.18, + "output": 0.45 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "gemini-3-flash-preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -164219,23 +154671,19 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.5, + "output": 3 } }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", + "gemini-2.5-flash-preview-05-20": { + "id": "gemini-2.5-flash-preview-05-20", + "name": "gemini-2.5-flash-preview-05-20", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -164245,23 +154693,21 @@ "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -164269,31 +154715,26 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 1048576, + "output": 200000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 0.135, + "output": 3.15 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "gemini-2.5-flash-lite-preview-06-17": { + "id": "gemini-2.5-flash-lite-preview-06-17", + "name": "gemini-2.5-flash-lite-preview-06-17", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -164302,15 +154743,15 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", + "video", "image", - "pdf" + "audio" ], "output": [ "text" @@ -164318,30 +154759,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 1048576, + "output": 65535 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.09, + "output": 0.36 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", + "gemini-2.5-pro-preview-06-05": { + "id": "gemini-2.5-pro-preview-06-05", + "name": "gemini-2.5-pro-preview-06-05", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", "high" @@ -164350,14 +154787,15 @@ ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" @@ -164365,50 +154803,30 @@ }, "open_weights": false, "limit": { - "context": 128000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" + "context": 1048576, + "output": 200000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 1.125, + "output": 9 } }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "gpt-5-codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -164416,78 +154834,63 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 400000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/openai" - }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.25, - "cache_write": 3.125 + "input": 1.125, + "output": 9 } - } - } - }, - "clarifai": { - "id": "clarifai", - "env": [ - "CLARIFAI_PAT" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.clarifai.com/v2/ext/openai/v1", - "name": "Clarifai", - "doc": "https://docs.clarifai.com/compute/inference/", - "models": { - "minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput": { - "id": "minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput", - "name": "MiniMax-M2.5 High Throughput", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + }, + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "gpt-5.2-codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-25", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 1.75, + "output": 14 } }, - "clarifai/main/models/mm-poly-8b": { - "id": "clarifai/main/models/mm-poly-8b", - "name": "MM Poly 8B", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "family": "mm-poly", + "gpt-5.1": { + "id": "gpt-5.1", + "name": "gpt-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-06", - "last_updated": "2026-02-25", + "release_date": "2026-02", + "last_updated": "2026-02", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -164495,29 +154898,29 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 4096 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.658, - "output": 1.11 + "input": 1.125, + "output": 9 } }, - "deepseek-ai/deepseek-ocr/models/DeepSeek-OCR": { - "id": "deepseek-ai/deepseek-ocr/models/DeepSeek-OCR", - "name": "DeepSeek OCR", - "description": "OCR model for extracting structured text from documents and screenshots", - "family": "deepseek", - "attachment": true, + "moonshotai/kimi-k2-instruct": { + "id": "moonshotai/kimi-k2-instruct", + "name": "Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-10-20", - "last_updated": "2026-02-25", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -164525,30 +154928,38 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 8192 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.7 + "input": 0.57, + "output": 2.3 } }, - "mistralai/completion/models/Ministral-3-3B-Reasoning-2512": { - "id": "mistralai/completion/models/Ministral-3-3B-Reasoning-2512", - "name": "Ministral 3 3B Reasoning 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 262143 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12", - "last_updated": "2026-02-25", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -164560,27 +154971,25 @@ "output": 262144 }, "cost": { - "input": 1.039, - "output": 0.54825 + "input": 0.6, + "output": 3 } }, - "mistralai/completion/models/Ministral-3-14B-Reasoning-2512": { - "id": "mistralai/completion/models/Ministral-3-14B-Reasoning-2512", - "name": "Ministral 3 14B Reasoning 2512", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "moonshotai/kimi-k2-0905": { + "id": "moonshotai/kimi-k2-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-01", - "last_updated": "2025-12-12", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -164592,31 +155001,26 @@ "output": 262144 }, "cost": { - "input": 2.5, - "output": 1.7 + "input": 0.6, + "output": 2.5 } }, - "moonshotai/chat-completion/models/Kimi-K2_6": { - "id": "moonshotai/chat-completion/models/Kimi-K2_6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "minimaxai/minimax-m1-80k": { + "id": "minimaxai/minimax-m1-80k", + "name": "MiniMax M1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -164624,30 +155028,31 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 40000 }, "cost": { - "input": 0.95, - "output": 4 + "input": 0.55, + "output": 2.2 } }, - "arcee_ai/AFM/models/trinity-mini": { - "id": "arcee_ai/AFM/models/trinity-mini", - "name": "Trinity Mini", - "description": "Reasoning-tuned 26B MoE model with 3B active parameters for agents, tools, and multi-step workloads", - "family": "trinity", - "attachment": false, + "baidu/ernie-4.5-vl-424b-a47b": { + "id": "baidu/ernie-4.5-vl-424b-a47b", + "name": "ERNIE 4.5 VL 424B A47B", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "ernie", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-01", - "last_updated": "2026-02-25", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -164655,26 +155060,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 123000, + "output": 16000 }, "cost": { - "input": 0.045, - "output": 0.15 + "input": 0.42, + "output": 1.25 } }, - "qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct": { - "id": "qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct", - "name": "Qwen3 Coder 30B A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "baidu/ernie-4.5-300b-a47b-paddle": { + "id": "baidu/ernie-4.5-300b-a47b-paddle", + "name": "ERNIE 4.5 300B A47B", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "ernie", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-31", - "last_updated": "2026-02-12", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -164685,26 +155090,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 123000, + "output": 12000 }, "cost": { - "input": 0.11458, - "output": 0.74812 + "input": 0.28, + "output": 1.1 } }, - "qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507": { - "id": "qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen3 30B A3B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "xiaomimimo/mimo-v2-flash": { + "id": "xiaomimimo/mimo-v2-flash", + "name": "XiaomiMiMo/MiMo-V2-Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-30", - "last_updated": "2026-02-25", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -164716,26 +155122,26 @@ "open_weights": true, "limit": { "context": 262144, - "output": 262144 + "output": 131072 }, "cost": { - "input": 0.3, - "output": 0.5 + "input": 0, + "output": 0 } }, - "qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507": { - "id": "qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507", - "name": "Qwen3 30B A3B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "zai-org/glm-4.7": { + "id": "zai-org/glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-07-31", - "last_updated": "2026-02-25", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -164746,38 +155152,32 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 204800, "output": 131072 }, "cost": { - "input": 0.36, - "output": 1.3 + "input": 0.6, + "output": 2.2 } }, - "openai/chat-completion/models/gpt-oss-120b-high-throughput": { - "id": "openai/chat-completion/models/gpt-oss-120b-high-throughput", - "name": "GPT OSS 120B High Throughput", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "zai-org/glm-4.5v": { + "id": "zai-org/glm-4.5v", + "name": "GLM 4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glmv", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2026-02-25", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -164785,35 +155185,31 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 65536, "output": 16384 }, "cost": { - "input": 0.09, - "output": 0.36 + "input": 0.6, + "output": 1.8 } }, - "openai/chat-completion/models/gpt-oss-20b": { - "id": "openai/chat-completion/models/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "zai-org/glm-4.5": { + "id": "zai-org/glm-4.5", + "name": "GLM-4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-12-12", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -164825,49 +155221,26 @@ "open_weights": true, "limit": { "context": 131072, - "output": 16384 + "output": 98304 }, "cost": { - "input": 0.045, - "output": 0.18 + "input": 0.6, + "output": 2.2 } - } - } - }, - "aiand": { - "id": "aiand", - "env": [ - "AIAND_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.aiand.com/v1", - "name": "ai&", - "doc": "https://docs.aiand.com/", - "models": { - "motif-technologies/motif-3": { - "id": "motif-technologies/motif-3", - "name": "Motif 3", - "description": "Motif 3 is a large-scale, decoder-only Mixture-of-Experts (MoE) language model with 314 billion total parameters and 13.2 billion parameters activated per token.", - "family": "motif", + }, + "zai-org/glm-4.7-flash": { + "id": "zai-org/glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-08-07", - "last_updated": "2026-08-07", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -164878,45 +155251,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.5, - "output": 2, - "cache_read": 0.2 + "input": 0.07, + "output": 0.4 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "qwen/qwen3-235b-a22b-instruct-2507": { + "id": "qwen/qwen3-235b-a22b-instruct-2507", + "name": "Qwen3 235B A22B Instruct 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -164924,41 +155281,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.15, + "output": 0.8 } }, - "deepseek-ai/deepseek-v4-pro": { - "id": "deepseek-ai/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "qwen/qwen3-next-80b-a3b-instruct": { + "id": "qwen/qwen3-next-80b-a3b-instruct", + "name": "Qwen3 Next 80B A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -164969,41 +155311,27 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 384000 + "context": 65536, + "output": 65536 }, "cost": { - "input": 1, - "output": 2.5, - "cache_read": 0.25 + "input": 0.15, + "output": 1.5 } }, - "deepseek-ai/deepseek-v4-flash": { - "id": "deepseek-ai/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -165014,45 +155342,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 384000 + "context": 65536, + "output": 65536 }, "cost": { "input": 0.15, - "output": 0.25, - "cache_read": 0.08 + "output": 1.5 } }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "qwen/qwen3-235b-a22b-thinking-2507": { + "id": "qwen/qwen3-235b-a22b-thinking-2507", + "name": "Qwen3 235B A22b Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -165060,44 +155373,30 @@ }, "open_weights": true, "limit": { - "context": 1048576, + "context": 131072, "output": 131072 }, "cost": { - "input": 3, - "output": 12.5, - "cache_read": 0.5 + "input": 0.3, + "output": 3 } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "qwen/qwen3-32b-fp8": { + "id": "qwen/qwen3-32b-fp8", + "name": "Qwen3 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -165105,40 +155404,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 40960, + "output": 20000 }, "cost": { - "input": 0.75, - "output": 3.5, - "cache_read": 0.2 + "input": 0.1, + "output": 0.45 } }, - "zai-org/glm-5.2": { - "id": "zai-org/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "qwen/qwen3-30b-a3b-fp8": { + "id": "qwen/qwen3-30b-a3b-fp8", + "name": "Qwen3 30B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -165149,40 +155435,26 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 40960, + "output": 20000 }, "cost": { - "input": 1, - "output": 4, - "cache_read": 0.3 + "input": 0.09, + "output": 0.45 } }, - "zai-org/glm-5.3": { - "id": "zai-org/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "qwen/qwen3-coder-next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-02", + "last_updated": "2026-02", "modalities": { "input": [ "text" @@ -165193,47 +155465,29 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1, - "output": 4, - "cache_read": 0.3 + "input": 0.2, + "output": 1.5 } }, - "qwen/qwen3.8-27b": { - "id": "qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "qwen/qwen3-coder-480b-a35b-instruct": { + "id": "qwen/qwen3-coder-480b-a35b-instruct", + "name": "Qwen3 Coder 480B A35B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -165242,44 +155496,29 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 65536 }, "cost": { - "input": 0.4, - "output": 3, - "cache_read": 0.2 + "input": 0.29, + "output": 1.2 } }, - "qwen/qwen3.6-27b": { - "id": "qwen/qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen/qwen3-235b-a22b-fp8": { + "id": "qwen/qwen3-235b-a22b-fp8", + "name": "Qwen3 235B A22B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -165287,40 +155526,27 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 40960, + "output": 20000 }, "cost": { - "input": 0.32, - "output": 3.2, - "cache_read": 0.2 + "input": 0.2, + "output": 0.8 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "deepseek/deepseek-r1-0528": { + "id": "deepseek/deepseek-r1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text" @@ -165331,223 +155557,148 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 163840, "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.08 + "input": 0.7, + "output": 2.5 } - } - } - }, - "frogbot": { - "id": "frogbot", - "env": [ - "FROGBOT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://app.frogbot.ai/api/v1", - "name": "FrogBot", - "doc": "https://docs.frogbot.ai", - "models": { - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "Grok 4.1 Fast (Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + }, + "deepseek/deepseek-v3-0324": { + "id": "deepseek/deepseek-v3-0324", + "name": "DeepSeek V3 0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 128000 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.28, + "output": 1.14 } }, - "gpt-5-4-mini": { - "id": "gpt-5-4-mini", - "name": "GPT-5.4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "deepseek/deepseek-v3.1": { + "id": "deepseek/deepseek-v3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024, + "max": 32767 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 163840, + "output": 32768 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.27, + "output": 1 } }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "Minimax M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "budget_tokens", - "min": 1024 + "min": 1024, + "max": 131071 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image", - "pdf" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "Grok 4.1 Fast (Non-Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-25", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.3, + "output": 1.2 } - }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + } + } + }, + "nova": { + "id": "nova", + "env": [ + "NOVA_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.nova.amazon.com/v1", + "name": "Nova", + "doc": "https://nova.amazon.com/dev/documentation", + "models": { + "nova-2-pro-v1": { + "id": "nova-2-pro-v1", + "name": "Nova 2 Pro", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "release_date": "2025-12-03", + "last_updated": "2026-01-03", "modalities": { "input": [ "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "minimax-m2-5": { - "id": "minimax-m2-5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-09", - "release_date": "2025-01-15", - "last_updated": "2025-02-22", - "modalities": { - "input": [ - "text" + "image", + "video", + "pdf" ], "output": [ "text" @@ -165555,23 +155706,26 @@ }, "open_weights": false, "limit": { - "context": 192000, - "output": 8192 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0, + "output": 0, + "reasoning": 0 } }, - "gpt-5-4-nano": { - "id": "gpt-5-4-nano", - "name": "GPT-5.4 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "nova-2-lite-v1": { + "id": "nova-2-lite-v1", + "name": "Nova 2 Lite", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "nova-lite", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -165582,15 +155736,15 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" @@ -165598,113 +155752,139 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0, + "output": 0, + "reasoning": 0 } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + } + } + }, + "alibaba-token-plan": { + "id": "alibaba-token-plan", + "env": [ + "ALIBABA_TOKEN_PLAN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://token-plan.ap-southeast-1.maas.aliyuncs.com/compatible-mode/v1", + "name": "Alibaba Token Plan", + "doc": "https://www.alibabacloud.com/help/en/model-studio/token-plan-overview", + "models": { + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 262144 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.07, - "output": 0.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "budget_tokens", - "min": 128, - "max": 32768 + "max": 262144 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-05", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -165712,69 +155892,81 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 1000000, "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.31 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 0, - "max": 24576 + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "temperature": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, "knowledge": "2025-01", - "release_date": "2025-07-17", - "last_updated": "2025-07-17", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "audio", - "video", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "minimax-m2-7": { - "id": "minimax-m2-7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": true, - "reasoning": false, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -165783,145 +155975,163 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 192000, - "output": 8192 + "context": 202752, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "kimi-k2-6": { - "id": "kimi-k2-6", - "name": "Kimi-K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 128000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "grok-code-fast-1": { - "id": "grok-code-fast-1", - "name": "Grok 4.1 Fast (Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "wan2.7-image-pro": { + "id": "wan2.7-image-pro", + "name": "Wan2.7 Image Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2023-10", - "release_date": "2025-08-28", - "last_updated": "2025-08-28", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 256000, - "output": 128000 + "context": 8192, + "output": 0 }, "cost": { - "input": 0.2, - "output": 1.5, - "cache_read": 0.02 + "input": 0, + "output": 0 } }, - "qwen-3-6-plus": { - "id": "qwen-3-6-plus", - "name": "Qwen 3.6 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-03", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 64000 + "output": 131072 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" }, { "type": "budget_tokens", - "min": 1024 + "max": 81920 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -165929,39 +156139,39 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "grok-4-3": { - "id": "grok-4-3", - "name": "Grok 4.3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-11", - "release_date": "2026-04-30", - "last_updated": "2026-04-30", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ "text", @@ -165972,256 +156182,207 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "gpt-5-5": { - "id": "gpt-5-5", - "name": "GPT-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "qwen-image-2.0": { + "id": "qwen-image-2.0", + "name": "Qwen Image 2.0", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 272000, - "output": 128000 + "context": 8192, + "output": 0 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 + "input": 0, + "output": 0 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 196608, + "input": 196601, + "output": 24576 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "gemini-3-1-pro-preview": { - "id": "gemini-3-1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-02-18", - "last_updated": "2026-02-18", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 16384 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek v4 Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", - "attachment": true, + "qwen-image-2.0-pro": { + "id": "qwen-image-2.0-pro", + "name": "Qwen Image 2.0 Pro", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 8192, + "output": 0 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.14 + "input": 0, + "output": 0 } }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, + "wan2.7-image": { + "id": "wan2.7-image", + "name": "Wan2.7 Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "text" + "image" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 8192, + "output": 0 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0, + "output": 0 } }, - "gpt-5-3-codex": { - "id": "gpt-5-3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt", - "attachment": true, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -166229,37 +156390,34 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 202752, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "1970-01-01", - "last_updated": "1970-01-01", + "knowledge": "2025-01", + "release_date": "2025-12-03", + "last_updated": "2025-12-05", "modalities": { "input": [ "text" @@ -166271,40 +156429,39 @@ "open_weights": true, "limit": { "context": 131072, - "output": 32768 + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0, + "output": 0 } }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "max": 81920 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -166312,75 +156469,84 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "zai-glm-5-1": { - "id": "zai-glm-5-1", - "name": "Z.AI GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": true, - "reasoning": false, + } + } + }, + "alibaba": { + "id": "alibaba", + "env": [ + "DASHSCOPE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1", + "name": "Alibaba", + "doc": "https://www.alibabacloud.com/help/en/model-studio/models", + "models": { + "qwen3-omni-flash": { + "id": "qwen3-omni-flash", + "name": "Qwen3-Omni Flash", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-01-20", - "last_updated": "2025-02-22", + "knowledge": "2024-04", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 198000, - "output": 8192 + "context": 65536, + "output": 16384 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.43, + "output": 1.66, + "input_audio": 3.81, + "output_audio": 15.11 } - } - } - }, - "stackit": { - "id": "stackit", - "env": [ - "STACKIT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.openai-compat.model-serving.eu01.onstackit.cloud/v1", - "name": "STACKIT", - "doc": "https://docs.stackit.cloud/products/data-and-ai/ai-model-serving/basics/available-shared-models", - "models": { - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma 3 27B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + }, + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2025-05-17", - "last_updated": "2025-05-17", + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -166388,26 +156554,34 @@ }, "open_weights": true, "limit": { - "context": 37000, - "output": 4096 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.53, - "output": 0.76 + "input": 1, + "output": 5 } }, - "intfloat/e5-mistral-7b-instruct": { - "id": "intfloat/e5-mistral-7b-instruct", - "name": "E5 Mistral 7B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "mistral", + "qwen-plus": { + "id": "qwen-plus", + "name": "Qwen Plus", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "release_date": "2023-12-11", - "last_updated": "2023-12-11", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-09-11", "modalities": { "input": [ "text" @@ -166416,32 +156590,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 4096, - "output": 4096 + "context": 1000000, + "output": 32768 }, "cost": { - "input": 0.02, - "output": 0.02 + "input": 0.4, + "output": 1.2, + "reasoning": 4 } }, - "Qwen/Qwen3-VL-235B-A22B-Instruct-FP8": { - "id": "Qwen/Qwen3-VL-235B-A22B-Instruct-FP8", - "name": "Qwen3-VL 235B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3-Coder 30B-A3B Instruct", + "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2024-11-01", - "last_updated": "2024-11-01", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -166449,57 +156623,70 @@ }, "open_weights": true, "limit": { - "context": 218000, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.76, - "output": 2.05 + "input": 0.45, + "output": 2.25 } }, - "Qwen/Qwen3-VL-Embedding-8B": { - "id": "Qwen/Qwen3-VL-Embedding-8B", - "name": "Qwen3-VL Embedding 8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "qwen3-omni-flash-realtime": { + "id": "qwen3-omni-flash-realtime", + "name": "Qwen3-Omni Flash Realtime", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", "family": "qwen", - "attachment": true, + "attachment": false, "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32000, - "output": 4096 + "context": 65536, + "output": 16384 }, "cost": { - "input": 0.09, - "output": 0.09 + "input": 0.52, + "output": 1.99, + "input_audio": 4.57, + "output_audio": 18.13 } }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3 32B", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -166510,57 +156697,62 @@ }, "open_weights": true, "limit": { - "context": 262144, + "context": 131072, "output": 16384 }, "cost": { - "input": 0.53, - "output": 0.76 + "input": 0.7, + "output": 2.8, + "reasoning": 8.4 } }, - "cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic": { - "id": "cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic", - "name": "Llama 3.3 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "qwen-omni-turbo-realtime": { + "id": "qwen-omni-turbo-realtime", + "name": "Qwen-Omni Turbo Realtime", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "knowledge": "2024-04", + "release_date": "2025-05-08", + "last_updated": "2025-05-08", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 32768, + "output": 2048 }, "cost": { - "input": 0.53, - "output": 0.76 + "input": 0.27, + "output": 1.07, + "input_audio": 4.44, + "output_audio": 8.89 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "qwen-plus-character-ja": { + "id": "qwen-plus-character-ja", + "name": "Qwen Plus Character (Japanese)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2024-04", + "release_date": "2024-01", + "last_updated": "2024-01", "modalities": { "input": [ "text" @@ -166569,38 +156761,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 8192, + "output": 512 }, "cost": { - "input": 0.18, - "output": 0.29 + "input": 0.5, + "output": 1.4 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ "text" @@ -166611,61 +156793,39 @@ }, "open_weights": true, "limit": { - "context": 131000, - "output": 8192 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.53, - "output": 0.76 + "input": 0.5, + "output": 2 } - } - } - }, - "anyapi": { - "id": "anyapi", - "env": [ - "ANYAPI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.anyapi.ai/v1", - "name": "AnyAPI", - "doc": "https://docs.anyapi.ai", - "models": { - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, + }, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-04", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -166673,15 +156833,39 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "output": 65536 + }, + "cost": { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } } }, - "anthropic/claude-sonnet-4-5": { - "id": "anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "qwen3.6-35b-a3b": { + "id": "qwen3.6-35b-a3b", + "name": "Qwen3.6 35B-A3B", + "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -166689,75 +156873,57 @@ "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 + "type": "budget_tokens" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.248, + "output": 1.485 } }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 + "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -166766,64 +156932,30 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 65536 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "cost": { + "input": 2.5, + "output": 7.5, + "cache_read": 0.5, + "cache_write": 3.125 } }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -166831,83 +156963,69 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 1.2, + "output": 6 } }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "qwen2-5-omni-7b": { + "id": "qwen2-5-omni-7b", + "name": "Qwen2.5-Omni 7B", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-12", + "last_updated": "2024-12", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 32768, + "output": 2048 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "cost": { + "input": 0.1, + "output": 0.4, + "input_audio": 6.76 } }, - "cohere/command-r-plus-08-2024": { - "id": "cohere/command-r-plus-08-2024", - "name": "Command R+", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", + "qwen3-8b": { + "id": "qwen3-8b", + "name": "Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -166918,22 +157036,27 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4000 + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.18, + "output": 0.7, + "reasoning": 2.1 } }, - "deepseek/deepseek-chat": { - "id": "deepseek/deepseek-chat", - "name": "DeepSeek Chat", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": true, + "qwen2-5-14b-instruct": { + "id": "qwen2-5-14b-instruct", + "name": "Qwen2.5 14B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -166944,26 +157067,31 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.35, + "output": 1.4 } }, - "deepseek/deepseek-r1": { - "id": "deepseek/deepseek-r1", - "name": "DeepSeek Reasoner", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": true, + "qwen3-next-80b-a3b-thinking": { + "id": "qwen3-next-80b-a3b-thinking", + "name": "Qwen3-Next 80B-A3B (Thinking)", + "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ "text" @@ -166974,84 +157102,99 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.5, + "output": 6 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "qvq-max": { + "id": "qvq-max", + "name": "QVQ Max", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qvq", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-04", + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 1.2, + "output": 4.8 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-04-27", + "last_updated": "2026-04-27", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 65536 + }, + "cost": { + "input": 0.1875, + "output": 1.125, + "cache_write": 0.234375 } }, - "perplexity/sonar-pro": { - "id": "perplexity/sonar-pro", - "name": "Sonar Pro", - "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", - "family": "sonar-pro", - "attachment": true, + "qwen2-5-vl-72b-instruct": { + "id": "qwen2-5-vl-72b-instruct", + "name": "Qwen2.5-VL 72B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text", @@ -167061,25 +157204,36 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, + "context": 131072, "output": 8192 + }, + "cost": { + "input": 2.8, + "output": 8.4 } }, - "perplexity/sonar-reasoning-pro": { - "id": "perplexity/sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", - "family": "sonar-reasoning", - "attachment": true, + "qwen3-vl-plus": { + "id": "qwen3-vl-plus", + "name": "Qwen3-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", + "knowledge": "2025-04", + "release_date": "2025-09-23", + "last_updated": "2025-09-23", "modalities": { "input": [ "text", @@ -167091,31 +157245,31 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262144, + "output": 32768 + }, + "cost": { + "input": 0.2, + "output": 1.6, + "reasoning": 4.8 } }, - "google/gemini-3-pro-preview": { - "id": "google/gemini-3-pro-preview", - "name": "Gemini 3 Pro Preview", - "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "qwen-vl-ocr": { + "id": "qwen-vl-ocr", + "name": "Qwen-VL OCR", + "description": "OCR model for extracting structured text from documents and screenshots", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", + "knowledge": "2024-04", + "release_date": "2024-10-28", + "last_updated": "2025-04-13", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -167123,31 +157277,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 34096, + "output": 4096 + }, + "cost": { + "input": 0.72, + "output": 0.72 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "qwen-mt-turbo": { + "id": "qwen-mt-turbo", + "name": "Qwen-MT Turbo", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-04", + "release_date": "2025-01", + "last_updated": "2025-01", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -167155,31 +157307,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 16384, + "output": 8192 + }, + "cost": { + "input": 0.16, + "output": 0.49 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "qwen-mt-plus": { + "id": "qwen-mt-plus", + "name": "Qwen-MT Plus", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-04", + "release_date": "2025-01", + "last_updated": "2025-01", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -167187,31 +157337,39 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 16384, + "output": 8192 + }, + "cost": { + "input": 2.46, + "output": 7.37 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, + "qwen3.5-plus": { + "id": "qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", "image", - "video", - "audio", - "pdf" + "video" ], "output": [ "text" @@ -167219,83 +157377,63 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 1000000, "output": 65536 + }, + "cost": { + "input": 0.4, + "output": 2.4, + "reasoning": 2.4 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "qwen-omni-turbo": { + "id": "qwen-omni-turbo", + "name": "Qwen-Omni Turbo", + "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "knowledge": "2024-04", + "release_date": "2025-01-19", + "last_updated": "2025-03-26", "modalities": { "input": [ "text", "image", "audio", - "video", - "pdf" + "video" ], "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - } - }, - "xai/grok-4.3": { - "id": "xai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ "text", - "image", - "pdf" - ], - "output": [ - "text" + "audio" ] }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 32768, + "output": 2048 + }, + "cost": { + "input": 0.07, + "output": 0.27, + "input_audio": 4.44, + "output_audio": 8.89 } }, - "mistralai/devstral-2512": { - "id": "mistralai/devstral-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", + "qwen2-5-72b-instruct": { + "id": "qwen2-5-72b-instruct", + "name": "Qwen2.5 72B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -167306,66 +157444,37 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated" - }, - "mistralai/mistral-large-2512": { - "id": "mistralai/mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] + "context": 131072, + "output": 8192 }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 + "cost": { + "input": 1.4, + "output": 5.6 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, + "qwen-flash": { + "id": "qwen-flash", + "name": "Qwen Flash", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2024-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -167373,50 +157482,31 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "output": 32768 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "cost": { + "input": 0.05, + "output": 0.4 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, + "qwen3-vl-235b-a22b": { + "id": "qwen3-vl-235b-a22b", + "name": "Qwen3-VL 235B-A22B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text", @@ -167426,36 +157516,34 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.7, + "output": 2.8, + "reasoning": 8.4 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, + "qwen3-vl-30b-a3b": { + "id": "qwen3-vl-30b-a3b", + "name": "Qwen3-VL 30B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text", @@ -167465,38 +157553,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 32768 + }, + "cost": { + "input": 0.2, + "output": 0.8, + "reasoning": 2.4 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", + "qwen-vl-max": { + "id": "qwen-vl-max", + "name": "Qwen-VL Max", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2024-04-08", + "last_updated": "2025-08-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -167504,76 +157587,70 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 131072, + "output": 8192 + }, + "cost": { + "input": 0.8, + "output": 3.2 } }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", + "qwen3.5-27b": { + "id": "qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.4 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "qwen-max": { + "id": "qwen-max", + "name": "Qwen Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "knowledge": "2024-04", + "release_date": "2024-04-03", + "last_updated": "2025-01-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -167581,173 +157658,183 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 32768, + "output": 8192 + }, + "cost": { + "input": 1.6, + "output": 6.4 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, + "qwen3-235b-a22b": { + "id": "qwen3-235b-a22b", + "name": "Qwen3 235B-A22B", + "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 131072, + "output": 16384 + }, + "cost": { + "input": 0.7, + "output": 2.8, + "reasoning": 8.4 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, + "qwen3-livetranslate-flash-realtime": { + "id": "qwen3-livetranslate-flash-realtime", + "name": "Qwen3-LiveTranslate Flash Realtime", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ - "text" + "text", + "audio" ] }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 53248, + "output": 4096 + }, + "cost": { + "input": 10, + "output": 10, + "input_audio": 10, + "output_audio": 38 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "qwen3.6-27b": { + "id": "qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.6, + "output": 3.6 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", + "qwen3.5-35b-a3b": { + "id": "qwen3.5-35b-a3b", + "name": "Qwen3.5 35B-A3B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.25, + "output": 2 } - } - } - }, - "crusoe": { - "id": "crusoe", - "env": [ - "CRUSOE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.inference.crusoecloud.com/v1", - "name": "Crusoe", - "doc": "https://docs.crusoecloud.com/managed-inference/overview", - "models": { - "zai/GLM-5.1": { - "id": "zai/GLM-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + }, + "qwen3-coder-480b-a35b-instruct": { + "id": "qwen3-coder-480b-a35b-instruct", + "name": "Qwen3-Coder 480B-A35B Instruct", + "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -167758,33 +157845,27 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 1.2, - "output": 4.4, - "cache_read": 0.25 + "input": 1.5, + "output": 7.5 } }, - "zai/GLM-5.2": { - "id": "zai/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "qwq-plus": { + "id": "qwq-plus", + "name": "QwQ Plus", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-04", + "release_date": "2025-03-05", + "last_updated": "2025-03-05", "modalities": { "input": [ "text" @@ -167793,29 +157874,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "output": 8192 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.8, + "output": 2.4 } }, - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "qwen2-5-32b-instruct": { + "id": "qwen2-5-32b-instruct", + "name": "Qwen2.5 32B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ "text" @@ -167826,36 +157906,40 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.13 + "input": 0.7, + "output": 2.8 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" @@ -167864,25 +157948,25 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.4, - "cache_read": 0.14 + "input": 0.6, + "output": 3.6 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B-A22B Instruct 2507", - "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", + "qwen3-coder-flash": { + "id": "qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -167891,28 +157975,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0.22, - "output": 0.8, - "cache_read": 0.11 + "input": 0.3, + "output": 1.5 } }, - "deepseek-ai/DeepSeek-V3-0324": { - "id": "deepseek-ai/DeepSeek-V3-0324", - "name": "DeepSeek V3 0324", - "description": "March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding", - "family": "deepseek", + "qwen3-14b": { + "id": "qwen3-14b", + "name": "Qwen3 14B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens" + } + ], "tool_call": true, "temperature": true, - "release_date": "2025-03-24", - "last_updated": "2025-03-24", + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ "text" @@ -167923,71 +158015,65 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.25 + "input": 0.35, + "output": 1.4, + "reasoning": 4.2 } }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "qwen3-asr-flash": { + "id": "qwen3-asr-flash", + "name": "Qwen3-ASR Flash", + "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": false, + "knowledge": "2024-04", + "release_date": "2025-09-08", + "last_updated": "2025-09-08", "modalities": { "input": [ - "text", - "image", - "video" + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 53248, + "output": 4096 }, "cost": { - "input": 0.7, - "output": 3.5, - "cache_read": 0.35 + "input": 0.035, + "output": 0.035 } }, - "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B": { - "id": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", + "qwen-turbo": { + "id": "qwen-turbo", + "name": "Qwen Turbo", + "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "knowledge": "2024-04", + "release_date": "2024-11-01", + "last_updated": "2025-04-28", "modalities": { "input": [ "text" @@ -167996,40 +158082,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 16384 }, "cost": { "input": 0.05, "output": 0.2, - "cache_read": 0.03 + "reasoning": 0.5 } }, - "nvidia/Nemotron-3-Nano-Omni-Reasoning-30B-A3B": { - "id": "nvidia/Nemotron-3-Nano-Omni-Reasoning-30B-A3B", - "name": "Nemotron 3 Nano Omni 30B A3B Reasoning", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen2-5-7b-instruct": { + "id": "qwen2-5-7b-instruct", + "name": "Qwen2.5 7B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": true, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -168037,35 +158115,30 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.83, - "cache_read": 0.3, - "input_audio": 0.5 + "input": 0.175, + "output": 0.7 } }, - "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B": { - "id": "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B", - "name": "Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "qwen2-5-vl-7b-instruct": { + "id": "qwen2-5-vl-7b-instruct", + "name": "Qwen2.5-VL 7B Instruct", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "knowledge": "2024-04", + "release_date": "2024-09", + "last_updated": "2024-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -168073,38 +158146,34 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 2.4, - "cache_read": 0.15 + "input": 0.35, + "output": 1.05 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "qwen3.6-max-preview": { + "id": "qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "family": "qwen", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": true, - "structured_output": true, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "knowledge": "2025-04", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ "text" @@ -168113,150 +158182,110 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.05 + "input": 1.3, + "output": 7.8, + "cache_read": 0.13, + "cache_write": 1.625 } - } - } - }, - "volcengine-coding-plan": { - "id": "volcengine-coding-plan", - "env": [ - "ARK_CODING_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://ark.cn-beijing.volces.com/api/coding/v3", - "name": "Volcengine Ark Coding Plan", - "doc": "https://www.volcengine.com/docs/82379/1928261", - "models": { - "doubao-seed-2.0-lite": { - "id": "doubao-seed-2.0-lite", - "name": "Seed 2.0 Lite", - "description": "Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation", - "family": "seed", + }, + "qwen3.5-122b-a10b": { + "id": "qwen3.5-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 32000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.4, + "output": 3.2 } }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "qwen-vl-plus": { + "id": "qwen-vl-plus", + "name": "Qwen-VL Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2024-04", + "release_date": "2024-01-25", + "last_updated": "2025-08-15", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.21, + "output": 0.63 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "budget_tokens" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -168267,47 +158296,72 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "doubao-seed-2.1-turbo": { - "id": "doubao-seed-2.1-turbo", - "name": "Seed 2.1 Turbo", - "description": "Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "cache_write": 0.625, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.2, + "cache_write": 2.5 + } + } + } + } + }, + "databricks": { + "id": "databricks", + "env": [ + "DATABRICKS_HOST", + "DATABRICKS_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://${DATABRICKS_HOST}/ai-gateway/mlflow/v1", + "name": "Databricks", + "doc": "https://docs.databricks.com/aws/en/machine-learning/foundation-models/", + "models": { + "databricks-claude-opus-4-7": { + "id": "databricks-claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -168315,85 +158369,143 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "databricks-gpt-5-4": { + "id": "databricks-gpt-5-4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 512000 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "doubao-seed-evolving": { - "id": "doubao-seed-evolving", - "name": "Seed Evolving", - "description": "Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities", - "family": "seed", + "databricks-gemini-3-flash": { + "id": "databricks-gemini-3-flash", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", + "medium", "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", - "video" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -168401,205 +158513,218 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "databricks-claude-opus-4-5": { + "id": "databricks-claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "databricks-gpt-5-nano": { + "id": "databricks-gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "databricks-gpt-5-mini": { + "id": "databricks-gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "databricks-gpt-5": { + "id": "databricks-gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } - } - } - }, - "jiekou": { - "id": "jiekou", - "env": [ - "JIEKOU_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.jiekou.ai/openai", - "name": "Jiekou.AI", - "doc": "https://docs.jiekou.ai/docs/support/quickstart?utm_source=github_models.dev", - "models": { - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "grok-4-1-fast-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + }, + "databricks-gemini-2-5-pro": { + "id": "databricks-gemini-2-5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -168607,30 +158732,61 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.18, - "output": 0.45 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "grok-4-1-fast-non-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "databricks-gemini-3-1-pro": { + "id": "databricks-gemini-3-1-pro", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -168638,31 +158794,58 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.18, - "output": 0.45 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "gpt-5.2-codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "databricks-gemini-2-5-flash": { + "id": "databricks-gemini-2-5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -168670,31 +158853,39 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "gpt-5.1-codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "databricks-claude-sonnet-4": { + "id": "databricks-claude-sonnet-4", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -168702,32 +158893,34 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.125, - "output": 9 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "gemini-3-pro-preview": { - "id": "gemini-3-pro-preview", - "name": "gemini-3-pro-preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "databricks-claude-haiku-4-5": { + "id": "databricks-claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video", - "audio" + "pdf" ], "output": [ "text" @@ -168735,26 +158928,39 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "output": 64000 }, "cost": { - "input": 1.8, - "output": 10.8 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "gpt-5-codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "databricks-gpt-5-4-nano": { + "id": "databricks-gpt-5-4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -168767,30 +158973,38 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 1.125, - "output": 9 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "gpt-5-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "databricks-claude-opus-4-6": { + "id": "databricks-claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -168798,27 +159012,60 @@ }, "open_weights": false, "limit": { - "context": 400000, + "context": 1000000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } + }, "cost": { - "input": 0.225, - "output": 1.8 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "gpt-5.2-pro": { - "id": "gpt-5.2-pro", - "name": "gpt-5.2-pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "databricks-gpt-5-1": { + "id": "databricks-gpt-5-1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -168831,29 +159078,38 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 18.9, - "output": 151.2 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "databricks-claude-sonnet-4-5": { + "id": "databricks-claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -168862,29 +159118,38 @@ "open_weights": false, "limit": { "context": 200000, - "output": 100000 + "output": 64000 }, "cost": { - "input": 1.1, - "output": 4.4 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "databricks-claude-opus-4-1": { + "id": "databricks-claude-opus-4-1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -168892,30 +159157,44 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1.1, - "output": 4.4 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "grok-4-fast-reasoning": { - "id": "grok-4-fast-reasoning", - "name": "grok-4-fast-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", + "databricks-gpt-5-5": { + "id": "databricks-gpt-5-5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -168923,30 +159202,78 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 0.18, - "output": 0.45 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "gpt-5-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "databricks-gemini-3-1-flash-lite": { + "id": "databricks-gemini-3-1-flash-lite", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -168954,26 +159281,27 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.045, - "output": 0.36 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "gemini-2.5-flash-lite-preview-06-17": { - "id": "gemini-2.5-flash-lite-preview-06-17", - "name": "gemini-2.5-flash-lite-preview-06-17", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "databricks-gemini-3-pro": { + "id": "databricks-gemini-3-pro", + "name": "Gemini 3 Pro Preview", + "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -168983,14 +159311,16 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", - "video", "image", - "audio" + "video", + "audio", + "pdf" ], "output": [ "text" @@ -168999,29 +159329,53 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 65535 + "output": 65536 }, "cost": { - "input": 0.09, - "output": 0.36 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "claude-opus-4-20250514": { - "id": "claude-opus-4-20250514", - "name": "claude-opus-4-20250514", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "databricks-claude-sonnet-4-6": { + "id": "databricks-claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -169029,89 +159383,120 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 13.5, - "output": 67.5 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "grok-4-fast-non-reasoning": { - "id": "grok-4-fast-non-reasoning", - "name": "grok-4-fast-non-reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": false, + "databricks-gpt-oss-20b": { + "id": "databricks-gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.18, - "output": 0.45 + "input": 0.05, + "output": 0.2 } }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "claude-opus-4-1-20250805", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": false, + "databricks-gpt-oss-120b": { + "id": "databricks-gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 32000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 13.5, - "output": 67.5 + "input": 0.072, + "output": 0.28 } }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "gpt-5-pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "databricks-gpt-5-2": { + "id": "databricks-gpt-5-2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -169124,25 +159509,38 @@ "open_weights": false, "limit": { "context": 400000, - "output": 272000 + "input": 272000, + "output": 128000 }, "cost": { - "input": 13.5, - "output": 108 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "gpt-5-chat-latest": { - "id": "gpt-5-chat-latest", - "name": "gpt-5-chat-latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", + "databricks-gpt-5-4-mini": { + "id": "databricks-gpt-5-4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -169155,61 +159553,140 @@ "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } + }, "cost": { - "input": 1.125, - "output": 9 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } - }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "claude-opus-4-5-20251101", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": false, + } + } + }, + "crof": { + "id": "crof", + "env": [ + "CROF_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://crof.ai/v1", + "name": "CrofAI", + "doc": "https://crof.ai/docs", + "models": { + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 65536 + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 4.5, - "output": 22.5 + "input": 0.12, + "output": 0.21, + "cache_read": 0.003 } }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "gpt-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "minimax-m2.5": { + "id": "minimax-m2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02", - "last_updated": "2026-02", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 0.11, + "output": 0.95, + "cache_read": 0.02, + "cache_write": 0.375 + } + }, + "greg-2-ultra": { + "id": "greg-2-ultra", + "name": "Greg 2 Ultra", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-06-14", + "last_updated": "2026-06-14", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -169217,20 +159694,60 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 229376, + "output": 229376 }, "cost": { - "input": 1.125, - "output": 9 + "input": 3, + "output": 10, + "cache_read": 0.5 } }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "gemini-2.5-pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 202752, + "output": 202752 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 0.25, + "output": 1.1, + "cache_read": 0.05, + "cache_write": 0 + } + }, + "deepseek-v4-pro-lightning": { + "id": "deepseek-v4-pro-lightning", + "name": "DeepSeek V4 Pro Lightning", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -169244,16 +159761,49 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" + }, + "cost": { + "input": 0.8, + "output": 1.6, + "cache_read": 0.02 + } + }, + "greg-rp": { + "id": "greg-rp", + "name": "Greg (Roleplay)", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "temperature": true, + "release_date": "2026-01-27", + "last_updated": "2026-01-27", + "modalities": { + "input": [ + "text" ], "output": [ "text" @@ -169261,19 +159811,20 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 229376, + "output": 229376 }, "cost": { - "input": 1.125, - "output": 9 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "gemini-2.5-flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "gemma-4-31b-it": { + "id": "gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -169290,34 +159841,36 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65535 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.27, - "output": 2.25 + "input": 0.1, + "output": 0.3, + "cache_read": 0.02 } }, - "gemini-2.5-flash-preview-05-20": { - "id": "gemini-2.5-flash-preview-05-20", - "name": "gemini-2.5-flash-preview-05-20", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -169332,141 +159885,197 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", "image", - "video", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 200000 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.135, - "output": 3.15 + "input": 0.55, + "output": 2.25, + "cache_read": 0.05 } }, - "gpt-5.1-codex-max": { - "id": "gpt-5.1-codex-max", - "name": "gpt-5.1-codex-max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 202752, + "output": 202752 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 1.125, - "output": 9 + "input": 0.45, + "output": 2.15, + "cache_read": 0.08, + "cache_write": 0 } }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "gpt-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 1.575, - "output": 12.6 + "input": 0.35, + "output": 0.8, + "cache_read": 0.003 } }, - "grok-code-fast-1": { - "id": "grok-code-fast-1", - "name": "grok-code-fast-1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": false, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.18, - "output": 1.35 + "input": 0.5, + "output": 2.2, + "cache_read": 0.08 } }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "claude-sonnet-4-5-20250929", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "greg-2-super": { + "id": "greg-2-super", + "name": "Greg 2 Super", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-06-14", + "last_updated": "2026-06-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -169474,131 +160083,169 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 229376, + "output": 229376 }, "cost": { - "input": 2.7, - "output": 13.5 + "input": 1.5, + "output": 5, + "cache_read": 0.25 } }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "claude-opus-4-6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "kimi-k2.5-lightning": { + "id": "kimi-k2.5-lightning", + "name": "Kimi K2.5 (Lightning)", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02", - "last_updated": "2026-02", + "temperature": false, + "release_date": "2026-02-06", + "last_updated": "2026-02-06", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "output": 32768 }, "cost": { - "input": 5, - "output": 25 + "input": 1, + "output": 3, + "cache_read": 0.2 } }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "gemini-3-flash-preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": false, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, + "temperature": false, + "knowledge": "2025-01", "release_date": "2026-01", "last_updated": "2026-01", "modalities": { "input": [ "text", "image", - "video", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.35, + "output": 1.7, + "cache_read": 0.07 } }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "claude-haiku-4-5-20251001", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 20000, - "output": 64000 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.9, - "output": 4.5 + "input": 0.5, + "output": 1.99, + "cache_read": 0.05 } }, - "gemini-2.5-flash-lite-preview-09-2025": { - "id": "gemini-2.5-flash-lite-preview-09-2025", - "name": "gemini-2.5-flash-lite-preview-09-2025", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "qwen3.6-27b": { + "id": "qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -169615,8 +160262,8 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", @@ -169628,21 +160275,25 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.09, - "output": 0.36 + "input": 0.2, + "output": 1.5, + "cache_read": 0.04 } }, - "gemini-2.5-pro-preview-06-05": { - "id": "gemini-2.5-pro-preview-06-05", - "name": "gemini-2.5-pro-preview-06-05", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -169657,10 +160308,13 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-03-13", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -169672,21 +160326,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 200000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.125, - "output": 9 + "input": 0.04, + "output": 0.15, + "cache_read": 0.008 } }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "gemini-2.5-flash-lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "qwen3.5-397b-a17b": { + "id": "qwen3.5-397b-a17b", + "name": "Qwen3.5 397B-A17B", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -169703,8 +160358,8 @@ "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", @@ -169716,125 +160371,172 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65535 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.09, - "output": 0.36 + "input": 0.35, + "output": 1.75, + "cache_read": 0.07 } }, - "claude-sonnet-4-20250514": { - "id": "claude-sonnet-4-20250514", - "name": "claude-sonnet-4-20250514", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": false, + "glm-4.7-flash": { + "id": "glm-4.7-flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 202752, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 2.7, - "output": 13.5 + "input": 0.04, + "output": 0.3, + "cache_read": 0.008, + "cache_write": 0 } }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "gpt-5.1-codex-mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 1048576, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 0.225, - "output": 1.8 + "input": 0.4, + "output": 0.8, + "cache_read": 0.003, + "tiers": [ + { + "input": 2, + "output": 6, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2, + "output": 6, + "cache_read": 0.4 + } } }, - "o3": { - "id": "o3", - "name": "o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "attachment": true, - "reasoning": false, + "glm-5": { + "id": "glm-5", + "name": "GLM-5", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 202752, + "output": 202752 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible" }, "cost": { - "input": 10, - "output": 40 + "input": 0.48, + "output": 1.9, + "cache_read": 0.1, + "cache_write": 0 } }, - "grok-4-0709": { - "id": "grok-4-0709", - "name": "grok-4-0709", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -169842,26 +160544,25 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 8192 + "context": 163840, + "output": 163840 }, "cost": { - "input": 2.7, - "output": 13.5 + "input": 0.18, + "output": 0.35, + "cache_read": 0.04 } }, - "deepseek/deepseek-v3-0324": { - "id": "deepseek/deepseek-v3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "greg-1-mini": { + "id": "greg-1-mini", + "name": "Greg 1 Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-01-27", + "last_updated": "2026-01-27", "modalities": { "input": [ "text" @@ -169870,66 +160571,66 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 229376, + "output": 229376 }, "cost": { - "input": 0.28, - "output": 1.14 + "input": 0.07, + "output": 0.15, + "cache_read": 0.01 } - }, - "deepseek/deepseek-v3.1": { - "id": "deepseek/deepseek-v3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 32767 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + } + } + }, + "fastrouter": { + "id": "fastrouter", + "env": [ + "FASTROUTER_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://go.fastrouter.ai/api/v1", + "name": "FastRouter", + "doc": "https://fastrouter.ai/models", + "models": { + "wanx/wan-v2-6": { + "id": "wanx/wan-v2-6", + "name": "Wan 2.6", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, "open_weights": true, "limit": { - "context": 163840, - "output": 32768 - }, - "cost": { - "input": 0.27, - "output": 1 + "context": 400000, + "output": 0 } }, - "deepseek/deepseek-r1-0528": { - "id": "deepseek/deepseek-r1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "moonshotai/kimi-k2": { + "id": "moonshotai/kimi-k2", + "name": "Kimi K2", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2024-10", + "release_date": "2025-07-11", + "last_updated": "2025-07-11", "modalities": { "input": [ "text" @@ -169940,62 +160641,37 @@ }, "open_weights": true, "limit": { - "context": 163840, + "context": 131072, "output": 32768 }, "cost": { - "input": 0.7, - "output": 2.5 - } - }, - "xiaomimimo/mimo-v2-flash": { - "id": "xiaomimimo/mimo-v2-flash", - "name": "XiaomiMiMo/MiMo-V2-Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 + "input": 0.55, + "output": 2.2 } }, - "baidu/ernie-4.5-vl-424b-a47b": { - "id": "baidu/ernie-4.5-vl-424b-a47b", - "name": "ERNIE 4.5 VL 424B A47B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ernie", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "structured_output": false, + "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -170003,194 +160679,210 @@ }, "open_weights": true, "limit": { - "context": 123000, - "output": 16000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.42, - "output": 1.25 + "input": 0.75, + "output": 3.5 } }, - "baidu/ernie-4.5-300b-a47b-paddle": { - "id": "baidu/ernie-4.5-300b-a47b-paddle", - "name": "ERNIE 4.5 300B A47B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "ernie", + "google/imagen-4.0-fast": { + "id": "google/imagen-4.0-fast", + "name": "Imagen 4 Fast", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "tool_call": false, + "temperature": false, + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 123000, - "output": 12000 - }, - "cost": { - "input": 0.28, - "output": 1.1 + "context": 480, + "output": 0 } }, - "minimaxai/minimax-m1-80k": { - "id": "minimaxai/minimax-m1-80k", - "name": "MiniMax M1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "google/gemini-2.5-pro": { + "id": "google/gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, - "structured_output": false, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 40000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.55, - "output": 2.2 + "input": 1.25, + "output": 10, + "cache_read": 0.31 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "google/gemini-2.5-flash": { + "id": "google/gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "budget_tokens", - "min": 1024, - "max": 262143 + "min": 0, + "max": 24576 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 3 + "input": 0.3, + "output": 2.5, + "cache_read": 0.0375 } }, - "moonshotai/kimi-k2-instruct": { - "id": "moonshotai/kimi-k2-instruct", - "name": "Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.57, - "output": 2.3 + "input": 1.5, + "output": 9 } }, - "moonshotai/kimi-k2-0905": { - "id": "moonshotai/kimi-k2-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, + "google/veo3.1-lite": { + "id": "google/veo3.1-lite", + "name": "Veo 3.1 Lite", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "tool_call": false, + "temperature": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 2.5 + "context": 400000, + "output": 0 } }, - "zai-org/glm-4.5": { - "id": "zai-org/glm-4.5", - "name": "GLM-4.5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -170198,250 +160890,267 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 98304 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 0.13, + "output": 0.38 } }, - "zai-org/glm-4.5v": { - "id": "zai-org/glm-4.5v", - "name": "GLM 4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glmv", + "google/veo3.1": { + "id": "google/veo3.1", + "name": "Veo 3.1", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 16384 - }, - "cost": { - "input": 0.6, - "output": 1.8 + "context": 400000, + "output": 0 } }, - "zai-org/glm-4.7-flash": { - "id": "zai-org/glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "google/imagen-4.0-ultra": { + "id": "google/imagen-4.0-ultra", + "name": "Imagen 4 Ultra", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "imagen", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.07, - "output": 0.4 + "context": 480, + "output": 0 } }, - "zai-org/glm-4.7": { - "id": "zai-org/glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "google/gemini-3-pro-image-preview": { + "id": "google/gemini-3-pro-image-preview", + "name": "Nano Banana Pro", + "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2025-11-20", + "last_updated": "2025-11-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 65536, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.2 + "input": 2, + "output": 12 } }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "Minimax M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "google/gemini-3.1-flash-image-preview": { + "id": "google/gemini-3.1-flash-image-preview", + "name": "Nano Banana 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 131071 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2026-02-26", + "last_updated": "2026-02-26", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 65536, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.5, + "output": 3 } }, - "qwen/qwen3-235b-a22b-instruct-2507": { - "id": "qwen/qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.8 + "input": 2, + "output": 12 } }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "google/veo3.1-fast": { + "id": "google/veo3.1-fast", + "name": "Veo 3.1 Fast", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "veo", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 65536 - }, - "cost": { - "input": 0.15, - "output": 1.5 + "context": 400000, + "output": 0 } }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "qwen/qwen3-coder-next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "Grok 4.3", + "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", + "family": "grok", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-02", - "last_updated": "2026-02", + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 30000 }, "cost": { - "input": 0.2, - "output": 1.5 + "input": 1.25, + "output": 2.5 } }, - "qwen/qwen3-235b-a22b-fp8": { - "id": "qwen/qwen3-235b-a22b-fp8", - "name": "Qwen3 235B A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "x-ai/grok-4": { + "id": "x-ai/grok-4", + "name": "Grok 4", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "tool_call": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2025-07", + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text" @@ -170450,59 +161159,77 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 20000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 3, + "output": 15, + "cache_read": 0.75, + "cache_write": 15 } }, - "qwen/qwen3-coder-480b-a35b-instruct": { - "id": "qwen/qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "x-ai/grok-build-0.1": { + "id": "x-ai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", + "family": "grok-build", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 256000, + "output": 256000 }, "cost": { - "input": 0.29, - "output": 1.2 + "input": 1, + "output": 2 } }, - "qwen/qwen3-235b-a22b-thinking-2507": { - "id": "qwen/qwen3-235b-a22b-thinking-2507", - "name": "Qwen3 235B A22b Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "z-ai/glm-5.1": { + "id": "z-ai/glm-5.1", + "name": "GLM-5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -170513,27 +161240,33 @@ }, "open_weights": true, "limit": { - "context": 131072, + "context": 200000, "output": 131072 }, "cost": { - "input": 0.3, - "output": 3 + "input": 1.05, + "output": 3.5 } }, - "qwen/qwen3-30b-a3b-fp8": { - "id": "qwen/qwen3-30b-a3b-fp8", - "name": "Qwen3 30B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "z-ai/glm-5": { + "id": "z-ai/glm-5", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -170544,92 +161277,95 @@ }, "open_weights": true, "limit": { - "context": 40960, - "output": 20000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.09, - "output": 0.45 + "input": 0.95, + "output": 3.15 } }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "knowledge": "2024-10-01", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 65536, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "qwen/qwen3-32b-fp8": { - "id": "qwen/qwen3-32b-fp8", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-01", - "last_updated": "2026-01", + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40960, - "output": 20000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.45 + "input": 0.2, + "output": 1.25 } - } - } - }, - "ollama-cloud": { - "id": "ollama-cloud", - "env": [ - "OLLAMA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://ollama.com/v1", - "name": "Ollama Cloud", - "doc": "https://docs.ollama.com/cloud", - "models": { - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + }, + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -170637,105 +161373,94 @@ "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 1.75, + "output": 14 } }, - "nemotron-3-ultra": { - "id": "nemotron-3-ultra", - "name": "nemotron-3-ultra", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "openai/gpt-realtime-1.5": { + "id": "openai/gpt-realtime-1.5", + "name": "GPT Realtime 1.5", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "release_date": "2025-06-01", + "last_updated": "2025-06-01", "modalities": { "input": [ - "text" + "text", + "audio", + "image" ], "output": [ - "text" + "text", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 128000 + "context": 32000, + "output": 4096 }, "cost": { - "input": 0.1, - "output": 3, - "cache_read": 0.1 + "input": 4, + "output": 16 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "kimi-k3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-27", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -170743,44 +161468,37 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.15, + "output": 0.6 } }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", @@ -170790,58 +161508,71 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 + "input": 0.75, + "output": 4.5 } }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "minimax-m2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "knowledge": "2025-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "kimi-k2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "temperature": true, + "knowledge": "2024-10-01", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -170851,35 +161582,39 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "qwen3.5:397b": { - "id": "qwen3.5:397b", - "name": "qwen3.5:397b", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "release_date": "2026-02-15", - "last_updated": "2026-02-17", + "temperature": true, + "knowledge": "2024-10-01", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -170889,22 +161624,23 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 } }, - "gpt-oss:20b": { - "id": "gpt-oss:20b", - "name": "gpt-oss:20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "openai/gpt-5.5-pro": { + "id": "openai/gpt-5.5-pro", + "name": "GPT-5.5 Pro", + "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -170917,30 +161653,35 @@ } ], "tool_call": true, - "release_date": "2025-08-05", - "last_updated": "2026-01-19", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.07, - "output": 0.3, - "cache_read": 0.035 + "input": 30, + "output": 180 } }, - "gpt-oss:120b": { - "id": "gpt-oss:120b", - "name": "gpt-oss:120b", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", "family": "gpt-oss", "attachment": false, @@ -170956,8 +161697,9 @@ } ], "tool_call": true, + "temperature": true, "release_date": "2025-08-05", - "last_updated": "2026-01-19", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -170969,295 +161711,284 @@ "open_weights": true, "limit": { "context": 131072, - "output": 32768 + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.014 + "input": 0.05, + "output": 0.2 } }, - "nemotron-3-nano:30b": { - "id": "nemotron-3-nano:30b", - "name": "nemotron-3-nano:30b", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "openai/gpt-image-2": { + "id": "openai/gpt-image-2", + "name": "GPT Image 2", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "gpt-image", + "attachment": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2026-01-19", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.06, - "output": 0.24 + "context": 128000, + "output": 0 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "kimi-k2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "structured_output": true, + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "cost": { + "input": 5, + "output": 30 } }, - "mistral-large-3:675b": { - "id": "mistral-large-3:675b", - "name": "mistral-large-3:675b", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", + "bytedance/seedance-2": { + "id": "bytedance/seedance-2", + "name": "Seedance 2", + "description": "Video model for prompt-guided generation, editing, and motion workflows", + "family": "seed", "attachment": true, "reasoning": false, - "tool_call": true, - "release_date": "2025-12-02", - "last_updated": "2026-01-19", + "tool_call": false, + "temperature": false, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", "image" ], "output": [ - "text" + "video" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 1.5 + "context": 4096, + "output": 0 } }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "minimax-m2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "anthropic/claude-sonnet-4": { + "id": "anthropic/claude-sonnet-4", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024, + "max": 32000 } ], "tool_call": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 196608 + "context": 200000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek-v4-pro:0813": { - "id": "deepseek-v4-pro:0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] + "type": "budget_tokens", + "min": 1024, + "max": 32000 } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.66, - "output": 1.98, - "cache_read": 0.022 + "input": 5, + "output": 25 } }, - "minimax-m3": { - "id": "minimax-m3", - "name": "minimax-m3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax-m3", + "anthropic/claude-opus-4.1": { + "id": "anthropic/claude-opus-4.1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] + "type": "budget_tokens", + "min": 1024, + "max": 32000 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-31", - "last_updated": "2026-05-31", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512000, - "output": 131072 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "deepseek-v4-flash:0731": { - "id": "deepseek-v4-flash:0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] + "type": "budget_tokens", + "min": 1024, + "max": 32000 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 + "input": 3, + "output": 15 } }, - "gemma4:31b": { - "id": "gemma4:31b", - "name": "gemma4:31b", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "sarvam/sarvam-105b": { + "id": "sarvam/sarvam-105b", + "name": "Sarvam 105B", + "description": "Flagship Indian-language reasoning model for enterprise multilingual applications", + "family": "sarvam", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-08", + "temperature": true, + "release_date": "2025-09-01", + "last_updated": "2025-09-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -171265,39 +161996,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.4, - "cache_read": 0.05 + "input": 0.04, + "output": 0.16 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "sarvam/sarvam-30b": { + "id": "sarvam/sarvam-30b", + "name": "Sarvam 30B", + "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", + "family": "sarvam", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-02-18", + "last_updated": "2026-02-18", "modalities": { "input": [ "text" @@ -171308,33 +162026,27 @@ }, "open_weights": true, "limit": { - "context": 976000, - "output": 131072 + "context": 128000, + "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.02, + "output": 0.1 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "glm-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "deepseek-ai/deepseek-r1-distill-llama-70b": { + "id": "deepseek-ai/deepseek-r1-distill-llama-70b", + "name": "DeepSeek R1 Distill Llama 70B", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "release_date": "2026-03-27", - "last_updated": "2026-04-07", + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2025-01-23", + "last_updated": "2025-01-23", "modalities": { "input": [ "text" @@ -171345,37 +162057,26 @@ }, "open_weights": true, "limit": { - "context": 202752, + "context": 131072, "output": 131072 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 0.03, + "output": 0.14 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "deepseek-v4-pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", + "qwen/qwen3-coder": { + "id": "qwen/qwen3-coder", + "name": "Qwen3 Coder", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-23", + "last_updated": "2025-07-23", "modalities": { "input": [ "text" @@ -171386,40 +162087,32 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 262144, + "output": 66536 }, "cost": { - "input": 0.66, - "output": 1.98, - "cache_read": 0.022 + "input": 0.3, + "output": 1.2 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -171430,37 +162123,29 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 1.74, + "output": 3.48 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "kimi-k2.7-code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, + "minimax/minimax-m2.7-highspeed": { + "id": "minimax/minimax-m2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -171468,31 +162153,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.6, + "output": 2.4 } }, - "nemotron-3-super": { - "id": "nemotron-3-super", - "name": "nemotron-3-super", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -171503,40 +162183,93 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.015, - "output": 0.6, - "cache_read": 0.015 + "input": 0.3, + "output": 1.2 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "deepseek-v4-flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "leonardo-ai/lucid-origin": { + "id": "leonardo-ai/lucid-origin", + "name": "Lucid Origin", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "lucid", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-01", + "last_updated": "2025-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 4096, + "output": 0 + } + }, + "leonardo-ai/lucid-realism": { + "id": "leonardo-ai/lucid-realism", + "name": "Lucid Realism", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "family": "lucid", + "attachment": true, + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-06-01", + "last_updated": "2025-06-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 4096, + "output": 0 + } + } + } + }, + "abliteration-ai": { + "id": "abliteration-ai", + "env": [ + "ABLIT_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.abliteration.ai/v1", + "name": "abliteration.ai", + "doc": "https://docs.abliteration.ai/models", + "models": { + "abliterated-model": { + "id": "abliterated-model", + "name": "Abliterated Model", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, + "reasoning": false, "tool_call": true, - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "structured_output": false, + "temperature": true, + "release_date": "2026-01-06", + "last_updated": "2026-01-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -171544,33 +162277,33 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 150000, + "input": 150000, + "output": 8192 }, "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 + "input": 3, + "output": 3 } } } }, - "agentrouter": { - "id": "agentrouter", + "xpersona": { + "id": "xpersona", "env": [ - "AGENTROUTER_API_KEY" + "XPERSONA_API_KEY" ], "npm": "@ai-sdk/openai-compatible", - "api": "https://agentrouter.org/v1", - "name": "AgentRouter", - "doc": "https://agentrouter.org/docs/opencode.html", + "api": "https://www.xpersona.co/v1", + "name": "Xpersona", + "doc": "https://www.xpersona.co/docs", "models": { - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, + "xpersona-gpt-5.5": { + "id": "xpersona-gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -171585,15 +162318,18 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "knowledge": "2025-12-30", + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -171604,17 +162340,18 @@ "context": 1000000, "output": 128000 }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://agentrouter.org/v1" + "cost": { + "input": 3, + "output": 18, + "reasoning": 18, + "cache_read": 0.3 } }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "xpersona-frieren-coder": { + "id": "xpersona-frieren-coder", + "name": "Xpersona Frieren 1", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -171629,15 +162366,15 @@ } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "knowledge": "2025-12-30", + "release_date": "2026-05-01", + "last_updated": "2026-05-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -171646,75 +162383,101 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 384000 }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://agentrouter.org/v1" + "cost": { + "input": 1.5, + "output": 6, + "reasoning": 6, + "cache_read": 0.15 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "low", + "medium", "high", + "xhigh", "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "output": 128000 + }, + "cost": { + "input": 3, + "output": 18, + "reasoning": 18, + "cache_read": 0.3 } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + } + } + }, + "azure-cognitive-services": { + "id": "azure-cognitive-services", + "env": [ + "AZURE_COGNITIVE_SERVICES_RESOURCE_NAME", + "AZURE_COGNITIVE_SERVICES_API_KEY" + ], + "npm": "@ai-sdk/azure", + "name": "Azure Cognitive Services", + "doc": "https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models", + "models": { + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-03-31", + "release_date": "2025-11-24", + "last_updated": "2025-08-01", "modalities": { "input": [ "text", @@ -171727,251 +162490,247 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-07-31", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + }, + "cost": { + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } - } - } - }, - "tencent-tokenhub": { - "id": "tencent-tokenhub", - "env": [ - "TENCENT_TOKENHUB_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://tokenhub.tencentmaas.com/v1", - "name": "Tencent TokenHub", - "doc": "https://cloud.tencent.com/document/product/1823/130050", - "models": { - "hy3-preview": { - "id": "hy3-preview", - "name": "Hy3 preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + }, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 Nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "hy3": { - "id": "hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "none", - "high" + "low", + "medium", + "high", + "max" ] } ], "tool_call": true, "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "knowledge": "2025-12-31", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 192000, + "context": 200000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" + }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "hy4-preview": { - "id": "hy4-preview", - "name": "Hy4 preview", - "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", - "family": "Hy", - "attachment": false, + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1024000, - "output": 64000 - }, - "cost": { - "input": 0.834, - "output": 2.501, - "cache_read": 0.042 - } - } - } - }, - "watsonx": { - "id": "watsonx", - "env": [ - "WATSONX_AI_APIKEY", - "WATSONX_AI_PROJECT_ID" - ], - "npm": "watsonx-ai-provider", - "name": "watsonx.ai", - "doc": "https://www.ibm.com/docs/en/watsonx/saas?topic=solutions-supported-foundation-models", - "models": { - "meta-llama/llama-4-maverick-17b-128e-instruct-fp8": { - "id": "meta-llama/llama-4-maverick-17b-128e-instruct-fp8", - "name": "Llama 4 Maverick 17B 128E Instruct FP8", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-03-31", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 200000, + "output": 32000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.371, - "output": 1.484 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "meta-llama/llama-3-3-70b-instruct": { - "id": "meta-llama/llama-3-3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -171979,156 +162738,197 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 4096 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 0.7526, - "output": 0.7526 + "input": 0.6, + "output": 3 } }, - "ibm/granite-4-h-small": { - "id": "ibm/granite-4-h-small", - "name": "Granite-4.0-H-Small", - "description": "Open-weight hybrid model for enterprise chat, coding, retrieval-augmented generation, and tool-calling workloads", - "family": "granite", - "attachment": false, - "reasoning": false, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-10-02", - "last_updated": "2025-10-02", + "knowledge": "2025-02-31", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.0636, - "output": 0.265 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "mistralai/mistral-small-3-1-24b-instruct-2503": { - "id": "mistralai/mistral-small-3-1-24b-instruct-2503", - "name": "Mistral Small 3.1 24B", - "description": "Efficient multimodal model for instruction following, coding, reasoning, and function calling", - "family": "mistral-small", + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-17", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.106, - "output": 0.318 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 Mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.159, - "output": 0.636 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } - } - } - }, - "ambient": { - "id": "ambient", - "env": [ - "AMBIENT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.ambient.xyz/v1", - "name": "Ambient", - "doc": "https://ambient.xyz", - "models": { - "deepseek/deepseek-v4-flash-0731": { - "id": "deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", + }, + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] } ], "tool_call": true, + "interleaved": true, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -172136,200 +162936,272 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/models", + "shape": "completions" }, "cost": { - "input": 0.08, - "output": 0.18, - "cache_read": 0.016, - "cache_write": 0 + "input": 0.95, + "output": 4 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", - "attachment": false, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", + "medium", "high", "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 200000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/anthropic", + "api": "https://${AZURE_COGNITIVE_SERVICES_RESOURCE_NAME}.services.ai.azure.com/anthropic/v1" }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028, - "cache_write": 0 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "gpt-5.4-pro": { + "id": "gpt-5.4-pro", + "name": "GPT-5.4 Pro", + "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "structured_output": false, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 202752 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2, - "cache_read": 0.15, - "cache_write": 0 + "input": 30, + "output": 180, + "tiers": [ + { + "input": 60, + "output": 270, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 60, + "output": 270 + } } }, - "ambient/large": { - "id": "ambient/large", - "name": "Ambient Large", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 202752 + "context": 1050000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 2, - "cache_read": 0.15, - "cache_write": 0 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text", - "image" + "image", + "audio" ], "output": [ - "text" + "text", + "image", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 272000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.2, - "cache_write": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "meta-llama-3-70b-instruct": { + "id": "meta-llama-3-70b-instruct", + "name": "Meta-Llama-3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2023-12", + "release_date": "2024-04-18", + "last_updated": "2024-04-18", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -172337,37 +163209,26 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 8192, + "output": 2048 }, "cost": { - "input": 0.69, - "output": 3.49, - "cache_read": 0.14, - "cache_write": 0 + "input": 2.68, + "output": 3.54 } }, - "zai-org/GLM-5.2-FP8": { - "id": "zai-org/GLM-5.2-FP8", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "phi-3-mini-4k-instruct": { + "id": "phi-3-mini-4k-instruct", + "name": "Phi-3-mini-instruct (4k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -172378,32 +163239,26 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 202752 + "context": 4096, + "output": 1024 }, "cost": { - "input": 1.2, - "output": 4.2, - "cache_read": 0.26, - "cache_write": 0 + "input": 0.13, + "output": 0.52 } }, - "zai-org/GLM-5.1-FP8": { - "id": "zai-org/GLM-5.1-FP8", - "name": "GLM 5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "phi-3-mini-128k-instruct": { + "id": "phi-3-mini-128k-instruct", + "name": "Phi-3-mini-instruct (128k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -172414,40 +163269,30 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0, - "cache_write": 0 + "input": 0.13, + "output": 0.52 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo-V2.5", - "description": "MiMo omni model for text, image, video, audio, and agents", - "family": "mimo", - "attachment": true, + "phi-4-mini-reasoning": { + "id": "phi-4-mini-reasoning", + "name": "Phi-4-mini-reasoning", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -172455,21 +163300,20 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08, - "cache_write": 0 + "input": 0.075, + "output": 0.3 } }, - "stepfun/step-3.7-flash": { - "id": "stepfun/step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": true, + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -172477,137 +163321,105 @@ "values": [ "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.19, - "output": 1.14, - "cache_read": 0.03, - "cache_write": 0 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } - } - } - }, - "model-oracle-ai": { - "id": "model-oracle-ai", - "env": [ - "MODEL_ORACLE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.modeloracle.com/api/v1", - "name": "Model Oracle AI", - "doc": "https://modeloracle.com/setup/", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + }, + "llama-4-maverick-17b-128e-instruct-fp8": { + "id": "llama-4-maverick-17b-128e-instruct-fp8", + "name": "Llama 4 Maverick 17B 128E Instruct FP8", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 0.25, + "output": 1 } }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek-V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "input": 272000, + "context": 128000, "output": 128000 + }, + "cost": { + "input": 0.58, + "output": 1.68 } }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, + "gpt-5-codex": { + "id": "gpt-5-codex", + "name": "GPT-5-Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -172620,11 +163432,10 @@ } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "knowledge": "2024-09-30", + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text", @@ -172636,39 +163447,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 400000, + "output": 128000 + }, + "cost": { + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "claude-opus-4.8": { - "id": "claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "model-router": { + "id": "model-router", + "name": "Model Router", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "family": "model-router", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "release_date": "2025-05-19", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -172676,39 +163477,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.14, + "output": 0 } }, - "claude-haiku-4.5": { - "id": "claude-haiku-4.5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "gpt-4o-mini": { + "id": "gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", + "family": "gpt-mini", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2023-09", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -172716,78 +163508,94 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.6, + "cache_read": 0.075 } }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "phi-4-multimodal": { + "id": "phi-4-multimodal", + "name": "Phi-4-multimodal", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "phi", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text", "image", - "pdf" + "audio" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.08, + "output": 0.32, + "input_audio": 4 } }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "phi-4-reasoning-plus": { + "id": "phi-4-reasoning-plus", + "name": "Phi-4-reasoning-plus", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "temperature": true, + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 32000, + "output": 4096 + }, + "cost": { + "input": 0.125, + "output": 0.5 + } + }, + "codestral-2501": { + "id": "codestral-2501", + "name": "Codestral 25.01", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "codestral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "knowledge": "2024-03", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -172795,110 +163603,100 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 256000, + "output": 256000 + }, + "cost": { + "input": 0.3, + "output": 0.9 } }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "cohere-embed-v3-english": { + "id": "cohere-embed-v3-english", + "name": "Embed v3 English", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", + "attachment": false, + "reasoning": false, + "tool_call": false, "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "release_date": "2023-11-07", + "last_updated": "2023-11-07", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 512, + "output": 1024 + }, + "cost": { + "input": 0.1, + "output": 0 } }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, + "phi-3-medium-4k-instruct": { + "id": "phi-3-medium-4k-instruct", + "name": "Phi-3-medium-instruct (4k)", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 4096, + "output": 1024 + }, + "cost": { + "input": 0.17, + "output": 0.68 } }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -172906,47 +163704,44 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 272000, "output": 128000 + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.01 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "xhigh" - ] - } - ], + "gpt-4-turbo": { + "id": "gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2023-12", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 10, + "output": 30 } }, "gpt-4.1-mini": { @@ -172957,7 +163752,6 @@ "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, "knowledge": "2024-04", "release_date": "2025-04-14", @@ -172965,8 +163759,7 @@ "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -172976,55 +163769,18 @@ "limit": { "context": 1047576, "output": 32768 - } - }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 + "cost": { + "input": 0.4, + "output": 1.6, + "cache_read": 0.1 } }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -173039,9 +163795,8 @@ } ], "tool_call": true, - "structured_output": true, "temperature": false, - "knowledge": "2024-09-30", + "knowledge": "2024-05-30", "release_date": "2025-08-07", "last_updated": "2025-08-07", "modalities": { @@ -173055,41 +163810,31 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 272000, "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2, + "cache_read": 0.03 } }, - "auto": { - "id": "auto", - "name": "Auto", - "description": "Model Oracle AI decision engine that selects and routes among configured coding-agent models", + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-29", - "last_updated": "2026-07-07", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -173097,150 +163842,118 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 } - } - } - }, - "xai": { - "id": "xai", - "env": [ - "XAI_API_KEY" - ], - "npm": "@ai-sdk/xai", - "name": "xAI", - "doc": "https://docs.x.ai/docs/models", - "models": { - "grok-4.7": { - "id": "grok-4.7", - "name": "Grok 4.7", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + }, + "ministral-3b": { + "id": "ministral-3b", + "name": "Ministral 3B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "knowledge": "2024-03", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "input": 0.04, + "output": 0.04 } }, - "grok-imagine-image": { - "id": "grok-imagine-image", - "name": "Grok Imagine Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", + "llama-3.2-90b-vision-instruct": { + "id": "llama-3.2-90b-vision-instruct", + "name": "Llama-3.2-90B-Vision-Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-01-28", - "last_updated": "2026-01-28", + "tool_call": true, + "temperature": true, + "knowledge": "2023-12", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "image", - "pdf" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16000, - "output": 0 + "context": 128000, + "output": 8192 + }, + "cost": { + "input": 2.04, + "output": 2.04 } }, - "grok-imagine-video": { - "id": "grok-imagine-video", - "name": "Grok Imagine Video", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-01-28", - "last_updated": "2026-01-28", + "deepseek-v3.1": { + "id": "deepseek-v3.1", + "name": "DeepSeek-V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "knowledge": "2024-07", + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1024, - "output": 0 + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 0.56, + "output": 1.68 } }, - "grok-4.3": { - "id": "grok-4.3", - "name": "Grok 4.3", - "description": "xAI's Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + "o1": { + "id": "o1", + "name": "o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -173248,15 +163961,14 @@ } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-12-05", + "last_updated": "2024-12-05", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -173264,49 +163976,64 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } + "input": 15, + "output": 60, + "cache_read": 7.5 + } + }, + "gpt-4.1-nano": { + "id": "gpt-4.1-nano", + "name": "GPT-4.1 nano", + "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", + "family": "gpt-nano", + "attachment": true, + "reasoning": false, + "tool_call": true, + "temperature": true, + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "grok-4.20-0309-reasoning": { - "id": "grok-4.20-0309-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", + "grok-4-fast-reasoning": { + "id": "grok-4-fast-reasoning", + "name": "Grok 4 Fast (Reasoning)", + "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "knowledge": "2025-07", + "release_date": "2025-09-19", + "last_updated": "2025-09-19", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -173314,65 +164041,51 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 2000000, "output": 30000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.2, + "output": 0.5, + "cache_read": 0.05 } }, - "grok-imagine-video-1.5": { - "id": "grok-imagine-video-1.5", - "name": "Grok Imagine Video 1.5", - "description": "Video model for image-to-video generation, editing, and extension workflows", - "family": "grok", - "attachment": true, + "phi-4": { + "id": "phi-4", + "name": "Phi-4", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-05-30", - "last_updated": "2026-05-30", + "temperature": true, + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ - "text", - "image", - "audio", - "pdf" + "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1024, - "output": 0 + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.125, + "output": 0.5 } }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -173386,14 +164099,14 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -173401,186 +164114,119 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 0.6 - } + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "grok-4.20-0309-non-reasoning": { - "id": "grok-4.20-0309-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "meta-llama-3.1-8b-instruct": { + "id": "meta-llama-3.1-8b-instruct", + "name": "Meta-Llama-3.1-8B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 30000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.3, + "output": 0.61 } }, - "grok-imagine-image-quality": { - "id": "grok-imagine-image-quality", - "name": "Grok Imagine Image Quality", - "description": "Higher-fidelity Grok Imagine image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", - "attachment": true, + "gpt-3.5-turbo-0301": { + "id": "gpt-3.5-turbo-0301", + "name": "GPT-3.5 Turbo 0301", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-04-03", - "last_updated": "2026-04-03", + "temperature": true, + "knowledge": "2021-08", + "release_date": "2023-03-01", + "last_updated": "2023-03-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ - "image", - "pdf" + "text" ] }, "open_weights": false, "limit": { - "context": 16000, - "output": 0 + "context": 4096, + "output": 4096 + }, + "cost": { + "input": 1.5, + "output": 2 } }, - "grok-build-0.1": { - "id": "grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, + "deepseek-v3.2-speciale": { + "id": "deepseek-v3.2-speciale", + "name": "DeepSeek-V3.2-Speciale", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "knowledge": "2024-07", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 128000, + "output": 128000 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 4, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 4, - "cache_read": 0.4 - } + "input": 0.58, + "output": 1.68 } }, - "grok-4.20-multi-agent-0309": { - "id": "grok-4.20-multi-agent-0309", - "name": "Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "text-embedding-3-small": { + "id": "text-embedding-3-small", + "name": "text-embedding-3-small", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", + "attachment": false, + "reasoning": false, "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -173588,42 +164234,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 8191, + "output": 1536 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.02, + "output": 0 } }, - "grok-4.6": { - "id": "grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -173633,15 +164263,14 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -173649,125 +164278,105 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } + "input": 1.75, + "output": 14, + "cache_read": 0.175 } - } - } - }, - "nebius": { - "id": "nebius", - "env": [ - "NEBIUS_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.tokenfactory.nebius.com/v1", - "name": "Nebius Token Factory", - "doc": "https://docs.tokenfactory.nebius.com/", - "models": { - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma-3-27b-it", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": true, + }, + "mistral-large-2411": { + "id": "mistral-large-2411", + "name": "Mistral Large 24.11", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-10", - "release_date": "2026-01-20", - "last_updated": "2026-02-04", + "knowledge": "2024-09", + "release_date": "2024-11-01", + "last_updated": "2024-11-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 110000, - "input": 100000, - "output": 8192 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01, - "cache_write": 0.125 + "input": 2, + "output": 6 } }, - "Qwen/Qwen3-30B-A3B-Instruct-2507": { - "id": "Qwen/Qwen3-30B-A3B-Instruct-2507", - "name": "Qwen3-30B-A3B-Instruct-2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, + "gpt-4-turbo-vision": { + "id": "gpt-4-turbo-vision", + "name": "GPT-4 Turbo Vision", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-01-28", - "last_updated": "2026-02-04", + "knowledge": "2023-11", + "release_date": "2023-11-06", + "last_updated": "2024-04-09", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 262144, - "output": 8192 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.01, - "cache_write": 0.125 + "input": 10, + "output": 30 } }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-25", - "last_updated": "2025-10-04", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -173775,63 +164384,37 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 8192 - }, - "cost": { - "input": 0.2, - "output": 0.6 - } - }, - "Qwen/Qwen3-Embedding-8B": { - "id": "Qwen/Qwen3-Embedding-8B", - "name": "Qwen3-Embedding-8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2025-10", - "release_date": "2026-01-10", - "last_updated": "2026-02-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "input": 40960, - "output": 0 + "context": 400000, + "output": 128000 }, "cost": { - "input": 0.01, - "output": 0 + "input": 1.75, + "output": 14, + "cache_read": 0.125 } }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen3.5-397B-A17B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "o3-mini": { + "id": "o3-mini", + "name": "o3-mini", + "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", + "family": "o-mini", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-15", - "last_updated": "2026-05-07", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2024-12-20", + "last_updated": "2025-01-29", "modalities": { "input": [ "text" @@ -173840,46 +164423,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "input": 250000, - "output": 8192 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.6, - "output": 3.6, - "cache_read": 0.06, - "cache_write": 0.75 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + "text-embedding-ada-002": { + "id": "text-embedding-ada-002", + "name": "text-embedding-ada-002", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "reasoning": false, + "tool_call": false, + "release_date": "2022-12-15", + "last_updated": "2022-12-15", "modalities": { "input": [ "text" @@ -173888,43 +164452,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1024000, - "output": 1024000 + "context": 8192, + "output": 1536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.14 + "input": 0.1, + "output": 0 } }, - "deepseek-ai/DeepSeek-V4-Pro-0813": { - "id": "deepseek-ai/DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "meta-llama-3.1-70b-instruct": { + "id": "meta-llama-3.1-70b-instruct", + "name": "Meta-Llama-3.1-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -173935,42 +164484,26 @@ }, "open_weights": true, "limit": { - "context": 979000, - "output": 979000 + "context": 128000, + "output": 32768 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 1.32 + "input": 2.68, + "output": 3.54 } }, - "deepseek-ai/DeepSeek-V4.1-Flash": { - "id": "deepseek-ai/DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "mistral-medium-2505": { + "id": "mistral-medium-2505", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ "text", @@ -173980,75 +164513,29 @@ "text" ] }, - "open_weights": true, - "limit": { - "context": 1048000, - "output": 1048000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.3 - } - }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 128000, + "output": 128000 }, "cost": { - "input": 1.75, - "output": 3.5, - "cache_read": 0.15 + "input": 0.4, + "output": 2 } }, - "MiniMaxAI/MiniMax-M3": { - "id": "MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "phi-4-reasoning": { + "id": "phi-4-reasoning", + "name": "Phi-4-reasoning", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "phi", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -174059,77 +164546,72 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 32000, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.125, + "output": 0.5 } }, - "moonshotai/Kimi-K3": { - "id": "moonshotai/Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt-codex", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ - "text" + "text", + "image", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 8000 + "context": 400000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 3 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "gpt-3.5-turbo-0613": { + "id": "gpt-3.5-turbo-0613", + "name": "GPT-3.5 Turbo 0613", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "reasoning": false, + "tool_call": false, + "temperature": true, + "knowledge": "2021-08", + "release_date": "2023-06-13", + "last_updated": "2023-06-13", "modalities": { "input": [ "text" @@ -174138,41 +164620,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 8000 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0.95, + "input": 3, "output": 4 } }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "cohere-embed-v3-multilingual": { + "id": "cohere-embed-v3-multilingual", + "name": "Embed v3 Multilingual", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2023-11-07", + "last_updated": "2023-11-07", "modalities": { "input": [ "text" @@ -174183,39 +164651,26 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 512, + "output": 1024 }, "cost": { - "input": 1.4, - "output": 4.4 + "input": 0.1, + "output": 0 } }, - "zai-org/GLM-5.3-Flash": { - "id": "zai-org/GLM-5.3-Flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "gpt-3.5-turbo-0125": { + "id": "gpt-3.5-turbo-0125", + "name": "GPT-3.5 Turbo 0125", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2021-08", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -174224,42 +164679,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1024000, - "output": 1024000 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.15 + "input": 0.5, + "output": 1.5 } }, - "zai-org/GLM-5.3": { - "id": "zai-org/GLM-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "phi-3-small-8k-instruct": { + "id": "phi-3-small-8k-instruct", + "name": "Phi-3-small-instruct (8k)", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -174270,69 +164711,27 @@ }, "open_weights": true, "limit": { - "context": 1024000, - "output": 1024000 + "context": 8192, + "output": 2048 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 1.4 + "input": 0.15, + "output": 0.6 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "Nemotron-3-Super-120B-A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek-R1", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-03-11", - "last_updated": "2026-03-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "input": 262144, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 0.9 - } - }, - "nvidia/Nemotron-3_5-Lightning": { - "id": "nvidia/Nemotron-3_5-Lightning", - "name": "Nemotron 3.5 Lightning 30B A3B", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "tool_call": false, "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "knowledge": "2024-07", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -174343,35 +164742,28 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 163840, + "output": 163840 }, "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.06 + "input": 1.35, + "output": 5.4 } }, - "nvidia/Nemotron-3-Ultra-550b-a55b": { - "id": "nvidia/Nemotron-3-Ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", + "kimi-k2-thinking": { + "id": "kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": true, "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "knowledge": "2024-08", + "release_date": "2025-11-06", + "last_updated": "2025-12-02", "modalities": { "input": [ "text" @@ -174382,78 +164774,71 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1, - "output": 3, - "cache_read": 1 + "input": 0.6, + "output": 2.5, + "cache_read": 0.15 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "attachment": false, + "gpt-5.1-chat": { + "id": "gpt-5.1-chat", + "name": "GPT-5.1 Chat", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high" + "medium" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-01-10", - "last_updated": "2026-02-04", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-14", + "last_updated": "2025-11-14", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ - "text" + "text", + "image", + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 124000, - "output": 8192 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.15, - "output": 0.6, - "reasoning": 0.6, - "cache_read": 0.015, - "cache_write": 0.18 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "NousResearch/Hermes-4-405B": { - "id": "NousResearch/Hermes-4-405B", - "name": "Hermes-4-405B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "llama-3.3-70b-instruct": { + "id": "llama-3.3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-11", - "release_date": "2026-01-30", - "last_updated": "2026-02-04", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ "text" @@ -174464,51 +164849,30 @@ }, "open_weights": true, "limit": { - "context": 131072, - "input": 120000, - "output": 8192 + "context": 128000, + "output": 32768 }, "cost": { - "input": 1, - "output": 3, - "reasoning": 3, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.71, + "output": 0.71 } - } - } - }, - "minimax-cn": { - "id": "minimax-cn", - "env": [ - "MINIMAX_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://api.minimax.cn/anthropic/v1", - "name": "MiniMax (minimax.cn)", - "doc": "https://platform.minimaxi.com/docs/guides/quickstart", - "models": { - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", + }, + "llama-4-scout-17b-16e-instruct": { + "id": "llama-4-scout-17b-16e-instruct", + "name": "Llama 4 Scout 17B 16E Instruct", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-25", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -174516,43 +164880,26 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 512000 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "tiers": [ - { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12, - "tier": { - "type": "context", - "size": 512000 - } - } - ], - "context_over_200k": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 - } + "input": 0.2, + "output": 0.78 } }, - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "gpt-3.5-turbo-1106": { + "id": "gpt-3.5-turbo-1106", + "name": "GPT-3.5 Turbo 1106", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", + "knowledge": "2021-08", + "release_date": "2023-11-06", + "last_updated": "2023-11-06", "modalities": { "input": [ "text" @@ -174561,30 +164908,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 16384, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 1, + "output": 2 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "phi-4-mini": { + "id": "phi-4-mini", + "name": "Phi-4-mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2023-10", + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text" @@ -174595,28 +164940,26 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.075, + "output": 0.3 } }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "cohere-command-r-plus-08-2024": { + "id": "cohere-command-r-plus-08-2024", + "name": "Command R+", + "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", + "family": "command-r", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -174627,28 +164970,26 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 4000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 2.5, + "output": 10 } }, - "MiniMax-M2.7-highspeed": { - "id": "MiniMax-M2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "meta-llama-3.1-405b-instruct": { + "id": "meta-llama-3.1-405b-instruct", + "name": "Meta-Llama-3.1-405B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2023-12", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -174659,28 +165000,26 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 5.33, + "output": 16 } }, - "MiniMax-M2": { - "id": "MiniMax-M2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "gpt-4-32k": { + "id": "gpt-4-32k", + "name": "GPT-4 32K", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "knowledge": "2023-11", + "release_date": "2023-03-14", + "last_updated": "2023-03-14", "modalities": { "input": [ "text" @@ -174689,28 +165028,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 32768, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 60, + "output": 120 } }, - "MiniMax-M2.5-highspeed": { - "id": "MiniMax-M2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "phi-3-medium-128k-instruct": { + "id": "phi-3-medium-128k-instruct", + "name": "Phi-3-medium-instruct (128k)", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -174721,87 +165060,68 @@ }, "open_weights": true, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "output": 4096 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.17, + "output": 0.68 } - } - } - }, - "scaleway": { - "id": "scaleway", - "env": [ - "SCALEWAY_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.scaleway.ai/v1", - "name": "Scaleway", - "doc": "https://www.scaleway.com/en/docs/generative-apis/", - "models": { - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + }, + "o4-mini": { + "id": "o4-mini", + "name": "o4-mini", + "description": "Fast o-series model for compact reasoning, coding, and tool use", + "family": "o-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 16384 + "context": 200000, + "output": 100000 }, - "status": "beta", "cost": { - "input": 0.468, - "output": 0.936, - "reasoning": 0.936, - "cache_read": 0.0936 + "input": 1.1, + "output": 4.4, + "cache_read": 0.275 } }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", - "family": "qwen", + "gpt-4": { + "id": "gpt-4", + "name": "GPT-4", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-01", - "last_updated": "2026-03-17", + "knowledge": "2023-11", + "release_date": "2023-03-14", + "last_updated": "2023-03-14", "modalities": { "input": [ "text" @@ -174810,61 +165130,32 @@ "text" ] }, - "open_weights": true, - "limit": { - "context": 260000, - "output": 16384 - }, - "cost": { - "input": 0.75, - "output": 2.25, - "reasoning": 8.4 - } - }, - "whisper-large-v3": { - "id": "whisper-large-v3", - "name": "Whisper Large v3", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2023-09-01", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 0, + "context": 8192, "output": 8192 }, "cost": { - "input": 0.003, - "output": 0 + "input": 60, + "output": 120 } }, - "bge-multilingual-gemma2": { - "id": "bge-multilingual-gemma2", - "name": "BGE Multilingual Gemma2", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, + "gpt-4o": { + "id": "gpt-4o", + "name": "GPT-4o", + "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", + "family": "gpt", + "attachment": true, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-07-26", - "last_updated": "2025-06-15", + "tool_call": true, + "temperature": true, + "knowledge": "2023-09", + "release_date": "2024-05-13", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -174872,26 +165163,27 @@ }, "open_weights": false, "limit": { - "context": 8191, - "output": 3072 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0 + "input": 2.5, + "output": 10, + "cache_read": 1.25 } }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", + "cohere-command-r-08-2024": { + "id": "cohere-command-r-08-2024", + "name": "Command R", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2026-03-17", + "knowledge": "2024-06-01", + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -174903,37 +165195,34 @@ "open_weights": true, "limit": { "context": 128000, - "output": 32768 + "output": 4000 }, "cost": { - "input": 0.2, - "output": 0.8 + "input": 0.15, + "output": 0.6 } }, - "gemma-4-26b-a4b-it": { - "id": "gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "gpt-5-pro": { + "id": "gpt-5-pro", + "name": "GPT-5 Pro", + "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", - "medium", "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-01", - "last_updated": "2026-05-22", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-10-06", + "last_updated": "2025-10-06", "modalities": { "input": [ "text", @@ -174943,32 +165232,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 16384 + "context": 400000, + "output": 272000 }, - "status": "beta", "cost": { - "input": 0.25, - "output": 0.5 + "input": 15, + "output": 120 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "llama-3.2-11b-vision-instruct": { + "id": "llama-3.2-11b-vision-instruct", + "name": "Llama-3.2-11B-Vision-Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", "family": "llama", "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2026-03-17", + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -174976,74 +165265,30 @@ }, "open_weights": true, "limit": { - "context": 100000, - "output": 16384 + "context": 128000, + "output": 8192 }, "cost": { - "input": 0.9, - "output": 0.9 + "input": 0.37, + "output": 0.37 } }, - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, + "cohere-command-a": { + "id": "cohere-command-a", + "name": "Command A", + "description": "Cohere command model for multilingual enterprise agents, tools, and chat", + "family": "command-a", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-05-01", - "last_updated": "2026-05-22", + "knowledge": "2024-06-01", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "status": "beta", - "cost": { - "input": 0.25, - "output": 1.5 - } - }, - "pixtral-12b-2409": { - "id": "pixtral-12b-2409", - "name": "Pixtral 12B 2409", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-09-25", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" @@ -175051,28 +165296,31 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 256000, + "output": 8000 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 2.5, + "output": 10 } }, - "qwen3-embedding-8b": { - "id": "qwen3-embedding-8b", - "name": "Qwen3 Embedding 8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gpt-5-chat": { + "id": "gpt-5-chat", + "name": "GPT-5 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": false, "temperature": false, - "release_date": "2025-06-05", - "last_updated": "2026-03-17", + "knowledge": "2024-10-24", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -175080,35 +165328,38 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 4096 + "context": 128000, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "mistral-medium-3.5-128b": { - "id": "mistral-medium-3.5-128b", - "name": "Mistral Medium 3.5 128B", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", + "low", + "medium", "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", @@ -175118,45 +165369,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 16384 + "context": 272000, + "output": 128000 }, "cost": { - "input": 1.5, - "output": 7.5 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "cohere-embed-v-4-0": { + "id": "cohere-embed-v-4-0", + "name": "Embed v4", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "cohere-embed", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "reasoning": false, + "tool_call": false, + "temperature": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -175164,37 +165402,26 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 128000, + "output": 1536 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.12, + "output": 0 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "mistral-nemo": { + "id": "mistral-nemo", + "name": "Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-07", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text" @@ -175205,30 +165432,29 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 16384 + "context": 128000, + "output": 128000 }, "cost": { - "input": 1.8, - "output": 5.5 + "input": 0.15, + "output": 0.15 } }, - "mistral-small-3.2-24b-instruct-2506": { - "id": "mistral-small-3.2-24b-instruct-2506", - "name": "Mistral Small 3.2 24B Instruct (2506)", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "phi-3.5-mini-instruct": { + "id": "phi-3.5-mini-instruct", + "name": "Phi-3.5-mini-instruct", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "phi", "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2026-03-17", + "knowledge": "2023-10", + "release_date": "2024-08-20", + "last_updated": "2024-08-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -175237,34 +165463,26 @@ "open_weights": true, "limit": { "context": 128000, - "output": 32768 + "output": 4096 }, "cost": { - "input": 0.15, - "output": 0.35 + "input": 0.13, + "output": 0.52 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT-OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": true, + "o1-mini": { + "id": "o1-mini", + "name": "o1-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2026-03-17", + "temperature": false, + "knowledge": "2023-09", + "release_date": "2024-09-12", + "last_updated": "2024-09-12", "modalities": { "input": [ "text" @@ -175273,63 +165491,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 32768 + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.6 - } - } - } - }, - "vercel": { - "id": "vercel", - "env": [ - "AI_GATEWAY_API_KEY" - ], - "npm": "@ai-sdk/gateway", - "name": "Vercel AI Gateway", - "doc": "https://github.com/vercel/ai/tree/5eb85cc45a259553501f535b8ac79a77d0e79223/packages/gateway", - "models": { - "voyage/voyage-law-2": { - "id": "voyage/voyage-law-2", - "name": "voyage-law-2", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-04-15", - "last_updated": "2024-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 1536 + "input": 1.1, + "output": 4.4, + "cache_read": 0.55 } }, - "voyage/rerank-2.5-lite": { - "id": "voyage/rerank-2.5-lite", - "name": "Voyage Rerank 2.5 Lite", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "voyage", + "text-embedding-3-large": { + "id": "text-embedding-3-large", + "name": "text-embedding-3-large", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "text-embedding", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -175340,24 +165522,30 @@ }, "open_weights": false, "limit": { - "context": 32000, - "output": 32000 + "context": 8191, + "output": 3072 + }, + "cost": { + "input": 0.13, + "output": 0 } }, - "voyage/voyage-3.5": { - "id": "voyage/voyage-3.5", - "name": "voyage-3.5", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "voyage", - "attachment": false, + "mistral-small-2503": { + "id": "mistral-small-2503", + "name": "Mistral Small 3.1", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "mistral-small", + "attachment": true, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "knowledge": "2024-09", + "release_date": "2025-03-01", + "last_updated": "2025-03-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -175365,21 +165553,26 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.3 } }, - "voyage/voyage-finance-2": { - "id": "voyage/voyage-finance-2", - "name": "voyage-finance-2", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "voyage", + "meta-llama-3-8b-instruct": { + "id": "meta-llama-3-8b-instruct", + "name": "Meta-Llama-3-8B-Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2024-06-03", - "last_updated": "2024-03", + "knowledge": "2023-12", + "release_date": "2024-04-18", + "last_updated": "2024-04-18", "modalities": { "input": [ "text" @@ -175388,48 +165581,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 8192, - "output": 1536 - } - }, - "voyage/voyage-4-large": { - "id": "voyage/voyage-4-large", - "name": "voyage-4-large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-15", - "last_updated": "2026-03-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "output": 2048 }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 0 + "cost": { + "input": 0.3, + "output": 0.61 } }, - "voyage/voyage-3.5-lite": { - "id": "voyage/voyage-3.5-lite", - "name": "voyage-3.5-lite", + "phi-3-small-128k-instruct": { + "id": "phi-3-small-128k-instruct", + "name": "Phi-3-small-instruct (128k)", "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "voyage", + "family": "phi", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "knowledge": "2023-10", + "release_date": "2024-04-23", + "last_updated": "2024-04-23", "modalities": { "input": [ "text" @@ -175438,23 +165611,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.15, + "output": 0.6 } }, - "voyage/voyage-4": { - "id": "voyage/voyage-4", - "name": "voyage-4", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "voyage", + "deepseek-v3-0324": { + "id": "deepseek-v3-0324", + "name": "DeepSeek-V3-0324", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, "temperature": true, - "release_date": "2026-01-15", - "last_updated": "2026-03-06", + "knowledge": "2024-07", + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ "text" @@ -175463,26 +165641,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "output": 0 + "context": 131072, + "output": 131072 + }, + "cost": { + "input": 1.14, + "output": 4.56 } }, - "voyage/voyage-3-large": { - "id": "voyage/voyage-3-large", - "name": "voyage-3-large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-01-07", - "last_updated": "2024-09", + "o3": { + "id": "o3", + "name": "o3", + "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", + "family": "o", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "temperature": false, + "knowledge": "2024-05", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -175490,24 +165684,40 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 + "context": 200000, + "output": 100000 + }, + "cost": { + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "voyage/voyage-code-2": { - "id": "voyage/voyage-code-2", - "name": "voyage-code-2", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "voyage", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2024-01", + "gpt-5.2-chat": { + "id": "gpt-5.2-chat", + "name": "GPT-5.2 Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "medium" + ] + } + ], + "tool_call": true, + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -175515,21 +165725,28 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 + "context": 128000, + "output": 16384 + }, + "cost": { + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "voyage/voyage-code-3": { - "id": "voyage/voyage-code-3", - "name": "voyage-code-3", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "voyage", + "deepseek-r1-0528": { + "id": "deepseek-r1-0528", + "name": "DeepSeek-R1-0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, "temperature": true, - "release_date": "2024-12-04", - "last_updated": "2024-09", + "knowledge": "2024-07", + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -175538,23 +165755,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 8192, - "output": 1536 + "context": 163840, + "output": 163840 + }, + "cost": { + "input": 1.35, + "output": 5.4 } }, - "voyage/rerank-2.5": { - "id": "voyage/rerank-2.5", - "name": "Voyage Rerank 2.5", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "voyage", + "gpt-3.5-turbo-instruct": { + "id": "gpt-3.5-turbo-instruct", + "name": "GPT-3.5 Turbo Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2025-08-11", - "last_updated": "2025-08-11", + "knowledge": "2021-08", + "release_date": "2023-09-21", + "last_updated": "2023-09-21", "modalities": { "input": [ "text" @@ -175565,21 +165787,26 @@ }, "open_weights": false, "limit": { - "context": 32000, - "output": 32000 + "context": 4096, + "output": 4096 + }, + "cost": { + "input": 1.5, + "output": 2 } }, - "voyage/voyage-4-lite": { - "id": "voyage/voyage-4-lite", - "name": "voyage-4-lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "voyage", + "phi-3.5-moe-instruct": { + "id": "phi-3.5-moe-instruct", + "name": "Phi-3.5-MoE-instruct", + "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", + "family": "phi", "attachment": false, "reasoning": false, "tool_call": false, "temperature": true, - "release_date": "2026-01-15", - "last_updated": "2026-03-06", + "knowledge": "2023-10", + "release_date": "2024-08-20", + "last_updated": "2024-08-20", "modalities": { "input": [ "text" @@ -175588,17 +165815,22 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "output": 0 + "context": 128000, + "output": 4096 + }, + "cost": { + "input": 0.16, + "output": 0.64 } }, - "interfaze/interfaze-beta": { - "id": "interfaze/interfaze-beta", - "name": "Interfaze Beta", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": false, + "codex-mini": { + "id": "codex-mini", + "name": "Codex Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -175610,48 +165842,11 @@ ] } ], - "tool_call": false, - "temperature": true, - "release_date": "2025-10-07", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32000 - }, - "cost": { - "input": 1.5, - "output": 3.5 - } - }, - "zai/glm-4.5": { - "id": "zai/glm-4.5", - "name": "GLM 4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "temperature": false, + "knowledge": "2024-04", + "release_date": "2025-05-16", + "last_updated": "2025-05-16", "modalities": { "input": [ "text" @@ -175660,22 +165855,34 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 96000 + "context": 200000, + "output": 100000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 1.5, + "output": 6, + "cache_read": 0.375 } - }, - "zai/glm-5v-turbo": { - "id": "zai/glm-5v-turbo", - "name": "GLM 5V Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", + } + } + }, + "baseten": { + "id": "baseten", + "env": [ + "BASETEN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://inference.baseten.co/v1", + "name": "Baseten", + "doc": "https://docs.baseten.co/inference/model-apis/overview", + "models": { + "moonshotai/Kimi-K2.6": { + "id": "moonshotai/Kimi-K2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -175684,52 +165891,55 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 1.2, + "input": 0.95, "output": 4, - "cache_read": 0.24 + "cache_read": 0.16 } }, - "zai/glm-5.3-flash": { - "id": "zai/glm-5.3-flash", - "name": "GLM 5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "moonshotai/Kimi-K2.5": { + "id": "moonshotai/Kimi-K2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "knowledge": "2025-12", + "release_date": "2026-01-30", + "last_updated": "2026-02-12", "modalities": { "input": [ "text", @@ -175741,21 +165951,21 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 0.6, + "output": 3, + "cache_read": 0.12 } }, - "zai/glm-4.6": { - "id": "zai/glm-4.6", - "name": "GLM 4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, + "moonshotai/Kimi-K2.7-Code": { + "id": "moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -175763,14 +165973,15 @@ } ], "tool_call": true, - "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -175778,39 +165989,38 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 96000 + "context": 262000, + "output": 262000 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "zai/glm-5.2-fast": { - "id": "zai/glm-5.2-fast", - "name": "GLM 5.2 Fast", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "OpenAI GPT 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "xhigh" + "low", + "medium", + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2025-08", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -175821,20 +166031,19 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128072, + "output": 128072 }, "cost": { - "input": 2.1, - "output": 6.6, - "cache_read": 0.21 + "input": 0.1, + "output": 0.5 } }, - "zai/glm-5": { - "id": "zai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "nvidia/Nemotron-120B-A12B": { + "id": "nvidia/Nemotron-120B-A12B", + "name": "Nemotron Super", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -175843,9 +166052,14 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "knowledge": "2026-02", + "release_date": "2026-03-11", + "last_updated": "2026-03-11", "modalities": { "input": [ "text" @@ -175857,19 +166071,20 @@ "open_weights": true, "limit": { "context": 202800, - "output": 131100 + "output": 202800 }, "cost": { - "input": 1, - "output": 3.2 + "input": 0.3, + "output": 0.75, + "cache_read": 0.06 } }, - "zai/glm-4.5v": { - "id": "zai/glm-4.5v", - "name": "GLM 4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, + "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B": { + "id": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B", + "name": "Nemotron Ultra", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -175877,56 +166092,16 @@ } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 66000, - "output": 16000 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11 - } - }, - "zai/glm-5.3-flashx": { - "id": "zai/glm-5.3-flashx", - "name": "GLM 5.3 FlashX", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -175934,20 +166109,20 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 202800, + "output": 202800 }, "cost": { - "input": 0.37, - "output": 1.25, - "cache_read": 0.075 + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 } }, - "zai/glm-4.7-flash": { - "id": "zai/glm-4.7-flash", - "name": "GLM 4.7 Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", + "zai-org/GLM-5": { + "id": "zai-org/GLM-5", + "name": "GLM 5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -175956,10 +166131,14 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "knowledge": "2026-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -175970,18 +166149,19 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 131000 + "context": 202800, + "output": 202800 }, "cost": { - "input": 0.07, - "output": 0.4 + "input": 0.95, + "output": 3.15, + "cache_read": 0.2 } }, - "zai/glm-4.7": { - "id": "zai/glm-4.7", + "zai-org/GLM-4.7": { + "id": "zai-org/GLM-4.7", "name": "GLM 4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, @@ -175991,7 +166171,10 @@ } ], "tool_call": true, - "interleaved": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, "knowledge": "2025-04", "release_date": "2025-12-22", @@ -176007,7 +166190,7 @@ "open_weights": true, "limit": { "context": 200000, - "output": 120000 + "output": 200000 }, "cost": { "input": 0.6, @@ -176015,8 +166198,8 @@ "cache_read": 0.12 } }, - "zai/glm-5.2": { - "id": "zai/glm-5.2", + "zai-org/GLM-5.2": { + "id": "zai-org/GLM-5.2", "name": "GLM 5.2", "description": "Open flagship GLM for long-horizon coding agents and million-token context work", "family": "glm", @@ -176025,16 +166208,12 @@ "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, "release_date": "2026-06-13", @@ -176049,21 +166228,21 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 202720, + "output": 202720 }, "cost": { - "input": 0.8, - "output": 2.55, - "cache_read": 0.16 + "input": 1.4, + "output": 4.4, + "cache_read": 0.3 } }, - "zai/glm-5.1": { - "id": "zai/glm-5.1", + "zai-org/GLM-5.1": { + "id": "zai-org/GLM-5.1", "name": "GLM 5.1", "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", "family": "glm", - "attachment": true, + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -176071,6 +166250,9 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, "release_date": "2026-04-07", @@ -176086,32 +166268,26 @@ "open_weights": true, "limit": { "context": 202800, - "output": 64000 + "output": 202800 }, "cost": { - "input": 1.4, - "output": 4.4, + "input": 1.3, + "output": 4.3, "cache_read": 0.26 } }, - "zai/glm-4.7-flashx": { - "id": "zai/glm-4.7-flashx", - "name": "GLM 4.7 FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek V3.1", + "description": "Legacy model retained for compatibility with older integrations", + "family": "deepseek", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", + "release_date": "2025-08-25", + "last_updated": "2025-08-25", "modalities": { "input": [ "text" @@ -176122,32 +166298,42 @@ }, "open_weights": true, "limit": { - "context": 200000, - "output": 128000 + "context": 164000, + "output": 131000 }, + "status": "deprecated", "cost": { - "input": 0.06, - "output": 0.4, - "cache_read": 0.01 + "input": 0.5, + "output": 1.5 } }, - "zai/glm-5-turbo": { - "id": "zai/glm-5-turbo", - "name": "GLM 5 Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", + "deepseek-ai/DeepSeek-V4-Pro": { + "id": "deepseek-ai/DeepSeek-V4-Pro", + "name": "Deepseek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -176156,34 +166342,33 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 202800, - "output": 131100 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 + "input": 1.74, + "output": 3.48, + "cache_read": 0.145 } }, - "zai/glm-4.5-air": { - "id": "zai/glm-4.5-air", - "name": "GLM 4.5 Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "MiniMaxAI/MiniMax-M2.5": { + "id": "MiniMaxAI/MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Legacy model retained for compatibility with older integrations", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2026-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -176194,37 +166379,38 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 96000 + "context": 204000, + "output": 204000 }, + "status": "deprecated", "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03 + "input": 0.3, + "output": 1.2 } - }, - "zai/glm-5.3": { - "id": "zai/glm-5.3", - "name": "GLM 5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + } + } + }, + "atomic-chat": { + "id": "atomic-chat", + "env": [ + "ATOMIC_CHAT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "http://127.0.0.1:1337/v1", + "name": "Atomic Chat", + "doc": "https://atomic.chat", + "models": { + "gemma-4-E4B-it-IQ4_XS": { + "id": "gemma-4-E4B-it-IQ4_XS", + "name": "Gemma 4 E4B Instruct (IQ4_XS)", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, + "reasoning": false, + "tool_call": false, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" @@ -176235,37 +166421,25 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 32768, + "output": 8192 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.14 + "input": 0, + "output": 0 } }, - "zai/glm-5.3-fast": { - "id": "zai/glm-5.3-fast", - "name": "GLM 5.3 Fast", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "Meta-Llama-3_1-8B-Instruct-GGUF": { + "id": "Meta-Llama-3_1-8B-Instruct-GGUF", + "name": "Meta Llama 3.1 8B Instruct (GGUF)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -176276,162 +166450,146 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 262144 + "context": 131072, + "output": 4096 }, "cost": { - "input": 2.1, - "output": 6.6, - "cache_read": 0.21 + "input": 0, + "output": 0 } }, - "spacexai/grok-4.20-multi-agent": { - "id": "spacexai/grok-4.20-multi-agent", - "name": "Grok 4.20 Multi-Agent", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "Qwen3_5-9B-MLX-4bit": { + "id": "Qwen3_5-9B-MLX-4bit", + "name": "Qwen 3.5 9B (MLX 4-bit)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-10", + "temperature": true, + "release_date": "2026-03-05", + "last_updated": "2026-04-04", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 32768, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200001 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0, + "output": 0 } }, - "spacexai/grok-tts": { - "id": "spacexai/grok-tts", - "name": "Grok TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "grok", + "gemma-4-E4B-it-MLX-4bit": { + "id": "gemma-4-E4B-it-MLX-4bit", + "name": "Gemma 4 E4B Instruct (MLX 4-bit)", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "temperature": true, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, + "context": 32768, + "output": 8192 + }, + "cost": { + "input": 0, "output": 0 } }, - "spacexai/grok-4.20-reasoning": { - "id": "spacexai/grok-4.20-reasoning", - "name": "Grok 4.20 Reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "Qwen3_5-9B-Q4_K_M": { + "id": "Qwen3_5-9B-Q4_K_M", + "name": "Qwen 3.5 9B (Q4_K_M)", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-10", + "temperature": true, + "release_date": "2026-03-05", + "last_updated": "2026-04-04", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 32768, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200001 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0, + "output": 0 } - }, - "spacexai/grok-4.7": { - "id": "spacexai/grok-4.7", - "name": "Grok 4.7", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", + } + } + }, + "meta": { + "id": "meta", + "env": [ + "META_MODEL_API_KEY" + ], + "npm": "@ai-sdk/openai", + "api": "https://api.meta.ai/v1", + "name": "Meta", + "doc": "https://dev.meta.ai/docs", + "models": { + "muse-spark-1.1": { + "id": "muse-spark-1.1", + "name": "Muse Spark 1.1", + "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", + "family": "muse", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", + "minimal", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "release_date": "2026-04-08", + "last_updated": "2026-07-09", "modalities": { "input": [ "text", - "image" + "image", + "pdf", + "video" ], "output": [ "text" @@ -176439,232 +166597,221 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 1000000, + "output": 32000 }, "cost": { - "input": 1.2, - "output": 3.6, - "cache_read": 0.3, - "tiers": [ - { - "input": 2.4, - "output": 7.2, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 200001 - } - } - ], - "context_over_200k": { - "input": 2.4, - "output": 7.2, - "cache_read": 0.6 - } - } - }, - "spacexai/grok-voice-think-fast-1.0": { - "id": "spacexai/grok-voice-think-fast-1.0", - "name": "Grok Voice Think Fast 1.0", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "input": 1.25, + "output": 4.25, + "cache_read": 0.15 } - }, - "spacexai/grok-4.20-non-reasoning": { - "id": "spacexai/grok-4.20-non-reasoning", - "name": "Grok 4.20 Non-Reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + } + } + }, + "routing-run": { + "id": "routing-run", + "env": [ + "ROUTING_RUN_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.routing.run/v1", + "name": "routing.run", + "doc": "https://docs.routing.run/api-reference/models", + "models": { + "kimi-k2.6-nitro": { + "id": "kimi-k2.6-nitro", + "name": "Kimi K2.6 Nitro", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-10", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200001 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.275, + "output": 1.1 } }, - "spacexai/grok-stt": { - "id": "spacexai/grok-stt", - "name": "Grok STT", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "grok", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.112, + "output": 0.224 } }, - "spacexai/grok-imagine-image": { - "id": "spacexai/grok-imagine-image", - "name": "Grok Imagine Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-01-28", - "last_updated": "2026-01-28", + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 0.275, + "output": 1.1 } }, - "spacexai/grok-imagine-video": { - "id": "spacexai/grok-imagine-video", - "name": "Grok Imagine", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-01-28", - "last_updated": "2026-01-28", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 64000 + }, + "cost": { + "input": 0.348, + "output": 0.696 } }, - "spacexai/grok-4.1-fast-reasoning": { - "id": "spacexai/grok-4.1-fast-reasoning", - "name": "Grok 4.1 Fast Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "input": 0.8, + "output": 2.4 } }, - "spacexai/grok-4.3": { - "id": "spacexai/grok-4.3", - "name": "Grok 4.3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -176678,176 +166825,134 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 1000000 + "output": 32000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200001 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 5, + "output": 25 } }, - "spacexai/grok-imagine-video-1.5": { - "id": "spacexai/grok-imagine-video-1.5", - "name": "Grok Imagine Video 1.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", + "glm-5.2-nitro": { + "id": "glm-5.2-nitro", + "name": "GLM 5.2 Nitro", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-05-30", - "last_updated": "2026-05-30", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 0.8, + "output": 2.4 } }, - "spacexai/grok-4.5": { - "id": "spacexai/grok-4.5", - "name": "Grok 4.5", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 500000, - "output": 500000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 200001 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 0.6 - } + "input": 0.275, + "output": 1.1 } }, - "spacexai/grok-voice-think-fast-2.0": { - "id": "spacexai/grok-voice-think-fast-2.0", - "name": "Grok Voice Think Fast 2.0", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "grok", + "qwen3.5-9b": { + "id": "qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text", - "audio" - ] + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "spacexai/grok-imagine-image-2.0": { - "id": "spacexai/grok-imagine-image-2.0", - "name": "Grok Imagine Image 2.0", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-08-07", - "last_updated": "2026-08-07", + "structured_output": true, + "temperature": true, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "output": 32000 + }, + "cost": { + "input": 0.16, + "output": 0.48 } }, - "spacexai/grok-4.20-multi-agent-beta": { - "id": "spacexai/grok-4.20-multi-agent-beta", - "name": "Grok 4.20 Multi Agent Beta", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -176860,233 +166965,208 @@ }, "open_weights": false, "limit": { - "context": 2000000, - "output": 2000000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200001 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 3, + "output": 15 } }, - "spacexai/grok-4.20-non-reasoning-beta": { - "id": "spacexai/grok-4.20-non-reasoning-beta", - "name": "Grok 4.20 Beta Non-Reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, + "nemotron-3-ultra": { + "id": "nemotron-3-ultra", + "name": "Nemotron 3 Ultra 550B A55B", + "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", + "family": "nemotron", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-06-04", + "last_updated": "2026-06-04", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 131072, + "output": 32000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.4, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200001 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.1, + "output": 0.1 } }, - "spacexai/grok-build-0.1": { - "id": "spacexai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Grok coding model for agentic engineering, edits, and codebase workflows", - "family": "grok-build", + "kimi-k2.7-code-nitro": { + "id": "kimi-k2.7-code-nitro", + "name": "Kimi K2.7 Code Nitro", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 256000, - "output": 256000 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 4, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200001 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 4, - "cache_read": 0.4 - } + "input": 0.275, + "output": 1.1 } - }, - "spacexai/grok-4.1-fast-non-reasoning": { - "id": "spacexai/grok-4.1-fast-non-reasoning", - "name": "Grok 4.1 Fast Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": false, + } + } + }, + "aihubmix": { + "id": "aihubmix", + "env": [ + "AIHUBMIX_API_KEY" + ], + "npm": "@aihubmix/ai-sdk-provider", + "name": "AIHubMix", + "doc": "https://docs.aihubmix.com", + "models": { + "coding-minimax-m2.7": { + "id": "coding-minimax-m2.7", + "name": "Coding MiniMax M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 1000000 + "context": 204800, + "output": 128100 }, "cost": { "input": 0.2, - "output": 0.5, - "cache_read": 0.05 + "output": 0.2 } }, - "spacexai/grok-4.20-reasoning-beta": { - "id": "spacexai/grok-4.20-reasoning-beta", - "name": "Grok 4.20 Beta Reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, + "alicloud-glm-5.1": { + "id": "alicloud-glm-5.1", + "name": "GLM-5.1 (Alibaba Cloud)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 2000000, - "output": 2000000 + "context": 200000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200001 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.84, + "output": 3.38, + "cache_read": 0.169, + "cache_write": 1.05625 } }, - "spacexai/grok-4.6": { - "id": "spacexai/grok-4.6", - "name": "Grok 4.6", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "claude-sonnet-4-6-think": { + "id": "claude-sonnet-4-6-think", + "name": "Claude Sonnet 4.6 Thinking", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -177094,296 +167174,377 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, "tiers": [ { - "input": 4, - "output": 12, - "cache_read": 1, + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, "tier": { "type": "context", - "size": 200001 + "size": 200000 } } ], "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 } } }, - "prodia/flux-fast-schnell": { - "id": "prodia/flux-fast-schnell", - "name": "Flux Schnell", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-08-02", - "last_updated": "2026-06-08", + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 512, - "output": 0 - } - }, - "recraft/recraft-v2": { - "id": "recraft/recraft-v2", - "name": "Recraft V2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-03-13", - "last_updated": "2024-03", - "modalities": { - "input": [ "text" - ], - "output": [ - "image" ] }, "open_weights": false, "limit": { - "context": 512, - "output": 0 - } - }, - "recraft/recraft-v4-pro": { - "id": "recraft/recraft-v4-pro", - "name": "Recraft V4 Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-02-17", - "last_updated": "2026-02-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] + "context": 1048576, + "output": 65536 }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "cost": { + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "cache_write": 1 } }, - "recraft/recraft-v4.1-utility": { - "id": "recraft/recraft-v4.1-utility", - "name": "Recraft V4.1 Utility", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", + "alicloud-deepseek-v4-flash": { + "id": "alicloud-deepseek-v4-flash", + "name": "DeepSeek V4 Flash (Alibaba Cloud)", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-05-14", - "last_updated": "2026-05-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "recraft/recraft-v4.1-pro": { - "id": "recraft/recraft-v4.1-pro", - "name": "Recraft V4.1 Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, - "reasoning": false, - "tool_call": false, + "structured_output": true, "temperature": true, - "release_date": "2026-05-14", - "last_updated": "2026-05-14", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" ], "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "recraft/recraft-v3": { - "id": "recraft/recraft-v3", - "name": "Recraft V3", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-10-30", - "last_updated": "2024-10", - "modalities": { - "input": [ "text" - ], - "output": [ - "image" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 512, - "output": 0 + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "recraft/recraft-v4.1": { - "id": "recraft/recraft-v4.1", - "name": "Recraft V4.1", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", + "xiaomi-mimo-v2.5-pro": { + "id": "xiaomi-mimo-v2.5-pro", + "name": "Xiaomi MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo-v2.5-pro", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-05-14", - "last_updated": "2026-05-14", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 1.1, + "output": 3.3, + "cache_read": 0.22, + "tiers": [ + { + "input": 2.2, + "output": 6.6, + "cache_read": 0.44, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 2.2, + "output": 6.6, + "cache_read": 0.44 + } } }, - "recraft/recraft-v4.1-utility-pro": { - "id": "recraft/recraft-v4.1-utility-pro", - "name": "Recraft V4.1 Utility Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", - "attachment": false, - "reasoning": false, - "tool_call": false, + "doubao-seed-2-0-code-preview": { + "id": "doubao-seed-2-0-code-preview", + "name": "Doubao Seed 2.0 Code Preview", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-05-14", - "last_updated": "2026-05-14", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "output": 128000 + }, + "cost": { + "input": 0.48, + "output": 2.41, + "cache_read": 0.09644, + "tiers": [ + { + "input": 0.72, + "output": 3.62, + "cache_read": 0.144656, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 1.45, + "output": 7.23, + "cache_read": 0.28932, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "recraft/recraft-v4": { - "id": "recraft/recraft-v4", - "name": "Recraft V4", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "recraft", + "coding-xiaomi-mimo-v2.5-pro": { + "id": "coding-xiaomi-mimo-v2.5-pro", + "name": "Coding Xiaomi MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo-v2.5-pro", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2026-02-17", - "last_updated": "2026-02-17", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1048576, + "output": 131072 + }, + "cost": { + "input": 0.2, + "output": 0.6, + "cache_read": 0.04, + "tiers": [ + { + "input": 0.4, + "output": 1.2, + "cache_read": 0.08, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.4, + "output": 1.2, + "cache_read": 0.08 + } } }, - "poolside/laguna-s-2.1-free": { - "id": "poolside/laguna-s-2.1-free", - "name": "Laguna S 2.1 Free", - "description": "Free provider route for experiments, demos, and cost-sensitive chat workloads", - "family": "laguna", - "attachment": false, + "doubao-seed-2-0-pro": { + "id": "doubao-seed-2-0-pro", + "name": "Doubao Seed 2.0 Pro", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 256000, - "output": 32768 + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 0.48, + "output": 2.41, + "cache_read": 0.09644, + "tiers": [ + { + "input": 0.72, + "output": 3.62, + "cache_read": 0.144656, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 1.45, + "output": 7.23, + "cache_read": 0.28932, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "poolside/laguna-s-2.1": { - "id": "poolside/laguna-s-2.1", - "name": "Laguna S 2.1", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", + "deep-deepseek-v4-flash": { + "id": "deep-deepseek-v4-flash", + "name": "DeepSeek V4 Flash (DeepSeek)", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -177392,10 +167553,14 @@ } ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -177407,20 +167572,20 @@ "open_weights": true, "limit": { "context": 1000000, - "output": 131072 + "output": 384000 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.01 + "input": 0.154, + "output": 0.308, + "cache_read": 0.0308 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -177428,20 +167593,21 @@ }, { "type": "budget_tokens", - "min": 1024 + "max": 262144 } ], "tool_call": true, - "interleaved": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "knowledge": "2025-04", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -177449,27 +167615,24 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 991000, "output": 64000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.282, + "output": 1.128, + "cache_read": 0.0564, + "cache_write": 0.3525 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ @@ -177477,22 +167640,20 @@ "medium", "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, - "interleaved": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-05", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -177501,47 +167662,58 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "grok-4.3": { + "id": "grok-4.3", + "name": "Grok 4.3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", + "family": "grok", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -177550,79 +167722,57 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 128000 + "output": 1000000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-3-haiku": { - "id": "anthropic/claude-3-haiku", - "name": "Claude Haiku 3", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-haiku", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-03-13", - "last_updated": "2024-03-13", - "modalities": { - "input": [ - "text", - "image" + "input": 1.25, + "output": 2.5, + "cache_read": 0.2, + "tiers": [ + { + "input": 2.5, + "output": 5, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 4096 - }, - "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.03, - "cache_write": 0.3 + "context_over_200k": { + "input": 2.5, + "output": 5, + "cache_read": 0.4 + } } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "family": "qwen", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" }, { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens", + "max": 262144 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -177630,44 +167780,53 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 991000, + "output": 64000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.69, + "output": 5.07, + "cache_read": 0.169, + "cache_write": 2.1125 } }, - "anthropic/claude-opus-5.5-fast": { - "id": "anthropic/claude-opus-5.5-fast", - "name": "Claude Opus 5.5 (Fast)", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2025-03-20", + "last_updated": "2025-06-05", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -177676,68 +167835,69 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 8, - "output": 40, - "cache_read": 0.4, - "cache_write": 10 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "xiaomi-mimo-v2.5-free": { + "id": "xiaomi-mimo-v2.5-free", + "name": "Xiaomi MiMo-V2.5 (free)", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo-v2.5", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, - "interleaved": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5", + "claude-opus-4-7-think": { + "id": "claude-opus-4-7-think", + "name": "Claude Opus 4.7 Thinking", "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ @@ -177750,10 +167910,14 @@ } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -177773,209 +167937,167 @@ "input": 5, "output": 25, "cache_read": 0.5, - "cache_write": 6.25 + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "anthropic/claude-fable-5.1": { - "id": "anthropic/claude-fable-5.1", - "name": "Claude Fable 5.1", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, + "coding-glm-5.1-free": { + "id": "coding-glm-5.1-free", + "name": "Coding GLM 5.1 (free)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm-free", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-04-11", + "last_updated": "2026-04-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 200000, "output": 128000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 0, + "output": 0 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, + "xiaomi-mimo-v2.5-pro-free": { + "id": "xiaomi-mimo-v2.5-pro-free", + "name": "Xiaomi MiMo-V2.5-Pro (free)", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo-v2.5-pro", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "anthropic/claude-opus-4.8-fast": { - "id": "anthropic/claude-opus-4.8-fast", - "name": "Claude Opus 4.8 (Fast)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "xiaomi-mimo-v2.5": { + "id": "xiaomi-mimo-v2.5", + "name": "Xiaomi MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo-v2.5", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, + "input": 0.44, + "output": 2.2, + "cache_read": 0.088, "tiers": [ { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, + "input": 0.88, + "output": 4.4, + "cache_read": 0.176, "tier": { "type": "context", - "size": 200001 + "size": 256000 } } ], "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 + "input": 0.88, + "output": 4.4, + "cache_read": 0.176 } } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -177988,15 +168110,18 @@ "low", "medium", "high", - "xhigh" + "xhigh", + "max" ] } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": false, "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -178013,47 +168138,99 @@ "output": 128000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "max" ] - }, + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 1000000, + "output": 128000 + }, + "cost": { + "input": 1.1268, + "output": 3.9438, + "cache_read": 0.2817 + } + }, + "qwen3.6-flash": { + "id": "qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.6", + "attachment": true, + "reasoning": true, + "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "toggle" } ], "tool_call": true, - "interleaved": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "knowledge": "2025-04", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -178061,85 +168238,97 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 991000, + "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, + "input": 0.17, + "output": 1.01, + "cache_read": 0.0169, + "cache_write": 0.21125, "tiers": [ { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, + "input": 0.68, + "output": 4.06, + "cache_read": 0.0676, + "cache_write": 0.845, "tier": { "type": "context", - "size": 200000 + "size": 256000 } } ], "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 + "input": 0.68, + "output": 4.06, + "cache_read": 0.0676, + "cache_write": 0.845 } } }, - "anthropic/claude-opus-5.5": { - "id": "anthropic/claude-opus-5.5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", + "coding-xiaomi-mimo-v2.5": { + "id": "coding-xiaomi-mimo-v2.5", + "name": "Coding Xiaomi MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo-v2.5", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-05-13", "modalities": { "input": [ "text", "image", - "pdf" + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 + "input": 0.08, + "output": 0.4, + "cache_read": 0.016, + "tiers": [ + { + "input": 0.16, + "output": 0.8, + "cache_read": 0.032, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 0.16, + "output": 0.8, + "cache_read": 0.032 + } } }, - "anthropic/claude-opus-5-fast": { - "id": "anthropic/claude-opus-5-fast", - "name": "Claude Opus 5 (Fast)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-5.1-codex": { + "id": "gpt-5.1-codex", + "name": "GPT-5.1 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -178148,22 +168337,20 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -178171,34 +168358,48 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "anthropic/claude-opus-4": { - "id": "anthropic/claude-opus-4", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "interleaved": true, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -178210,35 +168411,41 @@ "output": 32000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "gemini-3.1-pro-preview-customtools": { + "id": "gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ @@ -178247,79 +168454,65 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "cohere/rerank-v4-fast": { - "id": "cohere/rerank-v4-fast", - "name": "Cohere Rerank 4 Fast", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 - } - }, - "cohere/rerank-v4-pro": { - "id": "cohere/rerank-v4-pro", - "name": "Cohere Rerank 4 Pro", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text" + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "cohere/command-a": { - "id": "cohere/command-a", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command", - "attachment": false, - "reasoning": false, + "doubao-seed-2-0-mini-260428": { + "id": "doubao-seed-2-0-mini-260428", + "name": "Doubao Seed 2.0 Mini 260428", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -178328,109 +168521,124 @@ "open_weights": false, "limit": { "context": 256000, - "output": 8000 + "output": 128000 }, "cost": { - "input": 2.5, - "output": 10 - } - }, - "cohere/embed-v4.0": { - "id": "cohere/embed-v4.0", - "name": "Embed v4.0", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "cohere-embed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 1536 - } - }, - "cohere/rerank-v3.5": { - "id": "cohere/rerank-v3.5", - "name": "Cohere Rerank 3.5", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-12-02", - "last_updated": "2024-12-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" + "input": 0.03, + "output": 0.28, + "cache_read": 0.00564, + "input_audio": 0.423, + "tiers": [ + { + "input": 0.06, + "output": 0.56, + "cache_read": 0.01128, + "input_audio": 0.846, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.11, + "output": 1.13, + "cache_read": 0.02256, + "input_audio": 1.692, + "tier": { + "type": "context", + "size": 128000 + } + } ] - }, - "open_weights": false, - "limit": { - "context": 4096, - "output": 4096 } }, - "deepseek/deepseek-v4-flash-0731": { - "id": "deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "doubao-seed-2-0-lite-260428": { + "id": "doubao-seed-2-0-lite-260428", + "name": "Doubao Seed 2.0 Lite 260428", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "seed", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 256000, + "output": 128000 }, "cost": { - "input": 0.076, - "output": 0.153, - "cache_read": 0.014 + "input": 0.08, + "output": 0.51, + "cache_read": 0.01692, + "input_audio": 1.269, + "tiers": [ + { + "input": 0.13, + "output": 0.76, + "cache_read": 0.02536, + "input_audio": 1.902, + "tier": { + "type": "context", + "size": 32000 + } + }, + { + "input": 0.25, + "output": 1.52, + "cache_read": 0.05072, + "input_audio": 3.804, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "deepseek/deepseek-v4.1-flash": { - "id": "deepseek/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "low", + "medium", "high", "xhigh" ] @@ -178438,10 +168646,10 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", @@ -178451,31 +168659,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.007 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "deepseek/deepseek-v4-flash-vision-exp": { - "id": "deepseek/deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "low", + "medium", "high", "xhigh" ] @@ -178483,46 +168691,48 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "deepseek/deepseek-v3.1-terminus": { - "id": "deepseek/deepseek-v3.1-terminus", - "name": "DeepSeek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "minimax-m2.7": { + "id": "minimax-m2.7", + "name": "MiniMax M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-09-22", - "last_updated": "2025-09-22", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -178533,21 +168743,22 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 204800, + "output": 128000 }, "cost": { - "input": 0.27, - "output": 1, - "cache_read": 0.135 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "deepseek/deepseek-v4-pro-0813": { - "id": "deepseek/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", - "attachment": false, + "claude-opus-4-6-think": { + "id": "claude-opus-4-6-think", + "name": "Claude Opus 4.6 Thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -178556,52 +168767,93 @@ { "type": "effort", "values": [ + "low", + "medium", "high", - "xhigh" + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "output": 128000 }, "cost": { - "input": 0.66, - "output": 1.98, - "cache_read": 0.066 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "deepseek/deepseek-v3.2-thinking": { - "id": "deepseek/deepseek-v3.2-thinking", - "name": "DeepSeek V3.2 Thinking", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek-thinking", - "attachment": false, + "gpt-5.1-codex-mini": { + "id": "gpt-5.1-codex-mini", + "name": "GPT-5.1 Codex mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -178609,20 +168861,22 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8000 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.62, - "output": 1.85 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "deepseek/deepseek-v3.1": { - "id": "deepseek/deepseek-v3.1", - "name": "DeepSeek-V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "kimi-k2.5": { + "id": "kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -178630,12 +168884,19 @@ } ], "tool_call": true, - "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -178643,40 +168904,35 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 0.95, - "cache_read": 0.13 + "input": 0.6, + "output": 3, + "cache_read": 0.1 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "coding-glm-5.1": { + "id": "coding-glm-5.1", + "name": "Coding GLM 5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-04-11", + "last_updated": "2026-04-11", "modalities": { "input": [ "text" @@ -178687,28 +168943,31 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.66, - "output": 1.98, - "cache_read": 0.022 + "input": 0.06, + "output": 0.22, + "cache_read": 0.013 } }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "coding-minimax-m2.7-free": { + "id": "coding-minimax-m2.7-free", + "name": "Coding MiniMax M2.7 (Free)", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax-free", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -178719,28 +168978,28 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 8000 + "context": 204800, + "output": 128100 }, "cost": { - "input": 0.62, - "output": 1.85 + "input": 0, + "output": 0 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "low", + "medium", "high", "xhigh" ] @@ -178748,94 +169007,76 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.13, - "output": 0.26, - "cache_read": 0.028 - } - }, - "deepseek/deepseek-r1": { - "id": "deepseek/deepseek-r1", - "name": "DeepSeek-R1", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] + "context": 1050000, + "input": 922000, + "output": 128000 }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 5, + "output": 30, + "cache_read": 0.5 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 1.35, - "output": 5.4 - } - }, - "quiverai/arrow-1.1": { - "id": "quiverai/arrow-1.1", - "name": "Arrow 1.1", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text" + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 131072 + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "quiverai/arrow-2": { - "id": "quiverai/arrow-2", - "name": "Arrow 2", - "description": "Fast SVG generation model for creation, vectorization, editing, and animation", + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high", @@ -178844,359 +169085,120 @@ } ], "tool_call": true, - "release_date": "2026-09-16", - "last_updated": "2026-09-16", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "quiverai/arrow-2-telos": { - "id": "quiverai/arrow-2-telos", - "name": "Arrow 2 Telos", - "description": "High-fidelity SVG generation model for complex vector work and long-context refinement", + "kimi-k2.6": { + "id": "kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, - "release_date": "2026-09-16", - "last_updated": "2026-09-16", + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "text", - "image" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 32768 }, "cost": { - "input": 6, - "output": 30, - "cache_read": 0.6, - "cache_write": 7.5 + "input": 0.95, + "output": 4, + "cache_read": 0.16 } }, - "bfl/flux-kontext-max": { - "id": "bfl/flux-kontext-max", - "name": "FLUX.1 Kontext Max", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-05-29", - "last_updated": "2025-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 512, - "output": 0 - } - }, - "bfl/flux-pro-1.1-ultra": { - "id": "bfl/flux-pro-1.1-ultra", - "name": "FLUX1.1 [pro] Ultra", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-11-01", - "last_updated": "2024-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 512, - "output": 0 - } - }, - "bfl/flux-3-video": { - "id": "bfl/flux-3-video", - "name": "Flux 3", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-08-04", - "last_updated": "2026-08-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "video" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "bfl/flux-2-flex": { - "id": "bfl/flux-2-flex", - "name": "FLUX.2 [flex]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-11-25", - "last_updated": "2026-06-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "bfl/flux-2-klein-4b": { - "id": "bfl/flux-2-klein-4b", - "name": "FLUX.2 [klein] 4B", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-15", - "last_updated": "2026-06-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "bfl/flux-pro-1.0-fill": { - "id": "bfl/flux-pro-1.0-fill", - "name": "FLUX.1 Fill [pro]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-10-01", - "last_updated": "2024-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 512, - "output": 0 - } - }, - "bfl/flux-2-klein-9b": { - "id": "bfl/flux-2-klein-9b", - "name": "FLUX.2 [klein] 9B", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-15", - "last_updated": "2026-06-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "bfl/flux-2-pro": { - "id": "bfl/flux-2-pro", - "name": "FLUX.2 [pro]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-11-25", - "last_updated": "2026-06-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 67300, - "output": 67300 - } - }, - "bfl/flux-2-max": { - "id": "bfl/flux-2-max", - "name": "FLUX.2 [max]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2026-06-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 67300, - "output": 67300 - } - }, - "bfl/flux-pro-1.1": { - "id": "bfl/flux-pro-1.1", - "name": "FLUX1.1 [pro]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-10-02", - "last_updated": "2024-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 512, - "output": 0 - } - }, - "bfl/flux-kontext-pro": { - "id": "bfl/flux-kontext-pro", - "name": "FLUX.1 Kontext Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-05-29", - "last_updated": "2025-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 512, - "output": 0 - } - }, - "tencent/hy-mt2-lite": { - "id": "tencent/hy-mt2-lite", - "name": "Tencent Hy-MT2-Lite", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -179204,67 +169206,68 @@ }, "open_weights": false, "limit": { - "context": 8000, - "output": 4000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.044, - "output": 0.177 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "tencent/hy3": { - "id": "tencent/hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", "low", - "high" + "medium", + "high", + "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.14, - "output": 0.58, - "cache_read": 0.035 - } - }, - "tencent/hy-mt2-plus": { - "id": "tencent/hy-mt2-plus", - "name": "Tencent Hy-MT2-Plus", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -179272,62 +169275,55 @@ }, "open_weights": false, "limit": { - "context": 8000, - "output": 4000 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.074, - "output": 0.295 - } - }, - "tencent/hy-mt2-pro": { - "id": "tencent/hy-mt2-pro", - "name": "Tencent Hy-MT2-Pro", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8000, - "output": 4000 - }, - "cost": { - "input": 0.074, - "output": 0.295 + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "tencent/hy4-preview": { - "id": "tencent/hy4-preview", - "name": "Tencent Hy4 Preview", - "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", - "family": "Hy", + "deep-deepseek-v4-pro": { + "id": "deep-deepseek-v4-pro", + "name": "DeepSeek V4 Pro (DeepSeek)", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -179338,27 +169334,40 @@ }, "open_weights": true, "limit": { - "context": 1024000, - "output": 64000 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.834, - "output": 2.501, - "cache_read": 0.042 + "input": 0.478, + "output": 0.956, + "cache_read": 0.004302 } }, - "inference-net/schematron-v2-small": { - "id": "inference-net/schematron-v2-small", - "name": "Schematron V2 Small", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM 5 Vision Turbo", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glmv", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-05-09", + "last_updated": "2026-05-09", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -179366,24 +169375,35 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.05, - "output": 0.23, - "cache_read": 0.05 + "input": 0.7042, + "output": 3.09848, + "cache_read": 0.169008 } }, - "inference-net/schematron-v2-turbo": { - "id": "inference-net/schematron-v2-turbo", - "name": "Schematron V2 Turbo", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "zai-glm-5.1": { + "id": "zai-glm-5.1", + "name": "GLM-5.1 (Z.ai)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ "text" @@ -179392,42 +169412,49 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.03, - "output": 0.15, - "cache_read": 0.03 + "input": 0.845, + "output": 3.38, + "cache_read": 0.183112 } }, - "thinkingmachines/inkling-small": { - "id": "thinkingmachines/inkling-small", - "name": "Inkling Small", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ling", + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", "high", - "xhigh", "max" ] + }, + { + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, + "interleaved": true, + "structured_output": true, "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -179438,123 +169465,115 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 1000000 + "output": 64000 }, "cost": { - "input": 0.5, - "output": 1.2, - "cache_read": 0.1 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, + "tiers": [ + { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "thinkingmachines/inkling": { - "id": "thinkingmachines/inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.17 - } - }, - "perplexity/pplx-embed-v1-4b": { - "id": "perplexity/pplx-embed-v1-4b", - "name": "Embed v1 4b", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", - "modalities": { - "input": [ - "text" + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "tiers": [ + { + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "tier": { + "type": "context", + "size": 200000 + } + } ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 0 + "context_over_200k": { + "input": 0.5, + "output": 3, + "cache_read": 0.05 + } } }, - "perplexity/sonar-pro": { - "id": "perplexity/sonar-pro", - "name": "Sonar Pro", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", - "attachment": true, - "reasoning": false, + "coding-minimax-m2.7-highspeed": { + "id": "coding-minimax-m2.7-highspeed", + "name": "Coding MiniMax M2.7 Highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] + "interleaved": { + "field": "reasoning_content" }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8000 - } - }, - "perplexity/pplx-embed-v1-0.6b": { - "id": "perplexity/pplx-embed-v1-0.6b", - "name": "Embed v1 0.6b", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "v0", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "structured_output": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -179563,62 +169582,46 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32000, - "output": 0 - } - }, - "perplexity/sonar": { - "id": "perplexity/sonar", - "name": "Sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-02", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] + "context": 204800, + "output": 128100 }, - "open_weights": false, - "limit": { - "context": 127000, - "output": 8000 + "cost": { + "input": 0.2, + "output": 0.2 } }, - "perplexity/sonar-reasoning-pro": { - "id": "perplexity/sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded reasoning model for multi-step research and cited answers", - "family": "sonar-reasoning", - "attachment": false, + "claude-opus-4-8-think": { + "id": "claude-opus-4-8-think", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], - "tool_call": false, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -179630,40 +169633,40 @@ }, "open_weights": false, "limit": { - "context": 127000, - "output": 8000 + "context": 200000, + "output": 32000 + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "meta/muse-spark-1.3": { - "id": "meta/muse-spark-1.3", - "name": "Muse Spark 1.3", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "muse", - "attachment": true, + "qwen3.6-max-preview": { + "id": "qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "qwen3.6", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "knowledge": "2025-04", + "release_date": "2026-05-09", + "last_updated": "2026-05-09", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -179671,20 +169674,33 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 240000, + "output": 64000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 1.27, + "output": 7.61, + "cache_read": 0.1268, + "cache_write": 1.585, + "tiers": [ + { + "input": 2.11, + "output": 12.67, + "cache_read": 0.2112, + "cache_write": 2.64, + "tier": { + "type": "context", + "size": 128000 + } + } + ] } }, - "meta/muse-glimmer-30b": { - "id": "meta/muse-glimmer-30b", - "name": "Muse Glimmer 30B", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -179700,59 +169716,58 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.35, - "output": 1.5, - "cache_read": 0.04 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "meta/muse-spark-1.1": { - "id": "meta/muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "muse", - "attachment": false, + "qwen3.6-plus": { + "id": "qwen3.6-plus", + "name": "Qwen3.6 Plus", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "family": "qwen3.6", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", + "knowledge": "2025-04", + "release_date": "2026-05-09", + "last_updated": "2026-05-09", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -179760,27 +169775,55 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 991000, + "output": 64000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 0.28, + "output": 1.69, + "cache_read": 0.0282, + "cache_write": 0.3525, + "tiers": [ + { + "input": 1.13, + "output": 6.77, + "cache_read": 0.1128, + "cache_write": 1.41, + "tier": { + "type": "context", + "size": 256000 + } + } + ], + "context_over_200k": { + "input": 1.13, + "output": 6.77, + "cache_read": 0.1128, + "cache_write": 1.41 + } } }, - "meta/llama-3.1-8b": { - "id": "meta/llama-3.1-8b", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "alicloud-deepseek-v4-pro": { + "id": "alicloud-deepseek-v4-pro", + "name": "DeepSeek V4 Pro (Alibaba Cloud)", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -179789,31 +169832,45 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 8192 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.22, - "output": 0.22 + "input": 1.69, + "output": 3.38, + "cache_read": 0.13 } }, - "meta/llama-3.1-70b": { - "id": "meta/llama-3.1-70b", - "name": "Llama 3.1 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, + "gpt-5.1": { + "id": "gpt-5.1", + "name": "GPT-5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -179821,27 +169878,41 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.72, - "output": 0.72 + "input": 1.25, + "output": 10, + "cache_read": 0.13 } }, - "meta/muse-spark-1.2": { - "id": "meta/muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "muse", + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -179854,62 +169925,109 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 1050000, + "input": 922000, + "output": 128000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 12.5, + "output": 75, + "cache_read": 1.25 + }, + "provider": { + "body": { + "service_tier": "priority" + } + } + } + } }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } - }, - "meta/muse-image-1.0": { - "id": "meta/muse-image-1.0", - "name": "Muse Image 1.0", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "muse", - "attachment": true, + } + } + }, + "google-vertex": { + "id": "google-vertex", + "env": [ + "GOOGLE_VERTEX_PROJECT", + "GOOGLE_VERTEX_LOCATION", + "GOOGLE_APPLICATION_CREDENTIALS" + ], + "npm": "@ai-sdk/google-vertex", + "name": "Vertex", + "doc": "https://cloud.google.com/vertex-ai/generative-ai/docs/models", + "models": { + "gemini-2.5-pro-tts": { + "id": "gemini-2.5-pro-tts", + "name": "Gemini 2.5 Pro TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gemini-pro", + "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-09-30", + "last_updated": "2025-12-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "audio" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 32768, + "output": 16384 + }, + "cost": { + "input": 1, + "output": 20 } }, - "meta/muse-spark-1.2-contributor": { - "id": "meta/muse-spark-1.2-contributor", - "name": "Muse Spark 1.2 Contributor", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "muse", + "claude-haiku-4-5@20251001": { + "id": "claude-haiku-4-5@20251001", + "name": "Claude Haiku 4.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", @@ -179922,20 +170040,24 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "meta/muse-spark-1.3-contributor": { - "id": "meta/muse-spark-1.3-contributor", - "name": "Muse Spark 1.3 Contributor", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "muse", + "gemini-3.1-flash-lite": { + "id": "gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -179945,18 +170067,22 @@ "minimal", "low", "medium", - "high", - "xhigh" + "high" ] } ], "tool_call": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-07", + "last_updated": "2026-05-07", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -179966,158 +170092,215 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 1048576 + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "meta/llama-3.3-70b": { - "id": "meta/llama-3.3-70b", - "name": "Llama-3.3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "meta/llama-4-maverick": { - "id": "meta/llama-4-maverick", - "name": "Llama-4-Maverick-17B-128E-Instruct-FP8", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": true, + "gemini-2.5-flash-tts": { + "id": "gemini-2.5-flash-tts", + "name": "Gemini 2.5 Flash TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "gemini-flash", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "tool_call": false, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2025-09-30", + "last_updated": "2025-12-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 32768, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.5, + "output": 10 } }, - "meta/llama-4-scout": { - "id": "meta/llama-4-scout", - "name": "Llama-4-Scout-17B-16E-Instruct-FP8", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image" + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 2.5, + "cache_read": 0.075, + "cache_write": 0.383 } }, - "arcee-ai/trinity-large-thinking": { - "id": "arcee-ai/trinity-large-thinking", - "name": "Trinity Large Thinking", - "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", - "family": "trinity", - "attachment": false, + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-03", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262100, - "output": 80000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 0.8999999999999999 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 } }, - "alibaba/qwen3.5-flash": { - "id": "alibaba/qwen3.5-flash", - "name": "Qwen 3.5 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-opus-4@20250514": { + "id": "claude-opus-4@20250514", + "name": "Claude Opus 4", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "budget_tokens", - "min": 1, - "max": 81920 + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", @@ -180130,41 +170313,43 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0.125 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "alibaba/qwen-3-32b": { - "id": "alibaba/qwen-3-32b", - "name": "Qwen 3.32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "claude-opus-4-1@20250805": { + "id": "claude-opus-4-1@20250805", + "name": "Claude Opus 4.1", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "budget_tokens", - "min": 1, - "max": 38912 + "min": 1024 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-04", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -180172,94 +170357,118 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 8192 + "context": 200000, + "output": 32000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" }, "cost": { - "input": 0.16, - "output": 0.64 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "alibaba/wan-v2.6-t2v": { - "id": "alibaba/wan-v2.6-t2v", - "name": "Wan v2.6 Text-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", + "gemini-embedding-001": { + "id": "gemini-embedding-001", + "name": "Gemini Embedding 001", + "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", + "family": "gemini", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "temperature": false, + "knowledge": "2025-05", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, + "context": 2048, + "output": 1 + }, + "cost": { + "input": 0.15, "output": 0 } }, - "alibaba/qwen3-vl-235b-a22b-instruct": { - "id": "alibaba/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-opus-4-5@20251101": { + "id": "claude-opus-4-5@20251101", + "name": "Claude Opus 4.5", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2025-05", + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 129024 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" }, "cost": { - "input": 0.4, - "output": 1.6 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "alibaba/qwen3.7-max": { - "id": "alibaba/qwen3.7-max", - "name": "Qwen 3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "claude-3-5-haiku@20241022": { + "id": "claude-3-5-haiku@20241022", + "name": "Claude Haiku 3.5", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", + "family": "claude-haiku", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 - } - ], + "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2024-07-31", + "release_date": "2024-10-22", + "last_updated": "2024-10-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -180267,183 +170476,223 @@ }, "open_weights": false, "limit": { - "context": 991000, - "output": 64000 + "context": 200000, + "output": 8192 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 0.8, + "output": 4, + "cache_read": 0.08, + "cache_write": 1 } }, - "alibaba/qwen3.8-27b": { - "id": "alibaba/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-3.1-pro-preview-customtools": { + "id": "gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "xhigh" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1, - "cache_write": 0.625 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "alibaba/qwen3-vl-thinking": { - "id": "alibaba/qwen3-vl-thinking", - "name": "Qwen3 VL Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-flash-lite-latest": { + "id": "gemini-flash-lite-latest", + "name": "Gemini Flash-Lite Latest", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-23", - "last_updated": "2025-09-24", + "knowledge": "2025-01", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", "image", + "audio", + "video", "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 4 + "input": 0.1, + "output": 0.4, + "cache_read": 0.025 } }, - "alibaba/qwen3.8-2.4t-a95b": { - "id": "alibaba/qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, + "claude-sonnet-4@20250514": { + "id": "claude-sonnet-4@20250514", + "name": "Claude Sonnet 4", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] + "type": "budget_tokens", + "min": 1024 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 128000 + "context": 200000, + "output": 64000 + }, + "status": "deprecated", + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "alibaba/qwen3-next-80b-a3b-thinking": { - "id": "alibaba/qwen3-next-80b-a3b-thinking", - "name": "Qwen3 Next 80B A3B Thinking", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", - "attachment": false, + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash-Lite", + "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", + "family": "gemini-flash-lite", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "budget_tokens", - "min": 1 + "min": 512, + "max": 24576 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 1.2 + "input": 0.1, + "output": 0.4, + "cache_read": 0.01, + "input_audio": 0.3 } }, - "alibaba/qwen3.8-max": { - "id": "alibaba/qwen3.8-max", - "name": "Qwen 3.8 Max", - "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", - "family": "qwen", + "claude-opus-4-7@default": { + "id": "claude-opus-4-7@default", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -180452,23 +170701,22 @@ "values": [ "low", "medium", - "xhigh" + "high", + "xhigh", + "max" ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 } ], "tool_call": true, - "temperature": true, - "release_date": "2026-07-19", - "last_updated": "2026-07-19", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -180476,94 +170724,174 @@ }, "open_weights": false, "limit": { - "context": 262144, + "context": 1000000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" + }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "alibaba/qwen3-coder-next": { - "id": "alibaba/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 1.2 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "alibaba/qwen3-vl-instruct": { - "id": "alibaba/qwen3-vl-instruct", - "name": "Qwen3 VL Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-sonnet-4-5@20250929": { + "id": "claude-sonnet-4-5@20250929", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-24", + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 129024 + "context": 200000, + "output": 64000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" }, "cost": { - "input": 0.4, - "output": 1.6 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "alibaba/qwen3-coder": { - "id": "alibaba/qwen3-coder", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, + "claude-sonnet-5@default": { + "id": "claude-sonnet-5@default", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-22", - "last_updated": "2025-04", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -180571,61 +170899,49 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" }, "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.3, - "tiers": [ - { - "input": 2.7, - "output": 13.5, - "cache_read": 0.54, - "tier": { - "type": "context", - "size": 32001 - } - }, - { - "input": 4.5, - "output": 22.5, - "cache_read": 0.9, - "tier": { - "type": "context", - "size": 128001 - } - } - ] + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "alibaba/qwen3.5-plus": { - "id": "alibaba/qwen3.5-plus", - "name": "Qwen 3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 81920 + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text", "image", + "video", + "audio", "pdf" ], "output": [ @@ -180634,155 +170950,162 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 2.5, - "cache_read": 0.04, - "cache_write": 0.5, - "tiers": [ - { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tier": { - "type": "context", - "size": 256001 - } - } - ], - "context_over_200k": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625 - } + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 } }, - "alibaba/qwen-3.6-max-preview": { - "id": "alibaba/qwen-3.6-max-preview", - "name": "Qwen 3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "claude-opus-4-6@default": { + "id": "claude-opus-4-6@default", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] }, { "type": "budget_tokens", - "min": 1, - "max": 131072 + "min": 1024 } ], "tool_call": true, "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-24", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 240000, - "output": 64000 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" }, "cost": { - "input": 1.3, - "output": 7.8, - "cache_read": 0.13, - "cache_write": 1.625, + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, "tiers": [ { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, "tier": { "type": "context", - "size": 128000 + "size": 200000 } } - ] - } - }, - "alibaba/wan-v2.6-i2v-flash": { - "id": "alibaba/wan-v2.6-i2v-flash", - "name": "Wan v2.6 Image-to-Video Flash", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", - "modalities": { - "input": [ - "text" ], - "output": [ - "video" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "alibaba/wan-v2.6-r2v-flash": { - "id": "alibaba/wan-v2.6-r2v-flash", - "name": "Wan v2.6 Reference-to-Video Flash", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, + "gemini-flash-latest": { + "id": "gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "knowledge": "2025-01", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 2.5, + "cache_read": 0.075, + "cache_write": 0.383 } }, - "alibaba/qwen-3-14b": { - "id": "alibaba/qwen-3-14b", - "name": "Qwen3-14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "claude-opus-4-8@default": { + "id": "claude-opus-4-8@default", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-04", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -180790,27 +171113,64 @@ }, "open_weights": false, "limit": { - "context": 40960, - "output": 16384 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" }, "cost": { - "input": 0.12, - "output": 0.24 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25, + "tiers": [ + { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 37.5, + "cache_read": 1, + "cache_write": 12.5 + } } }, - "alibaba/qwen3.7-flash": { - "id": "alibaba/qwen3.7-flash", - "name": "Qwen 3.7 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-sonnet-4-6@default": { + "id": "claude-sonnet-4-6@default", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + }, + { + "type": "budget_tokens", + "min": 1024 + } + ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", @@ -180823,53 +171183,68 @@ }, "open_weights": false, "limit": { - "context": 991000, - "output": 64000 + "context": 1000000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/google-vertex/anthropic" }, "cost": { - "input": 0.03, - "output": 0.13, - "cache_read": 0.006, - "cache_write": 0.038, + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75, "tiers": [ { - "input": 0.1, - "output": 0.4, - "cache_read": 0.02, - "cache_write": 0.125, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.2, - "output": 0.8, - "cache_read": 0.04, - "cache_write": 0.25, + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5, "tier": { "type": "context", - "size": 256000 + "size": 200000 } } - ] + ], + "context_over_200k": { + "input": 6, + "output": 22.5, + "cache_read": 0.6, + "cache_write": 7.5 + } } }, - "alibaba/qwen3-max": { - "id": "alibaba/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gemini-3.1-flash-lite-preview": { + "id": "gemini-3.1-flash-lite-preview", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -180877,46 +171252,34 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1048576, + "output": 65536 }, + "status": "deprecated", "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24, - "tiers": [ - { - "input": 2.4, - "output": 12, - "cache_read": 0.48, - "tier": { - "type": "context", - "size": 32001 - } - }, - { - "input": 3, - "output": 15, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 128001 - } - } - ] + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "alibaba/qwen3-embedding-0.6b": { - "id": "alibaba/qwen3-embedding-0.6b", - "name": "Qwen3 Embedding 0.6B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", + "moonshotai/kimi-k2-thinking-maas": { + "id": "moonshotai/kimi-k2-thinking-maas", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-11-14", - "last_updated": "2025-11-14", + "knowledge": "2024-08", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text" @@ -180925,29 +171288,41 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 262144, + "output": 262144 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.6, + "output": 2.5 } }, - "alibaba/qwen-3-30b": { - "id": "alibaba/qwen-3-30b", - "name": "Qwen3-30B-A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "openai/gpt-oss-120b-maas": { + "id": "openai/gpt-oss-120b-maas", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-04", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -180956,102 +171331,118 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 40960, - "output": 16384 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.12, - "output": 0.5 + "input": 0.09, + "output": 0.36 } }, - "alibaba/qwen3.8-omni-flash": { - "id": "alibaba/qwen3.8-omni-flash", - "name": "Qwen 3.8 Omni Flash", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": true, + "openai/gpt-oss-20b-maas": { + "id": "openai/gpt-oss-20b-maas", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "minimal", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", + "temperature": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016 + "input": 0.07, + "output": 0.25 } }, - "alibaba/qwen3-coder-30b-a3b": { - "id": "alibaba/qwen3-coder-30b-a3b", - "name": "Qwen 3 Coder 30B A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "zai-org/glm-4.7-maas": { + "id": "zai-org/glm-4.7-maas", + "name": "GLM-4.7", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-07-31", - "last_updated": "2025-04", + "release_date": "2026-01-06", + "last_updated": "2026-01-06", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 8192 + "context": 200000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.6, + "output": 2.2 } }, - "alibaba/qwen3-embedding-4b": { - "id": "alibaba/qwen3-embedding-4b", - "name": "Qwen3 Embedding 4B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", + "zai-org/glm-5-maas": { + "id": "zai-org/glm-5-maas", + "name": "GLM-5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ "text" @@ -181060,99 +171451,107 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 32768 + "context": 202752, + "output": 131072 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 1, + "output": 3.2, + "cache_read": 0.1 } }, - "alibaba/qwen3-max-preview": { - "id": "alibaba/qwen3-max-preview", - "name": "Qwen3 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", + "deepseek-ai/deepseek-v3.1-maas": { + "id": "deepseek-ai/deepseek-v3.1-maas", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-05", - "last_updated": "2025-09-23", + "release_date": "2025-08-28", + "last_updated": "2025-08-28", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, + "context": 163840, "output": 32768 }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24, - "tiers": [ - { - "input": 2.4, - "output": 12, - "cache_read": 0.48, - "tier": { - "type": "context", - "size": 32001 - } - }, - { - "input": 3, - "output": 15, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 128001 - } - } - ] + "input": 0.6, + "output": 1.7 } }, - "alibaba/wan-v2.6-i2v": { - "id": "alibaba/wan-v2.6-i2v", - "name": "Wan v2.6 Image-to-Video", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "deepseek-ai/deepseek-v3.2-maas": { + "id": "deepseek-ai/deepseek-v3.2-maas", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "release_date": "2025-12-17", + "last_updated": "2026-04-04", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 163840, + "output": 65536 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.56, + "output": 1.68, + "cache_read": 0.056 } }, - "alibaba/qwen3-next-80b-a3b-instruct": { - "id": "alibaba/qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct", + "qwen/qwen3-235b-a22b-instruct-2507-maas": { + "id": "qwen/qwen3-235b-a22b-instruct-2507-maas", + "name": "Qwen3 235B A22B Instruct", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2025-08-13", + "last_updated": "2025-08-13", "modalities": { "input": [ "text" @@ -181163,60 +171562,70 @@ }, "open_weights": true, "limit": { - "context": 262114, - "output": 262114 + "context": 262144, + "output": 16384 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" }, "cost": { - "input": 0.15, - "output": 1.2 + "input": 0.22, + "output": 0.88 } }, - "alibaba/wan-v2.5-t2v-preview": { - "id": "alibaba/wan-v2.5-t2v-preview", - "name": "Wan v2.5 Text-to-Video Preview", - "description": "Video model for prompt-guided generation, editing, and motion workflows", + "meta/llama-3.3-70b-instruct-maas": { + "id": "meta/llama-3.3-70b-instruct-maas", + "name": "Llama 3.3 70B Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": false, + "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-09-24", - "last_updated": "2025-09-24", + "knowledge": "2023-12", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "output": 8192 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" + }, + "cost": { + "input": 0.72, + "output": 0.72 } }, - "alibaba/qwen3-max-thinking": { - "id": "alibaba/qwen3-max-thinking", - "name": "Qwen 3 Max Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1, - "max": 81920 - } - ], + "meta/llama-4-maverick-17b-128e-instruct-maas": { + "id": "meta/llama-4-maverick-17b-128e-instruct-maas", + "name": "Llama 4 Maverick 17B 128E Instruct", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-23", - "last_updated": "2025-01", + "knowledge": "2024-08", + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -181224,50 +171633,44 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 65536 + "context": 524288, + "output": 8192 + }, + "provider": { + "npm": "@ai-sdk/openai-compatible", + "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" }, "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24, - "tiers": [ - { - "input": 2.4, - "output": 12, - "cache_read": 0.48, - "tier": { - "type": "context", - "size": 32001 - } - }, - { - "input": 3, - "output": 15, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 128001 - } - } - ] + "input": 0.35, + "output": 1.15 } - }, - "alibaba/qwen3-coder-plus": { - "id": "alibaba/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", - "attachment": false, + } + } + }, + "nano-gpt": { + "id": "nano-gpt", + "env": [ + "NANO_GPT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://nano-gpt.com/api/v1", + "name": "NanoGPT", + "doc": "https://docs.nano-gpt.com", + "models": { + "step-3": { + "id": "step-3", + "name": "Step-3", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-31", + "last_updated": "2025-07-31", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -181275,68 +171678,37 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 65536, + "input": 65536, + "output": 8192 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.2, - "tiers": [ - { - "input": 1.8, - "output": 9, - "cache_read": 0.36, - "tier": { - "type": "context", - "size": 32001 - } - }, - { - "input": 3, - "output": 15, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 128001 - } - }, - { - "input": 6, - "output": 60, - "cache_read": 1.2, - "tier": { - "type": "context", - "size": 256001 - } - } - ] + "input": 0.2499, + "output": 0.6494 } }, - "alibaba/qwen3-235b-a22b-thinking": { - "id": "alibaba/qwen3-235b-a22b-thinking", - "name": "Qwen3 235B A22B Thinking 2507", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "qwen3.5-35b-a3b:thinking": { + "id": "qwen3.5-35b-a3b:thinking", + "name": "Qwen3.5 35B A3B Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "budget_tokens", - "min": 1, + "min": 1024, "max": 81920 } ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-04", + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -181344,38 +171716,29 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 260096, + "input": 260096, + "output": 65536 }, "cost": { - "input": 0.4, - "output": 4 + "input": 0.225, + "output": 1.8 } }, - "alibaba/qwen3.8-flash": { - "id": "alibaba/qwen3.8-flash", - "name": "Qwen 3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "glm-4.1v-thinking-flashx": { + "id": "glm-4.1v-thinking-flashx", + "name": "GLM 4.1V Thinking FlashX", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -181383,30 +171746,29 @@ }, "open_weights": false, "limit": { - "context": 991000, - "output": 128000 + "context": 64000, + "input": 64000, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016, - "cache_write": 0.2 + "input": 0.3, + "output": 0.3 } }, - "alibaba/qwen3-embedding-8b": { - "id": "alibaba/qwen3-embedding-8b", - "name": "Qwen3 Embedding 8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "qwen", - "attachment": false, + "ernie-x1.1-preview": { + "id": "ernie-x1.1-preview", + "name": "ERNIE X1.1", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", + "structured_output": false, + "release_date": "2025-09-10", + "last_updated": "2025-09-10", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" @@ -181414,87 +171776,108 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 64000, + "input": 64000, + "output": 8192 + }, + "cost": { + "input": 0.15, + "output": 0.6 } }, - "alibaba/wan-v3.0-video": { - "id": "alibaba/wan-v3.0-video", - "name": "Wan v3.0 Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", + "qwen25-vl-72b-instruct": { + "id": "qwen25-vl-72b-instruct", + "name": "Qwen25 VL 72b", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, "tool_call": false, - "release_date": "2026-08-23", - "last_updated": "2026-08-23", + "structured_output": false, + "release_date": "2025-05-10", + "last_updated": "2025-05-10", "modalities": { "input": [ "text", "image" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 32000, + "input": 32000, + "output": 32768 + }, + "cost": { + "input": 0.69989, + "output": 0.69989 } }, - "alibaba/wan-v2.7-r2v": { - "id": "alibaba/wan-v2.7-r2v", - "name": "Wan v2.7 Reference-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": false, + "gemini-2.0-pro-exp-02-05": { + "id": "gemini-2.0-pro-exp-02-05", + "name": "Gemini 2.0 Pro 0205", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "structured_output": false, + "release_date": "2025-02-05", + "last_updated": "2025-02-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 2097152, + "input": 2097152, + "output": 8192 + }, + "cost": { + "input": 1.989, + "output": 7.956 } }, - "alibaba/wan-v3.0-video-prime": { - "id": "alibaba/wan-v3.0-video-prime", - "name": "Wan v3.0 Video Prime", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": true, + "doubao-seed-2-0-lite-260215": { + "id": "doubao-seed-2-0-lite-260215", + "name": "Doubao Seed 2.0 Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", + "structured_output": false, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "input": 256000, + "output": 32000 + }, + "cost": { + "input": 0.1462, + "output": 0.8738 } }, - "alibaba/qwen3.6-27b": { - "id": "alibaba/qwen3.6-27b", - "name": "Qwen 3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "Qwen3.5-27B-Writer-V2-Derestricted": { + "id": "Qwen3.5-27B-Writer-V2-Derestricted", + "name": "Qwen3.5 27B Writer V2 Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -181503,51 +171886,51 @@ }, { "type": "budget_tokens", - "min": 1, - "max": 131072 + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-06", + "last_updated": "2026-04-06", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.306, + "output": 0.306 } }, - "alibaba/qwen-3-235b": { - "id": "alibaba/qwen-3-235b", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "claude-opus-4-thinking:8192": { + "id": "claude-opus-4-thinking:8192", + "name": "Claude 4 Opus Thinking (8K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-04", + "structured_output": true, + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -181555,41 +171938,36 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.22, - "output": 0.88 + "input": 14.994, + "output": 75.004 } }, - "alibaba/qwen3.6-plus": { - "id": "alibaba/qwen3.6-plus", - "name": "Qwen 3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", + "qwen3-vl-235b-a22b-thinking": { + "id": "qwen3-vl-235b-a22b-thinking", + "name": "Qwen3 VL 235B A22B Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "budget_tokens", - "min": 1, - "max": 131072 + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -181597,39 +171975,48 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "output": 6 } }, - "alibaba/qwen3.7-plus": { - "id": "alibaba/qwen3.7-plus", - "name": "Qwen 3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "glm-4-air-0111": { + "id": "glm-4-air-0111", + "name": "GLM 4 Air 0111", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-11", + "last_updated": "2025-01-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 128000, + "output": 4096 + }, + "cost": { + "input": 0.1394, + "output": 0.1394 + } + }, + "Qwen3.5-27B-Queen-Derestricted": { + "id": "Qwen3.5-27B-Queen-Derestricted", + "name": "Qwen3.5 27B Queen Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -181638,20 +172025,19 @@ }, { "type": "budget_tokens", - "min": 1, - "max": 262144 + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -181659,92 +172045,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "cache_write": 0.5, - "tiers": [ - { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24, - "cache_write": 1.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24, - "cache_write": 1.5 - } - } - }, - "alibaba/wan-v2.6-r2v": { - "id": "alibaba/wan-v2.6-r2v", - "name": "Wan v2.6 Reference-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "video" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "input": 0.306, + "output": 0.306 } }, - "alibaba/qwen3.8-max-0902": { - "id": "alibaba/qwen3.8-max-0902", - "name": "Qwen3.8 Max 0902", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "gemini-2.5-pro-preview-03-25": { + "id": "gemini-2.5-pro-preview-03-25", + "name": "Gemini 2.5 Pro Preview 0325", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "xhigh" + "high" ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 } ], - "tool_call": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "tool_call": false, + "structured_output": false, + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -181752,49 +172085,64 @@ }, "open_weights": false, "limit": { - "context": 991000, - "output": 128000 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 + "input": 2.5, + "output": 10 } }, - "alibaba/wan-v2.7-t2v": { - "id": "alibaba/wan-v2.7-t2v", - "name": "Wan v2.7 Text-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", + "brave-research": { + "id": "brave-research", + "name": "Brave (Research)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "structured_output": false, + "release_date": "2023-03-02", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 16384, + "input": 16384, + "output": 16384 + }, + "cost": { + "input": 5, + "output": 5 } }, - "mixedbread/toast-1": { - "id": "mixedbread/toast-1", - "name": "Toast 1", - "description": "Specialized search model for knowledge-intensive questions, multi-step retrieval, and evidence synthesis", + "qwen-plus": { + "id": "qwen-plus", + "name": "Qwen Plus", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": false, - "tool_call": true, - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-25", + "last_updated": "2024-01-25", "modalities": { "input": [ "text" @@ -181805,295 +172153,339 @@ }, "open_weights": false, "limit": { - "context": 131000, - "output": 4000 + "context": 995904, + "input": 995904, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 0.72, - "cache_read": 0.036 + "input": 0.3995, + "output": 1.2002 } }, - "klingai/kling-v2.6-t2v": { - "id": "klingai/kling-v2.6-t2v", - "name": "Kling v2.6 Text-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ling", + "doubao-seed-2-0-mini-260215": { + "id": "doubao-seed-2-0-mini-260215", + "name": "Doubao Seed 2.0 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-21", + "structured_output": false, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "input": 256000, + "output": 32000 + }, + "cost": { + "input": 0.0493, + "output": 0.4845 } }, - "klingai/kling-v2.6-i2v": { - "id": "klingai/kling-v2.6-i2v", - "name": "Kling v2.6 Image-to-Video", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ling", - "attachment": false, - "reasoning": false, + "Qwen3.5-27B-BlueStar-v3-Derestricted-Lite": { + "id": "Qwen3.5-27B-BlueStar-v3-Derestricted-Lite", + "name": "Qwen3.5 27B BlueStar v3 Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], "tool_call": false, - "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-21", + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "input": 262144, + "output": 16384 + }, + "cost": { + "input": 0.306, + "output": 0.306 } }, - "klingai/kling-v2.5-turbo-t2v": { - "id": "klingai/kling-v2.5-turbo-t2v", - "name": "Kling v2.5 Turbo Text-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ling", + "deepseek-v3-0324": { + "id": "deepseek-v3-0324", + "name": "DeepSeek Chat 0324", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "tool_call": true, + "structured_output": true, + "release_date": "2025-03-24", + "last_updated": "2025-03-24", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 8192 + }, + "cost": { + "input": 0.25, + "output": 0.7 } }, - "klingai/kling-v3.0-i2v": { - "id": "klingai/kling-v3.0-i2v", - "name": "Kling v3.0 Image-to-Video", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ling", - "attachment": false, - "reasoning": false, + "Gemma-4-31B-Musica-v1": { + "id": "Gemma-4-31B-Musica-v1", + "name": "Gemma 4 31B Musica v1", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": false, - "temperature": true, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "structured_output": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "input": 262144, + "output": 16384 + }, + "cost": { + "input": 0.306, + "output": 0.306 } }, - "klingai/kling-v3.0-motion-control": { - "id": "klingai/kling-v3.0-motion-control", - "name": "Kling v3.0 Motion Control", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ling", + "mirothinker-1-7-deepresearch-mini": { + "id": "mirothinker-1-7-deepresearch-mini", + "name": "MiroThinker 1.7 Deep Research Mini", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "temperature": true, - "release_date": "2026-03-04", - "last_updated": "2026-03-04", + "structured_output": false, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "input": 262144, + "output": 16384 + }, + "cost": { + "input": 1.25, + "output": 10 } }, - "klingai/kling-v3.0-t2v": { - "id": "klingai/kling-v3.0-t2v", - "name": "Kling v3.0 Text-to-Video", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ling", + "Baichuan4-Turbo": { + "id": "Baichuan4-Turbo", + "name": "Baichuan 4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "structured_output": false, + "release_date": "2025-08-19", + "last_updated": "2025-08-19", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 32768 + }, + "cost": { + "input": 2.42, + "output": 2.42 } }, - "klingai/kling-v2.6-motion-control": { - "id": "klingai/kling-v2.6-motion-control", - "name": "Kling v2.6 Motion Control", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "ling", + "doubao-seed-1-6-flash-250615": { + "id": "doubao-seed-1-6-flash-250615", + "name": "Doubao Seed 1.6 Flash", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-21", + "structured_output": false, + "release_date": "2025-06-15", + "last_updated": "2025-06-15", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "input": 256000, + "output": 16384 + }, + "cost": { + "input": 0.0374, + "output": 0.374 } }, - "klingai/kling-v2.5-turbo-i2v": { - "id": "klingai/kling-v2.5-turbo-i2v", - "name": "Kling v2.5 Turbo Image-to-Video", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "ling", - "attachment": false, + "kimi-k2-instruct-fast": { + "id": "kimi-k2-instruct-fast", + "name": "Kimi K2 0711 Fast", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "structured_output": false, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "input": 131072, + "output": 16384 + }, + "cost": { + "input": 0.1, + "output": 2 } }, - "bytedance/seedream-5.0-lite": { - "id": "bytedance/seedream-5.0-lite", - "name": "Seedream 5.0 Lite", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "seed", + "glm-4-plus": { + "id": "glm-4-plus", + "name": "GLM-4 Plus", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-01-28", + "structured_output": false, + "release_date": "2024-08-01", + "last_updated": "2024-08-01", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 4096 + }, + "cost": { + "input": 7.497, + "output": 7.497 } }, - "bytedance/seedream-4.5": { - "id": "bytedance/seedream-4.5", - "name": "Seedream 4.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "seed", + "v0-1.0-md": { + "id": "v0-1.0-md", + "name": "v0 1.0 MD", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-11-28", + "structured_output": false, + "release_date": "2025-07-04", + "last_updated": "2025-07-04", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "input": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15 } }, - "bytedance/seed-1.8": { - "id": "bytedance/seed-1.8", - "name": "Seed 1.8", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "qwen3-coder-30b-a3b-instruct": { + "id": "qwen3-coder-30b-a3b-instruct", + "name": "Qwen3 Coder 30B A3B Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-01", - "last_updated": "2025-10", + "structured_output": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -182101,54 +172493,31 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05, - "tiers": [ - { - "input": 0.5, - "output": 4, - "cache_read": 0.05, - "tier": { - "type": "context", - "size": 128001 - } - } - ] + "input": 0.1, + "output": 0.4 } }, - "bytedance/seed-2.1-turbo": { - "id": "bytedance/seed-2.1-turbo", - "name": "Seed 2.1 Turbo", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", + "gemini-2.5-flash-preview-09-2025-thinking": { + "id": "gemini-2.5-flash-preview-09-2025-thinking", + "name": "Gemini 2.5 Flash Preview (09/2025) – Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", + "structured_output": true, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -182156,263 +172525,259 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.1 + "input": 0.3, + "output": 2.5 } }, - "bytedance/seedream-4.0": { - "id": "bytedance/seedream-4.0", - "name": "Seedream 4.0", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "seed", + "v0-1.5-lg": { + "id": "v0-1.5-lg", + "name": "v0 1.5 LG", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-09-09", - "last_updated": "2025-08-28", + "structured_output": false, + "release_date": "2025-07-04", + "last_updated": "2025-07-04", "modalities": { "input": [ "text" ], "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "bytedance/seedance-v1.0-pro": { - "id": "bytedance/seedance-v1.0-pro", - "name": "Seedance v1.0 Pro", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-06-11", - "last_updated": "2025-06-11", - "modalities": { - "input": [ "text" - ], - "output": [ - "video" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "input": 1000000, + "output": 64000 + }, + "cost": { + "input": 15, + "output": 75 } }, - "bytedance/seedance-2.0": { - "id": "bytedance/seedance-2.0", - "name": "Seedance 2.0", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", + "qwen3.5-flash:thinking": { + "id": "qwen3.5-flash:thinking", + "name": "Qwen3.5 Flash Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], "tool_call": false, - "temperature": true, - "release_date": "2026-04-14", - "last_updated": "2026-04-14", + "structured_output": false, + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 991808, + "input": 991808, + "output": 65536 + }, + "cost": { + "input": 0.09, + "output": 0.36 } }, - "bytedance/seedance-2.0-fast": { - "id": "bytedance/seedance-2.0-fast", - "name": "Seedance 2.0 Fast", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", - "attachment": true, + "jamba-large": { + "id": "jamba-large", + "name": "Jamba Large", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-04-14", - "last_updated": "2026-04-14", + "structured_output": false, + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "input": 256000, + "output": 4096 + }, + "cost": { + "input": 1.989, + "output": 7.99 } }, - "bytedance/seedance-v1.5-pro": { - "id": "bytedance/seedance-v1.5-pro", - "name": "Seedance v1.5 Pro", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", - "attachment": false, + "kimi-thinking-preview": { + "id": "kimi-thinking-preview", + "name": "Kimi Thinking Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", + "structured_output": false, + "release_date": "2025-05-07", + "last_updated": "2025-05-07", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 16384 + }, + "cost": { + "input": 31.46, + "output": 31.46 } }, - "bytedance/seedance-2.0-mini": { - "id": "bytedance/seedance-2.0-mini", - "name": "Seedance 2.0 Mini", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", + "gemini-2.0-flash-thinking-exp-01-21": { + "id": "gemini-2.0-flash-thinking-exp-01-21", + "name": "Gemini 2.0 Flash Thinking 0121", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "release_date": "2026-06-22", - "last_updated": "2026-06-22", + "structured_output": false, + "release_date": "2025-01-21", + "last_updated": "2025-01-21", "modalities": { "input": [ "text", "image" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "input": 1000000, + "output": 8192 + }, + "cost": { + "input": 0.306, + "output": 1.003 } }, - "bytedance/seedance-2.5": { - "id": "bytedance/seedance-2.5", - "name": "Seedance 2.5", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-08-07", - "last_updated": "2026-08-07", + "claude-opus-4-1-thinking:1024": { + "id": "claude-opus-4-1-thinking:1024", + "name": "Claude 4.1 Opus Thinking (1K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "input": 200000, + "output": 32000 + }, + "cost": { + "input": 14.994, + "output": 75.004 } }, - "bytedance/seedance-v1.0-pro-fast": { - "id": "bytedance/seedance-v1.0-pro-fast", - "name": "Seedance v1.0 Pro Fast", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", - "attachment": false, - "reasoning": false, + "Qwen3.5-27B-NaNovel-Derestricted-Lite": { + "id": "Qwen3.5-27B-NaNovel-Derestricted-Lite", + "name": "Qwen3.5 27B NaNovel Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], "tool_call": false, - "temperature": true, - "release_date": "2025-10-24", - "last_updated": "2025-10-31", + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "input": 262144, + "output": 16384 + }, + "cost": { + "input": 0.306, + "output": 0.306 } }, - "bytedance/seedream-5.0-pro": { - "id": "bytedance/seedream-5.0-pro", - "name": "Seedream 5.0 Pro", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "seed", + "doubao-seed-1-6-250615": { + "id": "doubao-seed-1-6-250615", + "name": "Doubao Seed 1.6", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-07-11", - "last_updated": "2026-07-11", + "structured_output": false, + "release_date": "2025-06-15", + "last_updated": "2025-06-15", "modalities": { "input": [ "text" ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "bytedance/seed-1.6": { - "id": "bytedance/seed-1.6", - "name": "Seed 1.6", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-01", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text", - "image" - ], "output": [ "text" ] @@ -182420,74 +172785,60 @@ "open_weights": false, "limit": { "context": 256000, - "output": 32000 + "input": 256000, + "output": 16384 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05, - "tiers": [ - { - "input": 0.5, - "output": 4, - "cache_read": 0.05, - "tier": { - "type": "context", - "size": 128001 - } - } - ] + "input": 0.204, + "output": 0.51 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "ernie-5.0-thinking-preview": { + "id": "ernie-5.0-thinking-preview", + "name": "Ernie 5.0 Thinking Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.14, - "output": 0.4 + "input": 1.1, + "output": 2 } }, - "google/gemini-omni-flash-preview": { - "id": "google/gemini-omni-flash-preview", - "name": "Gemini Omni Flash Preview", - "description": "Omni-modal model for text, vision, audio, and multimodal agent tasks", - "family": "gemini", + "gemini-2.5-flash-preview-05-20:thinking": { + "id": "gemini-2.5-flash-preview-05-20:thinking", + "name": "Gemini 2.5 Flash 0520 Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": false, - "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "structured_output": false, + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -182495,69 +172846,69 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 57920 + "context": 1048000, + "input": 1048000, + "output": 65536 }, "cost": { - "input": 1.5, - "output": 9 + "input": 0.15, + "output": 3.5 } }, - "google/gemini-3.1-flash-image-preview": { - "id": "google/gemini-3.1-flash-image-preview", - "name": "Gemini 3.1 Flash Image Preview (Nano Banana 2)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", + "Qwen3.5-27B-Omega-Evolution-v2.2-Derestricted": { + "id": "Qwen3.5-27B-Omega-Evolution-v2.2-Derestricted", + "name": "Qwen3.5 27B Omega Evolution v2.2 Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "structured_output": false, + "release_date": "2026-05-02", + "last_updated": "2026-05-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 0.306, + "output": 0.306 } }, - "google/gemini-3.5-transcribe-live": { - "id": "google/gemini-3.5-transcribe-live", - "name": "Gemini 3.5 Transcribe Live", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "gemini", + "azure-o1": { + "id": "azure-o1", + "name": "Azure o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "structured_output": false, + "release_date": "2024-12-17", + "last_updated": "2024-12-17", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" @@ -182565,54 +172916,57 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "input": 200000, + "output": 100000 + }, + "cost": { + "input": 14.994, + "output": 59.993 } }, - "google/gemini-2.5-flash-image": { - "id": "google/gemini-2.5-flash-image", - "name": "Nano Banana (Gemini 2.5 Flash Image)", - "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", - "family": "gemini-flash", - "attachment": false, + "qwen3.7-plus": { + "id": "qwen3.7-plus", + "name": "Qwen3.7 Plus", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", + "structured_output": false, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 65535 + "context": 991808, + "input": 991808, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 0.4, + "output": 1.6, + "cache_read": 0.04 } }, - "google/text-multilingual-embedding-002": { - "id": "google/text-multilingual-embedding-002", - "name": "Text Multilingual Embedding 002", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "glm-4-airx": { + "id": "glm-4-airx", + "name": "GLM-4 AirX", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-03-01", - "last_updated": "2024-03", + "structured_output": false, + "release_date": "2024-06-05", + "last_updated": "2024-06-05", "modalities": { "input": [ "text" @@ -182623,22 +172977,25 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 + "context": 8000, + "input": 8000, + "output": 4096 + }, + "cost": { + "input": 2.006, + "output": 2.006 } }, - "google/gemini-embedding-2": { - "id": "google/gemini-embedding-2", - "name": "Gemini Embedding 2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "gemini", + "step-r1-v-mini": { + "id": "step-r1-v-mini", + "name": "Step R1 V Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2025-11", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "structured_output": false, + "release_date": "2025-04-08", + "last_updated": "2025-04-08", "modalities": { "input": [ "text" @@ -182649,34 +173006,66 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 - } - }, - "google/gemini-3.6-flash": { - "id": "google/gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "context": 128000, + "input": 128000, + "output": 65536 + }, + "cost": { + "input": 2.5, + "output": 11 + } + }, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", "low", "medium", "high" ] } ], + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048756, + "input": 1048756, + "output": 65536 + }, + "cost": { + "input": 2.5, + "output": 10 + } + }, + "claude-haiku-4-5-20251001-thinking": { + "id": "claude-haiku-4-5-20251001-thinking", + "name": "Claude Haiku 4.5 Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", @@ -182689,52 +173078,55 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 200000, + "input": 200000, "output": 64000 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 + "input": 1, + "output": 5, + "cache_read": 0.1 } }, - "google/veo-3.1-lite-generate-001": { - "id": "google/veo-3.1-lite-generate-001", - "name": "Veo 3.1 Lite Generate", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", + "jamba-mini-1.6": { + "id": "jamba-mini-1.6", + "name": "Jamba Mini 1.6", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "structured_output": false, + "release_date": "2025-03-01", + "last_updated": "2025-03-01", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "input": 256000, + "output": 4096 + }, + "cost": { + "input": 0.1989, + "output": 0.408 } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "claude-haiku-4-5-20251001": { + "id": "claude-haiku-4-5-20251001", + "name": "Claude Haiku 4.5", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", @@ -182745,41 +173137,59 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 64000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 1, + "output": 5 } }, - "google/gemini-3.5-flash-lite": { - "id": "google/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "claude-3-5-haiku-20241022": { + "id": "claude-3-5-haiku-20241022", + "name": "Claude 3.5 Haiku", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "release_date": "2024-10-22", + "last_updated": "2024-10-22", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "input": 200000, + "output": 8192 + }, + "cost": { + "input": 0.8, + "output": 4 + } + }, + "claude-sonnet-4-thinking:32768": { + "id": "claude-sonnet-4-thinking:32768", + "name": "Claude 4 Sonnet Thinking (32K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", @@ -182793,79 +173203,58 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 65000 + "input": 1000000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "input": 2.992, + "output": 14.994 } }, - "google/gemini-3.1-flash-image": { - "id": "google/gemini-3.1-flash-image", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", + "hermes-low": { + "id": "hermes-low", + "name": "Hermes Low", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", + "claude-sonnet-4-thinking": { + "id": "claude-sonnet-4-thinking", + "name": "Claude 4 Sonnet Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "release_date": "2025-02-24", + "last_updated": "2025-02-24", "modalities": { "input": [ "text", @@ -182879,83 +173268,57 @@ "open_weights": false, "limit": { "context": 1000000, + "input": 1000000, "output": 64000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200001 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 2.992, + "output": 14.994 } }, - "google/veo-3.1-fast-generate-001": { - "id": "google/veo-3.1-fast-generate-001", - "name": "Veo 3.1 Fast Generate", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7 Max", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2026-06-08", + "structured_output": false, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "input": 1000000, + "output": 65536 + }, + "cost": { + "input": 2.5, + "output": 7.5, + "cache_read": 0.25 } }, - "google/gemini-3-flash": { - "id": "google/gemini-3-flash", - "name": "Gemini 3 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "yi-medium-200k": { + "id": "yi-medium-200k", + "name": "Yi Medium 200k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-03-01", + "last_updated": "2024-03-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -182963,44 +173326,40 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65000 + "context": 200000, + "input": 200000, + "output": 4096 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 + "input": 2.499, + "output": 2.499 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "Qwen3.5-27B-BlueStar-v2-Derestricted-Lite": { + "id": "Qwen3.5-27B-BlueStar-v2-Derestricted-Lite", + "name": "Qwen3.5 27B BlueStar v2 Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-06", + "last_updated": "2026-04-06", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -183008,190 +173367,205 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 + "input": 0.306, + "output": 0.306 } }, - "google/gemini-3-pro-image": { - "id": "google/gemini-3-pro-image", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", - "attachment": false, - "reasoning": false, + "Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled": { + "id": "Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled", + "name": "Gemma 4 31B Claude 4.6 Opus Reasoning Distilled", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 32768 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 + "input": 0.306, + "output": 0.306, + "cache_read": 0.0306 } }, - "google/veo-3.0-fast-generate-001": { - "id": "google/veo-3.0-fast-generate-001", - "name": "Veo 3.0 Fast Generate", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": false, + "doubao-1-5-thinking-vision-pro-250428": { + "id": "doubao-1-5-thinking-vision-pro-250428", + "name": "Doubao 1.5 Thinking Vision Pro", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-07-31", - "last_updated": "2026-06-08", + "structured_output": false, + "release_date": "2025-05-15", + "last_updated": "2025-05-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 16384 + }, + "cost": { + "input": 0.55, + "output": 1.43 } }, - "google/gemini-3.8-live": { - "id": "google/gemini-3.8-live", - "name": "Gemini 3.8 Live", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gemini", - "attachment": false, - "reasoning": false, + "Qwen3.5-27B-BlueStar-Derestricted-Lite": { + "id": "Qwen3.5-27B-BlueStar-Derestricted-Lite", + "name": "Qwen3.5 27B BlueStar Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], "tool_call": false, - "release_date": "2026-09-15", - "last_updated": "2026-09-15", + "structured_output": false, + "release_date": "2026-04-06", + "last_updated": "2026-04-06", "modalities": { "input": [ "text", - "audio" + "image", + "video" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.75, - "output": 4.5 + "input": 0.306, + "output": 0.306 } }, - "google/veo-3.1-generate-001": { - "id": "google/veo-3.1-generate-001", - "name": "Veo 3.1", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", + "Qwen2.5-32B-EVA-v0.2": { + "id": "Qwen2.5-32B-EVA-v0.2", + "name": "Qwen 2.5 32b EVA", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2026-06-08", + "structured_output": false, + "release_date": "2024-09-01", + "last_updated": "2024-09-01", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 24576, + "input": 24576, + "output": 8192 + }, + "cost": { + "input": 0.493, + "output": 0.493 } }, - "google/gemini-3.1-flash-lite-image": { - "id": "google/gemini-3.1-flash-lite-image", - "name": "Gemini 3.1 Flash Lite Image (Nano Banana 2 Lite)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash-lite", + "gemini-2.5-flash": { + "id": "gemini-2.5-flash", + "name": "Gemini 2.5 Flash", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": false, "structured_output": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ "text", "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 4096 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.03 + "input": 0.3, + "output": 2.5 } }, - "google/gemini-3.7-flash": { - "id": "google/gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "qwen-long": { + "id": "qwen-long", + "name": "Qwen Long 10M", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-25", + "last_updated": "2025-01-25", "modalities": { "input": [ "text", - "image", "pdf" ], "output": [ @@ -183200,29 +173574,35 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65535 + "context": 10000000, + "input": 10000000, + "output": 8192 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 + "input": 0.1003, + "output": 0.408 } }, - "google/text-embedding-005": { - "id": "google/text-embedding-005", - "name": "Text Embedding 005", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, + "Gemma-4-31B-DarkIdol": { + "id": "Gemma-4-31B-DarkIdol", + "name": "Gemma 4 31B DarkIdol", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": false, - "temperature": true, - "release_date": "2024-08-01", - "last_updated": "2024-08", + "structured_output": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -183230,23 +173610,28 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 + "context": 262144, + "input": 262144, + "output": 16384 + }, + "cost": { + "input": 0.306, + "output": 0.306 } }, - "google/gemini-3.5-transcribe": { - "id": "google/gemini-3.5-transcribe", - "name": "Gemini 3.5 Transcribe", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "gemini", + "step-2-16k-exp": { + "id": "step-2-16k-exp", + "name": "Step-2 16k Exp", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "structured_output": false, + "release_date": "2024-07-05", + "last_updated": "2024-07-05", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" @@ -183254,32 +173639,40 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 16000, + "input": 16000, + "output": 8192 }, "cost": { - "input": 2, - "output": 12 + "input": 7.004, + "output": 19.992 } }, - "google/gemini-3.8-flash": { - "id": "google/gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", + "Qwen3.5-27B-Marvin-V2-Derestricted-Lite": { + "id": "Qwen3.5-27B-Marvin-V2-Derestricted-Lite", + "name": "Qwen3.5 27B Marvin V2 Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -183287,20 +173680,19 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65535 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 + "input": 0.306, + "output": 0.306 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", + "Qwen3.5-27B-NaNovel-Derestricted": { + "id": "Qwen3.5-27B-NaNovel-Derestricted", + "name": "Qwen3.5 27B NaNovel Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -183309,21 +173701,19 @@ }, { "type": "budget_tokens", - "min": 512, - "max": 24576 + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -183331,27 +173721,25 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65535 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 + "input": 0.306, + "output": 0.306 } }, - "google/gemini-embedding-001": { - "id": "google/gemini-embedding-001", - "name": "Gemini Embedding 001", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "gemini", + "ernie-4.5-8k-preview": { + "id": "ernie-4.5-8k-preview", + "name": "Ernie 4.5 8k Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "structured_output": false, + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ "text" @@ -183362,77 +173750,58 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 - } - }, - "google/gemini-3.8-live-extended-thinking": { - "id": "google/gemini-3.8-live-extended-thinking", - "name": "Gemini 3.8 Live Extended Thinking", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gemini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-09-15", - "last_updated": "2026-09-15", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "context": 8000, + "input": 8000, + "output": 16384 }, "cost": { - "input": 0.75, - "output": 4.5 + "input": 0.66, + "output": 2.6 } }, - "google/veo-3.0-generate-001": { - "id": "google/veo-3.0-generate-001", - "name": "Veo 3.0", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", + "gemini-2.0-flash-exp-image-generation": { + "id": "gemini-2.0-flash-exp-image-generation", + "name": "Gemini Text + Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-05-20", - "last_updated": "2026-06-08", + "structured_output": false, + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 32767, + "input": 32767, + "output": 8192 + }, + "cost": { + "input": 0.2, + "output": 0.8 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", + "gemini-2.5-flash-lite-preview-09-2025": { + "id": "gemini-2.5-flash-lite-preview-09-2025", + "name": "Gemini 2.5 Flash Lite Preview (09/2025)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", "high" @@ -183441,10 +173810,8 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", @@ -183457,45 +173824,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65000 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.03 + "input": 0.1, + "output": 0.4 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "deepseek-reasoner-cheaper": { + "id": "deepseek-reasoner-cheaper", + "name": "Deepseek R1 Cheaper", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -183503,43 +173853,28 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 128000, + "input": 128000, "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 + "input": 0.4, + "output": 1.7 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "venice-uncensored": { + "id": "venice-uncensored", + "name": "Venice Uncensored", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-24", + "last_updated": "2025-02-24", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -183547,51 +173882,25 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.4, + "output": 0.4 } }, - "sakana/fugu-max": { - "id": "sakana/fugu-max", - "name": "Fugu Max", - "description": "Multi-agent model for routing expert agents across complex analytical tasks", - "family": "fugu", + "gemini-2.0-flash-001": { + "id": "gemini-2.0-flash-001", + "name": "Gemini 2.0 Flash", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "structured_output": true, + "release_date": "2024-12-11", + "last_updated": "2024-12-11", "modalities": { "input": [ "text", @@ -183604,38 +173913,34 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 1000000 + "input": 1000000, + "output": 8192 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 0.1003, + "output": 0.408 } }, - "sakana/fugu-ultra-v2": { - "id": "sakana/fugu-ultra-v2", - "name": "Fugu Ultra v2", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", + "gemma-4-31B-Larkspur-v0.5": { + "id": "gemma-4-31B-Larkspur-v0.5", + "name": "Gemma 4 31B Larkspur v0.5", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], - "tool_call": true, - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-02", + "last_updated": "2026-05-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -183643,41 +173948,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272001 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.306, + "output": 0.306 } }, - "sakana/namazu": { - "id": "sakana/namazu", - "name": "Sakana Namazu", - "description": "Multi-agent model for routing expert agents across complex analytical tasks", + "claude-opus-4-1-20250805": { + "id": "claude-opus-4-1-20250805", + "name": "Claude 4.1 Opus", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "structured_output": true, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text", @@ -183690,28 +173979,30 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.15 + "input": 14.994, + "output": 75.004 } }, - "sakana/fugu-ultra": { - "id": "sakana/fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", + "holo3-35b-a3b": { + "id": "holo3-35b-a3b", + "name": "Holo3-35B-A3B", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text", @@ -183723,49 +174014,31 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 65536, + "input": 65536, + "output": 65536 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272001 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.25, + "output": 1.8 } }, - "inclusionai/ling-3.0-flash-fin-free": { - "id": "inclusionai/ling-3.0-flash-fin-free", - "name": "Ling 3.0 Flash Fin (Free)", - "description": "Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution", - "family": "ling", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "qwen3.5-omni-plus": { + "id": "qwen3.5-omni-plus", + "name": "Qwen3.5 Omni Plus", + "description": "Omni-modal model for text, vision, audio, and multimodal agent tasks", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-30", + "last_updated": "2026-03-30", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio" ], "output": [ "text" @@ -183773,19 +174046,19 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 983616, + "input": 983616, + "output": 65536 }, "cost": { "input": 0, "output": 0 } }, - "inclusionai/ling-3.0-flash-vl-free": { - "id": "inclusionai/ling-3.0-flash-vl-free", - "name": "Ling 3.0 Flash VL (Free)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ling", + "Gemma-4-31B-it": { + "id": "Gemma-4-31B-it", + "name": "Gemma 4 31B IT", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -183793,13 +174066,15 @@ "type": "toggle" } ], - "tool_call": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-09", + "last_updated": "2026-04-09", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -183807,28 +174082,40 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.306, + "output": 0.306 } }, - "inclusionai/ling-3.0-flash": { - "id": "inclusionai/ling-3.0-flash", - "name": "Ling 3.0 Flash", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", - "attachment": false, + "Qwen3.5-27B-Derestricted": { + "id": "Qwen3.5-27B-Derestricted", + "name": "Qwen3.5 27B Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "release_date": "2026-08-06", - "last_updated": "2026-08-06", + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -183836,26 +174123,26 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.021, - "output": 0.063, - "cache_read": 0.0042 + "input": 0.306, + "output": 0.306 } }, - "inclusionai/ling-3.0-flash-sante": { - "id": "inclusionai/ling-3.0-flash-sante", - "name": "Ling 3.0 Flash Sante", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", + "sonar-reasoning-pro": { + "id": "sonar-reasoning-pro", + "name": "Perplexity Reasoning Pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ "text" @@ -183866,33 +174153,28 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 127000, + "input": 127000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 2.006, + "output": 7.9985 } }, - "inclusionai/ling-3.0-flash-vl": { - "id": "inclusionai/ling-3.0-flash-vl", - "name": "Ling 3.0 Flash VL", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "venice-uncensored:web": { + "id": "venice-uncensored:web", + "name": "Venice Uncensored Web", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-05-01", + "last_updated": "2024-05-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -183900,29 +174182,25 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 80000, + "input": 80000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.4, + "output": 0.4 } }, - "inclusionai/ling-3.0-flash-fin": { - "id": "inclusionai/ling-3.0-flash-fin", - "name": "Ling 3.0 Flash Fin", - "description": "Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution", - "family": "ling", + "doubao-1.5-pro-32k": { + "id": "doubao-1.5-pro-32k", + "name": "Doubao 1.5 Pro 32k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-22", + "last_updated": "2025-01-22", "modalities": { "input": [ "text" @@ -183933,25 +174211,25 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 32000, + "input": 32000, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1343, + "output": 0.3349 } }, - "inclusionai/ling-3.0-flash-sante-free": { - "id": "inclusionai/ling-3.0-flash-sante-free", - "name": "Ling 3.0 Flash Sante (Free)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", + "qwen3-30b-a3b-instruct-2507": { + "id": "qwen3-30b-a3b-instruct-2507", + "name": "Qwen3 30B A3B Instruct 2507", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-20", + "last_updated": "2025-02-20", "modalities": { "input": [ "text" @@ -183963,104 +174241,88 @@ "open_weights": false, "limit": { "context": 256000, - "output": 32000 + "input": 256000, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 0.5 } }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k3", + "ernie-5.1:thinking": { + "id": "ernie-5.1:thinking", + "name": "ERNIE 5.1 Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-10", + "last_updated": "2026-05-10", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 119000, + "input": 119000, + "output": 64000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.75, + "output": 3, + "cache_read": 0.75 } }, - "moonshotai/kimi-k3-fast": { - "id": "moonshotai/kimi-k3-fast", - "name": "Kimi K3 Fast", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k3", - "attachment": true, + "mirothinker-1-7-deepresearch": { + "id": "mirothinker-1-7-deepresearch", + "name": "MiroThinker 1.7 Deep Research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 4.5, - "output": 22.5, - "cache_read": 0.45 + "input": 4, + "output": 25 } }, - "moonshotai/kimi-k2.7-code-highspeed": { - "id": "moonshotai/kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code High Speed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", + "hermes-high": { + "id": "hermes-high", + "name": "Hermes High", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ "text", @@ -184071,137 +174333,149 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 + "input": 4.998, + "output": 25.007 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "jamba-large-1.6": { + "id": "jamba-large-1.6", + "name": "Jamba Large 1.6", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 256000, + "input": 256000, + "output": 4096 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 1.989, + "output": 7.99 } }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", + "glm-4-long": { + "id": "glm-4-long", + "name": "GLM-4 Long", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-08-01", + "last_updated": "2024-08-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 1000000, + "output": 4096 + }, + "cost": { + "input": 0.2006, + "output": 0.2006 + } + }, + "claude-opus-4-thinking:32768": { + "id": "claude-opus-4-thinking:32768", + "name": "Claude 4 Opus Thinking (32K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "structured_output": true, + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 216144, - "output": 216144 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.47, - "output": 2, - "cache_read": 0.141 + "input": 14.994, + "output": 75.004 } }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "claude-sonnet-4-thinking:8192": { + "id": "claude-sonnet-4-thinking:8192", + "name": "Claude 4 Sonnet Thinking (8K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 1000000, + "input": 1000000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 3 + "input": 2.992, + "output": 14.994 } }, - "moonshotai/kimi-k2": { - "id": "moonshotai/kimi-k2", - "name": "Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "azure-o3-mini": { + "id": "azure-o3-mini", + "name": "Azure o3-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-07-11", - "last_updated": "2025-09-05", + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ "text" @@ -184212,90 +174486,91 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 65536 }, "cost": { - "input": 0.57, - "output": 2.3 + "input": 1.088, + "output": 4.3996 } }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "qvq-max": { + "id": "qvq-max", + "name": "Qwen: QvQ Max", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-03-28", + "last_updated": "2025-03-28", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 1.4, + "output": 5.3 } }, - "nvidia/nemotron-3-nano-30b-a3b": { - "id": "nvidia/nemotron-3-nano-30b-a3b", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, + "Qwen3.5-27B-BlueStar-v2-Derestricted": { + "id": "Qwen3.5-27B-BlueStar-v2-Derestricted", + "name": "Qwen3.5 27B BlueStar v2 Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], "tool_call": false, - "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "structured_output": false, + "release_date": "2026-04-06", + "last_updated": "2026-04-06", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 262144 + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.025 + "input": 0.306, + "output": 0.306 } }, - "nvidia/nemotron-3.5-lightning": { - "id": "nvidia/nemotron-3.5-lightning", - "name": "Nemotron 3.5 Lightning 30B", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, + "Qwen3.5-27B-Vivid-Durian": { + "id": "Qwen3.5-27B-Vivid-Durian", + "name": "Qwen3.5 27B Vivid Durian", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -184303,50 +174578,45 @@ }, { "type": "budget_tokens", - "min": 1, - "max": 32768 + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 131072 + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.01 + "input": 0.306, + "output": 0.306 } }, - "nvidia/nemotron-nano-9b-v2": { - "id": "nvidia/nemotron-nano-9b-v2", - "name": "Nvidia Nemotron Nano 9B V2", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", + "jamba-large-1.7": { + "id": "jamba-large-1.7", + "name": "Jamba Large 1.7", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-18", - "last_updated": "2025-08-18", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text" @@ -184355,32 +174625,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 4096 }, "cost": { - "input": 0.06, - "output": 0.23 + "input": 1.989, + "output": 7.99 } }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "NVIDIA Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", + "Baichuan-M2": { + "id": "Baichuan-M2", + "name": "Baichuan M2 32B Medical", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "structured_output": false, + "release_date": "2025-08-19", + "last_updated": "2025-08-19", "modalities": { "input": [ "text" @@ -184389,32 +174654,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 32000 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.65 + "input": 15.73, + "output": 15.73 } }, - "nvidia/nemotron-3-ultra-550b-a55b": { - "id": "nvidia/nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", + "Magistral-Small-2506": { + "id": "Magistral-Small-2506", + "name": "Magistral Small 2506", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text" @@ -184423,308 +174683,265 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 65000 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 + "input": 0.4, + "output": 1.4 } }, - "nvidia/nemotron-nano-12b-v2-vl": { - "id": "nvidia/nemotron-nano-12b-v2-vl", - "name": "Nvidia Nemotron Nano 12B V2 VL", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", - "attachment": true, + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek R1", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 0.4, + "output": 1.7 } }, - "xiaomi/mimo-v2.6-pro": { - "id": "xiaomi/mimo-v2.6-pro", - "name": "MiMo V2.6 Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "Qwen3.5-27B-earica-Derestricted-Lite": { + "id": "Qwen3.5-27B-earica-Derestricted-Lite", + "name": "Qwen3.5 27B earica Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "input": 0.306, + "output": 0.306 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo M2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "doubao-seed-2-0-pro-260215": { + "id": "doubao-seed-2-0-pro-260215", + "name": "Doubao Seed 2.0 Pro", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1050000, - "output": 131100 + "context": 256000, + "input": 256000, + "output": 128000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.782, + "output": 3.876 } }, - "xiaomi/mimo-v2.6-pro-ultraspeed": { - "id": "xiaomi/mimo-v2.6-pro-ultraspeed", - "name": "MiMo V2.6 Pro UltraSpeed", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-09-21", - "last_updated": "2026-09-21", + "step-2-mini": { + "id": "step-2-mini", + "name": "Step-2 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-05", + "last_updated": "2024-07-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 8000, + "input": 8000, + "output": 4096 }, "cost": { - "input": 4.35, - "output": 8.7, - "cache_read": 0.036 + "input": 0.2006, + "output": 0.408 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo V2.5 Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", + "gemini-exp-1206": { + "id": "gemini-exp-1206", + "name": "Gemini 2.0 Pro 1206", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1050000, - "output": 131000 + "context": 2097152, + "input": 2097152, + "output": 8192 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 + "input": 1.258, + "output": 4.998 } }, - "xiaomi/mimo-v2.6-flash": { - "id": "xiaomi/mimo-v2.6-flash", - "name": "MiMo V2.6 Flash", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", + "claude-opus-4-5-20251101:thinking": { + "id": "claude-opus-4-5-20251101:thinking", + "name": "Claude 4.5 Opus Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "structured_output": true, + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 4.998, + "output": 25.007 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, + "Qwen3.5-27B-BlueStar-v3-Derestricted": { + "id": "Qwen3.5-27B-BlueStar-v3-Derestricted", + "name": "Qwen3.5 27B BlueStar v3 Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.306, + "output": 0.306 } }, - "minimax/minimax-m2.1-lightning": { - "id": "minimax/minimax-m2.1-lightning", - "name": "MiniMax M2.1 Lightning", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, + "claude-opus-4-thinking": { + "id": "claude-opus-4-thinking", + "name": "Claude 4 Opus Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-23", - "last_updated": "2025-10-27", + "structured_output": true, + "release_date": "2025-07-15", + "last_updated": "2025-07-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -184732,53 +174949,56 @@ }, "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.3, - "output": 2.4, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 14.994, + "output": 75.004 } }, - "minimax/minimax-h3": { - "id": "minimax/minimax-h3", - "name": "MiniMax H3", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "minimax", + "qwen3.5-flash": { + "id": "qwen3.5-flash", + "name": "Qwen3.5 Flash", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, "tool_call": false, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "structured_output": false, + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 991808, + "input": 991808, + "output": 65536 + }, + "cost": { + "input": 0.09, + "output": 0.36 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "Minimax M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "exa-research-pro": { + "id": "exa-research-pro", + "name": "Exa (Research Pro)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-04", + "last_updated": "2025-06-04", "modalities": { "input": [ "text" @@ -184787,59 +175007,57 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131000 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 2.5, + "output": 2.5 } }, - "minimax/minimax-h3-max": { - "id": "minimax/minimax-h3-max", - "name": "MiniMax H3 Max", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "minimax", - "attachment": true, + "jamba-mini": { + "id": "jamba-mini", + "name": "Jamba Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "structured_output": false, + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "input": 256000, + "output": 4096 + }, + "cost": { + "input": 0.1989, + "output": 0.408 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "claude-opus-4-1-thinking:32000": { + "id": "claude-opus-4-1-thinking:32000", + "name": "Claude 4.1 Opus Thinking (32K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "structured_output": true, + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", @@ -184850,29 +175068,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512000, - "output": 512000 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 14.994, + "output": 75.004 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax M2.7 High Speed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "KAT-Coder-Air-V1": { + "id": "KAT-Coder-Air-V1", + "name": "KAT Coder Air V1", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-10-28", + "last_updated": "2025-10-28", "modalities": { "input": [ "text" @@ -184881,97 +175097,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131100 + "context": 128000, + "input": 128000, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.1, + "output": 0.2 } }, - "minimax/minimax-m2.5-highspeed": { - "id": "minimax/minimax-m2.5-highspeed", - "name": "MiniMax M2.5 High Speed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "exa-research": { + "id": "exa-research", + "name": "Exa (Research)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131000 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-10-27", - "last_updated": "2025-10-27", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-04", + "last_updated": "2025-06-04", "modalities": { "input": [ "text" @@ -184980,32 +175126,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 205000, - "output": 205000 + "context": 8192, + "input": 8192, + "output": 8192 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 2.5, + "output": 2.5 } }, - "stepfun/step-5-preview": { - "id": "stepfun/step-5-preview", - "name": "Step 5 Preview", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "deepclaude": { + "id": "deepclaude", + "name": "DeepClaude", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, "tool_call": false, - "structured_output": true, - "release_date": "2026-09-16", - "last_updated": "2026-09-20", + "structured_output": false, + "release_date": "2025-02-01", + "last_updated": "2025-02-01", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" @@ -185013,26 +175158,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 1, - "output": 2.7, - "cache_read": 0.05 + "input": 3, + "output": 15 } }, - "stepfun/step-3.5-flash": { - "id": "stepfun/step-3.5-flash", - "name": "StepFun 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", - "attachment": false, + "gemini-2.5-flash-preview-09-2025": { + "id": "gemini-2.5-flash-preview-09-2025", + "name": "Gemini 2.5 Flash Preview (09/2025)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "none", "low", "medium", "high" @@ -185040,83 +175188,116 @@ } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", + "structured_output": true, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262114, - "output": 262114 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 0.09, - "output": 0.3, - "cache_read": 0.02 + "input": 0.3, + "output": 2.5 } }, - "stepfun/step-3.7-flash": { - "id": "stepfun/step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "claude-opus-4-5-20251101": { + "id": "claude-opus-4-5-20251101", + "name": "Claude 4.5 Opus", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", "high" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", + "structured_output": true, + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 + "input": 4.998, + "output": 25.007 } }, - "inception/mercury-coder-small": { - "id": "inception/mercury-coder-small", - "name": "Mercury Coder Small Beta", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "mercury", + "GLM-4.6-Derestricted-v5": { + "id": "GLM-4.6-Derestricted-v5", + "name": "GLM 4.6 Derestricted v5", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "input": 131072, + "output": 8192 + }, + "cost": { + "input": 0.4, + "output": 1.5 + } + }, + "glm-z1-airx": { + "id": "glm-z1-airx", + "name": "GLM Z1 AirX", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2025-02-26", - "last_updated": "2025-02-26", + "structured_output": true, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text" @@ -185128,37 +175309,61 @@ "open_weights": false, "limit": { "context": 32000, + "input": 32000, "output": 16384 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0.7, + "output": 0.7 } }, - "inception/mercury-2.5": { - "id": "inception/mercury-2.5", - "name": "Mercury 2.5", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", - "attachment": false, + "claude-sonnet-4-thinking:1024": { + "id": "claude-sonnet-4-thinking:1024", + "name": "Claude 4 Sonnet Thinking (1K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "structured_output": true, + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ + "text", + "image", + "pdf" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 1000000, + "output": 64000 + }, + "cost": { + "input": 2.992, + "output": 14.994 + } + }, + "holo3-35b-a3b:thinking": { + "id": "holo3-35b-a3b:thinking", + "name": "Holo3-35B-A3B Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" @@ -185166,36 +175371,25 @@ }, "open_weights": false, "limit": { - "context": 260000, + "context": 65536, + "input": 65536, "output": 65536 }, "cost": { - "input": 0.04, - "output": 0.15, - "cache_read": 0.004 + "input": 0.25, + "output": 1.8 } }, - "inception/mercury-2": { - "id": "inception/mercury-2", - "name": "Mercury 2", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", + "owl": { + "id": "owl", + "name": "OWL", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-02-24", - "last_updated": "2026-03-06", + "structured_output": true, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ "text" @@ -185206,26 +175400,25 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 1048756, + "input": 1048756, + "output": 262144 }, "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.024999999999999998 + "input": 0.1, + "output": 0.3 } }, - "amazon/titan-embed-text-v2": { - "id": "amazon/titan-embed-text-v2", - "name": "Titan Text Embeddings V2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "titan-embed", + "gemini-2.0-pro-reasoner": { + "id": "gemini-2.0-pro-reasoner", + "name": "Gemini 2.0 Pro Reasoner", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-04-30", - "last_updated": "2024-04", + "structured_output": false, + "release_date": "2025-02-05", + "last_updated": "2025-02-05", "modalities": { "input": [ "text" @@ -185236,40 +175429,35 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 + "context": 128000, + "input": 128000, + "output": 65536 + }, + "cost": { + "input": 1.292, + "output": 4.998 } }, - "amazon/nova-2-lite": { - "id": "amazon/nova-2-lite", - "name": "Nova 2 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova", + "Gemma-4-31B-Cognitive-Unshackled": { + "id": "Gemma-4-31B-Cognitive-Unshackled", + "name": "Gemma 4 31B Cognitive Unshackled", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] } ], "tool_call": false, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-02", - "last_updated": "2024-12-01", + "structured_output": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -185277,33 +175465,40 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 1000000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075 + "input": 0.306, + "output": 0.306 } }, - "amazon/nova-lite": { - "id": "amazon/nova-lite", - "name": "Nova Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", + "Qwen3.5-27B-earica-Derestricted": { + "id": "Qwen3.5-27B-earica-Derestricted", + "name": "Qwen3.5 27B earica Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", "image", - "video", - "pdf" + "video" ], "output": [ "text" @@ -185311,34 +175506,39 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 10000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.015, - "cache_write": 0.06 + "input": 0.306, + "output": 0.306 } }, - "amazon/nova-pro": { - "id": "amazon/nova-pro", - "name": "Nova Pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", + "command-a-plus-05-2026": { + "id": "command-a-plus-05-2026", + "name": "Cohere Command A+ (05/2026)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": false, + "structured_output": true, + "release_date": "2026-05-22", + "last_updated": "2026-05-22", "modalities": { "input": [ "text", - "image", - "video", - "pdf" + "image" ], "output": [ "text" @@ -185346,28 +175546,25 @@ }, "open_weights": false, "limit": { - "context": 300000, - "output": 10000 + "context": 128000, + "input": 128000, + "output": 64000 }, "cost": { - "input": 0.8, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0.8 + "input": 2.5, + "output": 10 } }, - "amazon/nova-micro": { - "id": "amazon/nova-micro", - "name": "Nova Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", + "auto-model-premium": { + "id": "auto-model-premium", + "name": "Auto model (Premium)", + "description": "Automatic model router for matching prompts to suitable backends and budgets", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", + "tool_call": false, + "structured_output": false, + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ "text" @@ -185378,27 +175575,25 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 10000 + "context": 1000000, + "input": 1000000, + "output": 1000000 }, "cost": { - "input": 0.035, - "output": 0.14, - "cache_read": 0.00875, - "cache_write": 0.035 + "input": 9.996, + "output": 19.992 } }, - "mistral/codestral-embed": { - "id": "mistral/codestral-embed", - "name": "Codestral Embed", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "codestral-embed", + "learnlm-1.5-pro-experimental": { + "id": "learnlm-1.5-pro-experimental", + "name": "Gemini LearnLM Experimental", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "structured_output": false, + "release_date": "2024-05-14", + "last_updated": "2024-05-14", "modalities": { "input": [ "text" @@ -185409,22 +175604,25 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 + "context": 32767, + "input": 32767, + "output": 8192 + }, + "cost": { + "input": 3.502, + "output": 10.506 } }, - "mistral/mistral-nemo": { - "id": "mistral/mistral-nemo", - "name": "Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", + "deepseek-r1-sambanova": { + "id": "deepseek-r1-sambanova", + "name": "DeepSeek R1 Fast", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-07-18", - "last_updated": "2024-07-01", + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-20", + "last_updated": "2025-02-20", "modalities": { "input": [ "text" @@ -185435,38 +175633,29 @@ }, "open_weights": false, "limit": { - "context": 60288, - "output": 16000 + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 0.04, - "output": 0.17 + "input": 4.998, + "output": 6.987 } }, - "mistral/mistral-medium-3.5": { - "id": "mistral/mistral-medium-3.5", - "name": "Mistral Medium Latest", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", + "claw-medium": { + "id": "claw-medium", + "name": "Claw Medium", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-05-21", + "structured_output": true, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -185474,63 +175663,59 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 256000 + "context": 204800, + "input": 204800, + "output": 131072 }, "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15 + "input": 0.3, + "output": 1.2 } }, - "mistral/mistral-small": { - "id": "mistral/mistral-small", - "name": "Mistral Small (latest)", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", + "claude-opus-4-20250514": { + "id": "claude-opus-4-20250514", + "name": "Claude 4 Opus", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2024-09-17", - "last_updated": "2026-03-16", + "structured_output": true, + "release_date": "2025-05-14", + "last_updated": "2025-05-14", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 4000 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 14.994, + "output": 75.004 } }, - "mistral/mistral-large-3": { - "id": "mistral/mistral-large-3", - "name": "Mistral Large 3", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, + "yi-large": { + "id": "yi-large", + "name": "Yi Large", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", + "structured_output": false, + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -185538,26 +175723,25 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 256000 + "context": 32000, + "input": 32000, + "output": 4096 }, "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.05 + "input": 3.196, + "output": 3.196 } }, - "mistral/mistral-embed": { - "id": "mistral/mistral-embed", - "name": "Mistral Embed", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "mistral-embed", + "qwen3-max-2026-01-23": { + "id": "qwen3-max-2026-01-23", + "name": "Qwen3 Max 2026-01-23", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2023-12-11", - "last_updated": "2023-12-11", + "structured_output": false, + "release_date": "2026-01-26", + "last_updated": "2026-01-26", "modalities": { "input": [ "text" @@ -185568,22 +175752,54 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1536 + "context": 256000, + "input": 256000, + "output": 32768 + }, + "cost": { + "input": 1.2002, + "output": 6.001 } }, - "mistral/ministral-14b": { - "id": "mistral/ministral-14b", - "name": "Ministral 14B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "phi-4-mini-instruct": { + "id": "phi-4-mini-instruct", + "name": "Phi 4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 128000, + "output": 16384 + }, + "cost": { + "input": 0.17, + "output": 0.68 + } + }, + "ernie-x1-turbo-32k": { + "id": "ernie-x1-turbo-32k", + "name": "Ernie X1 Turbo 32k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-02", - "last_updated": "2025-12-01", + "structured_output": false, + "release_date": "2025-05-08", + "last_updated": "2025-05-08", "modalities": { "input": [ "text", @@ -185596,163 +175812,198 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 256000 + "context": 32000, + "input": 32000, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.02 + "input": 0.165, + "output": 0.66 } }, - "mistral/codestral": { - "id": "mistral/codestral", - "name": "Codestral (latest)", - "description": "Mistral code model for completions, refactors, and developer IDE workflows", - "family": "codestral", - "attachment": false, - "reasoning": false, + "claw-low": { + "id": "claw-low", + "name": "Claw Low", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-05-29", - "last_updated": "2025-01-04", + "structured_output": true, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 4096 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 0.9 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025 } }, - "mistral/ministral-8b": { - "id": "mistral/ministral-8b", - "name": "Ministral 8B (latest)", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, + "gemini-3-pro-image-preview": { + "id": "gemini-3-pro-image-preview", + "name": "Gemini 3 Pro Image", + "description": "Image model for prompt-driven generation, editing, and visual design workflows", + "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-04", + "tool_call": false, + "structured_output": false, + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 0.1, - "output": 0.1 + "input": 2, + "output": 12 } }, - "mistral/ministral-3b": { - "id": "mistral/ministral-3b", - "name": "Ministral 3B (latest)", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-04", + "gemma-4-31B-Garnet": { + "id": "gemma-4-31B-Garnet", + "name": "Gemma 4 31B Garnet", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-02", + "last_updated": "2026-05-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.04, - "output": 0.04 + "input": 0.306, + "output": 0.306 } }, - "fish-audio/s2.1-pro": { - "id": "fish-audio/s2.1-pro", - "name": "S2.1 Pro", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "attachment": false, + "doubao-1.5-vision-pro-32k": { + "id": "doubao-1.5-vision-pro-32k", + "name": "Doubao 1.5 Vision Pro 32k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": false, - "release_date": "2026-07-28", - "last_updated": "2026-07-28", + "structured_output": false, + "release_date": "2025-01-22", + "last_updated": "2025-01-22", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 32000, + "input": 32000, + "output": 8192 + }, + "cost": { + "input": 0.459, + "output": 1.377 } }, - "fish-audio/s1": { - "id": "fish-audio/s1", - "name": "S1", - "description": "Speech generation model for controllable voice, narration, and audio delivery", + "auto-model-standard": { + "id": "auto-model-standard", + "name": "Auto model (Standard)", + "description": "Automatic model router for matching prompts to suitable backends and budgets", "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2025-10-20", - "last_updated": "2025-10-20", + "structured_output": false, + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "input": 1000000, + "output": 1000000 + }, + "cost": { + "input": 9.996, + "output": 19.992 } }, - "fish-audio/transcribe-1": { - "id": "fish-audio/transcribe-1", - "name": "Transcribe-1", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "attachment": false, - "reasoning": false, + "Qwen3.5-27B-Marvin-DPO-V2-Derestricted-Lite": { + "id": "Qwen3.5-27B-Marvin-DPO-V2-Derestricted-Lite", + "name": "Qwen3.5 27B Marvin DPO V2 Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], "tool_call": false, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "audio" + "text", + "image", + "video" ], "output": [ "text" @@ -185760,47 +176011,69 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "input": 262144, + "output": 16384 + }, + "cost": { + "input": 0.306, + "output": 0.306 } }, - "fish-audio/s2-pro": { - "id": "fish-audio/s2-pro", - "name": "S2 Pro", - "description": "Speech generation model for controllable voice, narration, and audio delivery", + "glm-4": { + "id": "glm-4", + "name": "GLM-4", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", + "structured_output": false, + "release_date": "2024-01-16", + "last_updated": "2024-01-16", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 4096 + }, + "cost": { + "input": 14.994, + "output": 14.994 } }, - "morph/morph-v3-fast": { - "id": "morph/morph-v3-fast", - "name": "Morph v3 Fast", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "morph", - "attachment": false, - "reasoning": false, + "Qwen3.5-27B-Writer-Derestricted-Lite": { + "id": "Qwen3.5-27B-Writer-Derestricted-Lite", + "name": "Qwen3.5 27B Writer Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], "tool_call": false, - "temperature": false, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", + "structured_output": false, + "release_date": "2026-04-06", + "last_updated": "2026-04-06", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -185808,28 +176081,31 @@ }, "open_weights": false, "limit": { - "context": 16000, - "output": 16000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.8, - "output": 1.2 + "input": 0.306, + "output": 0.306 } }, - "morph/morph-v3-large": { - "id": "morph/morph-v3-large", - "name": "Morph v3 Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "morph", + "qwen-3.6-plus": { + "id": "qwen-3.6-plus", + "name": "Qwen 3.6 Plus", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "qwen3.6", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", + "structured_output": false, + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -185837,32 +176113,27 @@ }, "open_weights": false, "limit": { - "context": 32000, - "output": 32000 + "context": 991800, + "output": 65536 }, "cost": { - "input": 0.9, - "output": 1.9 + "input": 0.45, + "output": 2.7 } }, - "openai/gpt-4.1-mini-fast": { - "id": "openai/gpt-4.1-mini-fast", - "name": "GPT-4.1 mini (Fast)", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, + "brave-pro": { + "id": "brave-pro", + "name": "Brave (Pro)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "tool_call": false, + "structured_output": false, + "release_date": "2023-03-02", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -185870,43 +176141,28 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "input": 1014808, - "output": 32768 + "context": 8192, + "input": 8192, + "output": 8192 }, "cost": { - "input": 0.7, - "output": 2.8, - "cache_read": 0.175 + "input": 5, + "output": 5 } }, - "openai/o3-fast": { - "id": "openai/o3-fast", - "name": "o3 (Fast)", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "deepseek-chat-cheaper": { + "id": "deepseek-chat-cheaper", + "name": "DeepSeek V3/Chat Cheaper", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text", - "image", "pdf" ], "output": [ @@ -185915,46 +176171,35 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 100000, - "output": 100000 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 3.5, - "output": 14, - "cache_read": 0.875 - } + "input": 0.25, + "output": 0.7 + } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT 5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", + "Gemma-4-31B-Queen": { + "id": "Gemma-4-31B-Queen", + "name": "Gemma 4 31B Queen", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -185962,60 +176207,35 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT 5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", + "Gemma-4-31B-Gemopus": { + "id": "Gemma-4-31B-Gemopus", + "name": "Gemma 4 31B Gemopus", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], - "tool_call": true, + "tool_call": false, "structured_output": false, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -186023,83 +176243,98 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-oss-safeguard-120b": { - "id": "openai/gpt-oss-safeguard-120b", - "name": "GPT OSS Safeguard 120B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", + "mistral-code-latest": { + "id": "mistral-code-latest", + "name": "Mistral Code Latest", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "input": 256000, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 0.9 + } + }, + "gemini-2.5-flash-lite": { + "id": "gemini-2.5-flash-lite", + "name": "Gemini 2.5 Flash Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "none", "low", "medium", "high" ] } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 112000, - "output": 16000 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.1, + "output": 0.4 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5 Turbo", + "jamba-mini-1.7": { + "id": "jamba-mini-1.7", + "name": "Jamba Mini 1.7", "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", "attachment": false, "reasoning": false, "tool_call": false, "structured_output": false, - "temperature": true, - "knowledge": "2021-09", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "release_date": "2025-07-09", + "last_updated": "2025-07-09", "modalities": { "input": [ "text" @@ -186110,45 +176345,28 @@ }, "open_weights": false, "limit": { - "context": 16385, - "input": 12289, + "context": 256000, + "input": 256000, "output": 4096 }, "cost": { - "input": 0.5, - "output": 1.5 + "input": 0.1989, + "output": 0.408 } }, - "openai/gpt-5.2-fast": { - "id": "openai/gpt-5.2-fast", - "name": "GPT 5.2 (Fast)", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "universal-summarizer": { + "id": "universal-summarizer", + "name": "Universal Summarizer", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2023-05-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -186156,30 +176374,40 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 3.5, - "output": 28, - "cache_read": 0.35 + "input": 30, + "output": 30 } }, - "openai/gpt-4o-transcribe": { - "id": "openai/gpt-4o-transcribe", - "name": "GPT-4o Transcribe", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "gpt", - "attachment": false, - "reasoning": false, + "Qwen3.5-27B-Anko": { + "id": "Qwen3.5-27B-Anko", + "name": "Qwen3.5 27B Anko", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], "tool_call": false, - "temperature": true, - "release_date": "2024-03-13", - "last_updated": "2024-03-13", + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "audio" + "text", + "image", + "video" ], "output": [ "text" @@ -186187,37 +176415,25 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 2.5, - "output": 10 + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT 5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", + "claude-sonnet-4-20250514": { + "id": "claude-sonnet-4-20250514", + "name": "Claude 4 Sonnet", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", @@ -186230,40 +176446,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 872000, - "output": 128000 + "context": 200000, + "input": 200000, + "output": 64000 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 2.992, + "output": 14.994 } }, - "openai/text-embedding-3-small": { - "id": "openai/text-embedding-3-small", - "name": "text-embedding-3-small", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "ernie-x1-32k-preview": { + "id": "ernie-x1-32k-preview", + "name": "Ernie X1 32k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "structured_output": false, + "release_date": "2025-04-03", + "last_updated": "2025-04-03", "modalities": { "input": [ "text" @@ -186274,41 +176475,28 @@ }, "open_weights": false, "limit": { - "context": 8192, - "input": 6656, - "output": 1536 + "context": 32000, + "input": 32000, + "output": 16384 + }, + "cost": { + "input": 0.33, + "output": 1.32 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT 5.4 Nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "azure-gpt-4-turbo": { + "id": "azure-gpt-4-turbo", + "name": "Azure gpt-4-turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2023-11-06", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -186316,87 +176504,55 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 9.996, + "output": 30.005 } }, - "openai/gpt-realtime-2.1": { - "id": "openai/gpt-realtime-2.1", - "name": "gpt-realtime-2.1", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", + "Meta-Llama-3-1-8B-Instruct-FP8": { + "id": "Meta-Llama-3-1-8B-Instruct-FP8", + "name": "Llama 3.1 8B (decentralized)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": false, "structured_output": false, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "audio" + "text" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { "context": 128000, - "input": 96000, - "output": 32000 + "input": 128000, + "output": 16384 }, "cost": { - "input": 4, - "output": 24, - "cache_read": 0.4 + "input": 0.02, + "output": 0.03 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2-Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", + "claude-sonnet-4-5-20250929-thinking": { + "id": "claude-sonnet-4-5-20250929-thinking", + "name": "Claude Sonnet 4.5 Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", @@ -186409,33 +176565,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "input": 1000000, + "output": 64000 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 2.992, + "output": 14.994 } }, - "openai/gpt-4o-fast": { - "id": "openai/gpt-4o-fast", - "name": "GPT-4o (Fast)", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "asi1-mini": { + "id": "asi1-mini", + "name": "ASI1 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "tool_call": false, + "structured_output": false, + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ "text", - "image", "pdf" ], "output": [ @@ -186445,43 +176596,29 @@ "open_weights": false, "limit": { "context": 128000, - "input": 111616, + "input": 128000, "output": 16384 }, "cost": { - "input": 4.25, - "output": 17, - "cache_read": 2.125 + "input": 1, + "output": 1 } }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1-Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", + "qwen3.5-27b": { + "id": "qwen3.5-27b", + "name": "Qwen3.5 27B", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -186489,87 +176626,67 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 260096, + "input": 260096, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 - } - }, - "openai/gpt-live-1": { - "id": "openai/gpt-live-1", - "name": "GPT-Live 1", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "input": 0.27, + "output": 2.16 } }, - "openai/gpt-realtime-1.5": { - "id": "openai/gpt-realtime-1.5", - "name": "GPT-Realtime-1.5", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", - "attachment": false, - "reasoning": false, + "Qwen3.5-27B-Omega-Evolution-v2.2-Derestricted-Lite": { + "id": "Qwen3.5-27B-Omega-Evolution-v2.2-Derestricted-Lite", + "name": "Qwen3.5 27B Omega Evolution v2.2 Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], "tool_call": false, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "structured_output": false, + "release_date": "2026-05-02", + "last_updated": "2026-05-02", "modalities": { "input": [ "text", - "audio" + "image", + "video" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 4, - "output": 16, - "cache_read": 0.4 + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-4.1-nano-fast": { - "id": "openai/gpt-4.1-nano-fast", - "name": "GPT-4.1 nano (Fast)", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", + "claude-opus-4-1-thinking:32768": { + "id": "claude-opus-4-1-thinking:32768", + "name": "Claude 4.1 Opus Thinking (32K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", @@ -186582,44 +176699,35 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "input": 1014808, - "output": 32768 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.2, - "output": 0.8, - "cache_read": 0.05 + "input": 14.994, + "output": 75.004 } }, - "openai/gpt-5-codex": { - "id": "openai/gpt-5-codex", - "name": "GPT-5-Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "gemma-4-31B-K1-v5": { + "id": "gemma-4-31B-K1-v5", + "name": "Gemma 4 31B K1 v5", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-02", + "last_updated": "2026-05-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -186627,77 +176735,58 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-image-2": { - "id": "openai/gpt-image-2", - "name": "GPT Image 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "sonar": { + "id": "sonar", + "name": "Perplexity Simple", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "structured_output": false, + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 127000, + "input": 127000, + "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 1.25 + "input": 1.003, + "output": 1.003 } }, - "openai/gpt-5.6-luna-fast": { - "id": "openai/gpt-5.6-luna-fast", - "name": "GPT 5.6 Luna (Fast)", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, + "MiniMax-M2": { + "id": "MiniMax-M2", + "name": "MiniMax M2", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-10-25", + "last_updated": "2025-10-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -186705,65 +176794,71 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "input": 200000, + "output": 131072 }, "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.04, - "cache_write": 0.5, - "tiers": [ - { - "input": 0.8, - "output": 3.6, - "cache_read": 0.08, - "cache_write": 1, - "tier": { - "type": "context", - "size": 272000 - } - } + "input": 0.17, + "output": 1.53 + } + }, + "sarvam-105b": { + "id": "sarvam-105b", + "name": "Sarvam 105B", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": false, + "release_date": "2026-05-12", + "last_updated": "2026-05-12", + "modalities": { + "input": [ + "text" ], - "context_over_200k": { - "input": 0.8, - "output": 3.6, - "cache_read": 0.08, - "cache_write": 1 - } + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "input": 131072, + "output": 4096 + }, + "cost": { + "input": 0.045, + "output": 0.177, + "cache_read": 0.028 } }, - "openai/gpt-6-astra-fast": { - "id": "openai/gpt-6-astra-fast", - "name": "GPT-6 Astra (Fast)", - "description": "Fast variant of GPT-6 Astra for low-latency assistance and high-volume workloads.", - "family": "gpt-astra", + "Qwen3.5-27B-Writer-V2-Derestricted-Lite": { + "id": "Qwen3.5-27B-Writer-V2-Derestricted-Lite", + "name": "Qwen3.5 27B Writer V2 Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-06", + "last_updated": "2026-04-06", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -186771,63 +176866,40 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 20, - "output": 100, - "cache_read": 2, - "cache_write": 25, - "tiers": [ - { - "input": 40, - "output": 150, - "cache_read": 4, - "cache_write": 50, - "tier": { - "type": "context", - "size": 272001 - } - } - ], - "context_over_200k": { - "input": 40, - "output": 150, - "cache_read": 4, - "cache_write": 50 - } + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT 5.2 ", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", + "Qwen3.5-27B-Marvin-V2-Derestricted": { + "id": "Qwen3.5-27B-Marvin-V2-Derestricted", + "name": "Qwen3.5 27B Marvin V2 Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, + "tool_call": false, "structured_output": false, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -186835,26 +176907,26 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 21, - "output": 168 + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-5.1-thinking-fast": { - "id": "openai/gpt-5.1-thinking-fast", - "name": "GPT 5.1 Thinking (Fast)", + "claw-high": { + "id": "claw-high", + "name": "Claw High", "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "release_date": "2025-11-12", - "last_updated": "2025-11-12", + "structured_output": true, + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ "text", @@ -186867,27 +176939,25 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 1000000, + "input": 1000000, "output": 128000 }, "cost": { - "input": 2.5, - "output": 20, - "cache_read": 0.25 + "input": 4.998, + "output": 25.007 } }, - "openai/text-embedding-ada-002": { - "id": "openai/text-embedding-ada-002", - "name": "text-embedding-ada-002", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "qwen-max": { + "id": "qwen-max", + "name": "Qwen 2.5 Max", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2022-12-15", - "last_updated": "2022-12-15", + "structured_output": false, + "release_date": "2024-04-03", + "last_updated": "2024-04-03", "modalities": { "input": [ "text" @@ -186898,40 +176968,35 @@ }, "open_weights": false, "limit": { - "context": 8192, - "input": 6656, - "output": 1536 + "context": 32000, + "input": 32000, + "output": 8192 + }, + "cost": { + "input": 1.5997, + "output": 6.392 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT 5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "gemma-4-31B-Fabled": { + "id": "gemma-4-31B-Fabled", + "name": "Gemma 4 31B Fabled", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-02", + "last_updated": "2026-05-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -186939,30 +177004,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.306, + "output": 0.306 } }, - "openai/whisper-1": { - "id": "openai/whisper-1", - "name": "Whisper", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", + "v0-1.5-md": { + "id": "v0-1.5-md", + "name": "v0 1.5 MD", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2022-09-21", - "last_updated": "2022-09-21", + "structured_output": false, + "release_date": "2025-07-04", + "last_updated": "2025-07-04", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" @@ -186970,34 +177033,26 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 200000, + "input": 200000, + "output": 64000 + }, + "cost": { + "input": 3, + "output": 15 } }, - "openai/gpt-5.3-codex-fast": { - "id": "openai/gpt-5.3-codex-fast", - "name": "GPT 5.3 Codex (Fast)", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "claude-opus-4-thinking:1024": { + "id": "claude-opus-4-thinking:1024", + "name": "Claude 4 Opus Thinking (1K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", @@ -187010,46 +177065,43 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 3.5, - "output": 28, - "cache_read": 0.35 + "input": 14.994, + "output": 75.004 } }, - "openai/gpt-5.5-fast": { - "id": "openai/gpt-5.5-fast", - "name": "GPT 5.5 (Fast)", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", + "gemini-2.5-flash-preview-04-17": { + "id": "gemini-2.5-flash-preview-04-17", + "name": "Gemini 2.5 Flash Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "none", "low", "medium", - "high", - "xhigh" + "high" ] } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-17", + "last_updated": "2025-04-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -187057,76 +177109,69 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 872000, - "output": 128000 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 + "input": 0.15, + "output": 0.6 } }, - "openai/gpt-image-2.5-flare": { - "id": "openai/gpt-image-2.5-flare", - "name": "GPT Image 2.5 Flare", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": false, - "reasoning": false, + "gemini-2.5-pro-exp-03-25": { + "id": "gemini-2.5-pro-exp-03-25", + "name": "Gemini 2.5 Pro Experimental 0325", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": false, - "temperature": false, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "structured_output": false, + "release_date": "2025-03-25", + "last_updated": "2025-03-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 1.25 + "input": 2.5, + "output": 10 } }, - "openai/gpt-5.4-fast": { - "id": "openai/gpt-5.4-fast", - "name": "GPT 5.4 (Fast)", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, + "hermes-medium": { + "id": "hermes-medium", + "name": "Hermes Medium", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2026-05-11", + "last_updated": "2026-05-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -187134,46 +177179,57 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 204800, + "input": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2 + } + }, + "sonar-pro": { + "id": "sonar-pro", + "name": "Perplexity Pro", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-19", + "last_updated": "2025-02-19", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "input": 200000, "output": 128000 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 + "input": 2.992, + "output": 14.994 } }, - "openai/gpt-5.6-sol-fast": { - "id": "openai/gpt-5.6-sol-fast", - "name": "GPT 5.6 Sol (Fast)", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", + "doubao-1-5-thinking-pro-250415": { + "id": "doubao-1-5-thinking-pro-250415", + "name": "Doubao 1.5 Thinking Pro", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-17", + "last_updated": "2025-04-17", "modalities": { "input": [ "text", - "image", "pdf" ], "output": [ @@ -187182,63 +177238,35 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 8, - "output": 40, - "cache_read": 0.8, - "cache_write": 10, - "tiers": [ - { - "input": 16, - "output": 60, - "cache_read": 1.6, - "cache_write": 20, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 16, - "output": 60, - "cache_read": 1.6, - "cache_write": 20 - } + "input": 0.6, + "output": 2.4 } }, - "openai/o4-mini-fast": { - "id": "openai/o4-mini-fast", - "name": "o4-mini (Fast)", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, + "qwen3.7-max:thinking": { + "id": "qwen3.7-max:thinking", + "name": "Qwen3.7 Max Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024, + "max": 262144 } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-21", + "last_updated": "2026-05-21", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -187246,42 +177274,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 100000, - "output": 100000 + "context": 1000000, + "input": 1000000, + "output": 65536 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 2.5, + "output": 7.5, + "cache_read": 0.25 } }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5 pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", + "ernie-4.5-turbo-vl-32k": { + "id": "ernie-4.5-turbo-vl-32k", + "name": "Ernie 4.5 Turbo VL 32k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-08", + "last_updated": "2025-05-08", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -187289,72 +177305,65 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 128000, - "output": 272000 + "context": 32000, + "input": 32000, + "output": 16384 }, "cost": { - "input": 15, - "output": 120 + "input": 0.495, + "output": 1.43 } }, - "openai/gpt-5.1-thinking": { - "id": "openai/gpt-5.1-thinking", - "name": "GPT 5.1 Thinking", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", + "gemini-2.5-pro-preview-05-06": { + "id": "gemini-2.5-pro-preview-05-06", + "name": "Gemini 2.5 Pro Preview 0506", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" ] } ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-11-12", - "last_updated": "2025-08-07", + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-06", + "last_updated": "2025-05-06", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 2.5, + "output": 10 } }, - "openai/text-embedding-3-large": { - "id": "openai/text-embedding-3-large", - "name": "text-embedding-3-large", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "hunyuan-turbos-20250226": { + "id": "hunyuan-turbos-20250226", + "name": "Hunyuan Turbo S", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", + "structured_output": false, + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ "text" @@ -187365,37 +177374,26 @@ }, "open_weights": false, "limit": { - "context": 8192, - "input": 6656, - "output": 1536 + "context": 24000, + "input": 24000, + "output": 8192 + }, + "cost": { + "input": 0.187, + "output": 0.374 } }, - "openai/gpt-6-astra": { - "id": "openai/gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", + "claude-sonnet-4-thinking:64000": { + "id": "claude-sonnet-4-thinking:64000", + "name": "Claude 4 Sonnet Thinking (64K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", @@ -187408,65 +177406,29 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "input": 1000000, + "output": 64000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272001 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } + "input": 2.992, + "output": 14.994 } }, - "openai/gpt-5.4-mini-fast": { - "id": "openai/gpt-5.4-mini-fast", - "name": "GPT 5.4 Mini (Fast)", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "ernie-x1-32k": { + "id": "ernie-x1-32k", + "name": "Ernie X1 32k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-08", + "last_updated": "2025-05-08", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -187474,77 +177436,57 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 32000, + "input": 32000, + "output": 16384 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 + "input": 0.33, + "output": 1.32 } }, - "openai/gpt-realtime-mini": { - "id": "openai/gpt-realtime-mini", - "name": "GPT-Realtime mini", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", + "command-a-reasoning-08-2025": { + "id": "command-a-reasoning-08-2025", + "name": "Cohere Command A (08/2025)", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-10-10", - "last_updated": "2025-10-10", + "structured_output": false, + "release_date": "2025-08-22", + "last_updated": "2025-08-22", "modalities": { "input": [ - "text", - "audio" + "text" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "input": 256000, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06 + "input": 2.5, + "output": 10 } }, - "openai/gpt-5-mini-fast": { - "id": "openai/gpt-5-mini-fast", - "name": "GPT-5 mini (Fast)", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "doubao-seed-2-0-code-preview-260215": { + "id": "doubao-seed-2-0-code-preview-260215", + "name": "Doubao Seed 2.0 Code Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-14", + "last_updated": "2026-02-14", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -187552,44 +177494,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 256000, + "input": 256000, "output": 128000 }, "cost": { - "input": 0.45, - "output": 3.6, - "cache_read": 0.045 + "input": 0.782, + "output": 3.893 } }, - "openai/o3-pro": { - "id": "openai/o3-pro", - "name": "o3 Pro", - "description": "High-effort o3 tier for difficult technical reasoning and careful answers", - "family": "o-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-06-10", - "last_updated": "2025-06-10", + "doubao-1.5-pro-256k": { + "id": "doubao-1.5-pro-256k", + "name": "Doubao 1.5 Pro 256k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-03-12", + "last_updated": "2025-03-12", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -187597,153 +177523,123 @@ }, "open_weights": false, "limit": { - "context": 200000, - "input": 100000, - "output": 100000 + "context": 256000, + "input": 256000, + "output": 16384 }, "cost": { - "input": 20, - "output": 80 + "input": 0.799, + "output": 1.445 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "qwen3.5-35b-a3b": { + "id": "qwen3.5-35b-a3b", + "name": "Qwen3.5 35B A3B", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "input": 122880, - "output": 8192 + "context": 260096, + "input": 260096, + "output": 65536 }, "cost": { - "input": 0.03, - "output": 0.14 + "input": 0.225, + "output": 1.8 } }, - "openai/gpt-oss-safeguard-20b": { - "id": "openai/gpt-oss-safeguard-20b", - "name": "gpt-oss-safeguard-20b", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, + "qwen3.5-122b-a10b:thinking": { + "id": "qwen3.5-122b-a10b:thinking", + "name": "Qwen3.5 122B A10B Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "input": 112000, - "output": 16000 + "context": 260096, + "input": 260096, + "output": 65536 }, "cost": { - "input": 0.07, - "output": 0.2 + "input": 0.36, + "output": 2.88 } }, - "openai/gpt-image-2.5-sunburst": { - "id": "openai/gpt-image-2.5-sunburst", - "name": "GPT Image 2.5 Sunburst", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "yi-lightning": { + "id": "yi-lightning", + "name": "Yi Lightning", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-09-08", - "last_updated": "2026-09-08", + "structured_output": false, + "release_date": "2024-10-16", + "last_updated": "2024-10-16", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 12000, + "input": 12000, + "output": 4096 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 1.25 + "input": 0.2006, + "output": 0.2006 } }, - "openai/gpt-5.6-terra-fast": { - "id": "openai/gpt-5.6-terra-fast", - "name": "GPT 5.6 Terra (Fast)", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", + "claude-sonnet-4-5-20250929": { + "id": "claude-sonnet-4-5-20250929", + "name": "Claude Sonnet 4.5", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", @@ -187756,90 +177652,55 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "input": 1000000, + "output": 64000 }, "cost": { - "input": 4, - "output": 24, - "cache_read": 0.4, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 36, - "cache_read": 0.8, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 36, - "cache_read": 0.8, - "cache_write": 10 - } + "input": 2.992, + "output": 14.994 } }, - "openai/gpt-image-1": { - "id": "openai/gpt-image-1", - "name": "GPT Image 1", - "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", - "family": "gpt-image", + "deepseek-math-v2": { + "id": "deepseek-math-v2", + "name": "DeepSeek Math V2", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-04-24", - "last_updated": "2025-04-24", + "structured_output": false, + "release_date": "2025-12-03", + "last_updated": "2025-12-03", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 5, - "output": 40, - "cache_read": 1.25 + "input": 0.6, + "output": 2.2 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT 5.4 Mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", + "claude-opus-4-1-thinking:8192": { + "id": "claude-opus-4-1-thinking:8192", + "name": "Claude 4.1 Opus Thinking (8K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ "text", @@ -187852,47 +177713,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 14.994, + "output": 75.004 } }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT 5.6 Luna", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "deepseek-reasoner": { + "id": "deepseek-reasoner", + "name": "DeepSeek Reasoner", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -187900,59 +177742,25 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 64000, + "input": 64000, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } + "input": 0.4, + "output": 1.7 } }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT 5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "gemini-2.5-flash-nothinking": { + "id": "gemini-2.5-flash-nothinking", + "name": "Gemini 2.5 Flash (No Thinking)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ "text", @@ -187965,46 +177773,40 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.3, + "output": 2.5 } }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", + "Qwen3.5-27B-Omega-Evolution-v2.0-Derestricted-Lite": { + "id": "Qwen3.5-27B-Omega-Evolution-v2.0-Derestricted-Lite", + "name": "Qwen3.5 27B Omega Evolution v2.0 Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-06", + "last_updated": "2026-04-06", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -188012,46 +177814,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT 5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "fastgpt": { + "id": "fastgpt", + "name": "Web Answer", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2023-08-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -188059,63 +177843,40 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 872000, - "output": 128000 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 7.5, + "output": 7.5 } }, - "openai/gpt-6-sol-fast": { - "id": "openai/gpt-6-sol-fast", - "name": "GPT-6 Sol (Fast)", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", + "Qwen3.5-27B-Infracelestial": { + "id": "Qwen3.5-27B-Infracelestial", + "name": "Qwen3.5 27B Infracelestial", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -188123,90 +177884,58 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272001 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10 - } + "input": 0.306, + "output": 0.306 } }, - "openai/tts-1": { - "id": "openai/tts-1", - "name": "TTS-1", - "description": "Speech generation model for controllable voice, narration, and audio delivery", + "glm-4-flash": { + "id": "glm-4-flash", + "name": "GLM-4 Flash", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2023-11-06", - "last_updated": "2023-11-06", + "structured_output": false, + "release_date": "2024-08-01", + "last_updated": "2024-08-01", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 4096 + }, + "cost": { + "input": 0.1003, + "output": 0.1003 } }, - "openai/gpt-6-luna-fast": { - "id": "openai/gpt-6-luna-fast", - "name": "GPT-6 Luna (Fast)", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", + "azure-gpt-4o-mini": { + "id": "azure-gpt-4o-mini", + "name": "Azure gpt-4o-mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -188214,64 +177943,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 1, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.5, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272001 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.5, - "cache_read": 0.04, - "cache_write": 0.5 - } + "input": 0.1496, + "output": 0.595 } }, - "openai/gpt-5-fast": { - "id": "openai/gpt-5-fast", - "name": "GPT-5 (Fast)", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "sonar-deep-research": { + "id": "sonar-deep-research", + "name": "Perplexity Deep Research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-25", + "last_updated": "2025-02-25", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -188279,66 +177972,119 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, + "context": 60000, + "input": 60000, "output": 128000 }, "cost": { - "input": 2.5, - "output": 20, - "cache_read": 0.25 + "input": 3.4, + "output": 13.6 } }, - "openai/gpt-realtime-2": { - "id": "openai/gpt-realtime-2", - "name": "gpt-realtime-2", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", + "qwq-32b": { + "id": "qwq-32b", + "name": "Qwen: QwQ 32B", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "structured_output": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text", - "audio" + "text" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 32768 }, "cost": { - "input": 4, - "output": 24, - "cache_read": 0.4 + "input": 0.25599999, + "output": 0.30499999 } }, - "openai/gpt-4.1-fast": { - "id": "openai/gpt-4.1-fast", - "name": "GPT-4.1 (Fast)", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, + "mistral-code-agent-latest": { + "id": "mistral-code-agent-latest", + "name": "Mistral Code Agent Latest", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "release_date": "2026-06-02", + "last_updated": "2026-06-02", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "input": 262144, + "output": 32768 + }, + "cost": { + "input": 0.4, + "output": 2 + } + }, + "glm-4.1v-thinking-flash": { + "id": "glm-4.1v-thinking-flash", + "name": "GLM 4.1V Thinking Flash", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-09", + "last_updated": "2025-07-09", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 64000, + "input": 64000, + "output": 8192 + }, + "cost": { + "input": 0.3, + "output": 0.3 + } + }, + "qwen3.5-omni-flash": { + "id": "qwen3.5-omni-flash", + "name": "Qwen3.5 Omni Flash", + "description": "Omni-modal model for text, vision, audio, and multimodal agent tasks", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-30", + "last_updated": "2026-03-30", "modalities": { "input": [ "text", "image", - "pdf" + "video", + "audio" ], "output": [ "text" @@ -188346,47 +178092,35 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "input": 1014808, - "output": 32768 + "context": 49152, + "input": 49152, + "output": 16384 }, "cost": { - "input": 3.5, - "output": 14, - "cache_read": 0.875 + "input": 0, + "output": 0 } }, - "openai/gpt-6-luna": { - "id": "openai/gpt-6-luna", - "name": "GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", + "gemma-4-31B-MeroMero": { + "id": "gemma-4-31B-MeroMero", + "name": "Gemma 4 31B MeroMero", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-02", + "last_updated": "2026-05-02", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -188394,79 +178128,59 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 272001 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25 - } + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-image-1-mini": { - "id": "openai/gpt-image-1-mini", - "name": "GPT Image 1 Mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": false, - "reasoning": false, + "gemini-2.5-flash-preview-04-17:thinking": { + "id": "gemini-2.5-flash-preview-04-17:thinking", + "name": "Gemini 2.5 Flash Preview Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": false, - "temperature": true, - "release_date": "2025-10-06", - "last_updated": "2025-10-06", + "structured_output": false, + "release_date": "2025-04-17", + "last_updated": "2025-04-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.2 + "input": 0.15, + "output": 3.5 } }, - "openai/gpt-4o-mini-transcribe": { - "id": "openai/gpt-4o-mini-transcribe", - "name": "GPT-4o mini Transcribe", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "o-mini", + "glm-4-air": { + "id": "glm-4-air", + "name": "GLM-4 Air", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2024-03-13", - "last_updated": "2024-03-13", + "structured_output": false, + "release_date": "2024-06-05", + "last_updated": "2024-06-05", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" @@ -188474,75 +178188,57 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 1.25, - "output": 5 + "input": 0.2006, + "output": 0.2006 } }, - "openai/gpt-image-1.5": { - "id": "openai/gpt-image-1.5", - "name": "GPT Image 1.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "doubao-seed-1-6-thinking-250615": { + "id": "doubao-seed-1-6-thinking-250615", + "name": "Doubao Seed 1.6 Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2025-11-25", - "last_updated": "2025-11-25", + "structured_output": false, + "release_date": "2025-06-15", + "last_updated": "2025-06-15", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 256000, + "input": 256000, + "output": 16384 }, "cost": { - "input": 5, - "output": 32, - "cache_read": 1.25 + "input": 0.204, + "output": 2.04 } }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT 5.6 Terra", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "auto-model-basic": { + "id": "auto-model-basic", + "name": "Auto model (Basic)", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -188550,104 +178246,71 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "input": 1000000, + "output": 1000000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 9.996, + "output": 19.992 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, + "Qwen3.5-27B-RpRMax-v1": { + "id": "Qwen3.5-27B-RpRMax-v1", + "name": "Qwen3.5 27B RpRMax v1", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.1 + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "ernie-5.1": { + "id": "ernie-5.1", + "name": "ERNIE 5.1", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-10", + "last_updated": "2026-05-10", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -188655,47 +178318,41 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 119000, + "input": 119000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.03 + "input": 0.75, + "output": 3, + "cache_read": 0.75 } }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT 5.6 Sol", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-sol", + "Qwen3.5-27B-Marvin-DPO-V2-Derestricted": { + "id": "Qwen3.5-27B-Marvin-DPO-V2-Derestricted", + "name": "Qwen3.5 27B Marvin DPO V2 Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -188703,53 +178360,29 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10 - } + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-4o-mini-fast": { - "id": "openai/gpt-4o-mini-fast", - "name": "GPT-4o mini (Fast)", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", + "azure-gpt-4o": { + "id": "azure-gpt-4o", + "name": "Azure gpt-4o", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -188758,70 +178391,27 @@ "open_weights": false, "limit": { "context": 128000, - "input": 111616, + "input": 128000, "output": 16384 }, "cost": { - "input": 0.25, - "output": 1, - "cache_read": 0.125 - } - }, - "openai/gpt-realtime-whisper": { - "id": "openai/gpt-realtime-whisper", - "name": "gpt-realtime-whisper", - "description": "Streaming speech-to-text model for low-latency transcript deltas from live audio", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 + "input": 2.499, + "output": 9.996 } }, - "openai/gpt-6-sol": { - "id": "openai/gpt-6-sol", - "name": "GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", + "deepseek-chat": { + "id": "deepseek-chat", + "name": "DeepSeek V3/Deepseek Chat", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ "text", - "image", "pdf" ], "output": [ @@ -188830,87 +178420,59 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272001 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 0.25, + "output": 0.7 } }, - "openai/tts-1-hd": { - "id": "openai/tts-1-hd", - "name": "TTS-1 HD", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "attachment": false, + "gemini-2.5-flash-preview-05-20": { + "id": "gemini-2.5-flash-preview-05-20", + "name": "Gemini 2.5 Flash 0520", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": true, - "release_date": "2023-11-06", - "last_updated": "2023-11-06", + "structured_output": false, + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1048000, + "input": 1048000, + "output": 65536 + }, + "cost": { + "input": 0.15, + "output": 0.6 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, + "mercury-2": { + "id": "mercury-2", + "name": "Mercury 2", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -188918,44 +178480,38 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 50000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.25, + "output": 0.75, + "cache_read": 0.025 } }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", + "qwen3.7-plus:thinking": { + "id": "qwen3.7-plus:thinking", + "name": "Qwen3.7 Plus Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024, + "max": 262144 } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "tool_call": false, + "structured_output": false, + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -188963,32 +178519,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 983616, + "input": 983616, + "output": 65536 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.4, + "output": 1.6, + "cache_read": 0.04 } }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", + "gemini-2.0-flash-thinking-exp-1219": { + "id": "gemini-2.0-flash-thinking-exp-1219", + "name": "Gemini 2.0 Flash Thinking 1219", "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, + "attachment": false, "reasoning": false, - "tool_call": true, + "tool_call": false, "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", + "release_date": "2024-12-19", + "last_updated": "2024-12-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -188996,33 +178549,28 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 32767, + "input": 32767, + "output": 8192 }, - "status": "deprecated", "cost": { - "input": 10, - "output": 30 + "input": 0.1003, + "output": 0.408 } }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, + "glm-4-plus-0111": { + "id": "glm-4-plus-0111", + "name": "GLM 4 Plus 0111", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-19", + "last_updated": "2025-02-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -189030,33 +178578,28 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 9.996, + "output": 9.996 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, + "brave": { + "id": "brave", + "name": "Brave (Answers)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "tool_call": false, + "structured_output": false, + "release_date": "2023-03-02", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -189064,33 +178607,28 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 8192, + "input": 8192, + "output": 8192 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 5, + "output": 5 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, + "glm-zero-preview": { + "id": "glm-zero-preview", + "name": "GLM Zero Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -189098,43 +178636,43 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 8000, + "input": 8000, + "output": 4096 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 + "input": 1.802, + "output": 1.802 } }, - "openai/o1": { - "id": "openai/o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "gemini-2.5-flash-lite-preview-06-17": { + "id": "gemini-2.5-flash-lite-preview-06-17", + "name": "Gemini 2.5 Flash Lite Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ + "none", "low", "medium", "high" ] } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -189142,44 +178680,40 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1048756, + "input": 1048756, + "output": 65536 }, - "status": "deprecated", "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 0.15, + "output": 0.6 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", + "Qwen3.5-27B-Writer-Derestricted": { + "id": "Qwen3.5-27B-Writer-Derestricted", + "name": "Qwen3.5 27B Writer Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-06", + "last_updated": "2026-04-06", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -189187,33 +178721,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, + "KAT-Coder-Exp-72B-1010": { + "id": "KAT-Coder-Exp-72B-1010", + "name": "KAT Coder Exp 72B 1010", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "tool_call": false, + "structured_output": false, + "release_date": "2025-10-28", + "last_updated": "2025-10-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -189221,22 +178750,20 @@ }, "open_weights": false, "limit": { - "context": 1047576, + "context": 128000, + "input": 128000, "output": 32768 }, - "status": "deprecated", "cost": { "input": 0.1, - "output": 0.4, - "cache_read": 0.025 + "output": 0.2 } }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": false, + "gemini-2.5-pro-preview-06-05": { + "id": "gemini-2.5-pro-preview-06-05", + "name": "Gemini 2.5 Pro Preview 0605", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -189248,15 +178775,14 @@ ] } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-05", + "last_updated": "2025-06-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189264,43 +178790,40 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 1048756, + "input": 1048756, + "output": 65536 }, - "status": "deprecated", "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 2.5, + "output": 10 } }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", + "Qwen3.5-27B-Musica-v1": { + "id": "Qwen3.5-27B-Musica-v1", + "name": "Qwen3.5 27B Musica v1", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -189308,44 +178831,28 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 262144, + "input": 262144, + "output": 16384 }, - "status": "deprecated", "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "MiniMax-M1": { + "id": "MiniMax-M1", + "name": "MiniMax M1", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-16", + "last_updated": "2025-06-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -189353,34 +178860,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 1000000, + "input": 1000000, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.1394, + "output": 1.3328 } }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", + "doubao-1-5-thinking-pro-vision-250415": { + "id": "doubao-1-5-thinking-pro-vision-250415", + "name": "Doubao 1.5 Thinking Pro Vision", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ "text" @@ -189389,183 +178891,139 @@ "open_weights": false, "limit": { "context": 128000, + "input": 128000, "output": 16384 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.6, + "output": 2.4 } - } - } - }, - "zai-coding-plan": { - "id": "zai-coding-plan", - "env": [ - "ZHIPU_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.z.ai/api/coding/paas/v4", - "name": "Z.AI Coding Plan", - "doc": "https://docs.z.ai/devpack/overview", - "models": { - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + }, + "qwen3.5-27b:thinking": { + "id": "qwen3.5-27b:thinking", + "name": "Qwen3.5 27B Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", "image", - "video", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 260096, + "input": 260096, + "output": 65536 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.27, + "output": 2.16 } }, - "glm-5.3-highspeed": { - "id": "glm-5.3-highspeed", - "name": "GLM-5.3 Highspeed", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "Qwen3.5-27B-Omega-Evolution-v2.0-Derestricted": { + "id": "Qwen3.5-27B-Omega-Evolution-v2.0-Derestricted", + "name": "Qwen3.5 27B Omega Evolution v2.0 Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-06", + "last_updated": "2026-04-06", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.306, + "output": 0.306 } }, - "glm-5.2-highspeed": { - "id": "glm-5.2-highspeed", - "name": "GLM-5.2 Highspeed", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "Gemma-4-31B-GarnetV2": { + "id": "Gemma-4-31B-GarnetV2", + "name": "Gemma 4 31B Garnet V2", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "high", - "max" - ] + "type": "toggle" } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-01", + "last_updated": "2026-05-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.306, + "output": 0.306 } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "qwen-turbo": { + "id": "qwen-turbo", + "name": "Qwen Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-11-01", + "last_updated": "2024-11-01", "modalities": { "input": [ "text" @@ -189574,42 +179032,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "input": 1000000, + "output": 8192 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.04998, + "output": 0.2006 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "phi-4-multimodal-instruct": { + "id": "phi-4-multimodal-instruct", + "name": "Phi 4 Multimodal", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text" @@ -189618,41 +179061,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.07, + "output": 0.11 } }, - "glm-5-turbo": { - "id": "glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", + "mistral-small-31-24b-instruct": { + "id": "mistral-small-31-24b-instruct", + "name": "Mistral Small 31 24b Instruct", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -189660,159 +179093,128 @@ }, "open_weights": false, "limit": { - "context": 200000, + "context": 128000, + "input": 128000, "output": 131072 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.1, + "output": 0.3 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "ernie-4.5-turbo-128k": { + "id": "ernie-4.5-turbo-128k", + "name": "Ernie 4.5 Turbo 128k", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-08", + "last_updated": "2025-05-08", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 + "input": 0.132, + "output": 0.55 } - } - } - }, - "ebcloud": { - "id": "ebcloud", - "env": [ - "EBCLOUD_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://maas-api.ebcloud.com/v1", - "name": "EBCloud", - "doc": "https://docs.ebtech.com/ai/model-api.html", - "models": { - "Kimi-K2.6": { - "id": "Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": false, + }, + "Qwen3.5-27B-BlueStar-Derestricted": { + "id": "Qwen3.5-27B-BlueStar-Derestricted", + "name": "Qwen3.5 27B BlueStar Derestricted", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 + } + ], + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-06", + "last_updated": "2026-04-06", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 262144 + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.9286, - "output": 3.8571 + "input": 0.306, + "output": 0.306 } }, - "GLM-5.1": { - "id": "GLM-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "gemini-2.5-flash-lite-preview-09-2025-thinking": { + "id": "gemini-2.5-flash-lite-preview-09-2025-thinking", + "name": "Gemini 2.5 Flash Lite Preview (09/2025) – Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 0.8571, - "output": 3.4286 + "input": 0.1, + "output": 0.4 } }, - "DeepSeek-V4-Flash": { - "id": "DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "doubao-seed-1-8-251215": { + "id": "doubao-seed-1-8-251215", + "name": "Doubao Seed 1.8", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ "text" @@ -189821,33 +179223,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 0.143, - "output": 0.2857 + "input": 0.612, + "output": 6.12 } }, - "DeepSeek-V4-Pro": { - "id": "DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "qwen3.6-max-preview": { + "id": "qwen3.6-max-preview", + "name": "Qwen3.6 Max Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "qwen3.6", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-20", + "last_updated": "2026-04-21", "modalities": { "input": [ "text" @@ -189856,52 +179253,26 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 245800, + "output": 65536 }, "cost": { - "input": 0.4286, - "output": 0.8571 + "input": 1.3, + "output": 7.8 } - } - } - }, - "greenpt": { - "id": "greenpt", - "env": [ - "GREENPT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.greenpt.ai/v1", - "name": "GreenPT", - "doc": "https://docs.greenpt.ai", - "models": { - "glm-5.2-honey-ultra": { - "id": "glm-5.2-honey-ultra", - "name": "GLM-5.2 Honey Ultra", - "description": "glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.", - "family": "glm", + }, + "exa-answer": { + "id": "exa-answer", + "name": "Exa (Answer)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-04", + "last_updated": "2025-06-04", "modalities": { "input": [ "text" @@ -189910,41 +179281,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 4096, + "input": 4096, + "output": 4096 }, "cost": { - "input": 1.254, - "output": 5.016, - "cache_read": 0.3135 + "input": 2.5, + "output": 2.5 } }, - "glm-5.2-ponytail-lite": { - "id": "glm-5.2-ponytail-lite", - "name": "GLM-5.2 Ponytail Lite", - "description": "glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.", - "family": "glm", + "Baichuan4-Air": { + "id": "Baichuan4-Air", + "name": "Baichuan 4 Air", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-19", + "last_updated": "2025-08-19", "modalities": { "input": [ "text" @@ -189953,73 +179310,59 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 1.254, - "output": 5.016, - "cache_read": 0.3135 + "input": 0.157, + "output": 0.157 } }, - "gemma-3-27b-it": { - "id": "gemma-3-27b-it", - "name": "Gemma 3 27B", - "description": "Google Gemma 3 multimodal model for chat, reasoning, and image understanding", - "family": "gemma", + "qwen3.5-122b-a10b": { + "id": "qwen3.5-122b-a10b", + "name": "Qwen3.5 122B A10B", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 40000, - "output": 8192 + "context": 260096, + "input": 260096, + "output": 65536 }, "cost": { - "input": 0.342, - "output": 0.684 + "input": 0.36, + "output": 2.88 } }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "sarvam-30b": { + "id": "sarvam-30b", + "name": "Sarvam 30B", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "structured_output": false, + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ "text" @@ -190028,161 +179371,121 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 65536, + "input": 65536, + "output": 4096 }, "cost": { - "input": 0.127754, - "output": 0.511016, - "cache_read": 0.0255508 + "input": 0.028, + "output": 0.111, + "cache_read": 0.017 } }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "claude-opus-4-1-thinking": { + "id": "claude-opus-4-1-thinking", + "name": "Claude 4.1 Opus Thinking", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 0.1596, - "output": 0.399, - "cache_read": 0.0456 + "input": 14.994, + "output": 75.004 } }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen3 235B MoE instruct model for long-context multilingual chat and reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "claude-opus-4-thinking:32000": { + "id": "claude-opus-4-thinking:32000", + "name": "Claude 4 Opus Thinking (32K)", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-21", - "last_updated": "2025-07-21", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 200000, + "input": 200000, + "output": 32000 }, "cost": { - "input": 1.026, - "output": 3.078 + "input": 14.994, + "output": 75.004 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "auto-model": { + "id": "auto-model", + "name": "Auto model", + "description": "Automatic model router for matching prompts to suitable backends and budgets", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-06-01", + "last_updated": "2024-06-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 1000000, + "input": 1000000, + "output": 1000000 }, "cost": { - "input": 3.762, - "output": 18.81, - "cache_read": 0.9405 + "input": 0, + "output": 0 } }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "qwen3-vl-235b-a22b-instruct-original": { + "id": "qwen3-vl-235b-a22b-instruct-original", + "name": "Qwen3 VL 235B A22B Instruct Original", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text", @@ -190192,40 +179495,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.255552, - "output": 1.27776, - "cache_read": 0.0127776 + "input": 0.5, + "output": 1.2 } }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "glm-z1-air": { + "id": "glm-z1-air", + "name": "GLM Z1 Air", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "structured_output": true, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text" @@ -190234,75 +179524,71 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 32000, + "input": 32000, + "output": 16384 }, "cost": { - "input": 0.1938, - "output": 1.129, - "cache_read": 0.0627 + "input": 0.07, + "output": 0.07 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "Qwen3.5-27B-Queen-Derestricted-Lite": { + "id": "Qwen3.5-27B-Queen-Derestricted-Lite", + "name": "Qwen3.5 27B Queen Derestricted Lite", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 262144 + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.7524, - "output": 4.275, - "cache_read": 0.2508 + "input": 0.306, + "output": 0.306 } }, - "green-s": { - "id": "green-s", - "name": "Green S", - "description": "GreenPT speech-to-text model for pre-recorded and live transcription", + "inclusionai/ling-2.6-1t": { + "id": "inclusionai/ling-2.6-1t", + "name": "Ling 2.6 1T", + "description": "Tool-capable chat model for instruction following and agentic application workflows", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-01", - "last_updated": "2025-01", + "tool_call": true, + "structured_output": true, + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" @@ -190310,57 +179596,56 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 8192 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.00437, - "output": 0 + "input": 0.3, + "output": 2.5, + "cache_read": 0.06 } }, - "green-l": { - "id": "green-l", - "name": "Green L", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, - "reasoning": false, + "inclusionai/ring-2.6-1t": { + "id": "inclusionai/ring-2.6-1t", + "name": "Ring 2.6 1T", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", + "structured_output": true, + "release_date": "2026-05-08", + "last_updated": "2026-05-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 262144, + "input": 262144, + "output": 65536 }, "cost": { - "input": 0.285, - "output": 0.912 + "input": 1, + "output": 3 } }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", + "inclusionai/ling-2.6-flash": { + "id": "inclusionai/ling-2.6-flash", + "name": "Ling 2.6 Flash", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "structured_output": true, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text" @@ -190369,28 +179654,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 262144, + "input": 262144, "output": 32768 }, "cost": { - "input": 0.285, - "output": 1.083 + "input": 0.08, + "output": 0.24 } }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", + "Alibaba-NLP/Tongyi-DeepResearch-30B-A3B": { + "id": "Alibaba-NLP/Tongyi-DeepResearch-30B-A3B", + "name": "Tongyi DeepResearch 30B A3B", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "yi", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text" @@ -190399,157 +179684,118 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 100000, - "output": 16384 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 1.254, - "output": 1.254 + "input": 0.08, + "output": 0.24000000000000002 } }, - "gemma4": { - "id": "gemma4", - "name": "gemma4", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "ibm-granite/granite-4.1-8b": { + "id": "ibm-granite/granite-4.1-8b", + "name": "Granite 4.1 8B", + "description": "Tool-capable chat model for instruction following and agentic application workflows", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 131072, + "input": 131072, + "output": 131072 }, "cost": { - "input": 0.57, - "output": 1.71 + "input": 0.05, + "output": 0.1, + "cache_read": 0.05 } }, - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "Salesforce/Llama-xLAM-2-70b-fc-r": { + "id": "Salesforce/Llama-xLAM-2-70b-fc-r", + "name": "Llama-xLAM-2 70B fc-r", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-13", + "last_updated": "2025-04-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.342, - "output": 2.052 + "input": 2.5, + "output": 2.5 } }, - "pixtral-12b-2409": { - "id": "pixtral-12b-2409", - "name": "Pixtral 12B", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", - "attachment": true, + "THUDM/GLM-Z1-32B-0414": { + "id": "THUDM/GLM-Z1-32B-0414", + "name": "GLM Z1 32B 0414", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm-z", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-09-01", - "last_updated": "2024-09-01", + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 128000, - "output": 4096 + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.285, - "output": 0.285 + "input": 0.2, + "output": 0.2 } }, - "glm-5.2-honey": { - "id": "glm-5.2-honey", - "name": "GLM-5.2 Honey", - "description": "glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.", + "THUDM/GLM-4-32B-0414": { + "id": "THUDM/GLM-4-32B-0414", + "name": "GLM 4 32B 0414", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ "text" @@ -190558,111 +179804,91 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 1.254, - "output": 5.016, - "cache_read": 0.3135 + "input": 0.2, + "output": 0.2 } }, - "holo2-30b-a3b": { - "id": "holo2-30b-a3b", - "name": "Holo2 30B A3B", - "description": "H Company Holo2 vision model for GUI navigation and computer-use agents", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "THUDM/GLM-4-9B-0414": { + "id": "THUDM/GLM-4-9B-0414", + "name": "GLM 4 9B 0414", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": false, - "temperature": true, - "release_date": "2025-11", - "last_updated": "2025-11", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 22016, - "output": 16384 + "context": 32000, + "input": 32000, + "output": 8000 }, "cost": { - "input": 0.399, - "output": 0.969 + "input": 0.2, + "output": 0.2 } }, - "green-r-raw": { - "id": "green-r-raw", - "name": "Green R Raw", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "THUDM/GLM-Z1-9B-0414": { + "id": "THUDM/GLM-Z1-9B-0414", + "name": "GLM Z1 9B 0414", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm-z", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 32000, + "input": 32000, + "output": 8000 }, "cost": { - "input": 0.399, - "output": 1.083 + "input": 0.2, + "output": 0.2 } }, - "green-s-pro": { - "id": "green-s-pro", - "name": "Green S Pro", - "description": "GreenPT advanced speech-to-text model with multilingual transcription support", + "meta-llama/llama-3.1-8b-instruct": { + "id": "meta-llama/llama-3.1-8b-instruct", + "name": "Llama 3.1 8b Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-02", - "last_updated": "2025-02", + "structured_output": false, + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" @@ -190670,26 +179896,57 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 8192 + "context": 131072, + "input": 131072, + "output": 16384 }, "cost": { - "input": 0.00437, - "output": 0 + "input": 0.0544, + "output": 0.0544 } }, - "devstral-2-123b-instruct-2512": { - "id": "devstral-2-123b-instruct-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", + "meta-llama/llama-4-maverick": { + "id": "meta-llama/llama-4-maverick", + "name": "Llama 4 Maverick", + "description": "Open multimodal Llama model for strong reasoning and fast responses", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "release_date": "2025-09-05", + "last_updated": "2025-09-05", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048576, + "input": 1048576, + "output": 65536 + }, + "cost": { + "input": 0.18000000000000002, + "output": 0.8 + } + }, + "meta-llama/llama-3.3-70b-instruct": { + "id": "meta-llama/llama-3.3-70b-instruct", + "name": "Llama 3.3 70b Instruct", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", + "structured_output": true, + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ "text" @@ -190698,37 +179955,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, + "context": 131072, + "input": 131072, "output": 16384 }, "cost": { - "input": 0.57, - "output": 2.736 + "input": 0.05, + "output": 0.23 } }, - "mistral-medium-3.5-128b": { - "id": "mistral-medium-3.5-128b", - "name": "Mistral Medium 3.5", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", + "meta-llama/llama-4-scout": { + "id": "meta-llama/llama-4-scout", + "name": "Llama 4 Scout", + "description": "Open multimodal Llama model for long-context analysis and efficient agents", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text", @@ -190738,82 +179986,59 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 328000, + "input": 328000, + "output": 65536 }, "cost": { - "input": 2.052, - "output": 10.26 + "input": 0.085, + "output": 0.46 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "meta-llama/llama-3.2-3b-instruct": { + "id": "meta-llama/llama-3.2-3b-instruct", + "name": "Llama 3.2 3b Instruct", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-09-25", + "last_updated": "2024-09-25", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 131072, + "input": 131072, + "output": 8192 }, "cost": { - "input": 0.798, - "output": 4.959 + "input": 0.0306, + "output": 0.0493 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "featherless-ai/Qwerky-72B": { + "id": "featherless-ai/Qwerky-72B", + "name": "Qwerky 72B", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "qwerky", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-03-20", + "last_updated": "2025-03-20", "modalities": { "input": [ "text" @@ -190822,104 +180047,89 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 32000, + "input": 32000, + "output": 8192 }, "cost": { - "input": 1.254, - "output": 5.016, - "cache_read": 0.3135 + "input": 0.5, + "output": 0.5 } }, - "voxtral-small-24b-2507": { - "id": "voxtral-small-24b-2507", - "name": "Voxtral Small 24B", - "description": "Mistral Voxtral audio-understanding model for speech and transcription tasks", - "family": "mistral", - "attachment": true, + "moonshotai/kimi-k2-instruct-0711": { + "id": "moonshotai/kimi-k2-instruct-0711", + "name": "Kimi K2 0711", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-15", - "last_updated": "2025-07-15", + "structured_output": true, + "release_date": "2025-07-11", + "last_updated": "2025-07-11", "modalities": { "input": [ - "text", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 0.228, - "output": 0.513 + "input": 0.1, + "output": 2 } }, - "green-l-raw": { - "id": "green-l-raw", - "name": "Green L Raw", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, + "moonshotai/Kimi-K2-Instruct-0905": { + "id": "moonshotai/Kimi-K2-Instruct-0905", + "name": "Kimi K2 0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", + "structured_output": true, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 256000, + "input": 256000, + "output": 262144 }, "cost": { - "input": 0.285, - "output": 0.912 + "input": 0.4, + "output": 2 } }, - "glm-5.2-ponytail-ultra": { - "id": "glm-5.2-ponytail-ultra", - "name": "GLM-5.2 Ponytail Ultra", - "description": "glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.", - "family": "glm", + "moonshotai/kimi-k2-thinking-original": { + "id": "moonshotai/kimi-k2-thinking-original", + "name": "Kimi K2 Thinking Original", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ "text" @@ -190928,39 +180138,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 16384 }, "cost": { - "input": 1.254, - "output": 5.016, - "cache_read": 0.3135 + "input": 0.6, + "output": 2.5 } }, - "green-r": { - "id": "green-r", - "name": "Green R", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "moonshotai/kimi-k2.5:thinking": { + "id": "moonshotai/kimi-k2.5:thinking", + "name": "Kimi K2.5 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": false, + "release_date": "2026-01-26", + "last_updated": "2026-01-26", "modalities": { "input": [ "text", @@ -190970,29 +180170,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 256000, + "input": 256000, + "output": 65536 }, "cost": { - "input": 0.399, - "output": 1.083 + "input": 0.3, + "output": 1.9 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "moonshotai/kimi-k2-instruct": { + "id": "moonshotai/kimi-k2-instruct", + "name": "Kimi K2 Instruct", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ "text" @@ -191001,75 +180200,63 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 8192 }, - "status": "deprecated", "cost": { - "input": 1.756, - "output": 5.518 + "input": 0.1, + "output": 2 } }, - "mistral-small-3.2-24b-instruct-2506": { - "id": "mistral-small-3.2-24b-instruct-2506", - "name": "Mistral Small 3.2", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, + "moonshotai/kimi-k2-thinking": { + "id": "moonshotai/kimi-k2-thinking", + "name": "Kimi K2 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": false, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", + "structured_output": true, + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32768 + "context": 256000, + "input": 256000, + "output": 262144 }, "cost": { - "input": 0.228, - "output": 0.456 + "input": 0.3, + "output": 1.2 } }, - "glm-5.2-caveman": { - "id": "glm-5.2-caveman", - "name": "GLM-5.2 Caveman", - "description": "glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.", - "family": "glm", - "attachment": false, + "moonshotai/kimi-k2.6:thinking": { + "id": "moonshotai/kimi-k2.6:thinking", + "name": "Kimi K2.6 Thinking", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "structured_output": false, + "release_date": "2026-04-16", + "last_updated": "2026-04-21", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -191077,29 +180264,25 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 131072 + "context": 256000, + "output": 65536 }, "cost": { - "input": 1.254, - "output": 5.016, - "cache_read": 0.3135 + "input": 0.53, + "output": 2.73 } }, - "kimi-k2.6-fast": { - "id": "kimi-k2.6-fast", - "name": "Kimi K2.6 Fast", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "moonshotai/kimi-k2.5": { + "id": "moonshotai/kimi-k2.5", + "name": "Kimi K2.5", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "family": "kimi-k2", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "structured_output": false, + "release_date": "2026-01-26", + "last_updated": "2026-01-26", "modalities": { "input": [ "text", @@ -191109,39 +180292,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "input": 256000, + "output": 65536 }, - "status": "deprecated", "cost": { - "input": 1.655, - "output": 8.778 + "input": 0.3, + "output": 1.9 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": false, + "release_date": "2026-04-16", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -191153,81 +180325,58 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.228, - "output": 0.798 + "input": 0.53, + "output": 2.73 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "moonshotai/kimi-latest": { + "id": "moonshotai/kimi-latest", + "name": "Kimi Latest", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "structured_output": false, + "release_date": "2026-05-03", + "last_updated": "2026-05-03", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 65536 }, "cost": { - "input": 1.27754, - "output": 5.11016, - "cache_read": 0.319385 + "input": 0.5, + "output": 2.6, + "cache_read": 0.125 } }, - "glm-5.2-caveman-ultra": { - "id": "glm-5.2-caveman-ultra", - "name": "GLM-5.2 Caveman Ultra", - "description": "glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The most aggressive tier, close to answer-only. Same upstream model and price per token as glm-5.2, with fewer output tokens.", - "family": "glm", + "moonshotai/kimi-k2-thinking-turbo-original": { + "id": "moonshotai/kimi-k2-thinking-turbo-original", + "name": "Kimi K2 Thinking Turbo Original", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-11-06", + "last_updated": "2025-11-06", "modalities": { "input": [ "text" @@ -191236,128 +180385,91 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 16384 }, "cost": { - "input": 1.254, - "output": 5.016, - "cache_read": 0.3135 + "input": 1.15, + "output": 8 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "baidu/ernie-4.5-vl-28b-a3b": { + "id": "baidu/ernie-4.5-vl-28b-a3b", + "name": "ERNIE 4.5 VL 28B", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "ernie", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-30", + "last_updated": "2025-06-30", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 0.9006, - "output": 4.389, - "cache_read": 0.1881 + "input": 0.13999999999999999, + "output": 0.5599999999999999 } }, - "glm-5.2-ponytail": { - "id": "glm-5.2-ponytail", - "name": "GLM-5.2 Ponytail", - "description": "glm-5.2 carrying a built-in ruleset that compresses generated code, preferring platform features over custom code. The middle tier, and the ruleset as its authors wrote it. Same upstream model and price per token as glm-5.2, with fewer output tokens.", - "family": "glm", - "attachment": false, + "perceptron/perceptron-mk1": { + "id": "perceptron/perceptron-mk1", + "name": "Perceptron Mk1", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "reasoning_options": [], + "tool_call": false, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 1.254, - "output": 5.016, - "cache_read": 0.3135 + "input": 0.15, + "output": 1.5 } }, - "glm-5.2-honey-lite": { - "id": "glm-5.2-honey-lite", - "name": "GLM-5.2 Honey Lite", - "description": "glm-5.2 carrying a built-in ruleset that compresses both generated code and prose. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.", - "family": "glm", + "failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5": { + "id": "failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5", + "name": "Llama 3 70B abliterated", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text" @@ -191366,41 +180478,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 8192, + "input": 8192, + "output": 8192 }, "cost": { - "input": 1.254, - "output": 5.016, - "cache_read": 0.3135 + "input": 0.7, + "output": 0.7 } }, - "glm-5.2-caveman-lite": { - "id": "glm-5.2-caveman-lite", - "name": "GLM-5.2 Caveman Lite", - "description": "glm-5.2 carrying a built-in ruleset that compresses prose, keeping code and technical detail verbatim. The gentlest tier: it cuts filler only and keeps the explanation intact. Same upstream model and price per token as glm-5.2, with fewer output tokens.", - "family": "glm", + "nanogpt/coding-router:max": { + "id": "nanogpt/coding-router:max", + "name": "Coding Router Max", + "description": "Automatic model router for matching prompts to suitable backends and budgets", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ "text" @@ -191409,45 +180508,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "input": 1000000, + "output": 128000 }, "cost": { - "input": 1.254, - "output": 5.016, - "cache_read": 0.3135 + "input": 5, + "output": 30, + "cache_read": 0.5 } - } - } - }, - "mixlayer": { - "id": "mixlayer", - "env": [ - "MIXLAYER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://models.mixlayer.ai/v1", - "name": "Mixlayer", - "doc": "https://docs.mixlayer.com", - "models": { - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + }, + "nanogpt/coding-router:high": { + "id": "nanogpt/coding-router:high", + "name": "Coding Router High", + "description": "Automatic model router for matching prompts to suitable backends and budgets", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": true, + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ "text" @@ -191456,32 +180539,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 2.4 + "input": 1.1, + "output": 2.2, + "cache_read": 0.11 } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B A10B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "nanogpt/coding-router:low": { + "id": "nanogpt/coding-router:low", + "name": "Coding Router Low", + "description": "Automatic model router for matching prompts to suitable backends and budgets", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": true, + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ "text" @@ -191490,32 +180570,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 3.2 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "qwen/qwen3.5-9b": { - "id": "qwen/qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "nanogpt/coding-router:medium": { + "id": "nanogpt/coding-router:medium", + "name": "Coding Router Medium", + "description": "Automatic model router for matching prompts to suitable backends and budgets", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": true, + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ "text" @@ -191524,32 +180601,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5 35B A3B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "nanogpt/coding-router": { + "id": "nanogpt/coding-router", + "name": "Coding Router", + "description": "Automatic model router for matching prompts to suitable backends and budgets", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": true, + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ "text" @@ -191558,32 +180632,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.25, - "output": 1.3 + "input": 1.1, + "output": 2.2, + "cache_read": 0.11 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "GalrionSoftworks/MN-LooseCannon-12B-v1": { + "id": "GalrionSoftworks/MN-LooseCannon-12B-v1", + "name": "MN-LooseCannon-12B-v1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ "text" @@ -191592,51 +180663,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } - } - } - }, - "hyper": { - "id": "hyper", - "env": [ - "HYPER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://hyper.charm.land/v1", - "name": "Charm Hyper", - "doc": "https://hyper.charm.land", - "models": { - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + }, + "TheDrummer/Cydonia-24B-v4.3": { + "id": "TheDrummer/Cydonia-24B-v4.3", + "name": "The Drummer Cydonia 24B v4.3", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-05-28", - "last_updated": "2026-07-22", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-25", + "last_updated": "2025-12-25", "modalities": { "input": [ "text" @@ -191647,82 +180694,54 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5 + "input": 0.1003, + "output": 0.1207 } }, - "qwen3.8-27b": { - "id": "qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "TheDrummer/Cydonia-24B-v4.1": { + "id": "TheDrummer/Cydonia-24B-v4.1", + "name": "The Drummer Cydonia 24B v4.1", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-19", + "last_updated": "2025-08-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1 + "input": 0.1003, + "output": 0.1207 } }, - "qwen3.8-2.4t-a95b": { - "id": "qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", + "TheDrummer/Cydonia-24B-v2": { + "id": "TheDrummer/Cydonia-24B-v2", + "name": "The Drummer Cydonia 24B v2", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-17", + "last_updated": "2025-02-17", "modalities": { "input": [ "text" @@ -191731,128 +180750,90 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 32768 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 0.1003, + "output": 0.1207 } }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-31", + "TheDrummer/Anubis-70B-v1.1": { + "id": "TheDrummer/Anubis-70B-v1.1", + "name": "Anubis 70B v1.1", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 131072, + "input": 131072, + "output": 16384 }, "cost": { - "input": 0.16332, - "output": 0.5444, - "cache_read": 0.031575 + "input": 0.31, + "output": 0.31 } }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-08-02", - "last_updated": "2026-08-02", + "TheDrummer/UnslopNemo-12B-v4.1": { + "id": "TheDrummer/UnslopNemo-12B-v4.1", + "name": "UnslopNemo 12b v4", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.044 + "input": 0.493, + "output": 0.493 } }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max Preview", - "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", - "family": "qwen", + "TheDrummer/skyfall-36b-v2": { + "id": "TheDrummer/skyfall-36b-v2", + "name": "TheDrummer Skyfall 36B V2", + "description": "Multimodal model for analyzing text, images, documents, and rich media", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-19", - "last_updated": "2026-07-19", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-03-10", + "last_updated": "2025-03-10", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" @@ -191860,128 +180841,86 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 64000, + "input": 64000, + "output": 32768 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 0.493, + "output": 0.493 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-27", - "last_updated": "2026-07-30", + "TheDrummer/Rocinante-12B-v1.1": { + "id": "TheDrummer/Rocinante-12B-v1.1", + "name": "Rocinante 12b", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 16000 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 3.2664, - "output": 16.332, - "cache_read": 0.32664 + "input": 0.408, + "output": 0.595 } }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "TheDrummer/Magidonia-24B-v4.3": { + "id": "TheDrummer/Magidonia-24B-v4.3", + "name": "The Drummer Magidonia 24B v4.3", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-25", + "last_updated": "2025-12-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, + "context": 32768, + "input": 32768, "output": 32768 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 + "input": 0.1003, + "output": 0.1207 } }, - "qwen3.7-flash": { - "id": "qwen3.7-flash", - "name": "Qwen3.7 Flash", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-27", - "last_updated": "2026-07-30", + "TheDrummer/Cydonia-24B-v4": { + "id": "TheDrummer/Cydonia-24B-v4", + "name": "The Drummer Cydonia 24B v4", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-22", + "last_updated": "2025-07-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -191989,61 +180928,55 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 991000, - "output": 64000 + "context": 16384, + "input": 16384, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 0.8, - "cache_read": 0.04 + "input": 0.2006, + "output": 0.2414 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "TheDrummer/Skyfall-31B-v4.2": { + "id": "TheDrummer/Skyfall-31B-v4.2", + "name": "TheDrummer Skyfall 31B v4.2", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-26", + "last_updated": "2026-03-26", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 26214 + "context": 131072, + "input": 131072, + "output": 16384 }, "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.3 + "input": 0.55, + "output": 0.8 } }, - "gemma-4-26b-a4b-it": { - "id": "gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", + "TheDrummer/Anubis-70B-v1": { + "id": "TheDrummer/Anubis-70B-v1", + "name": "Anubis 70B v1", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-07-22", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" @@ -192052,29 +180985,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "output": 25600 + "context": 65536, + "input": 65536, + "output": 16384 }, "cost": { - "input": 0.098, - "output": 0.334, - "cache_read": 0.049 + "input": 0.31, + "output": 0.31 } }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "huihui-ai/Qwen2.5-32B-Instruct-abliterated": { + "id": "huihui-ai/Qwen2.5-32B-Instruct-abliterated", + "name": "Qwen 2.5 32B Abliterated", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-05", - "last_updated": "2026-07-22", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-06", + "last_updated": "2025-01-06", "modalities": { "input": [ "text" @@ -192083,39 +181015,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262100, - "output": 6553 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0.484, - "output": 1.852, - "cache_read": 0.242 + "input": 0.7, + "output": 0.7 } }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "huihui-ai/Llama-3.3-70B-Instruct-abliterated": { + "id": "huihui-ai/Llama-3.3-70B-Instruct-abliterated", + "name": "Llama 3.3 70B Instruct abliterated", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ "text" @@ -192124,129 +181045,93 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 262144 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 1.437216, - "output": 4.311648, - "cache_read": 0.047907 + "input": 0.7, + "output": 0.7 } }, - "inkling": { - "id": "inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, + "huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated": { + "id": "huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated", + "name": "DeepSeek R1 Qwen Abliterated", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-08-15", - "last_updated": "2026-09-05", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 32768 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 1.0888, - "output": 4.40964, - "cache_read": 0.185096 + "input": 1.4, + "output": 1.4 } }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, + "huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated": { + "id": "huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated", + "name": "DeepSeek R1 Llama 70B Abliterated", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-31", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512000, - "output": 512000 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.32664, - "output": 1.30656, - "cache_read": 0.064239 + "input": 0.7, + "output": 0.7 } }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B": { + "id": "Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B", + "name": "Llama 3.05 Storybreaker Ministral 70b", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -192254,36 +181139,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B": { + "id": "Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B", + "name": "Nemotron Tenyxchat Storybreaker 70b", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-07-22", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ "text" @@ -192292,39 +181167,37 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 1.52432, - "output": 4.79072, - "cache_read": 0.152432 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "stepfun-ai/step-3.5-flash-2603": { + "id": "stepfun-ai/step-3.5-flash-2603", + "name": "Step 3.5 Flash 2603", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "high", - "xhigh" + "low", + "medium", + "high" ] } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-06", - "last_updated": "2026-07-22", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-14", + "last_updated": "2026-04-14", "modalities": { "input": [ "text" @@ -192333,22 +181206,22 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 2.4, - "output": 4.8, - "cache_read": 0.2 + "input": 0.1, + "output": 0.3 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "stepfun-ai/step-3.5-flash": { + "id": "stepfun-ai/step-3.5-flash", + "name": "Step 3.5 Flash", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "family": "step", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -192361,11 +181234,10 @@ ] } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-13", - "last_updated": "2026-07-22", + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-02", + "last_updated": "2026-02-02", "modalities": { "input": [ "text" @@ -192374,72 +181246,63 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 13107 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.178, - "output": 0.68, - "cache_read": 0.089 + "input": 0.2, + "output": 0.5 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "mistral/mistral-medium-3.5": { + "id": "mistral/mistral-medium-3.5", + "name": "Mistral Medium 3.5", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-31", + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 1.52432, - "output": 4.79072, - "cache_read": 0.283088 + "input": 1.5, + "output": 7.5 } }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "mistral/mistral-medium-3.5:thinking": { + "id": "mistral/mistral-medium-3.5:thinking", + "name": "Mistral Medium 3.5 Thinking", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-07-03", - "last_updated": "2026-07-22", + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", @@ -192449,45 +181312,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 256000, - "output": 16000 + "input": 256000, + "output": 32768 }, "cost": { - "input": 1.03436, - "output": 4.3552, - "cache_read": 0.206872 + "input": 1.5, + "output": 7.5 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", + "Tongyi-Zhiwen/QwenLong-L1-32B": { + "id": "Tongyi-Zhiwen/QwenLong-L1-32B", + "name": "QwenLong L1 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-15", - "last_updated": "2026-07-22", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-25", + "last_updated": "2025-01-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -192495,37 +181344,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 128000, + "input": 128000, + "output": 40960 }, "cost": { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24 + "input": 0.13999999999999999, + "output": 0.6 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "google/gemini-flash-1.5": { + "id": "google/gemini-flash-1.5", + "name": "Gemini 1.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-06", - "last_updated": "2026-07-22", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-05-14", + "last_updated": "2024-05-14", "modalities": { "input": [ "text" @@ -192534,42 +181372,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 2000000, + "input": 2000000, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 0.4, - "cache_read": 0.04 + "input": 0.0748, + "output": 0.306 } - } - } - }, - "jalapeno": { - "id": "jalapeno", - "env": [ - "JALAPENO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.jalapeno-cloud.ai/v1", - "name": "Jalapeno Cloud", - "doc": "https://www.jalapeno-cloud.ai/docs/", - "models": { - "Qwen3-VL-235B-A22B-Instruct": { - "id": "Qwen3-VL-235B-A22B-Instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", - "family": "qwen", + }, + "google/gemini-3.1-pro-preview-high": { + "id": "google/gemini-3.1-pro-preview-high", + "name": "Gemini 3.1 Pro (Preview High)", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-02-21", + "last_updated": "2026-02-21", "modalities": { "input": [ "text", @@ -192579,219 +181403,307 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 129024, - "output": 32768 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.5 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "google/gemini-3.1-flash-lite": { + "id": "google/gemini-3.1-flash-lite", + "name": "Gemini 3.1 Flash Lite", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 524288, - "output": 512000 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025 } }, - "Qwen3-Next-80B-A3B-Instruct": { - "id": "Qwen3-Next-80B-A3B-Instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "google/gemini-3-flash-preview-thinking": { + "id": "google/gemini-3-flash-preview-thinking", + "name": "Gemini 3 Flash Thinking", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 129024, - "output": 32768 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 0.5, + "output": 3 } }, - "GLM-5.1": { - "id": "GLM-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, + "google/gemini-3.5-flash": { + "id": "google/gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "audio" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "output": 131072 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 1.38, - "output": 4.4 + "input": 1.5, + "output": 9, + "cache_read": 0.15 } }, - "Hy3": { - "id": "Hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, "structured_output": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 202752, - "input": 192000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.58 + "input": 0.1, + "output": 0.35 } }, - "GLM-5.2": { - "id": "GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "google/gemini-pro-latest": { + "id": "google/gemini-pro-latest", + "name": "Gemini Pro Latest", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "structured_output": true, + "release_date": "2026-03-29", + "last_updated": "2026-03-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048756, + "input": 1048756, + "output": 65536 }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2 + } + }, + "google/gemini-3.1-pro-preview-customtools": { + "id": "google/gemini-3.1-pro-preview-customtools", + "name": "Gemini 3.1 Pro (Preview Custom Tools)", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-02-27", + "last_updated": "2026-02-27", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048756, + "input": 1048756, + "output": 65536 + }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2 + } + }, + "google/gemini-flash-lite-latest": { + "id": "google/gemini-flash-lite-latest", + "name": "Gemini Flash Lite Latest", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-03-29", + "last_updated": "2026-03-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 131072 + "input": 1048576, + "output": 65536 }, "cost": { - "input": 1.4, - "output": 4.4 + "input": 0.25, + "output": 1.5, + "cache_read": 0.025 } }, - "Qwen3-VL-235B-A22B-Thinking": { - "id": "Qwen3-VL-235B-A22B-Thinking", - "name": "Qwen3 VL 235B A22B Thinking", - "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "google/gemma-4-26b-a4b-it:thinking": { + "id": "google/gemma-4-26b-a4b-it:thinking", + "name": "Gemma 4 26B A4B Thinking", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", @@ -192801,62 +181713,62 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 262144, + "input": 262144, + "output": 131072 }, "cost": { - "input": 0.98, - "output": 3.95 + "input": 0.13, + "output": 0.4 } }, - "Qwen3.5-27B": { - "id": "Qwen3.5-27B", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "google/gemini-3.1-pro-preview": { + "id": "google/gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro (Preview)", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, + "context": 1048756, + "input": 1048756, "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.4 + "input": 2, + "output": 12, + "cache_read": 0.2 } }, - "Kimi-K3": { - "id": "Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "google/gemma-4-26b-a4b-it": { + "id": "google/gemma-4-26b-a4b-it", + "name": "Gemma 4 26B A4B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -192864,129 +181776,219 @@ "type": "toggle" } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, + "tool_call": false, "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, + "context": 262144, + "input": 262144, "output": 131072 }, "cost": { - "input": 3, - "output": 15 + "input": 0.13, + "output": 0.4 } }, - "Qwen3-Next-80B-A3B-Thinking": { - "id": "Qwen3-Next-80B-A3B-Thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", - "attachment": false, + "google/gemini-3.1-pro-preview-low": { + "id": "google/gemini-3.1-pro-preview-low", + "name": "Gemini 3.1 Pro (Preview Low)", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "structured_output": true, + "release_date": "2026-02-21", + "last_updated": "2026-02-21", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1048756, + "input": 1048756, + "output": 65536 }, + "cost": { + "input": 2, + "output": 12, + "cache_read": 0.2 + } + }, + "google/gemma-4-31b-it:thinking": { + "id": "google/gemma-4-31b-it:thinking", + "name": "Gemma 4 31B Thinking", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 262144, + "input": 262144, + "output": 131072 + }, + "cost": { + "input": 0.1, + "output": 0.35 + } + }, + "google/gemini-3.5-flash-thinking": { + "id": "google/gemini-3.5-flash-thinking", + "name": "Gemini 3.5 Flash Thinking", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-19", + "last_updated": "2026-05-19", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1048576, + "input": 1048576, + "output": 65536 }, "cost": { - "input": 0.15, - "output": 1.5 + "input": 1.5, + "output": 9, + "cache_read": 0.15 } }, - "DeepSeek-V4-Flash": { - "id": "DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "google/gemini-3-flash-preview": { + "id": "google/gemini-3-flash-preview", + "name": "Gemini 3 Flash (Preview)", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 1048756, + "input": 1048756, + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28 + "input": 0.5, + "output": 3 } }, - "DeepSeek-V4-Pro": { - "id": "DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, + "google/gemini-flash-latest": { + "id": "google/gemini-flash-latest", + "name": "Gemini Flash Latest", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "minimal", + "low", + "medium", + "high" + ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-29", + "last_updated": "2026-03-29", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "open_weights": false, + "limit": { + "context": 1048756, + "input": 1048756, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15 + } + }, + "liquid/lfm-2-24b-a2b": { + "id": "liquid/lfm-2-24b-a2b", + "name": "LFM2 24B A2B", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-20", + "last_updated": "2025-12-20", "modalities": { "input": [ "text" @@ -192995,244 +181997,231 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 1.6, - "output": 3.38 + "input": 0.03, + "output": 0.12 } }, - "Qwen3.5-35B-A3B": { - "id": "Qwen3.5-35B-A3B", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "x-ai/grok-4.20": { + "id": "x-ai/grok-4.20", + "name": "Grok 4.20", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 2000000, + "input": 2000000, + "output": 131072 }, "cost": { - "input": 0.25, - "output": 2 + "input": 2, + "output": 6 } }, - "Qwen3.5-122B-A10B": { - "id": "Qwen3.5-122B-A10B", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "x-ai/grok-4.3": { + "id": "x-ai/grok-4.3", + "name": "Grok 4.3", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "release_date": "2026-04-30", + "last_updated": "2026-04-30", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 1000000, + "input": 1000000, + "output": 1000000 }, "cost": { - "input": 0.4, - "output": 3.2 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "Qwen3.5-397B-A17B": { - "id": "Qwen3.5-397B-A17B", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", + "x-ai/grok-4.20-multi-agent": { + "id": "x-ai/grok-4.20-multi-agent", + "name": "Grok 4.20 Multi-Agent", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "release_date": "2026-03-31", + "last_updated": "2026-03-31", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 2000000, + "input": 2000000, + "output": 131072 }, "cost": { - "input": 0.6, - "output": 3.6 + "input": 2, + "output": 6 } }, - "Kimi-K2.7-Code": { - "id": "Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", + "x-ai/grok-latest": { + "id": "x-ai/grok-latest", + "name": "Grok Latest", + "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "release_date": "2026-05-03", + "last_updated": "2026-05-03", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 271360, - "output": 262144 + "context": 1000000, + "input": 1000000, + "output": 1000000 }, "cost": { - "input": 0.95, - "output": 4 + "input": 1.25, + "output": 2.5, + "cache_read": 0.2 } }, - "Kimi-K2.5": { - "id": "Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", + "x-ai/grok-build-0.1": { + "id": "x-ai/grok-build-0.1", + "name": "Grok Build 0.1", + "description": "Grok coding model for agentic engineering, edits, and codebase workflows", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-05-20", + "last_updated": "2026-05-20", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 180224 + "context": 256000, + "input": 256000, + "output": 256000 }, "cost": { - "input": 0.6, - "output": 3 + "input": 1, + "output": 2, + "cache_read": 0.2 } - } - } - }, - "dinference": { - "id": "dinference", - "env": [ - "DINFERENCE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.dinference.com/v1", - "name": "DInference", - "doc": "https://dinference.com", - "models": { - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + }, + "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0": { + "id": "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0", + "name": "EVA Llama 3.33 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text" @@ -193241,28 +182230,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 0.75, - "output": 2.4 + "input": 2.006, + "output": 2.006 } }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2": { + "id": "EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2", + "name": "EVA-Qwen2.5-72B-v0.2", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text" @@ -193271,32 +182260,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.22, - "output": 0.88 + "input": 0.7989999999999999, + "output": 0.7989999999999999 } }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", + "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1": { + "id": "EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1", + "name": "EVA-LLaMA-3.33-70B-v0.1", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text" @@ -193305,32 +182290,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 0.45, - "output": 1.65 + "input": 2.006, + "output": 2.006 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2": { + "id": "EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2", + "name": "EVA-Qwen2.5-32B-v0.2", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text" @@ -193339,60 +182320,59 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 3.89 + "input": 0.7989999999999999, + "output": 0.7989999999999999 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "microsoft/wizardlm-2-8x22b": { + "id": "microsoft/wizardlm-2-8x22b", + "name": "WizardLM-2 8x22B", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 65536, + "input": 65536, + "output": 8192 }, "cost": { - "input": 1.25, - "output": 3.89 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "Doctor-Shotgun/MS3.2-24B-Magnum-Diamond": { + "id": "Doctor-Shotgun/MS3.2-24B-Magnum-Diamond", + "name": "MS3.2 24B Magnum Diamond", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-08", - "last_updated": "2025-08", + "tool_call": false, + "structured_output": false, + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text" @@ -193401,106 +182381,91 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 16384, + "input": 16384, "output": 32768 }, "cost": { - "input": 0.0675, - "output": 0.27 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } - } - } - }, - "neosmith": { - "id": "neosmith", - "env": [ - "NEOSMITH_API_KEY" - ], - "npm": "@ai-sdk/openai", - "api": "https://router.neosmith.ai/v1", - "name": "NeoSmith", - "doc": "https://neosmith.ai/docs", - "models": { - "neosmith.neolite": { - "id": "neosmith.neolite", - "name": "NeoSmith NeoLite", - "description": "Sealed single-model budget tier. 512K context, text and images, tool use, and no escalation of any kind.", - "attachment": true, + }, + "poolside/laguna-xs.2": { + "id": "poolside/laguna-xs.2", + "name": "Laguna XS.2", + "description": "Agentic coding model from Poolside in the XS size class for local deployment", + "family": "laguna", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-06-20", - "last_updated": "2026-08-03", + "release_date": "2026-04-28", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 512000, - "output": 64000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.08, - "cache_write": 0 + "input": 0.2, + "output": 0.4 } }, - "neosmith.intelligent-maestro": { - "id": "neosmith.intelligent-maestro", - "name": "NeoSmith Maestro", - "description": "Highest-accuracy coding tier. Hard, self-contained problems run NeoSmith's premium multi-model solver; everything else gets the strongest intelligence tier.", - "attachment": true, + "poolside/laguna-m.1": { + "id": "poolside/laguna-m.1", + "name": "Laguna M.1", + "description": "Poolside's flagship agentic coding model for long-horizon work", + "family": "laguna", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": false, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-08", + "release_date": "2026-04-28", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 32768 }, "cost": { - "input": 2.4, - "output": 12, - "cache_read": 0.35, - "cache_write": 0 + "input": 0.2, + "output": 0.4 } }, - "neosmith.intelligent-pro": { - "id": "neosmith.intelligent-pro", - "name": "NeoSmith Pro", - "description": "Default production tier. Intelligent NeoSmith routing with a Claude Opus ceiling on escalation.", + "z-ai/glm-4.5v:thinking": { + "id": "z-ai/glm-4.5v:thinking", + "name": "GLM 4.5V Thinking", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glmv", "attachment": true, "reasoning": true, "reasoning_options": [], - "tool_call": true, + "tool_call": false, "structured_output": false, - "temperature": true, - "release_date": "2026-04-26", - "last_updated": "2026-07-18", + "release_date": "2025-11-22", + "last_updated": "2025-11-22", "modalities": { "input": [ "text", @@ -193512,32 +182477,30 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 64000, + "input": 64000, + "output": 96000 }, "cost": { - "input": 1.81, - "output": 8.39, - "cache_read": 0.3, - "cache_write": 0 + "input": 0.6, + "output": 1.7999999999999998 } }, - "neosmith.intelligent-basic": { - "id": "neosmith.intelligent-basic", - "name": "NeoSmith Basic", - "description": "Cost-capped tier. Intelligent routing with a Claude Sonnet ceiling — Opus is never invoked.", - "attachment": true, + "z-ai/glm-4.6:thinking": { + "id": "z-ai/glm-4.6:thinking", + "name": "GLM 4.6 Thinking", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-26", - "last_updated": "2026-08-03", + "structured_output": true, + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -193545,33 +182508,20 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 200000, + "input": 200000, + "output": 65535 }, "cost": { - "input": 1.17, - "output": 4.37, - "cache_read": 0.22, - "cache_write": 0 + "input": 0.4, + "output": 1.5 } - } - } - }, - "fireworks-ai": { - "id": "fireworks-ai", - "env": [ - "FIREWORKS_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.fireworks.ai/inference/v1/", - "name": "Fireworks AI", - "doc": "https://fireworks.ai/docs/", - "models": { - "accounts/fireworks/models/kimi-k2p6": { - "id": "accounts/fireworks/models/kimi-k2p6", - "name": "Kimi K2.6", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", + }, + "z-ai/glm-4.5v": { + "id": "z-ai/glm-4.5v", + "name": "GLM 4.5V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glmv", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -193579,13 +182529,10 @@ "type": "toggle" } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "tool_call": false, + "structured_output": false, + "release_date": "2025-11-22", + "last_updated": "2025-11-22", "modalities": { "input": [ "text", @@ -193595,39 +182542,22 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 - }, - "status": "deprecated", - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 1.5, - "output": 6, - "cache_read": 0.22 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 64000, + "input": 64000, + "output": 96000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 0.6, + "output": 1.7999999999999998 } }, - "accounts/fireworks/models/nemotron-3-ultra-nvfp4": { - "id": "accounts/fireworks/models/nemotron-3-ultra-nvfp4", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", + "z-ai/glm-4.6": { + "id": "z-ai/glm-4.6", + "name": "GLM 4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -193636,9 +182566,9 @@ } ], "tool_call": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", + "structured_output": true, + "release_date": "2025-09-30", + "last_updated": "2025-09-30", "modalities": { "input": [ "text" @@ -193647,106 +182577,59 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 128000 + "context": 200000, + "input": 200000, + "output": 65535 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 + "input": 0.4, + "output": 1.5 } }, - "accounts/fireworks/models/deepseek-v4-flash-0731": { - "id": "accounts/fireworks/models/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "z-ai/glm-5v-turbo:thinking": { + "id": "z-ai/glm-5v-turbo:thinking", + "name": "GLM 5V Turbo Thinking", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "status": "deprecated", - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 0.275, - "output": 0.825, - "cache_read": 0.00875 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 202800, + "input": 202800, + "output": 131100 }, "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "accounts/fireworks/models/muse-glimmer-30b": { - "id": "accounts/fireworks/models/muse-glimmer-30b", - "name": "Muse Glimmer 30B", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", + "z-ai/glm-5v-turbo": { + "id": "z-ai/glm-5v-turbo", + "name": "GLM 5V Turbo", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", @@ -193756,92 +182639,84 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 202800, + "input": 202800, + "output": 131100 }, - "status": "deprecated", "cost": { - "input": 0.35, - "output": 1.5, - "cache_read": 0.04 + "input": 1.2, + "output": 4, + "cache_read": 0.24 } }, - "accounts/fireworks/models/kimi-k3": { - "id": "accounts/fireworks/models/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], + "z-ai/glm-5-turbo": { + "id": "z-ai/glm-5-turbo", + "name": "GLM 5 Turbo", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-07-27", - "last_updated": "2026-07-27", + "release_date": "2026-03-15", + "last_updated": "2026-03-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, + "context": 202800, + "input": 202800, "output": 131072 }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 3.75, - "output": 18.75, - "cache_read": 0.375 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "cost": { + "input": 1.2, + "output": 4, + "cache_read": 0.24 + } + }, + "openai/o3-mini-low": { + "id": "openai/o3-mini-low", + "name": "OpenAI o3-mini (Low)", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "release_date": "2025-01-31", + "last_updated": "2025-01-31", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 9.996, + "output": 19.992 } }, - "accounts/fireworks/models/qwen3p8-2p4t-a95b": { - "id": "accounts/fireworks/models/qwen3p8-2p4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", + "openai/gpt-oss-safeguard-20b": { + "id": "openai/gpt-oss-safeguard-20b", + "name": "GPT OSS Safeguard 20B", + "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -193850,18 +182725,44 @@ "values": [ "low", "medium", - "xhigh" + "high" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "tool_call": false, + "structured_output": false, + "release_date": "2025-10-29", + "last_updated": "2025-10-29", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "open_weights": false, + "limit": { + "context": 128000, + "input": 128000, + "output": 16384 + }, + "cost": { + "input": 0.075, + "output": 0.3 + } + }, + "openai/o3": { + "id": "openai/o3", + "name": "OpenAI o3", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text" @@ -193870,163 +182771,116 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { "input": 2, - "output": 6, - "cache_read": 0.25 + "output": 8 } }, - "accounts/fireworks/models/deepseek-v4-flash-vision-exp": { - "id": "accounts/fireworks/models/deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", - "attachment": true, + "openai/o4-mini-high": { + "id": "openai/o4-mini-high", + "name": "OpenAI o4-mini high", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 200000, + "input": 200000, + "output": 100000 }, - "status": "deprecated", "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 + "input": 1.1, + "output": 4.4 } }, - "accounts/fireworks/models/deepseek-v4p1-flash": { - "id": "accounts/fireworks/models/deepseek-v4p1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "openai/o3-pro-2025-06-10": { + "id": "openai/o3-pro-2025-06-10", + "name": "OpenAI o3-pro (2025-06-10)", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2025-06-10", + "last_updated": "2025-06-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 0.275, - "output": 0.825, - "cache_read": 0.00875 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 + "input": 9.996, + "output": 19.992 } }, - "accounts/fireworks/models/qwen3p7-plus": { - "id": "accounts/fireworks/models/qwen3p7-plus", - "name": "Qwen 3.7 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "openai/gpt-5.2-pro": { + "id": "openai/gpt-5.2-pro", + "name": "GPT 5.2 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] - }, - { - "type": "budget_tokens", - "min": 1 } ], "tool_call": true, - "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "structured_output": true, + "release_date": "2026-01-01", + "last_updated": "2026-01-01", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -194034,40 +182888,26 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08 + "input": 21, + "output": 168 } }, - "accounts/fireworks/models/glm-5p3": { - "id": "accounts/fireworks/models/glm-5p3", - "name": "GLM 5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "openai/gpt-4o-mini-search-preview": { + "id": "openai/gpt-4o-mini-search-preview", + "name": "GPT-4o mini Search Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-09-07", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ "text" @@ -194076,44 +182916,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048573, - "output": 262144 - }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 1.75, - "output": 5.5, - "cache_read": 0.325 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.088, + "output": 0.35 } }, - "accounts/fireworks/models/minimax-m2p7": { - "id": "accounts/fireworks/models/minimax-m2p7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, + "openai/gpt-5": { + "id": "openai/gpt-5", + "name": "GPT 5", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -194121,54 +182946,41 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "structured_output": true, + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 196608, - "output": 131072 - }, - "status": "deprecated", - "provider": { - "body": { - "service_tier": "priority" - } + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 1.2, - "output": 1.2, - "cache_read": 0.6 + "input": 1.25, + "output": 10 } }, - "accounts/fireworks/models/nemotron-lightning-3p5-30b-a3b": { - "id": "accounts/fireworks/models/nemotron-lightning-3p5-30b-a3b", - "name": "Nemotron 3.5 Lightning 30B A3B", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron", + "openai/gpt-3.5-turbo": { + "id": "openai/gpt-3.5-turbo", + "name": "GPT-3.5 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2022-11-30", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" @@ -194177,139 +182989,108 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 16385, + "input": 16385, + "output": 4096 }, "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.01 + "input": 0.5, + "output": 1.5 } }, - "accounts/fireworks/models/deepseek-v4-pro-0813": { - "id": "accounts/fireworks/models/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, + "openai/gpt-5-pro": { + "id": "openai/gpt-5-pro", + "name": "GPT 5 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", + "family": "gpt-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "max" + "high" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "status": "deprecated", - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 1.65, - "output": 4.95, - "cache_read": 0.055 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 + "input": 15, + "output": 120 } }, - "accounts/fireworks/models/inkling": { - "id": "accounts/fireworks/models/inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", + "openai/gpt-4o": { + "id": "openai/gpt-4o", + "name": "GPT-4o", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.17 + "input": 2.499, + "output": 9.996 } }, - "accounts/fireworks/models/glm-5p2": { - "id": "accounts/fireworks/models/glm-5p2", - "name": "GLM 5.2", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "openai/o4-mini": { + "id": "openai/o4-mini", + "name": "OpenAI o4-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "max" + "low", + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-16", - "last_updated": "2026-06-16", + "structured_output": true, + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text" @@ -194318,112 +183099,82 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048575, - "output": 131072 - }, - "status": "deprecated", - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 1.75, - "output": 5.5, - "cache_read": 0.175 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.14 + "input": 1.1, + "output": 4.4 } }, - "accounts/fireworks/models/minimax-m3": { - "id": "accounts/fireworks/models/minimax-m3", - "name": "MiniMax-M3", - "description": "Fireworks text-only MiniMax coding model for long-context reasoning and agent tasks", - "family": "minimax", - "attachment": false, + "openai/gpt-5.4-nano": { + "id": "openai/gpt-5.4-nano", + "name": "GPT 5.4 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "structured_output": true, + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512000, - "output": 512000 - }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 0.45, - "output": 1.8, - "cache_read": 0.09 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "accounts/fireworks/models/kimi-k2p7-code": { - "id": "accounts/fireworks/models/kimi-k2p7-code", - "name": "Kimi K2.7 Code", - "description": "Kimi coding model for software agents, refactors, and repository reasoning", - "family": "kimi-k2", + "openai/gpt-5.1-codex": { + "id": "openai/gpt-5.1-codex", + "name": "GPT 5.1 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-16", + "tool_call": false, + "structured_output": false, + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -194433,39 +183184,22 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 - }, - "status": "deprecated", - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 1.425, - "output": 6, - "cache_read": 0.285 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 1.25, + "output": 10 } }, - "accounts/fireworks/models/glm-5p3-flash": { - "id": "accounts/fireworks/models/glm-5p3-flash", - "name": "GLM 5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "openai/gpt-5.1-codex-max": { + "id": "openai/gpt-5.1-codex-max", + "name": "GPT 5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -194473,124 +183207,73 @@ "type": "effort", "values": [ "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-09-07", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048573, - "output": 131072 - }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 0.1875, - "output": 0.625, - "cache_read": 0.0375 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 2.5, + "output": 20 } }, - "accounts/fireworks/models/deepseek-v4-pro": { - "id": "accounts/fireworks/models/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "openai/gpt-4o-2024-08-06": { + "id": "openai/gpt-4o-2024-08-06", + "name": "GPT-4o (2024-08-06)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-08-06", + "last_updated": "2024-08-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "status": "deprecated", - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 1.2, - "output": 1.2, - "cache_read": 0.6 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 1.2, - "output": 1.2, - "cache_read": 0.6 + "input": 2.499, + "output": 9.996 } }, - "accounts/fireworks/models/gpt-oss-120b": { - "id": "accounts/fireworks/models/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "openai/o1-preview": { + "id": "openai/o1-preview", + "name": "OpenAI o1-preview", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -194603,10 +183286,10 @@ ] } ], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2026-06-16", + "tool_call": false, + "structured_output": false, + "release_date": "2024-09-12", + "last_updated": "2024-09-12", "modalities": { "input": [ "text" @@ -194615,53 +183298,41 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 128000, + "input": 128000, "output": 32768 }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 0.18, - "output": 0.72, - "cache_read": 0.018 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 + "input": 14.993999999999998, + "output": 59.993 } }, - "accounts/fireworks/models/qwen3p8-max": { - "id": "accounts/fireworks/models/qwen3p8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, + "openai/o3-mini": { + "id": "openai/o3-mini", + "name": "OpenAI o3-mini", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "structured_output": true, + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -194669,95 +183340,63 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 - }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 3, - "output": 9, - "cache_read": 0.375 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 1.1, + "output": 4.4 } }, - "accounts/fireworks/routers/minimax-latest": { - "id": "accounts/fireworks/routers/minimax-latest", - "name": "MiniMax Latest", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": false, + "openai/gpt-5.2": { + "id": "openai/gpt-5.2", + "name": "GPT 5.2", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-09-15", + "structured_output": true, + "release_date": "2026-01-01", + "last_updated": "2026-01-01", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512000, - "output": 512000 - }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 0.45, - "output": 1.8, - "cache_read": 0.09 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 1.75, + "output": 14 } }, - "accounts/fireworks/routers/glm-flash-latest": { - "id": "accounts/fireworks/routers/glm-flash-latest", - "name": "GLM Flash Latest (GLM 5.3 Flash)", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "openai/gpt-5.3-codex": { + "id": "openai/gpt-5.3-codex", + "name": "GPT 5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -194765,132 +183404,103 @@ "type": "effort", "values": [ "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-09-15", + "release_date": "2026-02-24", + "last_updated": "2026-02-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048573, - "output": 131072 - }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 0.1875, - "output": 0.625, - "cache_read": 0.0375 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "accounts/fireworks/routers/glm-fast-latest": { - "id": "accounts/fireworks/routers/glm-fast-latest", - "name": "GLM 5.3 Fast (Latest)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, + "openai/gpt-latest": { + "id": "openai/gpt-latest", + "name": "GPT Latest", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-09-15", + "release_date": "2026-03-29", + "last_updated": "2026-03-29", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048572, - "output": 262144 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 2.1, - "output": 6.6, - "cache_read": 0.39 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "accounts/fireworks/routers/kimi-k3-fast": { - "id": "accounts/fireworks/routers/kimi-k3-fast", - "name": "Kimi K3 Fast", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "openai/gpt-5.1-codex-mini": { + "id": "openai/gpt-5.1-codex-mini", + "name": "GPT 5.1 Codex Mini", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex-mini", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-27", - "last_updated": "2026-07-27", + "tool_call": false, + "structured_output": false, + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -194900,42 +183510,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 4.5, - "output": 22.5, - "cache_read": 0.45 + "input": 0.25, + "output": 2 } }, - "accounts/fireworks/routers/glm-5p3-fast": { - "id": "accounts/fireworks/routers/glm-5p3-fast", - "name": "GLM 5.3 Fast", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "openai/o4-mini-deep-research": { + "id": "openai/o4-mini-deep-research", + "name": "OpenAI o4-mini Deep Research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "family": "o-mini", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "high", - "max" + "medium" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-09-07", + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text" @@ -194944,86 +183548,100 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048572, - "output": 262144 + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 2.1, - "output": 6.6, - "cache_read": 0.39 + "input": 9.996, + "output": 19.992 } }, - "accounts/fireworks/routers/kimi-fast-latest": { - "id": "accounts/fireworks/routers/kimi-fast-latest", - "name": "Kimi Fast Latest", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "openai/gpt-4.1-nano": { + "id": "openai/gpt-4.1-nano", + "name": "GPT 4.1 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1047576, + "input": 1047576, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.4 + } + }, + "openai/gpt-oss-120b": { + "id": "openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-07-27", - "last_updated": "2026-09-15", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 4.5, - "output": 22.5, - "cache_read": 0.45 + "input": 0.05, + "output": 0.25 } }, - "accounts/fireworks/routers/qwen-max-latest": { - "id": "accounts/fireworks/routers/qwen-max-latest", - "name": "Qwen Max Latest (Qwen3.8 Max)", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "openai/gpt-4o-2024-11-20": { + "id": "openai/gpt-4o-2024-11-20", + "name": "GPT-4o (2024-11-20)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-09-16", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-11-20", + "last_updated": "2024-11-20", "modalities": { "input": [ "text", @@ -195035,237 +183653,258 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 131072 - }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 3, - "output": 9, - "cache_read": 0.375 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 + "input": 2.5, + "output": 10 } }, - "accounts/fireworks/routers/kimi-latest": { - "id": "accounts/fireworks/routers/kimi-latest", - "name": "Kimi Latest", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "openai/o1": { + "id": "openai/o1", + "name": "OpenAI o1", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", "medium", - "high", - "max" + "high" ] - }, - { - "type": "budget_tokens", - "min": 1024 } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-17", + "last_updated": "2024-12-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-27", - "last_updated": "2026-09-15", + "open_weights": false, + "limit": { + "context": 200000, + "input": 200000, + "output": 100000 + }, + "cost": { + "input": 14.993999999999998, + "output": 59.993 + } + }, + "openai/o1-pro": { + "id": "openai/o1-pro", + "name": "OpenAI o1 Pro", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-pro", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-25", + "last_updated": "2025-01-25", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 100000 }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 3.75, - "output": 18.75, - "cache_read": 0.375 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "cost": { + "input": 150, + "output": 600 + } + }, + "openai/gpt-chat-latest": { + "id": "openai/gpt-chat-latest", + "name": "GPT Chat Latest", + "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "release_date": "2026-05-03", + "last_updated": "2026-05-03", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "accounts/fireworks/routers/deepseek-flash-latest": { - "id": "accounts/fireworks/routers/deepseek-flash-latest", - "name": "DeepSeek Flash Latest", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "openai/gpt-5.4": { + "id": "openai/gpt-5.4", + "name": "GPT 5.4", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 0.275, - "output": 0.825, - "cache_read": 0.00875 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 922000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 + "input": 2.5, + "output": 15, + "cache_read": 0.25 } }, - "accounts/fireworks/routers/glm-5p2-fast": { - "id": "accounts/fireworks/routers/glm-5p2-fast", - "name": "GLM 5.2 Fast", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm", - "attachment": false, + "openai/gpt-5.4-mini": { + "id": "openai/gpt-5.4-mini", + "name": "GPT 5.4 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", + "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-26", - "last_updated": "2026-06-26", + "structured_output": true, + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ + "text", + "image", + "pdf" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 400000, + "output": 128000 + }, + "cost": { + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 + } + }, + "openai/gpt-4.1": { + "id": "openai/gpt-4.1", + "name": "GPT 4.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": true, + "structured_output": true, + "release_date": "2025-09-10", + "last_updated": "2025-09-10", + "modalities": { + "input": [ + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048575, - "output": 131072 + "context": 1047576, + "input": 1047576, + "output": 32768 }, "cost": { - "input": 2.1, - "output": 6.6, - "cache_read": 0.21 + "input": 2, + "output": 8 } }, - "accounts/fireworks/routers/deepseek-pro-latest": { - "id": "accounts/fireworks/routers/deepseek-pro-latest", - "name": "DeepSeek Pro Latest", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "openai/o3-deep-research": { + "id": "openai/o3-deep-research", + "name": "OpenAI o3 Deep Research", + "description": "Research model for long-horizon investigation, synthesis, and analytical reports", + "family": "o", "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ - "high", - "max" + "medium" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-16", + "last_updated": "2025-04-16", "modalities": { "input": [ "text" @@ -195274,133 +183913,177 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 - }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 1.65, - "output": 4.95, - "cache_read": 0.055 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 + "input": 9.996, + "output": 19.992 } }, - "accounts/fireworks/routers/glm-latest": { - "id": "accounts/fireworks/routers/glm-latest", - "name": "GLM Latest", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "openai/gpt-4-turbo-preview": { + "id": "openai/gpt-4-turbo-preview", + "name": "GPT-4 Turbo Preview", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2023-11-06", + "last_updated": "2024-01-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 128000, + "output": 4096 + }, + "cost": { + "input": 9.996, + "output": 30.004999999999995 + } + }, + "openai/gpt-5-mini": { + "id": "openai/gpt-5-mini", + "name": "GPT 5 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", - "high", - "max" + "medium", + "high" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-09-15", + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ + "text", + "image" + ], + "output": [ "text" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 400000, + "output": 128000 + }, + "cost": { + "input": 0.25, + "output": 2 + } + }, + "openai/gpt-4.1-mini": { + "id": "openai/gpt-4.1-mini", + "name": "GPT 4.1 Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-14", + "last_updated": "2025-04-14", + "modalities": { + "input": [ + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048573, - "output": 262144 + "context": 1047576, + "input": 1047576, + "output": 32768 }, - "experimental": { - "modes": { - "priority": { - "cost": { - "input": 1.75, - "output": 5.5, - "cache_read": 0.325 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "cost": { + "input": 0.4, + "output": 1.6 + } + }, + "openai/gpt-4-turbo": { + "id": "openai/gpt-4-turbo", + "name": "GPT-4 Turbo", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2023-11-06", + "last_updated": "2024-01-01", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 10, + "output": 30 } - } - } - }, - "stepfun-ai": { - "id": "stepfun-ai", - "env": [ - "STEPFUN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.stepfun.ai/v1", - "name": "StepFun (Global)", - "doc": "https://platform.stepfun.ai/docs/en/overview/concept", - "models": { - "step-5-preview": { - "id": "step-5-preview", - "name": "Step 5 Preview", - "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", + }, + "openai/gpt-5-nano": { + "id": "openai/gpt-5-nano", + "name": "GPT 5 Nano", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-nano", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "release_date": "2026-09-16", - "last_updated": "2026-09-20", + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -195408,85 +184091,71 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 1, - "output": 2.7, - "cache_read": 0.05 + "input": 0.05, + "output": 0.4 } }, - "step-3.7-flash": { - "id": "step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", + "openai/gpt-5.4-pro": { + "id": "openai/gpt-5.4-pro", + "name": "GPT 5.4 Pro", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-06-29", + "structured_output": true, + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 922000, + "input": 922000, + "output": 128000 }, "cost": { - "input": 0.185, - "output": 1.11, - "cache_read": 0.037 + "input": 30, + "output": 180, + "cache_read": 3 } }, - "step-3.5-flash": { - "id": "step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", + "openai/o3-mini-high": { + "id": "openai/o3-mini-high", + "name": "OpenAI o3-mini (High)", + "description": "O-series reasoning model for hard analysis, math, coding, and planning", + "family": "o-mini", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-06-15", + "structured_output": true, + "release_date": "2025-01-31", + "last_updated": "2025-01-31", "modalities": { "input": [ "text" @@ -195495,32 +184164,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 200000, + "input": 200000, + "output": 100000 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 0.64, + "output": 2.588 } }, - "stepaudio-2.5-asr": { - "id": "stepaudio-2.5-asr", - "name": "StepAudio 2.5 ASR", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "step", - "attachment": false, + "openai/gpt-4o-search-preview": { + "id": "openai/gpt-4o-search-preview", + "name": "GPT-4o Search Preview", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-04-24", - "last_updated": "2026-07-02", + "structured_output": false, + "release_date": "2024-05-13", + "last_updated": "2024-05-13", "modalities": { "input": [ - "audio" + "text", + "image" ], "output": [ "text" @@ -195528,49 +184197,60 @@ }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 128000, + "input": 128000, + "output": 16384 + }, + "cost": { + "input": 1.47, + "output": 5.88 } }, - "stepaudio-2.5-tts": { - "id": "stepaudio-2.5-tts", - "name": "StepAudio 2.5 TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "step", + "openai/gpt-5.1-2025-11-13": { + "id": "openai/gpt-5.1-2025-11-13", + "name": "GPT-5.1 (2025-11-13)", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-04-16", - "last_updated": "2026-07-02", + "structured_output": false, + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "input": 1000000, + "output": 32768 + }, + "cost": { + "input": 1.25, + "output": 10 } }, - "step-2-16k": { - "id": "step-2-16k", - "name": "Step 2 (16K)", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, + "openai/gpt-4o-mini": { + "id": "openai/gpt-4o-mini", + "name": "GPT-4o mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt-mini", + "attachment": true, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-01-01", - "last_updated": "2026-02-13", + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -195578,20 +184258,20 @@ }, "open_weights": false, "limit": { - "context": 16384, - "input": 16384, - "output": 8192 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 5.21, - "output": 16.44, - "cache_read": 1.04 + "input": 0.1496, + "output": 0.595 } }, - "step-3.5-flash-2603": { - "id": "step-3.5-flash-2603", - "name": "Step 3.5 Flash 2603", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "openai/gpt-oss-20b": { + "id": "openai/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -195599,18 +184279,15 @@ "type": "effort", "values": [ "low", + "medium", "high" ] } ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -195619,54 +184296,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 256000, - "output": 256000 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 + "input": 0.04, + "output": 0.15 } }, - "step-tts-2": { - "id": "step-tts-2", - "name": "Step TTS 2", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "step", + "openai/gpt-5-codex": { + "id": "openai/gpt-5-codex", + "name": "GPT-5 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-03-01", - "last_updated": "2026-07-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "step-1-32k": { - "id": "step-1-32k", - "name": "Step 1 (32K)", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-01-01", - "last_updated": "2026-02-13", + "structured_output": false, + "release_date": "2025-09-15", + "last_updated": "2025-09-15", "modalities": { "input": [ "text" @@ -195677,47 +184328,37 @@ }, "open_weights": false, "limit": { - "context": 32768, - "input": 32768, + "context": 256000, + "input": 256000, "output": 32768 }, "cost": { - "input": 2.05, - "output": 9.59, - "cache_read": 0.41 + "input": 9.996, + "output": 19.992 } - } - } - }, - "fastrouter": { - "id": "fastrouter", - "env": [ - "FASTROUTER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://go.fastrouter.ai/api/v1", - "name": "FastRouter", - "doc": "https://fastrouter.ai/models", - "models": { - "anthropic/claude-opus-4.1": { - "id": "anthropic/claude-opus-4.1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + }, + "openai/gpt-5.2-codex": { + "id": "openai/gpt-5.2-codex", + "name": "GPT 5.2 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 32000 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "structured_output": true, + "release_date": "2026-01-14", + "last_updated": "2026-01-14", "modalities": { "input": [ "text", @@ -195730,35 +184371,37 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 400000, + "input": 400000, + "output": 128000 }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 1.75, + "output": 14 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "openai/gpt-5.1": { + "id": "openai/gpt-5.1", + "name": "GPT 5.1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 32000 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", @@ -195771,33 +184414,37 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 400000, "output": 128000 }, "cost": { - "input": 5, - "output": 25 + "input": 1.25, + "output": 10 } }, - "anthropic/claude-sonnet-4": { - "id": "anthropic/claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "openai/gpt-5.5": { + "id": "openai/gpt-5.5", + "name": "GPT 5.5", + "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens", - "min": 1024, - "max": 32000 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "structured_output": true, + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -195810,40 +184457,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 5, + "output": 30, + "cache_read": 0.5 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 32000 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "Gryphe/MythoMax-L2-13b": { + "id": "Gryphe/MythoMax-L2-13b", + "name": "MythoMax 13B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -195851,32 +184488,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 4000, + "input": 4000, + "output": 4096 }, "cost": { - "input": 3, - "output": 15 + "input": 0.1003, + "output": 0.1003 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "Unbabel/M-Prometheus-14B": { + "id": "Unbabel/M-Prometheus-14B", + "name": "M-Prometheus 14B", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-05-29", + "last_updated": "2026-05-29", "modalities": { "input": [ "text" @@ -195885,35 +184515,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 1.74, - "output": 3.48 + "input": 0.2, + "output": 0.2 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "LLM360/K2-Think": { + "id": "LLM360/K2-Think", + "name": "K2-Think", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text" @@ -195922,33 +184545,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 32768 }, "cost": { - "input": 0.95, - "output": 3.15 + "input": 0.17, + "output": 0.68 } }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", + "NousResearch/hermes-4-405b": { + "id": "NousResearch/hermes-4-405b", + "name": "Hermes 4 Large", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2025-08-26", + "last_updated": "2025-08-26", "modalities": { "input": [ "text" @@ -195957,28 +184574,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 200000, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 1.05, - "output": 3.5 + "input": 0.3, + "output": 1.2 } }, - "sarvam/sarvam-30b": { - "id": "sarvam/sarvam-30b", - "name": "Sarvam 30B", - "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", - "family": "sarvam", + "NousResearch/hermes-3-llama-3.1-70b": { + "id": "NousResearch/hermes-3-llama-3.1-70b", + "name": "Hermes 3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-18", - "last_updated": "2026-02-18", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-01-07", + "last_updated": "2026-01-07", "modalities": { "input": [ "text" @@ -195987,28 +184603,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 128000 + "context": 65536, + "input": 65536, + "output": 8192 }, "cost": { - "input": 0.02, - "output": 0.1 + "input": 0.408, + "output": 0.408 } }, - "sarvam/sarvam-105b": { - "id": "sarvam/sarvam-105b", - "name": "Sarvam 105B", - "description": "Flagship Indian-language reasoning model for enterprise multilingual applications", - "family": "sarvam", + "NousResearch/Hermes-4-70B:thinking": { + "id": "NousResearch/Hermes-4-70B:thinking", + "name": "Hermes 4 (Thinking)", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-09-01", - "last_updated": "2025-09-01", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-17", + "last_updated": "2025-09-17", "modalities": { "input": [ "text" @@ -196017,29 +184632,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 0.04, - "output": 0.16 + "input": 0.2006, + "output": 0.3995 } }, - "x-ai/grok-4": { - "id": "x-ai/grok-4", - "name": "Grok 4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", + "NousResearch/hermes-4-70b": { + "id": "NousResearch/hermes-4-70b", + "name": "Hermes 4 Medium", + "description": "General-purpose chat model for instruction following, writing, and analysis", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-09", - "last_updated": "2025-07-09", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-03", + "last_updated": "2025-07-03", "modalities": { "input": [ "text" @@ -196050,43 +184663,28 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75, - "cache_write": 15 + "input": 0.2006, + "output": 0.3995 } }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "NousResearch/DeepHermes-3-Mistral-24B-Preview": { + "id": "NousResearch/DeepHermes-3-Mistral-24B-Preview", + "name": "DeepHermes-3 Mistral 24B (Preview)", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-10", + "last_updated": "2025-05-10", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -196094,41 +184692,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 128000, + "input": 128000, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 2.5 + "input": 0.3, + "output": 0.3 } }, - "x-ai/grok-build-0.1": { - "id": "x-ai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "NousResearch/hermes-4-405b:thinking": { + "id": "NousResearch/hermes-4-405b:thinking", + "name": "Hermes 4 Large (Thinking)", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -196136,255 +184721,243 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 256000 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 1, - "output": 2 + "input": 0.3, + "output": 1.2 } }, - "bytedance/seedance-2": { - "id": "bytedance/seedance-2", - "name": "Seedance 2", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "seed", + "unsloth/gemma-3-12b-it": { + "id": "unsloth/gemma-3-12b-it", + "name": "Gemma 3 12B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "unsloth", "attachment": true, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", + "structured_output": false, + "release_date": "2025-03-10", + "last_updated": "2025-03-10", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 4096, - "output": 0 + "context": 128000, + "input": 128000, + "output": 131072 + }, + "cost": { + "input": 0.272, + "output": 0.272 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", + "unsloth/gemma-3-4b-it": { + "id": "unsloth/gemma-3-4b-it", + "name": "Gemma 3 4B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "unsloth", "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-03-10", + "last_updated": "2025-03-10", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 0.13, - "output": 0.38 + "input": 0.2006, + "output": 0.2006 } }, - "google/gemini-3.1-flash-image-preview": { - "id": "google/gemini-3.1-flash-image-preview", - "name": "Nano Banana 2 Preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", + "unsloth/gemma-3-27b-it": { + "id": "unsloth/gemma-3-27b-it", + "name": "Gemma 3 27B IT", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "unsloth", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", + "structured_output": false, + "release_date": "2025-03-10", + "last_updated": "2025-03-10", "modalities": { "input": [ "text", - "image", "pdf" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 65536 + "context": 128000, + "input": 128000, + "output": 96000 }, "cost": { - "input": 0.5, - "output": 3 + "input": 0.2992, + "output": 0.2992 } }, - "google/veo3.1": { - "id": "google/veo3.1", - "name": "Veo 3.1", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, + "NeverSleep/Lumimaid-v0.2-70B": { + "id": "NeverSleep/Lumimaid-v0.2-70B", + "name": "Lumimaid v0.2", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "structured_output": false, + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 0 + "context": 16384, + "input": 16384, + "output": 8192 + }, + "cost": { + "input": 1, + "output": 1.5 } }, - "google/imagen-4.0-ultra": { - "id": "google/imagen-4.0-ultra", - "name": "Imagen 4 Ultra", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", + "mistralai/mixtral-8x7b-instruct-v0.1": { + "id": "mistralai/mixtral-8x7b-instruct-v0.1", + "name": "Mixtral 8x7B", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mixtral", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "structured_output": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 480, - "output": 0 + "context": 32768, + "input": 32768, + "output": 32768 + }, + "cost": { + "input": 0.27, + "output": 0.27 } }, - "google/veo3.1-lite": { - "id": "google/veo3.1-lite", - "name": "Veo 3.1 Lite", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", + "mistralai/mistral-small-4-119b-2603:thinking": { + "id": "mistralai/mistral-small-4-119b-2603:thinking", + "name": "Mistral Small 4 119B Thinking", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "structured_output": true, + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", "image" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 0 + "context": 262144, + "input": 262144, + "output": 16384 + }, + "cost": { + "input": 0.4, + "output": 1.4 } }, - "google/gemini-3-pro-image-preview": { - "id": "google/gemini-3-pro-image-preview", - "name": "Nano Banana Pro Preview", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], + "mistralai/mixtral-8x22b-instruct-v0.1": { + "id": "mistralai/mixtral-8x22b-instruct-v0.1", + "name": "Mixtral 8x22B", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mixtral", + "attachment": false, + "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", + "structured_output": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "text", - "image" + "text" ] }, "open_weights": false, "limit": { "context": 65536, + "input": 65536, "output": 32768 }, "cost": { - "input": 2, - "output": 12 + "input": 0.8999999999999999, + "output": 0.8999999999999999 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "mistralai/Devstral-Small-2505": { + "id": "mistralai/Devstral-Small-2505", + "name": "Mistral Devstral Small 2505", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-02", + "last_updated": "2025-08-02", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -196392,44 +184965,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 2, - "output": 12 + "input": 0.060000000000000005, + "output": 0.060000000000000005 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "mistralai/ministral-8b-2512": { + "id": "mistralai/ministral-8b-2512", + "name": "Ministral 8B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-04", + "last_updated": "2025-12-04", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -196437,38 +184995,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 1.5, - "output": 9 + "input": 0.15, + "output": 0.15 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "mistralai/mistral-saba": { + "id": "mistralai/mistral-saba", + "name": "Mistral Saba", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-17", + "last_updated": "2025-02-17", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -196476,68 +185025,59 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 32000, + "input": 32000, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.31 + "input": 0.1989, + "output": 0.595 } }, - "google/veo3.1-fast": { - "id": "google/veo3.1-fast", - "name": "Veo 3.1 Fast", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "family": "veo", - "attachment": true, + "mistralai/mistral-large": { + "id": "mistralai/mistral-large", + "name": "Mistral Large 2411", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", + "structured_output": false, + "release_date": "2024-02-26", + "last_updated": "2024-02-26", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, "open_weights": false, "limit": { - "context": 400000, - "output": 0 + "context": 128000, + "input": 128000, + "output": 256000 + }, + "cost": { + "input": 2.006, + "output": 6.001 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "mistralai/mistral-medium-3.1": { + "id": "mistralai/mistral-medium-3.1", + "name": "Mistral Medium 3.1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -196545,76 +185085,80 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.0375 + "input": 0.4, + "output": 2 } }, - "google/imagen-4.0-fast": { - "id": "google/imagen-4.0-fast", - "name": "Imagen 4 Fast", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "imagen", + "mistralai/ministral-3b-2512": { + "id": "mistralai/ministral-3b-2512", + "name": "Ministral 3B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", + "structured_output": false, + "release_date": "2025-12-04", + "last_updated": "2025-12-04", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 480, - "output": 0 + "context": 131072, + "input": 131072, + "output": 32768 + }, + "cost": { + "input": 0.1, + "output": 0.1 } }, - "deepseek-ai/deepseek-r1-distill-llama-70b": { - "id": "deepseek-ai/deepseek-r1-distill-llama-70b", - "name": "DeepSeek R1 Distill Llama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], + "mistralai/ministral-14b-instruct-2512": { + "id": "mistralai/ministral-14b-instruct-2512", + "name": "Ministral 3 14B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, + "reasoning": false, "tool_call": false, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-01-23", - "last_updated": "2025-01-23", + "structured_output": false, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 32768 }, "cost": { - "input": 0.03, - "output": 0.14 + "input": 0.1, + "output": 0.4 } }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", + "mistralai/mistral-small-4-119b-2603": { + "id": "mistralai/mistral-small-4-119b-2603", + "name": "Mistral Small 4 119B", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -196624,42 +185168,39 @@ ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 262144 + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.75, - "output": 3.5 + "input": 0.4, + "output": 1.4 } }, - "moonshotai/kimi-k2": { - "id": "moonshotai/kimi-k2", - "name": "Kimi K2", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "mistralai/ministral-14b-2512": { + "id": "mistralai/ministral-14b-2512", + "name": "Ministral 14B", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-07-11", - "last_updated": "2025-07-11", + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-04", + "last_updated": "2025-12-04", "modalities": { "input": [ "text" @@ -196668,113 +185209,120 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 262144, + "input": 262144, "output": 32768 }, "cost": { - "input": 0.55, - "output": 2.2 + "input": 0.2, + "output": 0.2 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "mistralai/mistral-large-3-675b-instruct-2512": { + "id": "mistralai/mistral-large-3-675b-instruct-2512", + "name": "Mistral Large 3 675B", + "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", + "family": "mistral-large", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 262144, + "input": 262144, + "output": 256000 }, "cost": { - "input": 0.3, - "output": 1.2 + "input": 1, + "output": 3 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "mistralai/mistral-medium-3": { + "id": "mistralai/mistral-medium-3", + "name": "Mistral Medium 3", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-medium", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 0.6, - "output": 2.4 + "input": 0.4, + "output": 2 } }, - "wanx/wan-v2-6": { - "id": "wanx/wan-v2-6", - "name": "Wan 2.6", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": true, + "mistralai/devstral-2-123b-instruct-2512": { + "id": "mistralai/devstral-2-123b-instruct-2512", + "name": "Devstral 2 123B", + "description": "Mistral coding agent model for repository tasks and software engineering workflows", + "family": "devstral", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", + "structured_output": false, + "release_date": "2025-12-09", + "last_updated": "2025-12-09", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "video" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 400000, - "output": 0 + "context": 262144, + "input": 262144, + "output": 65536 + }, + "cost": { + "input": 0.4, + "output": 1.4 } }, - "qwen/qwen3-coder": { - "id": "qwen/qwen3-coder", - "name": "Qwen3 Coder", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "mistralai/codestral-2508": { + "id": "mistralai/codestral-2508", + "name": "Codestral 2508", + "description": "Mistral coding model for code completion, generation, and developer workflows", + "family": "codestral", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-01", + "last_updated": "2025-08-01", "modalities": { "input": [ "text" @@ -196783,96 +185331,90 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 66536 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { "input": 0.3, - "output": 1.2 + "output": 0.8999999999999999 } }, - "leonardo-ai/lucid-realism": { - "id": "leonardo-ai/lucid-realism", - "name": "Lucid Realism", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "lucid", - "attachment": true, + "mistralai/Mistral-Nemo-Instruct-2407": { + "id": "mistralai/Mistral-Nemo-Instruct-2407", + "name": "Mistral Nemo", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-06-01", - "last_updated": "2025-06-01", + "structured_output": false, + "release_date": "2024-07-18", + "last_updated": "2024-07-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 4096, - "output": 0 + "context": 16384, + "input": 16384, + "output": 8192 + }, + "cost": { + "input": 0.1003, + "output": 0.1207 } }, - "leonardo-ai/lucid-origin": { - "id": "leonardo-ai/lucid-origin", - "name": "Lucid Origin", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "lucid", - "attachment": true, + "bytedance-seed/seed-2.0-lite": { + "id": "bytedance-seed/seed-2.0-lite", + "name": "ByteDance Seed 2.0 Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-06-01", - "last_updated": "2025-06-01", + "structured_output": true, + "release_date": "2026-03-10", + "last_updated": "2026-03-10", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 4096, - "output": 0 + "context": 262144, + "input": 262144, + "output": 131072 + }, + "cost": { + "input": 0.25, + "output": 2 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "anthracite-org/magnum-v2-72b": { + "id": "anthracite-org/magnum-v2-72b", + "name": "Magnum V2 72B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -196880,42 +185422,30 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 30, - "output": 180 + "input": 2.006, + "output": 2.992 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", + "anthracite-org/magnum-v4-72b": { + "id": "anthracite-org/magnum-v4-72b", + "name": "Magnum v4 72B", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", - "image" + "pdf" ], "output": [ "text" @@ -196923,69 +185453,58 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 1.25 + "input": 2.006, + "output": 2.992 } }, - "openai/gpt-realtime-1.5": { - "id": "openai/gpt-realtime-1.5", - "name": "GPT Realtime 1.5", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gpt", - "attachment": true, + "inflatebot/MN-12B-Mag-Mell-R1": { + "id": "inflatebot/MN-12B-Mag-Mell-R1", + "name": "Mag Mell R1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-06-01", - "last_updated": "2025-06-01", + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ - "text", - "audio", - "image" + "text" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 32000, - "output": 4096 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 4, - "output": 16 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning": { + "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning", + "name": "Nvidia Nemotron 3 Nano Omni", + "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", + "family": "nemotron", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": false, "temperature": true, - "knowledge": "2024-10-01", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2026-04-28", + "last_updated": "2026-04-28", "modalities": { "input": [ "text", @@ -196997,69 +185516,61 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 256000, + "input": 256000, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 + "input": 0.105, + "output": 0.42 } }, - "openai/gpt-image-2": { - "id": "openai/gpt-image-2", - "name": "GPT Image 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, + "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF": { + "id": "nvidia/Llama-3.1-Nemotron-70B-Instruct-HF", + "name": "Nvidia Nemotron 70b", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, "reasoning": false, "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 128000, - "output": 0 + "context": 16384, + "input": 16384, + "output": 8192 + }, + "cost": { + "input": 0.357, + "output": 0.408 } }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, + "nvidia/nvidia-nemotron-nano-9b-v2": { + "id": "nvidia/nvidia-nemotron-nano-9b-v2", + "name": "Nvidia Nemotron Nano 9B v2", + "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2025-08-18", + "last_updated": "2025-08-18", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -197067,41 +185578,30 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 1.75, - "output": 14 + "input": 0.17, + "output": 0.68 } }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "nvidia/Llama-3_3-Nemotron-Super-49B-v1_5": { + "id": "nvidia/Llama-3_3-Nemotron-Super-49B-v1_5", + "name": "Nvidia Nemotron Super 49B v1.5", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "knowledge": "2024-10-01", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "release_date": "2025-08-08", + "last_updated": "2025-08-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -197109,36 +185609,27 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { "input": 0.05, - "output": 0.4, - "cache_read": 0.005 + "output": 0.25 } }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "nvidia/nemotron-3-nano-30b-a3b": { + "id": "nvidia/nemotron-3-nano-30b-a3b", + "name": "Nvidia Nemotron 3 Nano 30B", + "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", + "family": "nemotron", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2025-12-15", + "last_updated": "2025-12-15", "modalities": { "input": [ "text" @@ -197147,43 +185638,64 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 65536 + "context": 256000, + "input": 256000, + "output": 262144 }, "cost": { - "input": 0.05, - "output": 0.2 + "input": 0.17, + "output": 0.68 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, + "nvidia/Llama-3.3-Nemotron-Super-49B-v1": { + "id": "nvidia/Llama-3.3-Nemotron-Super-49B-v1", + "name": "Nvidia Nemotron Super 49B", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "temperature": true, + "release_date": "2025-08-08", + "last_updated": "2025-08-08", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 128000, + "output": 16384 + }, + "cost": { + "input": 0.15, + "output": 0.15 + } + }, + "nvidia/nemotron-3-super-120b-a12b:thinking": { + "id": "nvidia/nemotron-3-super-120b-a12b:thinking", + "name": "Nvidia Nemotron 3 Super 120B Thinking", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, + "structured_output": false, "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -197191,43 +185703,35 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.75, - "output": 4.5 + "input": 0.05, + "output": 0.25 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, + "nvidia/nemotron-3-super-120b-a12b": { + "id": "nvidia/nemotron-3-super-120b-a12b", + "name": "Nvidia Nemotron 3 Super 120B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "structured_output": false, + "temperature": true, + "release_date": "2026-03-01", + "last_updated": "2026-03-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -197235,31 +185739,29 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 5, - "output": 30 + "input": 0.05, + "output": 0.25 } }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, + "cognitivecomputations/dolphin-2.9.2-qwen2-72b": { + "id": "cognitivecomputations/dolphin-2.9.2-qwen2-72b", + "name": "Dolphin 72b", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -197267,36 +185769,31 @@ }, "open_weights": false, "limit": { - "context": 1047576, - "output": 32768 + "context": 8192, + "input": 8192, + "output": 4096 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.306, + "output": 0.306 } }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "xiaomi/mimo-v2-flash-original": { + "id": "xiaomi/mimo-v2-flash-original", + "name": "MiMo V2 Flash Original", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ "text" @@ -197305,33 +185802,36 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 256000, + "input": 256000, "output": 32768 }, "cost": { - "input": 0.15, - "output": 0.6 + "input": 0.102, + "output": 0.306 } }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, + "xiaomi/mimo-v2-flash-thinking-original": { + "id": "xiaomi/mimo-v2-flash-thinking-original", + "name": "MiMo V2 Flash (Thinking) Original", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", + "attachment": false, "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-10-01", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -197339,55 +185839,35 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.102, + "output": 0.306 } - } - } - }, - "orcarouter": { - "id": "orcarouter", - "env": [ - "ORCAROUTER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.orcarouter.ai/v1", - "name": "OrcaRouter", - "doc": "https://docs.orcarouter.ai", - "models": { - "grok/grok-4.3": { - "id": "grok/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + }, + "xiaomi/mimo-v2.5": { + "id": "xiaomi/mimo-v2.5", + "name": "MiMo V2.5", + "description": "MiMo omni model for text, image, video, audio, and agents", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" @@ -197395,57 +185875,37 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 30000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "grok/grok-4.5": { - "id": "grok/grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", + "xiaomi/mimo-v2-omni": { + "id": "xiaomi/mimo-v2-omni", + "name": "MiMo V2 Omni", + "description": "MiMo omni model for text, image, video, audio, and agents", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", + "structured_output": false, + "release_date": "2026-03-19", + "last_updated": "2026-03-19", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio" ], "output": [ "text" @@ -197453,43 +185913,35 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 262144, + "input": 262144, + "output": 65536 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 0.4, + "output": 2, + "cache_read": 0.08 } }, - "grok/grok-4.6": { - "id": "grok/grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, + "xiaomi/mimo-v2-flash": { + "id": "xiaomi/mimo-v2-flash", + "name": "MiMo V2 Flash", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -197497,42 +185949,33 @@ }, "open_weights": false, "limit": { - "context": 500000, - "output": 500000 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 + "input": 0.102, + "output": 0.306 } }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "xiaomi/mimo-v2-pro": { + "id": "xiaomi/mimo-v2-pro", + "name": "MiMo V2 Pro", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "structured_output": true, + "release_date": "2026-03-19", + "last_updated": "2026-03-19", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -197540,43 +185983,34 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1, + "output": 3, + "cache_read": 0.2 } }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, + "xiaomi/mimo-v2.5-pro": { + "id": "xiaomi/mimo-v2.5-pro", + "name": "MiMo V2.5 Pro", + "description": "MiMo pro model for strong multimodal reasoning and agent execution", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "structured_output": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -197584,58 +186018,71 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 1048576, + "input": 1048576, + "output": 131072 }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.0036 + } + }, + "xiaomi/mimo-v2-flash-thinking": { + "id": "xiaomi/mimo-v2-flash-thinking", + "name": "MiMo V2 Flash (Thinking)", + "description": "MiMo flash model for fast multimodal assistance and agent workflows", + "family": "mimo", + "attachment": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" } + ], + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-17", + "last_updated": "2025-12-17", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.102, + "output": 0.306 } }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", + "anthropic/claude-haiku-latest": { + "id": "anthropic/claude-haiku-latest", + "name": "Claude Haiku Latest", + "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "structured_output": true, + "release_date": "2026-03-29", + "last_updated": "2026-03-29", "modalities": { "input": [ "text", @@ -197648,58 +186095,27 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } + "context": 200000, + "input": 200000, + "output": 64000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1, + "output": 5, + "cache_read": 0.1 } }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "anthropic/claude-opus-4.7:thinking": { + "id": "anthropic/claude-opus-4.7:thinking", + "name": "Claude 4.7 Opus Thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", + "structured_output": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -197713,37 +186129,27 @@ "open_weights": false, "limit": { "context": 1000000, + "input": 1000000, "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 4.998, + "output": 25.007, + "cache_read": 0.4998 } }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", + "anthropic/claude-opus-4.6:thinking:max": { + "id": "anthropic/claude-opus-4.6:thinking:max", + "name": "Claude 4.6 Opus Thinking Max", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "structured_output": true, + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", @@ -197756,38 +186162,27 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 64000 + "context": 1000000, + "input": 1000000, + "output": 128000 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 4.998, + "output": 25.007 } }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", + "anthropic/claude-opus-4.6:thinking:low": { + "id": "anthropic/claude-opus-4.6:thinking:low", + "name": "Claude 4.6 Opus Thinking Low", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "structured_output": true, + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", @@ -197801,37 +186196,27 @@ "open_weights": false, "limit": { "context": 1000000, + "input": 1000000, "output": 128000 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 10 + "input": 4.998, + "output": 25.007 } }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5 (latest)", + "anthropic/claude-sonnet-4.6:thinking": { + "id": "anthropic/claude-sonnet-4.6:thinking", + "name": "Claude Sonnet 4.6 Thinking", "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", "family": "claude-sonnet", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "structured_output": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ "text", @@ -197845,37 +186230,36 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 + "input": 1000000, + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 2.992, + "output": 14.993999999999998 } }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", + "anthropic/claude-opus-4.7": { + "id": "anthropic/claude-opus-4.7", + "name": "Claude 4.7 Opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", + "structured_output": true, + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", @@ -197889,37 +186273,27 @@ "open_weights": false, "limit": { "context": 1000000, + "input": 1000000, "output": 128000 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 4.998, + "output": 25.007, + "cache_read": 0.4998 } }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", + "anthropic/claude-opus-4.6:thinking:medium": { + "id": "anthropic/claude-opus-4.6:thinking:medium", + "name": "Claude 4.6 Opus Thinking Medium", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "structured_output": true, + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", @@ -197933,37 +186307,25 @@ "open_weights": false, "limit": { "context": 1000000, + "input": 1000000, "output": 128000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 4.998, + "output": 25.007 } }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", + "anthropic/claude-opus-4.8:thinking": { + "id": "anthropic/claude-opus-4.8:thinking", + "name": "Claude Opus 4.8 Thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "structured_output": true, + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ "text", @@ -197977,251 +186339,278 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 64000 - }, + "input": 1000000, + "output": 128000 + }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 4.998, + "output": 25.007, + "cache_read": 0.4998 } }, - "deepseek/deepseek-v4-flash-0731": { - "id": "deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, + "anthropic/claude-sonnet-latest": { + "id": "anthropic/claude-sonnet-latest", + "name": "Claude Sonnet Latest", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "structured_output": true, + "release_date": "2026-03-01", + "last_updated": "2026-03-01", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 1000000, + "input": 1000000, + "output": 128000 + }, + "cost": { + "input": 2.992, + "output": 14.994, + "cache_read": 0.2992 + } + }, + "anthropic/claude-opus-4.8": { + "id": "anthropic/claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "attachment": true, + "reasoning": true, + "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.147, - "output": 0.295, - "cache_read": 0.02 + "input": 4.998, + "output": 25.007, + "cache_read": 0.4998 } }, - "deepseek/deepseek-v4-flash-vision-exp": { - "id": "deepseek/deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", + "anthropic/claude-opus-latest": { + "id": "anthropic/claude-opus-latest", + "name": "Claude Opus Latest", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", + "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", + "release_date": "2026-03-29", + "last_updated": "2026-03-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.147, - "output": 0.295, - "cache_read": 0.02 + "input": 4.998, + "output": 25.007, + "cache_read": 0.4998 } }, - "deepseek/deepseek-chat": { - "id": "deepseek/deepseek-chat", - "name": "DeepSeek Chat", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "anthropic/claude-sonnet-4.6": { + "id": "anthropic/claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": false, "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "structured_output": true, + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-02-17", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.147, - "output": 0.295, - "cache_read": 0.02 + "input": 2.992, + "output": 14.993999999999998 } }, - "deepseek/deepseek-v4-pro-0813": { - "id": "deepseek/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, + "anthropic/claude-opus-4.6:thinking": { + "id": "anthropic/claude-opus-4.6:thinking", + "name": "Claude 4.6 Opus Thinking", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0.442, - "output": 0.884, - "cache_read": 0.06 + "input": 4.998, + "output": 25.007 } }, - "deepseek/deepseek-v4-flash-free": { - "id": "deepseek/deepseek-v4-flash-free", - "name": "DeepSeek V4 Flash (free)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, + "anthropic/claude-opus-4.6": { + "id": "anthropic/claude-opus-4.6", + "name": "Claude 4.6 Opus", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ + "none", "low", + "medium", "high", - "max" + "xhigh" ] + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1000000, - "output": 384000 + "input": 1000000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 4.998, + "output": 25.007 } }, - "deepseek/deepseek-reasoner": { - "id": "deepseek/deepseek-reasoner", - "name": "DeepSeek Reasoner", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", + "tencent/Hunyuan-MT-7B": { + "id": "tencent/Hunyuan-MT-7B", + "name": "Hunyuan MT 7B", + "description": "Translation model for multilingual conversion, localization, and cross-language workflows", + "family": "hunyuan", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-18", + "last_updated": "2025-09-18", "modalities": { "input": [ "text" @@ -198230,45 +186619,27 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 8192, + "input": 8192, + "output": 8192 }, "cost": { - "input": 0.147, - "output": 0.295, - "cache_read": 0.028 + "input": 10, + "output": 20 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "tencent/hy3-preview": { + "id": "tencent/hy3-preview", + "name": "Tencent: Hy3 preview", + "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text" @@ -198277,46 +186648,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 262144, + "input": 262144, + "output": 262144 }, "cost": { - "input": 0.442, - "output": 0.884, - "cache_read": 0.06 + "input": 0.066, + "output": 0.26, + "cache_read": 0.029 } }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "chutesai/Mistral-Small-3.2-24B-Instruct-2506": { + "id": "chutesai/Mistral-Small-3.2-24B-Instruct-2506", + "name": "Mistral Small 3.2 24b Instruct", + "description": "Efficient Mistral model for fast chat, extraction, and production assistants", + "family": "chutesai", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text" @@ -198325,38 +186679,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 128000, + "input": 128000, + "output": 131072 }, "cost": { - "input": 0.147, - "output": 0.295, - "cache_read": 0.02 + "input": 0.2, + "output": 0.4 } }, - "tencent/hy3-free": { - "id": "tencent/hy3-free", - "name": "Hy3 (free)", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + "dmind/dmind-1-mini": { + "id": "dmind/dmind-1-mini", + "name": "DMind-1-Mini", + "description": "Compact GPT model for low-latency assistance and high-volume workloads", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-01", + "last_updated": "2025-06-01", "modalities": { "input": [ "text" @@ -198365,38 +186709,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 192000, - "output": 128000 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 0.4 } }, - "tencent/hy3": { - "id": "tencent/hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", + "dmind/dmind-1": { + "id": "dmind/dmind-1", + "name": "DMind-1", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-01", + "last_updated": "2025-06-01", "modalities": { "input": [ "text" @@ -198405,35 +186739,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256000, - "input": 192000, - "output": 128000 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0.18, - "output": 0.59, - "cache_read": 0.059 + "input": 0.3, + "output": 0.6 } }, - "z-ai/glm-4.5": { - "id": "z-ai/glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", + "MarinaraSpaghetti/NemoMix-Unleashed-12B": { + "id": "MarinaraSpaghetti/NemoMix-Unleashed-12B", + "name": "NemoMix 12B Unleashed", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ "text" @@ -198442,81 +186769,58 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 98304 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "z-ai/glm-5.3-flash": { - "id": "z-ai/glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "deepcogito/cogito-v1-preview-qwen-32B": { + "id": "deepcogito/cogito-v1-preview-qwen-32B", + "name": "Cogito v1 Preview Qwen 32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-10", + "last_updated": "2025-05-10", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 32768 }, "cost": { - "input": 0.075, - "output": 0.25 + "input": 1.7999999999999998, + "output": 1.7999999999999998 } }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", + "cohere/command-r": { + "id": "cohere/command-r", + "name": "Cohere: Command R", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-03-11", + "last_updated": "2024-03-11", "modalities": { "input": [ "text" @@ -198525,37 +186829,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 0.476, + "output": 1.428 } }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", + "cohere/command-r-plus-08-2024": { + "id": "cohere/command-r-plus-08-2024", + "name": "Cohere: Command R+", + "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", + "family": "command-r", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "structured_output": false, + "release_date": "2024-08-30", + "last_updated": "2024-08-30", "modalities": { "input": [ "text" @@ -198564,38 +186859,61 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 4096 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.26, - "cache_write": 0 + "input": 2.856, + "output": 14.246 } }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, + "stepfun/step-3.7-flash:thinking": { + "id": "stepfun/step-3.7-flash:thinking", + "name": "Step 3.7 Flash Thinking", + "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "structured_output": true, + "release_date": "2026-05-29", + "last_updated": "2026-05-29", + "modalities": { + "input": [ + "text", + "image", + "video" + ], + "output": [ + "text" + ] }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", + "open_weights": false, + "limit": { + "context": 256000, + "input": 256000, + "output": 256000 + }, + "cost": { + "input": 0.2, + "output": 1.15, + "cache_read": 0.04 + } + }, + "nex-agi/deepseek-v3.1-nex-n1": { + "id": "nex-agi/deepseek-v3.1-nex-n1", + "name": "DeepSeek V3.1 Nex N1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-10", + "last_updated": "2025-12-10", "modalities": { "input": [ "text" @@ -198604,88 +186922,89 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 + "input": 0.27999999999999997, + "output": 0.42000000000000004 } }, - "z-ai/glm-5.3-flash-free": { - "id": "z-ai/glm-5.3-flash-free", - "name": "GLM-5.3-Flash (free)", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "undi95/remm-slerp-l2-13b": { + "id": "undi95/remm-slerp-l2-13b", + "name": "ReMM SLERP 13B", + "description": "Open Llama multimodal model for image understanding and text reasoning", + "family": "llama", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", - "image", - "video", "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 6144, + "input": 6144, + "output": 4096 }, "cost": { - "input": 0, - "output": 0 + "input": 0.7989999999999999, + "output": 1.2069999999999999 } }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "nothingiisreal/L3.1-70B-Celeste-V0.1-BF16": { + "id": "nothingiisreal/L3.1-70B-Celeste-V0.1-BF16", + "name": "Llama 3.1 70B Celeste v0.1", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-23", + "last_updated": "2024-07-23", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "open_weights": false, + "limit": { + "context": 16384, + "input": 16384, + "output": 16384 + }, + "cost": { + "input": 0.49299999999999994, + "output": 0.49299999999999994 + } + }, + "zai-org/glm-4.7-flash-original:thinking": { + "id": "zai-org/glm-4.7-flash-original:thinking", + "name": "GLM 4.7 Flash Original Thinking", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -198694,21 +187013,21 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.07, + "output": 0.4 } }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "zai-org/glm-4.7": { + "id": "zai-org/glm-4.7", + "name": "GLM 4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, @@ -198718,13 +187037,9 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "release_date": "2026-01-29", + "last_updated": "2026-01-29", "modalities": { "input": [ "text" @@ -198736,32 +187051,25 @@ "open_weights": true, "limit": { "context": 200000, - "output": 131072 + "input": 200000, + "output": 128000 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 + "input": 0.15, + "output": 0.8 } }, - "z-ai/glm-4.5-air": { - "id": "z-ai/glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", + "zai-org/GLM-4.5-Air:thinking": { + "id": "zai-org/GLM-4.5-Air:thinking", + "name": "GLM 4.5 Air (Thinking)", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "structured_output": true, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text" @@ -198770,43 +187078,57 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 128000, + "input": 128000, "output": 98304 }, "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03, - "cache_write": 0 + "input": 0.12, + "output": 0.8 } }, - "z-ai/glm-5.3": { - "id": "z-ai/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "zai-org/GLM-4.5:thinking": { + "id": "zai-org/GLM-4.5:thinking", + "name": "GLM 4.5 (Thinking)", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "open_weights": false, + "limit": { + "context": 128000, + "input": 128000, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 1.3 + } + }, + "zai-org/glm-4.5": { + "id": "zai-org/glm-4.5", + "name": "GLM 4.5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ "text" @@ -198815,47 +187137,35 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 1.26, - "output": 3.96, - "cache_read": 0.234 + "input": 0.3, + "output": 1.3 } }, - "meta/muse-spark-1.1": { - "id": "meta/muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", - "family": "muse", - "attachment": true, + "zai-org/glm-5-original": { + "id": "zai-org/glm-5-original", + "name": "GLM 5 Original", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image", - "pdf", - "video" + "text" ], "output": [ "text" @@ -198863,92 +187173,69 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "meta/muse-spark-1.2": { - "id": "meta/muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", - "attachment": true, + "zai-org/glm-5.1": { + "id": "zai-org/glm-5.1", + "name": "GLM 5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ - "text", - "image", - "video", - "pdf", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 200000, + "input": 200000, "output": 131072 }, "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 + "input": 0.3, + "output": 2.55 } }, - "google/gemini-flash-latest": { - "id": "google/gemini-flash-latest", - "name": "Gemini Flash Latest", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, + "zai-org/glm-4.7-original": { + "id": "zai-org/glm-4.7-original", + "name": "GLM 4.7 Original", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -198956,79 +187243,65 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 65535 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, + "zai-org/glm-4.7-original:thinking": { + "id": "zai-org/glm-4.7-original:thinking", + "name": "GLM 4.7 Original Thinking", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "input": 200000, + "output": 65535 }, "cost": { - "input": 0.13, - "output": 0.38, - "cache_read": 0.02 + "input": 0.6, + "output": 2.2, + "cache_read": 0.11 } }, - "google/gemini-flash-lite-latest": { - "id": "google/gemini-flash-lite-latest", - "name": "Gemini Flash-Lite Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, + "zai-org/glm-4.7-flash-original": { + "id": "zai-org/glm-4.7-flash-original", + "name": "GLM 4.7 Flash Original", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -199036,47 +187309,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 - }, + "context": 200000, + "input": 200000, + "output": 128000 + }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 0.07, + "output": 0.4 } }, - "google/gemini-3.6-flash": { - "id": "google/gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "zai-org/glm-4.7:thinking": { + "id": "zai-org/glm-4.7:thinking", + "name": "GLM 4.7 Thinking", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -199084,75 +187339,61 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 65535 }, "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15 + "input": 0.2, + "output": 0.8 } }, - "google/gemini-robotics-er-1.6-preview": { - "id": "google/gemini-robotics-er-1.6-preview", - "name": "Gemini Robotics-ER 1.6 Preview", - "description": "Vision-language model for embodied reasoning: spatial understanding, task planning, and physical-world agentic robotics", - "family": "gemini", - "attachment": true, + "zai-org/glm-5.1:thinking": { + "id": "zai-org/glm-5.1:thinking", + "name": "GLM 5.1 Thinking", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0 - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-14", - "last_updated": "2026-04-14", + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 131072 }, "cost": { - "input": 1, - "output": 5 + "input": 0.3, + "output": 2.55 } }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, + "zai-org/glm-5:thinking": { + "id": "zai-org/glm-5:thinking", + "name": "GLM 5 Thinking", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -199160,46 +187401,29 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 0.06, - "output": 0.33, - "cache_read": 0.0075 + "input": 0.3, + "output": 2.55 } }, - "google/gemini-3.5-flash-lite": { - "id": "google/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", + "zai-org/glm-4.6v": { + "id": "zai-org/glm-4.6v", + "name": "GLM 4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -199207,45 +187431,33 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "input": 128000, + "output": 24000 }, "cost": { "input": 0.3, - "output": 2.5, - "cache_read": 0.03 + "output": 0.9 } }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, + "zai-org/glm-4.6-original": { + "id": "zai-org/glm-4.6-original", + "name": "GLM 4.6 Original", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -199253,63 +187465,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "input": 256000, + "output": 65535 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "input_audio": 2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.35, + "output": 1.4 } }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, + "zai-org/glm-5-original:thinking": { + "id": "zai-org/glm-5-original:thinking", + "name": "GLM 5 Original Thinking", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -199317,47 +187495,30 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 0.08333, - "input_audio": 3 + "input": 1, + "output": 3.2, + "cache_read": 0.2 } }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, + "zai-org/glm-4.7-flash:thinking": { + "id": "zai-org/glm-4.7-flash:thinking", + "name": "GLM 4.7 Flash Thinking", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" @@ -199365,61 +187526,30 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } + "input": 0.07, + "output": 0.4 } }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", + "zai-org/glm-4.6v-flash-original": { + "id": "zai-org/glm-4.6v-flash-original", + "name": "GLM 4.6V Flash", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", "image", - "audio", - "video", - "pdf" + "video" ], "output": [ "text" @@ -199427,46 +187557,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "input": 128000, + "output": 24000 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 + "input": 0.1, + "output": 0.4 } }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", + "zai-org/glm-4.6v-original": { + "id": "zai-org/glm-4.6v-original", + "name": "GLM 4.6V Original", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", - "image", - "video", - "audio", - "pdf" + "image" ], "output": [ "text" @@ -199474,62 +187587,37 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "input": 128000, + "output": 24000 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 0.6, + "output": 0.9 } }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, + "zai-org/glm-latest": { + "id": "zai-org/glm-latest", + "name": "GLM Latest", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", - "low", - "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "release_date": "2026-05-03", + "last_updated": "2026-05-03", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -199537,94 +187625,65 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 131072 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 + "input": 0.75, + "output": 2.6, + "cache_read": 0.15 } }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, + "zai-org/glm-4.7-flash": { + "id": "zai-org/glm-4.7-flash", + "name": "GLM 4.7 Flash", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image", - "audio", - "video", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 + "input": 0.07, + "output": 0.4 } }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, + "zai-org/GLM-4.6-turbo:thinking": { + "id": "zai-org/GLM-4.6-turbo:thinking", + "name": "GLM 4.6 Turbo (Thinking)", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-10-02", + "last_updated": "2025-10-02", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -199632,47 +187691,28 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 200000, + "input": 200000, + "output": 204800 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 1, + "output": 3 } }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "zai-org/GLM-4.5-Air": { + "id": "zai-org/GLM-4.5-Air", + "name": "GLM 4.5 Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", + "release_date": "2025-04-15", + "last_updated": "2025-04-15", "modalities": { "input": [ - "text", - "image", - "video", - "audio", - "pdf" + "text" ], "output": [ "text" @@ -199680,113 +187720,50 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 128000, + "input": 128000, + "output": 98304 }, "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 + "input": 0.12, + "output": 0.8 } }, - "kimi/kimi-k3": { - "id": "kimi/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "zai-org/GLM-4.6-turbo": { + "id": "zai-org/GLM-4.6-turbo", + "name": "GLM 4.6 Turbo", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-10-02", + "last_updated": "2025-10-02", "modalities": { "input": [ - "text", - "image", - "video" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33 - } - }, - "kimi/kimi-k2.6": { - "id": "kimi/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "input": 200000, + "output": 204800 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 + "input": 1, + "output": 3 } }, - "kimi/kimi-k2.5": { - "id": "kimi/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, + "zai-org/glm-5": { + "id": "zai-org/glm-5", + "name": "GLM 5", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -199794,19 +187771,12 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -199814,81 +187784,59 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 200000, + "input": 200000, + "output": 128000 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1, - "cache_write": 0 + "input": 0.3, + "output": 2.55 } }, - "kimi/kimi-k2.7-code": { - "id": "kimi/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "Infermatic/MN-12B-Inferor-v0.0": { + "id": "Infermatic/MN-12B-Inferor-v0.0", + "name": "Mistral Nemo Inferor 12B", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 + "input": 0.25499999999999995, + "output": 0.49299999999999994 } }, - "orcarouter/fusion-mini": { - "id": "orcarouter/fusion-mini", - "name": "OrcaRouter Fusion Mini", - "description": "Leaner two-model Fusion panel that runs Claude Opus 4.8 and GPT-5.5 in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Easy requests fall through to a cheaper default and bill as one call.", - "family": "model-router", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "shisa-ai/shisa-v2.1-llama3.3-70b": { + "id": "shisa-ai/shisa-v2.1-llama3.3-70b", + "name": "Shisa V2.1 Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -199896,23 +187844,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 32768, + "input": 32768, + "output": 4096 + }, + "cost": { + "input": 0.5, + "output": 0.5 } }, - "orcarouter/free": { - "id": "orcarouter/free", - "name": "OrcaRouter Free", - "description": "Built-in router over the free tier that scores each request's difficulty and sends light work to the smaller free model and harder work to the stronger one. Priced at zero and never falls back to a paid model.", - "family": "auto", + "shisa-ai/shisa-v2-llama3.3-70b": { + "id": "shisa-ai/shisa-v2-llama3.3-70b", + "name": "Shisa V2 Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text" @@ -199923,41 +187874,29 @@ }, "open_weights": false, "limit": { - "context": 65536, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.5, + "output": 0.5 } }, - "orcarouter/fusion": { - "id": "orcarouter/fusion", - "name": "OrcaRouter Fusion", - "description": "Curated fan-out router that runs Claude Opus 4.8, GPT-5.5 and Gemini 3.1 Pro in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Easy requests fall through to a cheaper default and bill as one call.", - "family": "model-router", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "abacusai/Dracarys-72B-Instruct": { + "id": "abacusai/Dracarys-72B-Instruct", + "name": "Llama 3.1 70B Dracarys 2", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-02", + "last_updated": "2025-08-02", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -199965,32 +187904,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 8192 + }, + "cost": { + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "orcarouter/fusion-flash": { - "id": "orcarouter/fusion-flash", - "name": "OrcaRouter Fusion Flash", - "description": "Budget Fusion panel that runs Gemini 3.5 Flash, MiniMax M2.7 and GLM 5.1 in parallel on hard requests, then has a Claude Opus 4.8 judge return the strongest single answer verbatim. Cost-sensitive fan-out over a 200K window.", - "family": "model-router", + "deepseek-ai/DeepSeek-V3.1:thinking": { + "id": "deepseek-ai/DeepSeek-V3.1:thinking", + "name": "DeepSeek V3.1 Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek-thinking", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ "text" @@ -200001,26 +187934,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 65536 + }, + "cost": { + "input": 0.2, + "output": 0.7 } }, - "orcarouter/auto": { - "id": "orcarouter/auto", - "name": "OrcaRouter Auto", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": true, + "deepseek-ai/DeepSeek-V3.1-Terminus": { + "id": "deepseek-ai/DeepSeek-V3.1-Terminus", + "name": "DeepSeek V3.1 Terminus", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2026-05-14", + "release_date": "2025-08-02", + "last_updated": "2025-08-02", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -200029,25 +187965,26 @@ "open_weights": false, "limit": { "context": 128000, - "output": 16384 + "input": 128000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 0.25, + "output": 0.7 } }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", + "deepseek-ai/deepseek-v3.2-exp-thinking": { + "id": "deepseek-ai/deepseek-v3.2-exp-thinking", + "name": "DeepSeek V3.2 Exp Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek-thinking", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text" @@ -200056,30 +187993,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 163840, + "input": 163840, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.27999999999999997, + "output": 0.42000000000000004 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "deepseek-ai/DeepSeek-R1-0528": { + "id": "deepseek-ai/DeepSeek-R1-0528", + "name": "DeepSeek R1 0528", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek", "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-28", + "last_updated": "2025-05-28", "modalities": { "input": [ "text" @@ -200088,67 +188024,59 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 163840 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.4, + "output": 1.7 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", + "deepseek-ai/DeepSeek-V3.1": { + "id": "deepseek-ai/DeepSeek-V3.1", + "name": "DeepSeek V3.1", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-26", + "last_updated": "2025-07-26", "modalities": { "input": [ "text", - "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 512000 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.2, + "output": 0.7 } }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", + "deepseek-ai/deepseek-v3.2-exp": { + "id": "deepseek-ai/deepseek-v3.2-exp", + "name": "DeepSeek V3.2 Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text" @@ -200157,30 +188085,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 163840, + "input": 163840, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 0.27999999999999997, + "output": 0.42000000000000004 } }, - "minimax/minimax-m2.5-highspeed": { - "id": "minimax/minimax-m2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", + "deepseek-ai/DeepSeek-V3.1-Terminus:thinking": { + "id": "deepseek-ai/DeepSeek-V3.1-Terminus:thinking", + "name": "DeepSeek V3.1 Terminus (Thinking)", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek-thinking", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", + "structured_output": true, + "release_date": "2025-09-22", + "last_updated": "2025-09-22", "modalities": { "input": [ "text" @@ -200189,45 +188115,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 128000, + "input": 128000, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.03, - "cache_write": 0.375 + "input": 0.25, + "output": 0.7 } }, - "qwen/qwen3.5-flash": { - "id": "qwen/qwen3.5-flash", - "name": "Qwen3.5 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "arcee-ai/trinity-large-thinking": { + "id": "arcee-ai/trinity-large-thinking", + "name": "Trinity Large Thinking", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 81920 - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "structured_output": false, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -200235,65 +188147,56 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 262144, + "input": 262144, + "output": 80000 }, "cost": { - "input": 0.1, - "output": 0.4 + "input": 0.25, + "output": 0.9 } }, - "qwen/qwen3-vl-235b-a22b-instruct": { - "id": "qwen/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "arcee-ai/trinity-mini": { + "id": "arcee-ai/trinity-mini", + "name": "Trinity Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "trinity-mini", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 131072, - "output": 32768 + "input": 131072, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 1.6 + "input": 0.045000000000000005, + "output": 0.15 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", + "soob3123/GrayLine-Qwen3-8B": { + "id": "soob3123/GrayLine-Qwen3-8B", + "name": "Grayline Qwen3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-25", + "last_updated": "2025-09-25", "modalities": { "input": [ "text" @@ -200304,177 +188207,119 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 16384, + "input": 16384, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 3.75, - "cache_read": 0.25, - "cache_write": 1.563 + "input": 0.3, + "output": 0.3 } }, - "qwen/qwen3.8-27b": { - "id": "qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "soob3123/Veiled-Calla-12B": { + "id": "soob3123/Veiled-Calla-12B", + "name": "Veiled Calla 12B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-13", + "last_updated": "2025-04-13", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0.33, - "output": 2.4 + "input": 0.3, + "output": 0.3 } }, - "qwen/qwen3-vl-235b-a22b-thinking": { - "id": "qwen/qwen3-vl-235b-a22b-thinking", - "name": "Qwen3 VL 235B A22B Thinking", - "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 81920 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "soob3123/amoral-gemma3-27B-v2": { + "id": "soob3123/amoral-gemma3-27B-v2", + "name": "Amoral Gemma3 27B v2", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-05-23", + "last_updated": "2025-05-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 40960 + "context": 32768, + "input": 32768, + "output": 8192 }, "cost": { - "input": 0.4, - "output": 4 + "input": 0.3, + "output": 0.3 } }, - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "meganova-ai/manta-mini-1.0": { + "id": "meganova-ai/manta-mini-1.0", + "name": "Manta Mini 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-20", + "last_updated": "2025-12-20", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 8192, + "input": 8192, + "output": 8192 }, "cost": { - "input": 0.086, - "output": 0.688 + "input": 0.02, + "output": 0.16 } }, - "qwen/qwen3.8-max": { - "id": "qwen/qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "meganova-ai/manta-flash-1.0": { + "id": "meganova-ai/manta-flash-1.0", + "name": "Manta Flash 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-20", + "last_updated": "2025-12-20", "modalities": { "input": [ - "text", - "image", - "video", - "pdf" + "text" ], "output": [ "text" @@ -200482,41 +188327,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 + "input": 0.02, + "output": 0.16 } }, - "qwen/qwen3.5-plus": { - "id": "qwen/qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "meganova-ai/manta-pro-1.0": { + "id": "meganova-ai/manta-pro-1.0", + "name": "Manta Pro 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-20", + "last_updated": "2025-12-20", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -200524,35 +188357,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.115, - "output": 0.688, - "reasoning": 2.4 + "input": 0.060000000000000005, + "output": 0.5 } }, - "qwen/qwen3.7-flash": { - "id": "qwen/qwen3.7-flash", - "name": "Qwen3.7 Flash", - "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", - "family": "qwen", + "qwen/Qwen3.6-35B-A3B:thinking": { + "id": "qwen/Qwen3.6-35B-A3B:thinking", + "name": "Qwen3.6 35B A3B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-19", + "last_updated": "2026-04-19", "modalities": { "input": [ "text", @@ -200565,29 +188389,25 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 991000, - "output": 65536 + "context": 262144, + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.03, - "output": 0.13, - "cache_read": 0.006, - "cache_write": 0.038 + "input": 0.112, + "output": 0.8 } }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", + "qwen/qwen3-coder-plus": { + "id": "qwen/qwen3-coder-plus", + "name": "Qwen3 Coder Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-17", + "last_updated": "2025-09-17", "modalities": { "input": [ "text" @@ -200598,121 +188418,60 @@ }, "open_weights": false, "limit": { - "context": 262144, + "context": 128000, + "input": 128000, "output": 65536 }, "cost": { - "input": 0.359, - "output": 1.434 + "input": 1, + "output": 5 } }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", + "qwen/Qwen3.6-35B-A3B": { + "id": "qwen/Qwen3.6-35B-A3B", + "name": "Qwen3.6 35B A3B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.115, - "output": 0.917 - } - }, - "qwen/qwen3.6-35b-a3b": { - "id": "qwen/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, "release_date": "2026-04-17", "last_updated": "2026-04-17", "modalities": { "input": [ "text", "image", - "video", - "audio" + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 65536 + "input": 262144, + "output": 16384 }, "cost": { - "input": 0.248, - "output": 1.485 + "input": 0.112, + "output": 0.8 } }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen3.6 Flash", + "qwen/qwen3-32b": { + "id": "qwen/qwen3-32b", + "name": "Qwen 3 32b", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text", - "image", - "video" + "pdf" ], "output": [ "text" @@ -200720,121 +188479,58 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 41000, + "input": 41000, + "output": 32768 }, "cost": { - "input": 0.25, - "output": 1.5 + "input": 0.1, + "output": 0.3 } }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5 35B-A3B", + "qwen/Qwen3-VL-235B-A22B-Instruct": { + "id": "qwen/Qwen3-VL-235B-A22B-Instruct", + "name": "Qwen3 VL 235B A22B Instruct", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ "text", - "image", - "video", - "audio" + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 128000, + "input": 128000, + "output": 262144 }, "cost": { - "input": 0.057, - "output": 0.459 + "input": 0.3, + "output": 1.2 } }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", + "qwen/qwen3-max": { + "id": "qwen/qwen3-max", + "name": "Qwen3 Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.172, - "output": 1.032 - } - }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" ], "output": [ "text" @@ -200842,54 +188538,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 1.08018, + "output": 5.4009 } }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", + "qwen/qwen3.5-397b-a17b-thinking": { + "id": "qwen/qwen3.5-397b-a17b-thinking", + "name": "Qwen3.5 397B A17B Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens" + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", @@ -200902,45 +188576,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 258048, + "input": 258048, + "output": 65536 }, "cost": { - "input": 0.35, - "output": 1.42, - "cache_read": 0.071 + "input": 0.6, + "output": 3.6 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen/Qwen3-Next-80B-A3B-Instruct": { + "id": "qwen/Qwen3-Next-80B-A3B-Instruct", + "name": "Qwen3 Next 80B A3B (Instruct)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "release_date": "2025-09-11", + "last_updated": "2025-09-11", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -200948,75 +188605,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } + "context": 256000, + "input": 256000, + "output": 262144 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0.15, + "output": 0.65 } }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], + "qwen/Qwen3-235B-A22B-Instruct-2507-TEE": { + "id": "qwen/Qwen3-235B-A22B-Instruct-2507-TEE", + "name": "Qwen 3 235b A22B 2507 (TEE)", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "structured_output": true, + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -201024,42 +188634,25 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 256000, + "input": 256000, + "output": 262144 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 0.13, + "output": 0.5 } }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", + "qwen/qwq-32b-preview": { + "id": "qwen/qwq-32b-preview", + "name": "Qwen QwQ 32B Preview", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", "attachment": false, "reasoning": false, "tool_call": false, "structured_output": false, - "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ "text" @@ -201070,43 +188663,28 @@ }, "open_weights": false, "limit": { - "context": 16385, - "output": 4096 + "context": 32768, + "input": 32768, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0 + "input": 0.2, + "output": 0.2 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "qwen/qwen3-next-80b-a3b-thinking": { + "id": "qwen/qwen3-next-80b-a3b-thinking", + "name": "Qwen3 Next 80B A3B (Thinking)", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" @@ -201114,92 +188692,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 100000 + "context": 256000, + "input": 256000, + "output": 32768 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 0.15, + "output": 0.65 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "qwen/qwen3-coder": { + "id": "qwen/qwen3-coder", + "name": "Qwen 3 Coder 480B", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", "release_date": "2026-03-17", "last_updated": "2026-03-17", "modalities": { "input": [ - "text", - "image" - ], - "output": [ "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "openai/gpt-4o-2024-05-13": { - "id": "openai/gpt-4o-2024-05-13", - "name": "GPT-4o (2024-05-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", - "modalities": { - "input": [ - "text", - "image" ], "output": [ "text" @@ -201207,115658 +188721,155 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 262000, + "input": 262000, + "output": 65536 }, "cost": { - "input": 5, - "output": 15 + "input": 0.13, + "output": 0.5 } }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "openai/gpt-5.1-codex": { - "id": "openai/gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2 Pro", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 21, - "output": 168 - } - }, - "openai/gpt-4": { - "id": "openai/gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 30, - "output": 60 - } - }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, + "qwen/Qwen3-235B-A22B-Thinking-2507": { + "id": "qwen/Qwen3-235B-A22B-Thinking-2507", + "name": "Qwen 3 235b A22B 2507 Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 - } - }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 - } - }, - "openai/gpt-5.2-chat-latest": { - "id": "openai/gpt-5.2-chat-latest", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "openai/gpt-5.1-chat-latest": { - "id": "openai/gpt-5.1-chat-latest", - "name": "GPT-5.1 Chat", - "description": "Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 272000 - }, - "cost": { - "input": 15, - "output": 120 - } - }, - "openai/gpt-5-chat-latest": { - "id": "openai/gpt-5-chat-latest", - "name": "GPT-5 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [], "tool_call": false, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 100000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-4o-2024-08-06": { - "id": "openai/gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-08-06", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 - } - }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 - } - }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "openai/gpt-4o-2024-11-20": { - "id": "openai/gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 - } - }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 10, - "output": 30 - } - }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.03, - "output": 0.17 - } - }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5 - } - } - } - }, - "friendli": { - "id": "friendli", - "env": [ - "FRIENDLI_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.friendli.ai/serverless/v1", - "name": "Friendli", - "doc": "https://friendli.ai/docs/guides/serverless_endpoints/introduction", - "models": { - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.14, - "output": 0.4 - } - }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "DeepSeek V3.2", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 64000 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.25 - } - }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "zai-org/GLM-5.3-Flash": { - "id": "zai-org/GLM-5.3-Flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 - } - }, - "zai-org/GLM-5.3": { - "id": "zai-org/GLM-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.26, - "output": 3.96, - "cache_read": 0.234 - } - } - } - }, - "kimi-code-plan-cn": { - "id": "kimi-code-plan-cn", - "env": [ - "KIMI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.kimi.com/coding/v1", - "name": "Kimi For Coding (kimi.com)", - "doc": "https://www.kimi.com/code/docs/en/kimi-code/models.html", - "models": { - "kimi-for-coding-highspeed": { - "id": "kimi-for-coding-highspeed", - "name": "Kimi For Coding HighSpeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "kimi-for-coding": { - "id": "kimi-for-coding", - "name": "kimi-for-coding", - "description": "Kimi coding model with more efficient thinking and up to 1M context, available through Kimi Code", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "k3-256k": { - "id": "k3-256k", - "name": "Kimi K3-256K", - "description": "256K-context version of Kimi K3, reducing token consumption for shorter coding sessions", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "k3": { - "id": "k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - } - } - }, - "inco": { - "id": "inco", - "env": [ - "INCO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.inco.ai/v1", - "name": "Inco", - "doc": "https://platform.inco.ai/docs", - "models": { - "kimi-k3:fast": { - "id": "kimi-k3:fast", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 6, - "output": 30 - } - }, - "glm-5.3:fast": { - "id": "glm-5.3:fast", - "name": "GLM-5.3 Fast", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2.8, - "output": 8.8 - } - }, - "glm-5.3-flash:fast": { - "id": "glm-5.3-flash:fast", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.5 - } - }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 512000 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "deepseek-v4.1-flash:fast": { - "id": "deepseek-v4.1-flash:fast", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.6, - "output": 2.4 - } - }, - "minimax-m3:fast": { - "id": "minimax-m3:fast", - "name": "MiniMax M3 Fast", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 512000 - }, - "cost": { - "input": 0.6, - "output": 2.4 - } - }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4 - } - } - } - }, - "sakana": { - "id": "sakana", - "env": [ - "SAKANA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.sakana.ai/v1", - "name": "Sakana AI", - "doc": "https://console.sakana.ai/models", - "models": { - "fugu-ultra": { - "id": "fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "provider": { - "shape": "responses" - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "fugu-ultra-20260615": { - "id": "fugu-ultra-20260615", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "provider": { - "shape": "responses" - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "sakana-namazu": { - "id": "sakana-namazu", - "name": "Sakana Namazu", - "description": "Japanese-specialized reasoning model based on Kimi K2.6 and tuned for Japanese language, culture, and business workflows", - "family": "sakana-namazu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.15 - } - }, - "fugu": { - "id": "fugu", - "name": "Fugu", - "description": "Multi-agent model for routing expert agents across complex analytical tasks", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "provider": { - "shape": "responses" - } - } - } - }, - "scx-ai": { - "id": "scx-ai", - "env": [ - "SCX_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.scx.ai/v1", - "name": "SCX.ai", - "doc": "https://platform.scx.ai/docs", - "models": { - "GLM-5.2": { - "id": "GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.55, - "output": 1.784, - "cache_read": 0.111 - } - }, - "Qwen3.8-Max": { - "id": "Qwen3.8-Max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 983616, - "output": 131072 - }, - "cost": { - "input": 1.815, - "output": 5.4461, - "cache_read": 0.17, - "cache_write": 2.5 - } - }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 196608 - }, - "cost": { - "input": 0.48, - "output": 1.79, - "cache_read": 0.05 - } - }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.17, - "output": 0.55 - } - } - } - }, - "zenifra": { - "id": "zenifra", - "env": [ - "ZENIFRA_AI_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://ai.zenifra.com/v1", - "name": "Zenifra", - "doc": "https://docs.zenifra.com", - "models": { - "alibaba/qwen3.6-35b-a3b": { - "id": "alibaba/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "provider": { - "shape": "completions" - }, - "cost": { - "input": 0.19, - "output": 0.48 - } - } - } - }, - "tokenrouter": { - "id": "tokenrouter", - "env": [ - "TOKENROUTER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.tokenrouter.com/v1", - "name": "TokenRouter", - "doc": "https://www.tokenrouter.com/docs/tokenrouter-feature-guide/", - "models": { - "z-ai/glm-5.3-free": { - "id": "z-ai/glm-5.3-free", - "name": "GLM-5.3 (free)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - } - } - }, - "google-vertex-anthropic": { - "id": "google-vertex-anthropic", - "env": [ - "GOOGLE_VERTEX_PROJECT", - "GOOGLE_VERTEX_LOCATION", - "GOOGLE_APPLICATION_CREDENTIALS" - ], - "npm": "@ai-sdk/google-vertex/anthropic", - "name": "Vertex (Anthropic)", - "doc": "https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude", - "models": { - "claude-opus-4-8@default": { - "id": "claude-opus-4-8@default", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "claude-sonnet-4-5@20250929": { - "id": "claude-sonnet-4-5@20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "claude-opus-4-6@default": { - "id": "claude-opus-4-6@default", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "claude-opus-4@20250514": { - "id": "claude-opus-4@20250514", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "claude-opus-5-5@default": { - "id": "claude-opus-5-5@default", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 - } - }, - "claude-opus-5@default": { - "id": "claude-opus-5@default", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "claude-sonnet-4-6@default": { - "id": "claude-sonnet-4-6@default", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } - } - }, - "claude-haiku-4-5@20251001": { - "id": "claude-haiku-4-5@20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "claude-opus-4-1@20250805": { - "id": "claude-opus-4-1@20250805", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "claude-opus-4-5@20251101": { - "id": "claude-opus-4-5@20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "claude-opus-4-7@default": { - "id": "claude-opus-4-7@default", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "claude-fable-5-1@default": { - "id": "claude-fable-5-1@default", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "claude-fable-5@default": { - "id": "claude-fable-5@default", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "claude-sonnet-4@20250514": { - "id": "claude-sonnet-4@20250514", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "status": "deprecated", - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "claude-sonnet-5@default": { - "id": "claude-sonnet-5@default", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - } - } - }, - "moonshotai": { - "id": "moonshotai", - "env": [ - "MOONSHOT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.moonshot.ai/v1", - "name": "Moonshot AI", - "doc": "https://platform.moonshot.ai/docs/api/chat", - "models": { - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "kimi-k2.7-code-highspeed": { - "id": "kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code HighSpeed", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 - } - }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 - } - }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - } - } - }, - "ainetcafe": { - "id": "ainetcafe", - "env": [ - "AINETCAFE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://microquickjs.com/v1", - "name": "ainetcafe", - "doc": "https://ainetcafe.com/k3/guides/", - "models": { - "Kimi-K3": { - "id": "Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 2.1, - "output": 10.5, - "cache_read": 0.3 - } - } - } - }, - "wallaby": { - "id": "wallaby", - "env": [ - "WALLABY_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.wallabytoken.com/v1", - "name": "Wallaby", - "doc": "https://wallabytoken.com/docs", - "models": { - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 2.7, - "output": 13.5, - "cache_read": 0.27 - } - } - } - }, - "scnet-token-plan": { - "id": "scnet-token-plan", - "env": [ - "SCNET_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.scnet.cn/api/llm/v1", - "name": "SCNet Token Plan", - "doc": "https://www.scnet.cn/ac/openapi/doc/2.0/moduleapi/plans/token-plan.html", - "models": { - "DeepSeek-V4-Flash-0731": { - "id": "DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "Kimi-K2.6": { - "id": "Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 512000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "GLM-5.1": { - "id": "GLM-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "GLM-5.2": { - "id": "GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "Qwen3.8-Max": { - "id": "Qwen3.8-Max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "Kimi-K3": { - "id": "Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "DeepSeek-V4-Pro-0813": { - "id": "DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "DeepSeek-V4.1-Flash": { - "id": "DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "DeepSeek-V4-Flash": { - "id": "DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "GLM-5": { - "id": "GLM-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "GLM-5.3-Flash": { - "id": "GLM-5.3-Flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "DeepSeek-V4-Pro": { - "id": "DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "Qwen3.8-Flash": { - "id": "Qwen3.8-Flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "GLM-5.3": { - "id": "GLM-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "Kimi-K2.7-Code": { - "id": "Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "Kimi-K2.5": { - "id": "Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - } - } - }, - "ofox": { - "id": "ofox", - "env": [ - "OFOX_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.ofox.ai/v1", - "name": "Ofox", - "doc": "https://ofox.ai/docs", - "models": { - "bailian/qwen-flash": { - "id": "bailian/qwen-flash", - "name": "Qwen Flash", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 0.022, - "output": 0.22, - "cache_read": 0.0043, - "cache_write": 0.027 - } - }, - "bailian/qwen3.5-flash": { - "id": "bailian/qwen3.5-flash", - "name": "Qwen3.5 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0.125 - } - }, - "bailian/qwen3.7-max": { - "id": "bailian/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 - } - }, - "bailian/qwen3.8-27b": { - "id": "bailian/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1131072, - "output": 131072 - }, - "cost": { - "input": 0.45, - "output": 3.2, - "cache_read": 0.05, - "cache_write": 0.5625 - } - }, - "bailian/qwen3.5-27b": { - "id": "bailian/qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.29, - "output": 2.05, - "cache_read": 0.29 - } - }, - "bailian/qwen-max": { - "id": "bailian/qwen-max", - "name": "Qwen Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 0.35, - "output": 1.38, - "cache_read": 0.069 - } - }, - "bailian/qwen3.8-max": { - "id": "bailian/qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 - } - }, - "bailian/qwen3-coder-next": { - "id": "bailian/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.2, - "output": 1.5 - } - }, - "bailian/qwen3.5-plus": { - "id": "bailian/qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.04, - "cache_write": 0.4 - } - }, - "bailian/qwen-plus": { - "id": "bailian/qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 0.12, - "output": 0.29, - "cache_read": 0.023 - } - }, - "bailian/qwen3-max": { - "id": "bailian/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.36, - "output": 1.43, - "cache_read": 0.072 - } - }, - "bailian/qwen-turbo": { - "id": "bailian/qwen-turbo", - "name": "Qwen Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11-01", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.05, - "output": 0.09, - "cache_read": 0.0086 - } - }, - "bailian/qwen-vl-max": { - "id": "bailian/qwen-vl-max", - "name": "Qwen-VL Max", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-08", - "last_updated": "2025-08-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.23, - "output": 0.58, - "cache_read": 0.046 - } - }, - "bailian/qwen3.5-122b-a10b": { - "id": "bailian/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.29, - "output": 2.29, - "cache_read": 0.29 - } - }, - "bailian/qwen3.6-flash": { - "id": "bailian/qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 0.31 - } - }, - "bailian/qwen3-coder-flash": { - "id": "bailian/qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.06, - "cache_write": 0.27 - } - }, - "bailian/qwen3-coder-plus": { - "id": "bailian/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.8, - "output": 9, - "cache_read": 0.2, - "cache_write": 1 - } - }, - "bailian/qwen3.8-flash": { - "id": "bailian/qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016, - "cache_write": 0.2 - } - }, - "bailian/qwen3.5-35b-a3b": { - "id": "bailian/qwen3.5-35b-a3b", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.29, - "output": 1.83, - "cache_read": 0.29 - } - }, - "bailian/qwen3.6-max-preview": { - "id": "bailian/qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 2.15, - "output": 12.86, - "cache_read": 0.2, - "cache_write": 1.17 - } - }, - "bailian/qwen3.5-397b-a17b": { - "id": "bailian/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.55, - "output": 3.5, - "cache_read": 0.55 - } - }, - "bailian/qwen3.6-27b": { - "id": "bailian/qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.6, - "output": 3.6 - } - }, - "bailian/qwen3.6-plus": { - "id": "bailian/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625 - } - }, - "bailian/qwen3.7-plus": { - "id": "bailian/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "cache_write": 0.5 - } - }, - "bailian/qwen3.8-max-0902": { - "id": "bailian/qwen3.8-max-0902", - "name": "Qwen3.8 Max 0902", - "description": "2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 - } - }, - "anthropic/claude-opus-4.5": { - "id": "anthropic/claude-opus-4.5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-opus-4.6": { - "id": "anthropic/claude-opus-4.6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-opus-4.8": { - "id": "anthropic/claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-haiku-4.5": { - "id": "anthropic/claude-haiku-4.5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-fable-5.1": { - "id": "anthropic/claude-fable-5.1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "anthropic/claude-sonnet-4.5": { - "id": "anthropic/claude-sonnet-4.5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "anthropic/claude-sonnet-4.6": { - "id": "anthropic/claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic/claude-opus-5.5": { - "id": "anthropic/claude-opus-5.5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic", - "api": "https://api.ofox.ai/anthropic/v1" - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 - } - }, - "deepseek/deepseek-v4-flash-0731": { - "id": "deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.308, - "output": 0.924, - "cache_read": 0.0098 - } - }, - "deepseek/deepseek-v4-flash-0423": { - "id": "deepseek/deepseek-v4-flash-0423", - "name": "DeepSeek V4 Flash 0423", - "description": "Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.19, - "output": 0.51, - "cache_read": 0.028 - } - }, - "deepseek/deepseek-v4.1-flash": { - "id": "deepseek/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.21, - "output": 0.84, - "cache_read": 0.0042 - } - }, - "deepseek/deepseek-v4-flash-vision-exp": { - "id": "deepseek/deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "status": "beta", - "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.014 - } - }, - "deepseek/deepseek-v4-pro-0813": { - "id": "deepseek/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.924, - "output": 2.772, - "cache_read": 0.0308 - } - }, - "deepseek/deepseek-v4-pro-0423": { - "id": "deepseek/deepseek-v4-pro-0423", - "name": "DeepSeek V4 Pro 0423", - "description": "DeepSeek V4 Pro initial snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.15 - } - }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 - } - }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32000 - }, - "cost": { - "input": 0.29, - "output": 0.43, - "cache_read": 0.06 - } - }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.014 - } - }, - "z-ai/glm-5v-turbo": { - "id": "z-ai/glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 - } - }, - "z-ai/glm-5.3-flash": { - "id": "z-ai/glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 - } - }, - "z-ai/glm-4.6": { - "id": "z-ai/glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 - } - }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 - } - }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.4, - "output": 2.2, - "cache_read": 0.11 - } - }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "z-ai/glm-4.7-flashx": { - "id": "z-ai/glm-4.7-flashx", - "name": "GLM-4.7-FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.072, - "output": 0.4, - "cache_read": 0.01 - } - }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 - } - }, - "z-ai/glm-5.3": { - "id": "z-ai/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "x-ai/grok-4.7": { - "id": "x-ai/grok-4.7", - "name": "Grok 4.7", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 30000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 - } - }, - "x-ai/grok-4.5": { - "id": "x-ai/grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3 - } - }, - "x-ai/grok-4.20": { - "id": "x-ai/grok-4.20", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 12, - "cache_read": 0.4 - } - }, - "x-ai/grok-4.1-fast": { - "id": "x-ai/grok-4.1-fast", - "name": "Grok 4.1 Fast", - "description": "xAI's fast agentic tool-calling model with a 2M context window; non-reasoning variant for low-latency responses", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 30000 - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } - }, - "x-ai/grok-4.6": { - "id": "x-ai/grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "volcengine/doubao-seed-2.0-mini": { - "id": "volcengine/doubao-seed-2.0-mini", - "name": "Seed 2.0 Mini", - "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 32000 - }, - "cost": { - "input": 0.06, - "output": 0.56, - "cache_read": 0.02, - "cache_write": 0.0024 - } - }, - "volcengine/doubao-seed-1-6-flash": { - "id": "volcengine/doubao-seed-1-6-flash", - "name": "Seed 1.6 Flash", - "description": "Low-latency ByteDance Seed model for high-throughput chat, extraction, and lightweight tool use", - "family": "seed", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-28", - "last_updated": "2025-08-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 32000 - }, - "cost": { - "input": 0.03, - "output": 0.22, - "cache_read": 0.0043 - } - }, - "volcengine/doubao-seed-2.0-lite": { - "id": "volcengine/doubao-seed-2.0-lite", - "name": "Seed 2.0 Lite", - "description": "Cost-efficient ByteDance Seed 2.0 model for production chat, analysis, and structured generation", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 32000 - }, - "cost": { - "input": 0.13, - "output": 0.76, - "cache_read": 0.03, - "cache_write": 0.0024 - } - }, - "volcengine/doubao-seed-1-8": { - "id": "volcengine/doubao-seed-1-8", - "name": "Seed 1.8", - "description": "ByteDance Seed model for multimodal reasoning, long-context analysis, and agent workflows", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-28", - "last_updated": "2025-12-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.12, - "output": 0.29, - "cache_read": 0.023 - } - }, - "volcengine/doubao-seed-1-6": { - "id": "volcengine/doubao-seed-1-6", - "name": "Seed 1.6", - "description": "ByteDance Seed model for long-context reasoning, instruction following, and tool-assisted tasks", - "family": "seed", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.12, - "output": 0.29, - "cache_read": 0.023 - } - }, - "volcengine/doubao-seed-2.0-pro": { - "id": "volcengine/doubao-seed-2.0-pro", - "name": "Seed 2.0 Pro", - "description": "Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.67, - "output": 3.36, - "cache_read": 0.14, - "cache_write": 0.0024 - } - }, - "volcengine/doubao-seed-1-6-vision": { - "id": "volcengine/doubao-seed-1-6-vision", - "name": "Seed 1.6 Vision", - "description": "ByteDance Seed multimodal model for image understanding, visual reasoning, and tool-assisted tasks", - "family": "seed", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-15", - "last_updated": "2025-08-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 32000 - }, - "cost": { - "input": 0.12, - "output": 1.15, - "cache_read": 0.023 - } - }, - "volcengine/doubao-seed-2.1-turbo": { - "id": "volcengine/doubao-seed-2.1-turbo", - "name": "Seed 2.1 Turbo", - "description": "Faster ByteDance Seed 2.1 model for multimodal reasoning and latency-sensitive agent workflows", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.3536, - "output": 1.7696, - "cache_read": 0.068, - "cache_write": 0.0019 - } - }, - "volcengine/doubao-seed-2.1-pro": { - "id": "volcengine/doubao-seed-2.1-pro", - "name": "Seed 2.1 Pro", - "description": "Flagship ByteDance Seed 2.1 model for complex multimodal reasoning, coding, and agents", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.7072, - "output": 3.536, - "cache_read": 0.1416, - "cache_write": 0.002 - } - }, - "volcengine/doubao-seed-character": { - "id": "volcengine/doubao-seed-character", - "name": "Seed Character", - "description": "ByteDance Seed model optimized for character-driven dialogue and consistent conversational behavior", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.177, - "output": 0.884, - "cache_read": 0.024, - "cache_write": 0.0025 - } - }, - "volcengine/doubao-seed-evolving": { - "id": "volcengine/doubao-seed-evolving", - "name": "Seed Evolving", - "description": "Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.884, - "output": 4.42, - "cache_read": 0.177, - "cache_write": 0.0025 - } - }, - "volcengine/doubao-seed-2.0-code": { - "id": "volcengine/doubao-seed-2.0-code", - "name": "Seed 2.0 Code", - "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.67, - "output": 3.36, - "cache_read": 0.14, - "cache_write": 0.0024 - } - }, - "google/gemini-3.6-flash": { - "id": "google/gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google", - "api": "https://api.ofox.ai/gemini/v1beta" - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.0415, - "input_audio": 1.5 - } - }, - "google/gemini-3.5-flash-lite": { - "id": "google/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google", - "api": "https://api.ofox.ai/gemini/v1beta" - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.083 - } - }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google", - "api": "https://api.ofox.ai/gemini/v1beta" - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 4.5 - } - }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google", - "api": "https://api.ofox.ai/gemini/v1beta" - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 0.083, - "input_audio": 3 - } - }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "cache_write": 4.5 - } - }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 1, - "input_audio": 1 - } - }, - "google/gemini-3.7-flash": { - "id": "google/gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google", - "api": "https://api.ofox.ai/gemini/v1beta" - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.0415, - "input_audio": 1.5 - } - }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 1 - } - }, - "google/gemini-3.8-flash": { - "id": "google/gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google", - "api": "https://api.ofox.ai/gemini/v1beta" - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.0415, - "input_audio": 0.75 - } - }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 1, - "input_audio": 0.3 - } - }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 1, - "input_audio": 0.5 - } - }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "moonshotai/kimi-k2.7-code-highspeed": { - "id": "moonshotai/kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 - } - }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 - } - }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 - } - }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - }, - "minimax/minimax-m2.5-lightning": { - "id": "minimax/minimax-m2.5-lightning", - "name": "MiniMax-M2.5 Lightning", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 2.4, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "minimax/minimax-m2.1-lightning": { - "id": "minimax/minimax-m2.1-lightning", - "name": "MiniMax-M2.1 Lightning", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 2.4, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 512000 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 - } - }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "minimax/m2-her": { - "id": "minimax/m2-her", - "name": "MiniMax-M2 Her", - "description": "MiniMax M2 variant tuned for conversational and character-driven agent interactions", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-01-23", - "last_updated": "2026-01-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 2048 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "qwen/qwen-flash": { - "id": "qwen/qwen-flash", - "name": "Qwen Flash", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32000 - }, - "cost": { - "input": 0.022, - "output": 0.22, - "cache_read": 0.0043, - "cache_write": 0.027 - } - }, - "qwen/qwen3.5-flash": { - "id": "qwen/qwen3.5-flash", - "name": "Qwen3.5 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0.125 - } - }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1064000, - "output": 64000 - }, - "cost": { - "input": 1.71, - "output": 5.14, - "cache_read": 0.17, - "cache_write": 2.14 - } - }, - "qwen/qwen3.8-27b": { - "id": "qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1131072, - "output": 131072 - }, - "cost": { - "input": 0.5, - "output": 1.71, - "cache_read": 0.043, - "cache_write": 0.63 - } - }, - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.29, - "output": 2.05, - "cache_read": 0.29 - } - }, - "qwen/qwen-max": { - "id": "qwen/qwen-max", - "name": "Qwen Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 8000 - }, - "cost": { - "input": 0.35, - "output": 1.38, - "cache_read": 0.069 - } - }, - "qwen/qwen3.8-max": { - "id": "qwen/qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.71, - "output": 5.14, - "cache_read": 0.17, - "cache_write": 2.14 - } - }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.2, - "output": 1.5 - } - }, - "qwen/qwen3.5-plus": { - "id": "qwen/qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.04, - "cache_write": 0.4 - } - }, - "qwen/qwen-plus": { - "id": "qwen/qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32000 - }, - "cost": { - "input": 0.12, - "output": 0.29, - "cache_read": 0.023 - } - }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.36, - "output": 1.43, - "cache_read": 0.072 - } - }, - "qwen/qwen-turbo": { - "id": "qwen/qwen-turbo", - "name": "Qwen Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11-01", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16000 - }, - "cost": { - "input": 0.043, - "output": 0.09, - "cache_read": 0.0086 - } - }, - "qwen/qwen-vl-max": { - "id": "qwen/qwen-vl-max", - "name": "Qwen-VL Max", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-08", - "last_updated": "2025-08-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 8000 - }, - "cost": { - "input": 0.23, - "output": 0.58, - "cache_read": 0.023 - } - }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.29, - "output": 2.29, - "cache_read": 0.29 - } - }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 0.31 - } - }, - "qwen/qwen3-coder-flash": { - "id": "qwen/qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.06, - "cache_write": 0.27 - } - }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 1.8, - "output": 9, - "cache_read": 0.2, - "cache_write": 1 - } - }, - "qwen/qwen3.8-flash": { - "id": "qwen/qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.11, - "output": 0.39, - "cache_read": 0.011, - "cache_write": 0.14 - } - }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.29, - "output": 1.83, - "cache_read": 0.29 - } - }, - "qwen/qwen3.6-max-preview": { - "id": "qwen/qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 2.15, - "output": 12.86, - "cache_read": 0.2, - "cache_write": 1.17 - } - }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.55, - "output": 3.5, - "cache_read": 0.55 - } - }, - "qwen/qwen3.6-27b": { - "id": "qwen/qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 64000 - }, - "cost": { - "input": 0.43, - "output": 2.57 - } - }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625 - } - }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1064000, - "output": 64000 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "cache_write": 0.5 - } - }, - "qwen/qwen3.8-max-0902": { - "id": "qwen/qwen3.8-max-0902", - "name": "Qwen3.8 Max 0902", - "description": "2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 262144 - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.71, - "output": 5.14, - "cache_read": 0.17, - "cache_write": 2.14 - } - }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 - } - }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 24, - "output": 144 - } - }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 0.16, - "output": 1, - "cache_read": 0.016 - } - }, - "openai/gpt-5.2-codex": { - "id": "openai/gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 1.4, - "output": 11.2, - "cache_read": 0.144 - } - }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 1 - } - }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 32768 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 0.2, - "output": 1.6, - "cache_read": 0.024 - } - }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 1.4, - "output": 11.2, - "cache_read": 0.144 - } - }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 0.04, - "output": 0.32, - "cache_read": 0.008 - } - }, - "openai/gpt-6-astra": { - "id": "openai/gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 8, - "output": 40, - "cache_read": 0.8, - "cache_write": 10 - } - }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 1, - "output": 8, - "cache_read": 0.104 - } - }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 0.12, - "output": 0.48, - "cache_read": 0.06 - } - }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 0.6, - "output": 3.6, - "cache_read": 0.06 - } - }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 - } - }, - "openai/gpt-5.1-codex-max": { - "id": "openai/gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 1, - "output": 8, - "cache_read": 0.104 - } - }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 1.4, - "output": 11.2, - "cache_read": 0.144 - } - }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 4, - "output": 24, - "cache_read": 0.4 - } - }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 1.6, - "output": 6.4, - "cache_read": 0.4 - } - }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 0.32, - "output": 1.28, - "cache_read": 0.08 - } - }, - "openai/gpt-6-luna": { - "id": "openai/gpt-6-luna", - "name": "GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 0.08, - "output": 0.4, - "cache_read": 0.008, - "cache_write": 0.1 - } - }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "openai/gpt-5.1-codex-mini": { - "id": "openai/gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 0.2, - "output": 1.6, - "cache_read": 0.024 - } - }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 1, - "output": 8, - "cache_read": 0.104 - } - }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125 - } - }, - "openai/gpt-6-sol": { - "id": "openai/gpt-6-sol", - "name": "GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "https://api.ofox.ai/v1" - }, - "cost": { - "input": 1.6, - "output": 8, - "cache_read": 0.16, - "cache_write": 2 - } - } - } - }, - "neon": { - "id": "neon", - "env": [ - "NEON_AI_GATEWAY_BASE_URL", - "NEON_AI_GATEWAY_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "${NEON_AI_GATEWAY_BASE_URL}/v1", - "name": "Neon", - "doc": "https://neon.com/docs", - "models": { - "gpt-5-4-mini": { - "id": "gpt-5-4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "gpt-5-6-luna": { - "id": "gpt-5-6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } - } - }, - "glm-5-3-flash": { - "id": "glm-5-3-flash", - "name": "GLM-5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 - } - }, - "gpt-5-6-terra": { - "id": "gpt-5-6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "llama-4-maverick": { - "id": "llama-4-maverick", - "name": "Llama 4 Maverick 17B Instruct", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 8192 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 - } - }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "claude-fable-5-1": { - "id": "claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "gpt-5-4-nano": { - "id": "gpt-5-4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "gpt-5-6-sol": { - "id": "gpt-5-6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "gpt-5-5-pro": { - "id": "gpt-5-5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } - } - }, - "gemini-3-1-pro": { - "id": "gemini-3-1-pro", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "glm-5-2": { - "id": "glm-5-2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } - } - }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 10000 - }, - "cost": { - "input": 0.15, - "output": 1.2 - } - }, - "gemini-3-flash": { - "id": "gemini-3-flash", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 - } - }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 25000 - }, - "cost": { - "input": 0.07, - "output": 0.3 - } - }, - "gpt-5-4": { - "id": "gpt-5-4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "inkling": { - "id": "inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.17 - } - }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gemini-3-1-flash-lite": { - "id": "gemini-3-1-flash-lite", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 - } - }, - "gpt-5-5": { - "id": "gpt-5-5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "grok-4-6": { - "id": "grok-4-6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 524288 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "qwen35-122b-a10b": { - "id": "qwen35-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 25000 - }, - "cost": { - "input": 0.22, - "output": 2.2 - } - }, - "gemini-3-6-flash": { - "id": "gemini-3-6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "meta-llama-3-3-70b-instruct": { - "id": "meta-llama-3-3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "gpt-5-2": { - "id": "gpt-5-2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "gpt-5-3-codex": { - "id": "gpt-5-3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "gemma-3-12b": { - "id": "gemma-3-12b", - "name": "Gemma 3 12B", - "description": "Google's open-weight Gemma 3 vision-language model for text and image understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08-31", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.5 - } - }, - "gemini-3-5-flash-lite": { - "id": "gemini-3-5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 - } - }, - "gemini-3-5-flash": { - "id": "gemini-3-5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "gpt-5-1": { - "id": "gpt-5-1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 25000 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai", - "api": "${NEON_AI_GATEWAY_BASE_URL}/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "meta-llama-3-1-8b-instruct": { - "id": "meta-llama-3-1-8b-instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Meta's compact open-weight Llama 3.1 model for fast, low-cost text generation", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.45 - } - } - } - }, - "aihubmix": { - "id": "aihubmix", - "env": [ - "AIHUBMIX_API_KEY" - ], - "npm": "@aihubmix/ai-sdk-provider", - "name": "AIHubMix", - "doc": "https://docs.aihubmix.com", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } - } - }, - "coding-minimax-m2.7-free": { - "id": "coding-minimax-m2.7-free", - "name": "Coding MiniMax M2.7 (Free)", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 128100 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 991000, - "output": 64000 - }, - "cost": { - "input": 1.69, - "output": 5.07, - "cache_read": 0.169, - "cache_write": 2.1125 - } - }, - "muse-spark-1.3": { - "id": "muse-spark-1.3", - "name": "Muse Spark 1.3", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.375, - "output": 4.675, - "cache_read": 0.165 - } - }, - "deep-deepseek-v4-pro": { - "id": "deep-deepseek-v4-pro", - "name": "DeepSeek V4 Pro (DeepSeek)", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.478, - "output": 0.956, - "cache_read": 0.004302 - } - }, - "qwen3.8-2.4t-a95b": { - "id": "qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM 5 Vision Turbo", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glmv", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-09", - "last_updated": "2026-05-09", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.7042, - "output": 3.09848, - "cache_read": 0.169008 - } - }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0.11268, - "output": 0.39438, - "cache_read": 0.02817 - } - }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.142, - "output": 0.284, - "cache_read": 0.0284 - } - }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 991000, - "output": 128000 - }, - "cost": { - "input": 1.69, - "output": 5.07, - "cache_read": 0.169, - "cache_write": 2.1125 - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "doubao-seed-2-0-lite-260428": { - "id": "doubao-seed-2-0-lite-260428", - "name": "Doubao Seed 2.0 Lite 260428", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.08, - "output": 0.51, - "cache_read": 0.01692, - "input_audio": 1.269, - "tiers": [ - { - "input": 0.13, - "output": 0.76, - "cache_read": 0.02536, - "input_audio": 1.902, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.25, - "output": 1.52, - "cache_read": 0.05072, - "input_audio": 3.804, - "tier": { - "type": "context", - "size": 128000 - } - } - ] - } - }, - "claude-sonnet-4-6-think": { - "id": "claude-sonnet-4-6-think", - "name": "Claude Sonnet 4.6 Thinking", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } - } - }, - "deepseek-v4-flash-0731-fast": { - "id": "deepseek-v4-flash-0731-fast", - "name": "DeepSeek V4 Flash 0731 Fast", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.28, - "output": 1.4, - "cache_read": 0.07 - } - }, - "hy3-preview": { - "id": "hy3-preview", - "name": "Hy3 Preview", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.17, - "output": 0.566661, - "cache_read": 0.051 - } - }, - "alicloud-deepseek-v4-pro": { - "id": "alicloud-deepseek-v4-pro", - "name": "DeepSeek V4 Pro (Alibaba Cloud)", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.69, - "output": 3.38, - "cache_read": 0.13 - } - }, - "zai-glm-5.1": { - "id": "zai-glm-5.1", - "name": "GLM-5.1 (Z.ai)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.845, - "output": 3.38, - "cache_read": 0.183112 - } - }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.155, - "output": 0.62, - "cache_read": 0.0031 - } - }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "coding-minimax-m2.7-highspeed": { - "id": "coding-minimax-m2.7-highspeed", - "name": "Coding MiniMax M2.7 Highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 128100 - }, - "cost": { - "input": 0.2, - "output": 0.2 - } - }, - "muse-spark-1.1": { - "id": "muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.375, - "output": 4.675 - } - }, - "deepseek-v4-flash-vision-exp": { - "id": "deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.155, - "output": 0.62, - "cache_read": 0.0031 - } - }, - "kimi-k2.7-code-highspeed": { - "id": "kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 1.9, - "output": 7.999, - "cache_read": 0.32167 - } - }, - "qwen3.7-flash": { - "id": "qwen3.7-flash", - "name": "Qwen3.7 Flash", - "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 991000, - "input": 991000, - "output": 64000 - }, - "cost": { - "input": 0.0282, - "output": 0.1128, - "cache_read": 0.00564, - "cache_write": 0.03525 - } - }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 - } - }, - "claude-opus-4-8-think": { - "id": "claude-opus-4-8-think", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "doubao-seed-2-0-pro": { - "id": "doubao-seed-2-0-pro", - "name": "Doubao Seed 2.0 Pro", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.48, - "output": 2.41, - "cache_read": 0.09644, - "tiers": [ - { - "input": 0.72, - "output": 3.62, - "cache_read": 0.144656, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.45, - "output": 7.23, - "cache_read": 0.28932, - "tier": { - "type": "context", - "size": 128000 - } - } - ] - } - }, - "xiaomi-mimo-v2.5": { - "id": "xiaomi-mimo-v2.5", - "name": "Xiaomi MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo-v2.5", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.44, - "output": 2.2, - "cache_read": 0.088, - "tiers": [ - { - "input": 0.88, - "output": 4.4, - "cache_read": 0.176, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.88, - "output": 4.4, - "cache_read": 0.176 - } - } - }, - "claude-fable-5-1": { - "id": "claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 11, - "output": 55, - "cache_read": 0.275, - "cache_write": 13.75 - } - }, - "gemini-3.6-flash": { - "id": "gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15 - } - }, - "longcat-2.0": { - "id": "longcat-2.0", - "name": "LongCat-2.0", - "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", - "family": "longcat", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.7746, - "output": 3.0984, - "cache_read": 0.015492 - } - }, - "doubao-seed-2-0-code-preview": { - "id": "doubao-seed-2-0-code-preview", - "name": "Doubao Seed 2.0 Code Preview", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.48, - "output": 2.41, - "cache_read": 0.09644, - "tiers": [ - { - "input": 0.72, - "output": 3.62, - "cache_read": 0.144656, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.45, - "output": 7.23, - "cache_read": 0.28932, - "tier": { - "type": "context", - "size": 128000 - } - } - ] - } - }, - "grok-4.3": { - "id": "grok-4.3", - "name": "Grok 4.3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-01", - "last_updated": "2026-05-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "qwen3.8-omni-flash": { - "id": "qwen3.8-omni-flash", - "name": "Qwen3.8 Omni Flash", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.1126, - "output": 0.380025, - "cache_read": 0.014075, - "cache_write": 0.175937 - } - }, - "doubao-seed-2-0-mini-260428": { - "id": "doubao-seed-2-0-mini-260428", - "name": "Doubao Seed 2.0 Mini 260428", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.03, - "output": 0.28, - "cache_read": 0.00564, - "input_audio": 0.423, - "tiers": [ - { - "input": 0.06, - "output": 0.56, - "cache_read": 0.01128, - "input_audio": 0.846, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.11, - "output": 1.13, - "cache_read": 0.02256, - "input_audio": 1.692, - "tier": { - "type": "context", - "size": 128000 - } - } - ] - } - }, - "xiaomi-mimo-v2.5-pro-free": { - "id": "xiaomi-mimo-v2.5-pro-free", - "name": "Xiaomi MiMo-V2.5-Pro (free)", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo-v2.5-pro", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.499999, - "cache_read": 0.03 - } - }, - "gemini-3.1-flash-image": { - "id": "gemini-3.1-flash-image", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, - "interleaved": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 3 - } - }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } - } - }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 - } - }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.13 - } - }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "coding-minimax-m2.7": { - "id": "coding-minimax-m2.7", - "name": "Coding MiniMax M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 128100 - }, - "cost": { - "input": 0.2, - "output": 0.2 - } - }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 128000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 1.5 - } - }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-05", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } - } - }, - "deep-deepseek-v4-flash": { - "id": "deep-deepseek-v4-flash", - "name": "DeepSeek V4 Flash (DeepSeek)", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.154, - "output": 0.308, - "cache_read": 0.0308 - } - }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 991000, - "output": 64000 - }, - "cost": { - "input": 0.17, - "output": 1.01, - "cache_read": 0.0169, - "cache_write": 0.21125, - "tiers": [ - { - "input": 0.68, - "output": 4.06, - "cache_read": 0.0676, - "cache_write": 0.845, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.68, - "output": 4.06, - "cache_read": 0.0676, - "cache_write": 0.845 - } - } - }, - "coding-glm-5.1": { - "id": "coding-glm-5.1", - "name": "Coding GLM 5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-11", - "last_updated": "2026-04-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.06, - "output": 0.22, - "cache_read": 0.013 - } - }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-20", - "last_updated": "2025-06-05", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 - } - }, - "xiaomi-mimo-v2.5-free": { - "id": "xiaomi-mimo-v2.5-free", - "name": "Xiaomi MiMo-V2.5 (free)", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo-v2.5", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 11, - "output": 55, - "cache_read": 1.1, - "cache_write": 13.75 - } - }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.6918, - "output": 2.0754, - "cache_read": 0.023058 - } - }, - "muse-spark-1.2": { - "id": "muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.375, - "output": 4.675 - } - }, - "gemini-3.1-flash-lite-image": { - "id": "gemini-3.1-flash-lite-image", - "name": "Nano Banana 2 Lite", - "description": "Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 4096 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025 - } - }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 512000 - }, - "cost": { - "input": 0.288, - "output": 1.152 - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.1126, - "output": 0.380025, - "cache_read": 0.014075, - "cache_write": 0.175937 - } - }, - "claude-opus-4-6-think": { - "id": "claude-opus-4-6-think", - "name": "Claude Opus 4.6 Thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "coding-xiaomi-mimo-v2.5": { - "id": "coding-xiaomi-mimo-v2.5", - "name": "Coding Xiaomi MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo-v2.5", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.08, - "output": 0.4, - "cache_read": 0.016, - "tiers": [ - { - "input": 0.16, - "output": 0.8, - "cache_read": 0.032, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.16, - "output": 0.8, - "cache_read": 0.032 - } - } - }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "ox-alpha": { - "id": "ox-alpha", - "name": "Ox Alpha", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen3.6-max-preview": { - "id": "qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "qwen3.6", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-05-09", - "last_updated": "2026-05-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 240000, - "output": 64000 - }, - "cost": { - "input": 1.27, - "output": 7.61, - "cache_read": 0.1268, - "cache_write": 1.585, - "tiers": [ - { - "input": 2.11, - "output": 12.67, - "cache_read": 0.2112, - "cache_write": 2.64, - "tier": { - "type": "context", - "size": 128000 - } - } - ] - } - }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 1.1268, - "output": 3.9438, - "cache_read": 0.2817 - } - }, - "hy3": { - "id": "hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 192000, - "output": 128000 - }, - "cost": { - "input": 0.1562, - "output": 0.6248, - "cache_read": 0.03905 - } - }, - "gemini-3.1-pro-preview-customtools": { - "id": "gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "tiers": [ - { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 - } - } - }, - "grok-build-0.1": { - "id": "grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 - } - }, - "claude-opus-4-7-think": { - "id": "claude-opus-4-7-think", - "name": "Claude Opus 4.7 Thinking", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "alicloud-deepseek-v4-flash": { - "id": "alicloud-deepseek-v4-flash", - "name": "DeepSeek V4 Flash (Alibaba Cloud)", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 - } - }, - "coding-xiaomi-mimo-v2.5-pro": { - "id": "coding-xiaomi-mimo-v2.5-pro", - "name": "Coding Xiaomi MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo-v2.5-pro", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.2, - "output": 0.6, - "cache_read": 0.04, - "tiers": [ - { - "input": 0.4, - "output": 1.2, - "cache_read": 0.08, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.2, - "cache_read": 0.08 - } - } - }, - "gemini-3.8-flash": { - "id": "gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "step-3.7-flash": { - "id": "step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 256000, - "output": 256000 - }, - "cost": { - "input": 0.22, - "output": 1.32, - "cache_read": 0.044 - } - }, - "alicloud-glm-5.1": { - "id": "alicloud-glm-5.1", - "name": "GLM-5.1 (Alibaba Cloud)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.84, - "output": 3.38, - "cache_read": 0.169, - "cache_write": 1.05625 - } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } - } - }, - "coding-glm-5.1-free": { - "id": "coding-glm-5.1-free", - "name": "Coding GLM 5.1 (free)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-11", - "last_updated": "2026-04-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-05-09", - "last_updated": "2026-05-09", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 991000, - "output": 64000 - }, - "cost": { - "input": 0.28, - "output": 1.69, - "cache_read": 0.0282, - "cache_write": 0.3525, - "tiers": [ - { - "input": 1.13, - "output": 6.77, - "cache_read": 0.1128, - "cache_write": 1.41, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.13, - "output": 6.77, - "cache_read": 0.1128, - "cache_write": 1.41 - } - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125 - } - }, - "hy4-preview": { - "id": "hy4-preview", - "name": "Hy4 preview", - "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1024000, - "output": 64000 - }, - "cost": { - "input": 0.845, - "output": 2.535, - "cache_read": 0.04225 - } - }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 1.1268, - "output": 3.9438, - "cache_read": 0.2817 - } - }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.95, - "output": 3.9995, - "cache_read": 0.160835 - } - }, - "grok-4.6": { - "id": "grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "xiaomi-mimo-v2.5-pro": { - "id": "xiaomi-mimo-v2.5-pro", - "name": "Xiaomi MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo-v2.5-pro", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-05-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.1, - "output": 3.3, - "cache_read": 0.22, - "tiers": [ - { - "input": 2.2, - "output": 6.6, - "cache_read": 0.44, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2.2, - "output": 6.6, - "cache_read": 0.44 - } - } - }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 991000, - "output": 64000 - }, - "cost": { - "input": 0.282, - "output": 1.128, - "cache_read": 0.0564, - "cache_write": 0.3525 - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 1 - } - } - } - }, - "merge-gateway": { - "id": "merge-gateway", - "env": [ - "MERGE_GATEWAY_API_KEY" - ], - "npm": "merge-gateway-ai-sdk-provider", - "api": "https://api-gateway.merge.dev/v1/ai-sdk", - "name": "Merge Gateway", - "doc": "https://docs.merge.dev/merge-gateway", - "models": { - "zai/glm-4.5": { - "id": "zai/glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 98304 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 - } - }, - "zai/glm-5.3-flash": { - "id": "zai/glm-5.3-flash", - "name": "GLM-5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.015, - "output": 0.05, - "cache_read": 0.003 - } - }, - "zai/glm-4.6": { - "id": "zai/glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 - } - }, - "zai/glm-5": { - "id": "zai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0 - } - }, - "zai/glm-4.5v": { - "id": "zai/glm-4.5v", - "name": "Glm 4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32000 - }, - "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11, - "cache_write": 0 - } - }, - "zai/glm-4.7-flash": { - "id": "zai/glm-4.7-flash", - "name": "GLM 4.7 Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.07, - "output": 0.4 - } - }, - "zai/glm-4.7": { - "id": "zai/glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 - } - }, - "zai/glm-5.2": { - "id": "zai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 50000 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.05, - "output": 3.3, - "cache_read": 0.195 - } - }, - "zai/glm-5.1": { - "id": "zai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 - } - }, - "zai/glm-4.7-flashx": { - "id": "zai/glm-4.7-flashx", - "name": "GLM-4.7 FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0 - } - }, - "zai/glm-5-turbo": { - "id": "zai/glm-5-turbo", - "name": "GLM-5 Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24, - "cache_write": 0 - } - }, - "zai/glm-4.5-air": { - "id": "zai/glm-4.5-air", - "name": "GLM-4.5 Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 98304 - }, - "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03, - "cache_write": 0 - } - }, - "zai/glm-5.3": { - "id": "zai/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.7, - "output": 2.2, - "cache_read": 0.13 - } - }, - "anthropic/claude-3-7-sonnet-20250219": { - "id": "anthropic/claude-3-7-sonnet-20250219", - "name": "Claude 3.7 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic/claude-opus-5-5": { - "id": "anthropic/claude-opus-5-5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 - } - }, - "anthropic/claude-opus-4-20250514": { - "id": "anthropic/claude-opus-4-20250514", - "name": "Claude Opus 4 (20250514)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "anthropic/claude-fable-5-1": { - "id": "anthropic/claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "anthropic/claude-opus-4-1-20250805": { - "id": "anthropic/claude-opus-4-1-20250805", - "name": "Claude Opus 4.1 (20250805)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "anthropic/claude-opus-4-5-20251101": { - "id": "anthropic/claude-opus-4-5-20251101", - "name": "Claude Opus 4.5 (20251101)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "anthropic/claude-opus-4-8": { - "id": "anthropic/claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 128000 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-sonnet-4-5-20250929": { - "id": "anthropic/claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5 (20250929)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 127999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-haiku-4-5-20251001": { - "id": "anthropic/claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5 (20251001)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic/claude-sonnet-4-20250514": { - "id": "anthropic/claude-sonnet-4-20250514", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "cohere/command-r-08-2024": { - "id": "cohere/command-r-08-2024", - "name": "Command R 08-2024", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "cohere/command-a-03-2025": { - "id": "cohere/command-a-03-2025", - "name": "Command A 03-2025", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8000 - }, - "cost": { - "input": 2.5, - "output": 10 - } - }, - "cohere/command-r7b-12-2024": { - "id": "cohere/command-r7b-12-2024", - "name": "Command R7B 12-2024", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-12-02", - "last_updated": "2024-12-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 0.0375, - "output": 0.15 - } - }, - "cohere/command-r-plus-08-2024": { - "id": "cohere/command-r-plus-08-2024", - "name": "Command R+ 08-2024", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 2.5, - "output": 10 - } - }, - "deepseek/deepseek-v4-flash-0731": { - "id": "deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.035, - "output": 0.07, - "cache_read": 0.007 - } - }, - "deepseek/deepseek-v4-flash-0423": { - "id": "deepseek/deepseek-v4-flash-0423", - "name": "DeepSeek V4 Flash 0423", - "description": "Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.139, - "output": 0.278 - } - }, - "deepseek/deepseek-v4-flash-0731-fast": { - "id": "deepseek/deepseek-v4-flash-0731-fast", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.28, - "output": 0.56, - "cache_read": 0.07 - } - }, - "deepseek/deepseek-v4.1-flash": { - "id": "deepseek/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 - } - }, - "deepseek/deepseek-v3": { - "id": "deepseek/deepseek-v3", - "name": "DeepSeek V3", - "description": "Open DeepSeek MoE chat model for coding, math, and general reasoning", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2024-12-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 81920 - }, - "cost": { - "input": 0.58, - "output": 1.68 - } - }, - "deepseek/deepseek-v4-pro-0813": { - "id": "deepseek/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.003625 - } - }, - "deepseek/deepseek-v4-pro-0423": { - "id": "deepseek/deepseek-v4-pro-0423", - "name": "DeepSeek V4 Pro 0423", - "description": "DeepSeek V4 Pro initial snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 393216 - }, - "cost": { - "input": 1.65, - "output": 3.3 - } - }, - "deepseek/deepseek-v3.1": { - "id": "deepseek/deepseek-v3.1", - "name": "DeepSeek V3.1", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 164000, - "output": 41000 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "deepseek/deepseek-r1": { - "id": "deepseek/deepseek-r1", - "name": "DeepSeek R1", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 40960 - }, - "cost": { - "input": 1.35, - "output": 5.4 - } - }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 262000 - }, - "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.145 - } - }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 40960 - }, - "cost": { - "input": 0.28, - "output": 0.45, - "cache_read": 0.14 - } - }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.035, - "output": 0.07, - "cache_read": 0.007 - } - }, - "thinkingmachines/inkling": { - "id": "thinkingmachines/inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048000, - "output": 32000 - }, - "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.17 - } - }, - "meta/muse-spark-1.1": { - "id": "meta/muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 262144 - }, - "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 - } - }, - "meta/llama-3.3-70b-instruct": { - "id": "meta/llama-3.3-70b-instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.22, - "output": 0.5, - "cache_read": 0.11 - } - }, - "meta/muse-spark-1.2": { - "id": "meta/muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 262144 - }, - "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 - } - }, - "meta/llama-3.1-8b-instruct": { - "id": "meta/llama-3.1-8b-instruct", - "name": "Llama 3.1 8B", - "description": "Compact open Llama model for lightweight chat, drafting, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 2048 - }, - "cost": { - "input": 0.22, - "output": 0.22 - } - }, - "meta/llama-3.1-70b-instruct": { - "id": "meta/llama-3.1-70b-instruct", - "name": "Llama 3.1 70B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 2048 - }, - "cost": { - "input": 0.99, - "output": 0.99 - } - }, - "bytedance/dola-seed-2.0-code-preview": { - "id": "bytedance/dola-seed-2.0-code-preview", - "name": "Dola Seed 2.0 Code (preview)", - "description": "Preview coding model for repository understanding, refactors, and engineering tasks", - "family": "seed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-03-28", - "last_updated": "2026-03-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 3 - } - }, - "bytedance/dola-seed-2.0-code": { - "id": "bytedance/dola-seed-2.0-code", - "name": "Seed 2.0 Code", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "family": "seed", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.4, - "output": 2.4 - } - }, - "bytedance/dola-seed-2.0-mini": { - "id": "bytedance/dola-seed-2.0-mini", - "name": "Seed 2.0 Mini", - "description": "Low-cost Seed model for general chat, extraction, and lightweight production tasks", - "family": "seed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.4 - } - }, - "bytedance/dola-seed-2.0-lite": { - "id": "bytedance/dola-seed-2.0-lite", - "name": "Seed 2.0 Lite", - "description": "Efficient Seed model for general chat, analysis, and lightweight production tasks", - "family": "seed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-28", - "last_updated": "2026-02-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.25, - "output": 2 - } - }, - "bytedance/dola-seed-2.0-pro": { - "id": "bytedance/dola-seed-2.0-pro", - "name": "Seed 2.0 Pro", - "description": "Higher-capability Seed model for complex chat, analysis, and production tasks", - "family": "seed", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-03-28", - "last_updated": "2026-03-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 3 - } - }, - "google/gemini-flash-latest": { - "id": "google/gemini-flash-latest", - "name": "Gemini Flash Latest", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B It", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.14, - "output": 0.4 - } - }, - "google/gemma-3-27b-it": { - "id": "google/gemma-3-27b-it", - "name": "Gemma 3 27B IT", - "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", - "family": "gemma", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.08, - "output": 0.45, - "cache_read": 0.04 - } - }, - "google/gemini-3-pro-preview": { - "id": "google/gemini-3-pro-preview", - "name": "Gemini 3 Pro Preview", - "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "google/gemini-2.5-flash-image": { - "id": "google/gemini-2.5-flash-image", - "name": "Gemini 2.5 Flash Image", - "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 2.5 - } - }, - "google/gemini-flash-lite-latest": { - "id": "google/gemini-flash-lite-latest", - "name": "Gemini Flash-Lite Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 - } - }, - "google/gemini-3.6-flash": { - "id": "google/gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15 - } - }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.13, - "output": 0.4 - } - }, - "google/gemini-3.5-flash-lite": { - "id": "google/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash-Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 - } - }, - "google/gemini-3.1-flash-image": { - "id": "google/gemini-3.1-flash-image", - "name": "Gemini 3.1 Flash Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 3 - } - }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "google/gemma-3-12b-it": { - "id": "google/gemma-3-12b-it", - "name": "Gemma 3 12B", - "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.09, - "output": 0.29 - } - }, - "google/gemini-2.5-computer-use-preview-10-2025": { - "id": "google/gemini-2.5-computer-use-preview-10-2025", - "name": "Gemini 2.5 Computer Use Preview (10-2025)", - "description": "Specialized Gemini 2.5 model for browser-control agents that automate UI tasks", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-10-07", - "last_updated": "2025-10-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 64000 - }, - "cost": { - "input": 1.25, - "output": 10 - } - }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "google/gemini-2.5-pro": { - "id": "google/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } - } - }, - "google/gemini-3-pro-image": { - "id": "google/gemini-3-pro-image", - "name": "Gemini 3 Pro Image", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12 - } - }, - "google/gemini-2.5-flash": { - "id": "google/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 - } - }, - "google/gemini-3.7-flash": { - "id": "google/gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 - } - }, - "google/gemini-3.8-flash": { - "id": "google/gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "google/gemma-3-4b-it": { - "id": "google/gemma-3-4b-it", - "name": "Gemma 3 4B", - "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.04, - "output": 0.08 - } - }, - "google/gemini-2.5-flash-lite": { - "id": "google/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 - } - }, - "google/gemini-embedding-001": { - "id": "google/gemini-embedding-001", - "name": "Gemini Embedding 001", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "gemini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2025-05", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2048, - "output": 4096 - }, - "cost": { - "input": 0.15, - "output": 0 - } - }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 - } - }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash-Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 - } - }, - "writer/palmyra-x4": { - "id": "writer/palmyra-x4", - "name": "Palmyra X4", - "description": "Enterprise language model for writing, analysis, and tool-assisted workflows", - "family": "palmyra", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2024-10-09", - "last_updated": "2024-10-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 32000 - }, - "cost": { - "input": 2.5, - "output": 10 - } - }, - "writer/palmyra-x5": { - "id": "writer/palmyra-x5", - "name": "Palmyra X5", - "description": "Enterprise multimodal model for writing, analysis, and tool-assisted workflows", - "family": "palmyra", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 250000 - }, - "cost": { - "input": 0.6, - "output": 6 - } - }, - "xai/grok-4.7": { - "id": "xai/grok-4.7", - "name": "Grok 4.7", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "xai/grok-4.3": { - "id": "xai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "xai/grok-4.20-0309-reasoning": { - "id": "xai/grok-4.20-0309-reasoning", - "name": "Grok 4.20", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "xai/grok-4.5": { - "id": "xai/grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "xai/grok-4.20-0309-non-reasoning": { - "id": "xai/grok-4.20-0309-non-reasoning", - "name": "Grok 4.20 Non-Reasoning", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 - } - }, - "xai/grok-build-0.1": { - "id": "xai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 - } - }, - "xai/grok-4.6": { - "id": "xai/grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "sakana/fugu-ultra": { - "id": "sakana/fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - } - }, - "sakana/sakana-namazu": { - "id": "sakana/sakana-namazu", - "name": "Sakana Namazu", - "description": "Japanese-specialized reasoning model based on Kimi K2.6 and tuned for Japanese language, culture, and business workflows", - "family": "sakana-namazu", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.15 - } - }, - "moonshotai/kimi-k2-thinking": { - "id": "moonshotai/kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 32768 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.6, - "output": 2.5 - } - }, - "nvidia/nemotron-3.5-lightning-30b-a3b": { - "id": "nvidia/nemotron-3.5-lightning-30b-a3b", - "name": "Nemotron 3.5 Lightning 30B A3B", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nemotron-nano-9b-v2": { - "id": "nvidia/nemotron-nano-9b-v2", - "name": "Nemotron Nano 9B", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-18", - "last_updated": "2025-08-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.06, - "output": 0.23 - } - }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 8192 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 8192 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 128000 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "minimax/minimax-m2.7-highspeed": { - "id": "minimax/minimax-m2.7-highspeed", - "name": "MiniMax M2.7 Highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 8192 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "minimax/minimax-m2.5-highspeed": { - "id": "minimax/minimax-m2.5-highspeed", - "name": "MiniMax M2.5 Highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 8192 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "minimax/minimax-m2.1": { - "id": "minimax/minimax-m2.1", - "name": "MiniMax M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 8192 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "minimax/minimax-m2": { - "id": "minimax/minimax-m2", - "name": "MiniMax M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 8192 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "mistral/codestral-latest": { - "id": "mistral/codestral-latest", - "name": "Codestral (latest)", - "description": "Mistral code model for completions, refactors, and developer IDE workflows", - "family": "codestral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-05-29", - "last_updated": "2025-01-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 4096 - }, - "cost": { - "input": 0.3, - "output": 0.9 - } - }, - "mistral/mistral-large-2411": { - "id": "mistral/mistral-large-2411", - "name": "Mistral Large 2.1", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-18", - "last_updated": "2024-11-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 2, - "output": 6 - } - }, - "mistral/mistral-large-latest": { - "id": "mistral/mistral-large-latest", - "name": "Mistral Large (latest)", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "mistral/mistral-small-latest": { - "id": "mistral/mistral-small-latest", - "name": "Mistral Small (latest)", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "mistral/devstral-medium-latest": { - "id": "mistral/devstral-medium-latest", - "name": "Devstral 2 (latest)", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2 - } - }, - "mistral/devstral-2512": { - "id": "mistral/devstral-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.04 - } - }, - "mistral/mistral-medium-2505": { - "id": "mistral/mistral-medium-2505", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.04 - } - }, - "mistral/magistral-medium-latest": { - "id": "mistral/magistral-medium-latest", - "name": "Magistral Medium (latest)", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-medium", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2, - "output": 5 - } - }, - "mistral/devstral-medium-2507": { - "id": "mistral/devstral-medium-2507", - "name": "Devstral Medium", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-07-10", - "last_updated": "2025-07-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 128000 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2 - } - }, - "mistral/mistral-large-2512": { - "id": "mistral/mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.05 - } - }, - "mistral/devstral-small-2507": { - "id": "mistral/devstral-small-2507", - "name": "Devstral Small", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-07-10", - "last_updated": "2025-07-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "status": "deprecated", - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "mistral/pixtral-large-latest": { - "id": "mistral/pixtral-large-latest", - "name": "Pixtral Large (latest)", - "description": "Mistral's larger vision model for document-heavy image understanding and chat", - "family": "pixtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2024-11-04", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 6 - } - }, - "mistral/mistral-medium-latest": { - "id": "mistral/mistral-medium-latest", - "name": "Mistral Medium (latest)", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.4, - "output": 2 - } - }, - "qwen/qwen-flash": { - "id": "qwen/qwen-flash", - "name": "Qwen Flash", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 250000 - }, - "cost": { - "input": 0.022, - "output": 0.216, - "cache_read": 0.0044 - } - }, - "qwen/qwen3.5-flash": { - "id": "qwen/qwen3.5-flash", - "name": "Qwen3.5 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 250000 - }, - "cost": { - "input": 0.029, - "output": 0.287, - "cache_read": 0.0058 - } - }, - "qwen/qwen3-vl-235b-a22b-instruct": { - "id": "qwen/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3-VL 235B A22B Instruct", - "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.287, - "output": 1.147, - "cache_read": 0.15785 - } - }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 250000 - }, - "cost": { - "input": 0.825, - "output": 2.4755, - "cache_read": 0.165 - } - }, - "qwen/qwen3-vl-235b-a22b-thinking": { - "id": "qwen/qwen3-vl-235b-a22b-thinking", - "name": "Qwen3-VL 235B A22B Thinking", - "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.287, - "output": 2.867, - "cache_read": 0.0574 - } - }, - "qwen/qwen3.8-2.4t-a95b": { - "id": "qwen/qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 1010000 - }, - "cost": { - "input": 2.5, - "output": 6.25, - "cache_read": 0.5 - } - }, - "qwen/qwen3-vl-plus": { - "id": "qwen/qwen3-vl-plus", - "name": "Qwen3-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.143, - "output": 1.434, - "cache_read": 0.0286 - } - }, - "qwen/qwen3.5-27b": { - "id": "qwen/qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.086, - "output": 0.688, - "cache_read": 0.0172 - } - }, - "qwen/qwen3-235b-a22b-instruct-2507": { - "id": "qwen/qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.6, - "cache_read": 0.05 - } - }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 1.2 - } - }, - "qwen/qwen3.8-max": { - "id": "qwen/qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 - } - }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.15, - "output": 0.8, - "cache_read": 0.075 - } - }, - "qwen/qwen3.5-plus": { - "id": "qwen/qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 250000 - }, - "cost": { - "input": 0.115, - "output": 0.688, - "cache_read": 0.023 - } - }, - "qwen/qwen3-32b": { - "id": "qwen/qwen3-32b", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "qwen/qwen-plus": { - "id": "qwen/qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 250000 - }, - "cost": { - "input": 0.115, - "output": 0.287, - "cache_read": 0.023 - } - }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.359, - "output": 1.434, - "cache_read": 0.0718 - } - }, - "qwen/qwen3-coder-480b-a35b-instruct": { - "id": "qwen/qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.22, - "output": 1.8 - } - }, - "qwen/qwen3-235b-a22b": { - "id": "qwen/qwen3-235b-a22b", - "name": "Qwen3 235B A22B", - "description": "Large open Qwen MoE for multilingual reasoning, coding, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.287, - "output": 1.147, - "cache_read": 0.0574 - } - }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.115, - "output": 0.917, - "cache_read": 0.023 - } - }, - "qwen/qwen3.6-35b-a3b": { - "id": "qwen/qwen3.6-35b-a3b", - "name": "Qwen3.6 35B A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.248, - "output": 1.485, - "cache_read": 0.0496 - } - }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.144, - "output": 0.574, - "cache_read": 0.0288 - } - }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 250000 - }, - "cost": { - "input": 0.165, - "output": 0.99, - "cache_read": 0.033 - } - }, - "qwen/qwen3-coder-flash": { - "id": "qwen/qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 250000 - }, - "cost": { - "input": 0.144, - "output": 0.574, - "cache_read": 0.0288 - } - }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 250000 - }, - "cost": { - "input": 0.574, - "output": 2.294, - "cache_read": 0.1148 - } - }, - "qwen/qwen3.5-9b": { - "id": "qwen/qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.09, - "output": 0.13 - } - }, - "qwen/qwen3.5-35b-a3b": { - "id": "qwen/qwen3.5-35b-a3b", - "name": "Qwen3.5 35B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.057, - "output": 0.459, - "cache_read": 0.020357 - } - }, - "qwen/qwen3.6-max-preview": { - "id": "qwen/qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 65536 - }, - "cost": { - "input": 1.31, - "output": 7.88 - } - }, - "qwen/qwen3-30b-a3b": { - "id": "qwen/qwen3-30b-a3b", - "name": "Qwen3 30B A3B", - "description": "Sparse MoE Qwen model with 3B active parameters for efficient chat and reasoning", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.108, - "output": 1.076, - "cache_read": 0.0216 - } - }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.172, - "output": 1.032, - "cache_read": 0.0344 - } - }, - "qwen/qwen3.6-27b": { - "id": "qwen/qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.289, - "output": 2.4 - } - }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 250000 - }, - "cost": { - "input": 0.276, - "output": 1.651, - "cache_read": 0.0552 - } - }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.4, - "output": 1.6 - } - }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } - } - }, - "openai/gpt-oss-safeguard-120b": { - "id": "openai/gpt-oss-safeguard-120b", - "name": "GPT OSS Safeguard 120B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 4096, - "output": 4096 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16385, - "output": 4096 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 Nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "openai/gpt-4o-2024-05-13": { - "id": "openai/gpt-4o-2024-05-13", - "name": "GPT-4o (2024-05-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 5, - "output": 15 - } - }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4 Mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 - } - }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3 Mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 - } - }, - "openai/gpt-4": { - "id": "openai/gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 30, - "output": 60 - } - }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 Nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 - } - }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 - } - }, - "openai/gpt-5.2-chat-latest": { - "id": "openai/gpt-5.2-chat-latest", - "name": "GPT-5.2 Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "openai/gpt-5.1-chat-latest": { - "id": "openai/gpt-5.1-chat-latest", - "name": "GPT-5.1 Chat Latest", - "description": "Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations", - "family": "gpt-codex", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/o1": { - "id": "openai/o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 - } - }, - "openai/gpt-5-chat-latest": { - "id": "openai/gpt-5-chat-latest", - "name": "GPT-5 Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-4o-2024-08-06": { - "id": "openai/gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-08-06", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "openai/gpt-6-astra": { - "id": "openai/gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o Mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 - } - }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT-OSS 20B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.04, - "output": 0.2, - "cache_read": 0.02 - } - }, - "openai/gpt-oss-safeguard-20b": { - "id": "openai/gpt-oss-safeguard-20b", - "name": "GPT OSS Safeguard 20B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 4096, - "output": 4096 - }, - "cost": { - "input": 0.07, - "output": 0.2, - "cache_read": 0, - "cache_write": 0 - } - }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 Mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 - } - }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "openai/gpt-4o-2024-11-20": { - "id": "openai/gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 Mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 - } - }, - "openai/gpt-6-luna": { - "id": "openai/gpt-6-luna", - "name": "GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125 - } - }, - "openai/gpt-5.3-chat-latest": { - "id": "openai/gpt-5.3-chat-latest", - "name": "GPT-5.3 Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 10, - "output": 30 - } - }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT-OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.09, - "output": 0.36 - } - }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 5 - } - }, - "openai/gpt-6-sol": { - "id": "openai/gpt-6-sol", - "name": "GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "moonshot/kimi-k3": { - "id": "moonshot/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 2.9, - "output": 14, - "cache_read": 0.3 - } - }, - "moonshot/kimi-k2.7-code-highspeed": { - "id": "moonshot/kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 32768 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 - } - }, - "moonshot/kimi-k2.6": { - "id": "moonshot/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 - } - }, - "moonshot/kimi-k2.5": { - "id": "moonshot/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 262144 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 - } - }, - "moonshot/kimi-k2.7-code": { - "id": "moonshot/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1, - "max": 32768 - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - } - } - }, - "opper": { - "id": "opper", - "env": [ - "OPPER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.opper.ai/v3/compat", - "name": "Opper", - "doc": "https://opper.ai/models", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } - } - }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1.1, - "output": 5.5, - "cache_read": 0.11, - "cache_write": 1.375 - } - }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } - } - }, - "gemma-4-31b-it": { - "id": "gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.46488, - "output": 2.44062 - } - }, - "muse-spark-1.3": { - "id": "muse-spark-1.3", - "name": "Muse Spark 1.3", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 - } - }, - "qwen3.8-27b": { - "id": "qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.5811, - "output": 3 - } - }, - "gpt-5.5-pro": { - "id": "gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 180 - } - }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "qwen3.8-2.4t-a95b": { - "id": "qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 2.5, - "output": 6, - "cache_read": 0.63 - } - }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.07 - } - }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 983616, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15 - } - }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.5811, - "output": 2.3244 - } - }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125, - "tiers": [ - { - "input": 6.6, - "output": 24.75, - "cache_read": 0.66, - "cache_write": 8.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6.6, - "output": 24.75, - "cache_read": 0.66, - "cache_write": 8.25 - } - } - }, - "claude-fable-5-1": { - "id": "claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "devstral-2512": { - "id": "devstral-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.4, - "output": 2 - } - }, - "sonar-pro": { - "id": "sonar-pro", - "name": "Sonar Pro", - "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", - "family": "sonar-pro", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8000 - }, - "cost": { - "input": 3, - "output": 15 - } - }, - "grok-4.3": { - "id": "grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 30000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 - } - }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 - } - }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } - } - }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.248, - "output": 1.485 - } - }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.11622, - "output": 0.488124 - } - }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 131072 - }, - "cost": { - "input": 0.69732, - "output": 2.78928 - } - }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "muse-spark-1.2": { - "id": "muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 - } - }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12, - "tiers": [ - { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24, - "tier": { - "type": "context", - "size": 524288 - } - } - ], - "context_over_200k": { - "input": 1.2, - "output": 4.8, - "cache_read": 0.24 - } - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } - } - }, - "mistral-small-2603": { - "id": "mistral-small-2603", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.5811, - "output": 2.44062 - } - }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 2.2, - "output": 11, - "cache_read": 0.22, - "cache_write": 2.75 - } - }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.62708, - "output": 5.811 - } - }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 - } - }, - "grok-build-0.1": { - "id": "grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 - } - }, - "sonar": { - "id": "sonar", - "name": "Sonar", - "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", - "family": "sonar", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 1, - "output": 1 - } - }, - "mistral-large-2512": { - "id": "mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "gemini-3.8-flash": { - "id": "gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.825, - "output": 4.125, - "cache_read": 0.0825 - } - }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 1.78812, - "output": 3.57624 - } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 - } - }, - "gpt-5.3-chat-latest": { - "id": "gpt-5.3-chat-latest", - "name": "GPT-5.3 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "status": "deprecated", - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "sonar-reasoning-pro": { - "id": "sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", - "family": "sonar-reasoning", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 2, - "output": 8 - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } - } - }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 1.1622, - "output": 4.88124 - } - }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.75, - "output": 4.6488, - "cache_read": 0.44 - } - }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "grok-4.6": { - "id": "grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.25, - "output": 0.66 - } - } - } - }, - "nvidia": { - "id": "nvidia", - "env": [ - "NVIDIA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://integrate.api.nvidia.com/v1", - "name": "Nvidia", - "doc": "https://docs.api.nvidia.com/nim/", - "models": { - "baai/bge-m3": { - "id": "baai/bge-m3", - "name": "BGE M3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "bge", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-01-30", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 1024 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "poolside/laguna-xs-2.1": { - "id": "poolside/laguna-xs-2.1", - "name": "Laguna XS 2.1", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-02", - "last_updated": "2026-07-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "abacusai/dracarys-llama-3.1-70b-instruct": { - "id": "abacusai/dracarys-llama-3.1-70b-instruct", - "name": "dracarys-llama-3.1-70b-instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-09-11", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "z-ai/glm-5.3-flash": { - "id": "z-ai/glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "z-ai/glm-5.3": { - "id": "z-ai/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "thinkingmachines/inkling": { - "id": "thinkingmachines/inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "black-forest-labs/flux_1-schnell": { - "id": "black-forest-labs/flux_1-schnell", - "name": "FLUX.1-schnell", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2024-07", - "release_date": "2024-08-01", - "last_updated": "2026-02-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": true, - "limit": { - "context": 77, - "input": 77, - "output": 0 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "black-forest-labs/flux.1-dev": { - "id": "black-forest-labs/flux.1-dev", - "name": "FLUX.1-dev", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2024-08-01", - "last_updated": "2025-09-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 4096, - "output": 0 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "black-forest-labs/flux_1-kontext-dev": { - "id": "black-forest-labs/flux_1-kontext-dev", - "name": "FLUX.1-Kontext-dev", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-08-12", - "last_updated": "2025-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "output": 40960 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "black-forest-labs/flux_2-klein-4b": { - "id": "black-forest-labs/flux_2-klein-4b", - "name": "FLUX.2 Klein 4B", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-01-14", - "last_updated": "2026-01-31", - "modalities": { - "input": [ - "image", - "text" - ], - "output": [ - "image" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "output": 40960 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/llama-guard-4-12b": { - "id": "meta/llama-guard-4-12b", - "name": "Llama Guard 4 12B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-04-05", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/muse-glimmer-30b": { - "id": "meta/muse-glimmer-30b", - "name": "Muse Glimmer 30B", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/esmfold": { - "id": "meta/esmfold", - "name": "esmfold", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-03-15", - "last_updated": "2025-06-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/llama-3.3-70b-instruct": { - "id": "meta/llama-3.3-70b-instruct", - "name": "Llama 3.3 70b Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-11-26", - "last_updated": "2024-11-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/llama-3.2-90b-vision-instruct": { - "id": "meta/llama-3.2-90b-vision-instruct", - "name": "Llama-3.2-90B-Vision-Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/llama-4-maverick-17b-128e-instruct": { - "id": "meta/llama-4-maverick-17b-128e-instruct", - "name": "Llama 4 Maverick 17b 128e Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-02", - "release_date": "2025-04-01", - "last_updated": "2025-04-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/llama-3.1-8b-instruct": { - "id": "meta/llama-3.1-8b-instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 16000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/llama-3.2-11b-vision-instruct": { - "id": "meta/llama-3.2-11b-vision-instruct", - "name": "Llama 3.2 11b Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-18", - "last_updated": "2024-09-18", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/llama-3.2-1b-instruct": { - "id": "meta/llama-3.2-1b-instruct", - "name": "Llama 3.2 1b Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-18", - "last_updated": "2024-09-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/llama-3.2-3b-instruct": { - "id": "meta/llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2024-09-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/llama-3.1-70b-instruct": { - "id": "meta/llama-3.1-70b-instruct", - "name": "Llama 3.1 70b Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-07-16", - "last_updated": "2024-07-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "meta/esm2-650m": { - "id": "meta/esm2-650m", - "name": "esm2-650m", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-08-29", - "last_updated": "2025-03-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "upstage/solar-10.7b-instruct": { - "id": "upstage/solar-10.7b-instruct", - "name": "solar-10.7b-instruct", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-06-05", - "last_updated": "2025-04-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "bytedance/seed-oss-36b-instruct": { - "id": "bytedance/seed-oss-36b-instruct", - "name": "ByteDance-Seed/Seed-OSS-36B-Instruct", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "seed", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-04", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma-4-31B-IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "google/gemma-2-2b-it": { - "id": "google/gemma-2-2b-it", - "name": "Gemma 2 2b It", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-07-16", - "last_updated": "2024-07-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "google/gemma-3-12b-it": { - "id": "google/gemma-3-12b-it", - "name": "Gemma 3 12B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "google/gemma-3n-e2b-it": { - "id": "google/gemma-3n-e2b-it", - "name": "Gemma 3n E2b It", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-06-12", - "last_updated": "2025-06-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "google/google-paligemma": { - "id": "google/google-paligemma", - "name": "paligemma", - "description": "Gemini multimodal model for text, image, audio, video, and document tasks", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-05-14", - "last_updated": "2024-08-26", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "google/gemma-3n-e4b-it": { - "id": "google/gemma-3n-e4b-it", - "name": "Gemma 3n E4b It", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-06-03", - "last_updated": "2025-06-03", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "google/gemma-3-4b-it": { - "id": "google/gemma-3-4b-it", - "name": "Gemma 3 4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "minimaxai/minimax-m2.7": { - "id": "minimaxai/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-04-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "minimaxai/minimax-m3": { - "id": "minimaxai/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "deepseek-ai/deepseek-v4-flash-0731": { - "id": "deepseek-ai/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "deepseek-ai/deepseek-v4-pro-0813": { - "id": "deepseek-ai/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "deepseek-ai/deepseek-v4-pro": { - "id": "deepseek-ai/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 393216 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 - } - }, - "deepseek-ai/deepseek-v4-flash": { - "id": "deepseek-ai/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 393216 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 - } - }, - "stepfun-ai/step-3.5-flash": { - "id": "stepfun-ai/step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-02", - "last_updated": "2026-02-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "stepfun-ai/step-3.7-flash": { - "id": "stepfun-ai/step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistralai/mistral-nemotron": { - "id": "mistralai/mistral-nemotron", - "name": "mistral-nemotron", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-06-11", - "last_updated": "2025-06-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistralai/mistral-7b-instruct-v0.3": { - "id": "mistralai/mistral-7b-instruct-v0.3", - "name": "Mistral-7B-Instruct-v0.3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-01", - "last_updated": "2025-04-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 65536, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistralai/mistral-large-3-675b-instruct-2512": { - "id": "mistralai/mistral-large-3-675b-instruct-2512", - "name": "Mistral Large 3 675B Instruct 2512", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistralai/magistral-small-2506": { - "id": "mistralai/magistral-small-2506", - "name": "Magistral Small 2506", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "input": 32768, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistralai/mistral-small-4-119b-2603": { - "id": "mistralai/mistral-small-4-119b-2603", - "name": "mistral-small-4-119b-2603", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistralai/mixtral-8x7b-instruct": { - "id": "mistralai/mixtral-8x7b-instruct", - "name": "Mistral: Mixtral 8x7B Instruct", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2023-12-10", - "last_updated": "2026-03-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistralai/mistral-medium-3.5-128b": { - "id": "mistralai/mistral-medium-3.5-128b", - "name": "Mistral Medium 3.5", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistralai/ministral-14b-instruct-2512": { - "id": "mistralai/ministral-14b-instruct-2512", - "name": "Ministral 3 14B Instruct 2512", - "description": "Compact Mistral VLM for chat and instruction-based workloads", - "family": "ministral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistralai/mixtral-8x22b-instruct": { - "id": "mistralai/mixtral-8x22b-instruct", - "name": "Mistral: Mixtral 8x22B Instruct", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-04-17", - "last_updated": "2024-04-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 65536, - "output": 13108 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistralai/mistral-medium-3-instruct": { - "id": "mistralai/mistral-medium-3-instruct", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-09-25", - "last_updated": "2025-09-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "input": 131072, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0 - } - }, - "moonshotai/kimi-k2-instruct-0905": { - "id": "moonshotai/kimi-k2-instruct-0905", - "name": "Kimi K2 0905", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0 - } - }, - "sarvamai/sarvam-m": { - "id": "sarvamai/sarvam-m", - "name": "sarvam-m", - "description": "Efficient Indian-language reasoning model for chat, coding, and multilingual work", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-07-25", - "last_updated": "2025-07-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nvidia-nemotron-nano-9b-v2": { - "id": "nvidia/nvidia-nemotron-nano-9b-v2", - "name": "nvidia-nemotron-nano-9b-v2", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-09", - "release_date": "2025-08-18", - "last_updated": "2025-08-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/llama-3.1-nemotron-ultra-253b-v1": { - "id": "nvidia/llama-3.1-nemotron-ultra-253b-v1", - "name": "Llama 3.1 Nemotron Ultra 253B", - "description": "Flagship Nemotron model for high-throughput reasoning and complex agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-04-07", - "last_updated": "2025-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/sparsedrive": { - "id": "nvidia/sparsedrive", - "name": "sparsedrive", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-03-18", - "last_updated": "2025-07-20", - "modalities": { - "input": [ - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nemotron-3-nano-30b-a3b": { - "id": "nvidia/nemotron-3-nano-30b-a3b", - "name": "nemotron-3-nano-30b-a3b", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-12", - "last_updated": "2024-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nemotron-3.5-lightning-30b-a3b": { - "id": "nvidia/nemotron-3.5-lightning-30b-a3b", - "name": "Nemotron 3.5 Lightning 30B A3B", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/llama-nemotron-embed-vl-1b-v2": { - "id": "nvidia/llama-nemotron-embed-vl-1b-v2", - "name": "llama-nemotron-embed-vl-1b-v2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "nemotron", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-02-10", - "last_updated": "2026-02-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 2048 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nv-embedcode-7b-v1": { - "id": "nvidia/nv-embedcode-7b-v1", - "name": "nv-embedcode-7b-v1", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-03-17", - "last_updated": "2025-05-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 2048 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nemotron-voicechat": { - "id": "nvidia/nemotron-voicechat", - "name": "nemotron-voicechat", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/bevformer": { - "id": "nvidia/bevformer", - "name": "bevformer", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-03-18", - "last_updated": "2025-07-20", - "modalities": { - "input": [ - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/cosmos-transfer1-7b": { - "id": "nvidia/cosmos-transfer1-7b", - "name": "cosmos-transfer1-7b", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-06-13", - "last_updated": "2025-06-30", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "video" - ] - }, - "open_weights": true, - "limit": { - "context": 0, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/cosmos-predict1-5b": { - "id": "nvidia/cosmos-predict1-5b", - "name": "cosmos-predict1-5b", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-03-18", - "last_updated": "2025-03-18", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "video" - ] - }, - "open_weights": true, - "limit": { - "context": 0, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/llama-3.1-nemotron-nano-vl-8b-v1": { - "id": "nvidia/llama-3.1-nemotron-nano-vl-8b-v1", - "name": "Llama 3.1 Nemotron Nano VL 8B v1", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-04-10", - "last_updated": "2025-04-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nemotron-3-super-120b-a12b": { - "id": "nvidia/nemotron-3-super-120b-a12b", - "name": "Nemotron 3 Super", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2026-03-11", - "last_updated": "2026-03-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.2, - "output": 0.8 - } - }, - "nvidia/studiovoice": { - "id": "nvidia/studiovoice", - "name": "studiovoice", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-10-03", - "last_updated": "2025-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nemotron-content-safety-reasoning-4b": { - "id": "nvidia/nemotron-content-safety-reasoning-4b", - "name": "nemotron-content-safety-reasoning-4b", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": false, - "release_date": "2026-01-22", - "last_updated": "2026-01-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/llama-3.1-nemotron-safety-guard-8b-v3": { - "id": "nvidia/llama-3.1-nemotron-safety-guard-8b-v3", - "name": "llama-3.1-nemotron-safety-guard-8b-v3", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nv-embed-v1": { - "id": "nvidia/nv-embed-v1", - "name": "nv-embed-v1", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-06-07", - "last_updated": "2025-07-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 2048 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/synthetic-video-detector": { - "id": "nvidia/synthetic-video-detector", - "name": "synthetic-video-detector", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 0, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/cosmos-transfer2_5-2b": { - "id": "nvidia/cosmos-transfer2_5-2b", - "name": "cosmos-transfer2.5-2b", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-02-26", - "last_updated": "2026-02-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "video" - ] - }, - "open_weights": true, - "limit": { - "context": 0, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nemotron-3-ultra-550b-a55b": { - "id": "nvidia/nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.15 - } - }, - "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning": { - "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning", - "name": "Nemotron 3 Nano Omni", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": -1, - "max": 32768 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/llama-3.1-nemotron-nano-8b-v1": { - "id": "nvidia/llama-3.1-nemotron-nano-8b-v1", - "name": "Llama 3.1 Nemotron Nano 8B v1", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-03-18", - "last_updated": "2025-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/active-speaker-detection": { - "id": "nvidia/active-speaker-detection", - "name": "Active Speaker Detection", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 0, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/streampetr": { - "id": "nvidia/streampetr", - "name": "streampetr", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/llama-nemotron-rerank-vl-1b-v2": { - "id": "nvidia/llama-nemotron-rerank-vl-1b-v2", - "name": "llama-nemotron-rerank-vl-1b-v2", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "nemotron", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-03-31", - "last_updated": "2026-03-31", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nemotron-mini-4b-instruct": { - "id": "nvidia/nemotron-mini-4b-instruct", - "name": "nemotron-mini-4b-instruct", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-08-21", - "last_updated": "2024-08-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/cosmos-reason2-8b": { - "id": "nvidia/cosmos-reason2-8b", - "name": "Cosmos Reason2 8B", - "description": "Vision language model for physical-world understanding with structured reasoning on video and images", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-12-01", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nemotron-3-content-safety": { - "id": "nvidia/nemotron-3-content-safety", - "name": "nemotron-3-content-safety", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/llama-3.3-nemotron-super-49b-v1": { - "id": "nvidia/llama-3.3-nemotron-super-49b-v1", - "name": "Llama 3.3 Nemotron Super 49B v1", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-04-07", - "last_updated": "2025-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/magpie-tts-zeroshot": { - "id": "nvidia/magpie-tts-zeroshot", - "name": "magpie-tts-zeroshot", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-05-22", - "last_updated": "2025-06-12", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "audio" - ] - }, - "open_weights": true, - "limit": { - "context": 0, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/riva-translate-4b-instruct-v1.1": { - "id": "nvidia/riva-translate-4b-instruct-v1.1", - "name": "riva-translate-4b-instruct-v1_1", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-12-12", - "last_updated": "2025-12-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/nemotron-nano-12b-v2-vl": { - "id": "nvidia/nemotron-nano-12b-v2-vl", - "name": "Nemotron Nano 12B v2 VL", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/llama-3.3-nemotron-super-49b-v1.5": { - "id": "nvidia/llama-3.3-nemotron-super-49b-v1.5", - "name": "Llama 3.3 Nemotron Super 49B v1.5", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-07-25", - "last_updated": "2025-07-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/llama-3_2-nemoretriever-300m-embed-v1": { - "id": "nvidia/llama-3_2-nemoretriever-300m-embed-v1", - "name": "llama-3_2-nemoretriever-300m-embed-v1", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-07-24", - "last_updated": "2025-07-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 2048 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/gliner-pii": { - "id": "nvidia/gliner-pii", - "name": "gliner-pii", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/usdcode": { - "id": "nvidia/usdcode", - "name": "usdcode", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-01", - "last_updated": "2026-01-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/usdvalidate": { - "id": "nvidia/usdvalidate", - "name": "usdvalidate", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-07-24", - "last_updated": "2025-01-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 0, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/llama-3.1-nemotron-70b-instruct": { - "id": "nvidia/llama-3.1-nemotron-70b-instruct", - "name": "Llama 3.1 Nemotron 70B Instruct", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nvidia/rerank-qa-mistral-4b": { - "id": "nvidia/rerank-qa-mistral-4b", - "name": "rerank-qa-mistral-4b", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-03-17", - "last_updated": "2025-01-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen/qwen-image-edit": { - "id": "qwen/qwen-image-edit", - "name": "Qwen Image Edit", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-19", - "last_updated": "2025-08-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen/qwen3-coder-480b-a35b-instruct": { - "id": "qwen/qwen3-coder-480b-a35b-instruct", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 66536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next-80B-A3B-Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-01", - "last_updated": "2025-09-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5-397B-A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen/qwen2.5-coder-32b-instruct": { - "id": "qwen/qwen2.5-coder-32b-instruct", - "name": "Qwen2.5 Coder 32b Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-11-06", - "last_updated": "2024-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen/qwen-image": { - "id": "qwen/qwen-image", - "name": "Qwen Image", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "openai/whisper-large-v3": { - "id": "openai/whisper-large-v3", - "name": "Whisper Large v3", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2023-09-01", - "last_updated": "2025-09-05", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 0, - "output": 4096 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT-OSS-120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-04", - "last_updated": "2025-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "microsoft/phi-4-mini-instruct": { - "id": "microsoft/phi-4-mini-instruct", - "name": "Phi-4-Mini", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "phi", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-01", - "last_updated": "2025-09-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "microsoft/phi-4-multimodal-instruct": { - "id": "microsoft/phi-4-multimodal-instruct", - "name": "Phi 4 Multimodal", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "input": 128000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - } - } - }, - "pioneer": { - "id": "pioneer", - "env": [ - "PIONEER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.pioneer.ai/v1", - "name": "Pioneer", - "doc": "https://agent.pioneer.ai/llms.txt", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 2.5 - } - }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 991000, - "output": 64000 - }, - "cost": { - "input": 1.25, - "output": 3.75, - "cache_read": 0.25, - "cache_write": 1.5625 - } - }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02, - "cache_write": 0.2 - } - }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25, - "cache_write": 2.5 - } - }, - "ministral-3b": { - "id": "ministral-3b", - "name": "Ministral 3B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-10-16", - "last_updated": "2024-10-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.1, - "cache_write": 0.1 - } - }, - "magistral-medium": { - "id": "magistral-medium", - "name": "Magistral Medium (latest)", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-medium", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 64000 - }, - "cost": { - "input": 2, - "output": 5, - "cache_read": 2, - "cache_write": 2 - } - }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025, - "cache_write": 0.25 - } - }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "claude-3-7-sonnet-latest": { - "id": "claude-3-7-sonnet-latest", - "name": "Claude Sonnet 3.7", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175, - "cache_write": 1.75 - } - }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.05, - "cache_write": 0.1 - } - }, - "mistral-medium-3.5": { - "id": "mistral-medium-3.5", - "name": "Mistral Medium 3.5", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 131072 - }, - "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 1.5, - "cache_write": 1.5 - } - }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005, - "cache_write": 0.05 - } - }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "gemini-3.6-flash": { - "id": "gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15, - "cache_write": 1.5 - } - }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65000 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.3 - } - }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "cache_write": 2 - } - }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 131072 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "cache_write": 1.25 - } - }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075, - "cache_write": 0.15 - } - }, - "gemini-3-flash": { - "id": "gemini-3-flash", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65000 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.083333 - } - }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 0.083333 - } - }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.1875, - "output": 1.125, - "cache_read": 0.0375, - "cache_write": 0.234375 - } - }, - "mistral-large-3": { - "id": "mistral-large-3", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 131072 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.5, - "cache_write": 0.5 - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 11, - "output": 55, - "cache_read": 1.1, - "cache_write": 13.75 - } - }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075, - "cache_write": 0.75 - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "devstral-small-2": { - "id": "devstral-small-2", - "name": "Devstral Small 2", - "description": "Compact multimodal coding model for repository exploration, file editing, and software agents", - "family": "devstral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 131072 - }, - "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.1, - "cache_write": 0.1 - } - }, - "devstral-2": { - "id": "devstral-2", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 131072 - }, - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.4, - "cache_write": 0.4 - } - }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 5 - } - }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 1, - "cache_write": 2 - } - }, - "mistral-medium": { - "id": "mistral-medium", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 64000 - }, - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.4, - "cache_write": 0.4 - } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "qwen3.6-max-preview": { - "id": "qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 240000, - "output": 64000 - }, - "cost": { - "input": 1.04, - "output": 6.24, - "cache_read": 0.208, - "cache_write": 1.3 - } - }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.2, - "cache_write": 0.4 - } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "claude-opus-5-fast": { - "id": "claude-opus-5-fast", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.325, - "output": 1.95, - "cache_read": 0.065, - "cache_write": 0.40625 - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125 - } - }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "ministral-14b": { - "id": "ministral-14b", - "name": "Ministral 14B", - "description": "Compact multimodal Mistral model for local assistants, edge agents, and efficient tool use", - "family": "ministral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 131072 - }, - "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.2, - "cache_write": 0.2 - } - }, - "gemini-3.1-pro": { - "id": "gemini-3.1-pro", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 0.375 - } - }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.32, - "output": 1.28, - "cache_read": 0.064, - "cache_write": 0.4 - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65000 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.03, - "cache_write": 0.25 - } - }, - "poolside/laguna-s-2.1": { - "id": "poolside/laguna-s-2.1", - "name": "Laguna S 2.1", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.01, - "cache_write": 0.1 - } - }, - "meta-llama/Llama-3.2-1B-Instruct": { - "id": "meta-llama/Llama-3.2-1B-Instruct", - "name": "Llama 3.2 1B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-08-31", - "last_updated": "2024-09-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 60000 - }, - "cost": { - "input": 0.1, - "output": 0.201, - "cache_read": 0.1, - "cache_write": 0.1 - } - }, - "meta-llama/Llama-3.2-3B-Instruct": { - "id": "meta-llama/Llama-3.2-3B-Instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-08-31", - "last_updated": "2024-09-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 80000 - }, - "cost": { - "input": 0.1, - "output": 0.335, - "cache_read": 0.1, - "cache_write": 0.1 - } - }, - "meta-llama/Llama-3.1-8B-Instruct": { - "id": "meta-llama/Llama-3.1-8B-Instruct", - "name": "Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2023-12-31", - "release_date": "2024-06-30", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.2, - "cache_write": 0.2 - } - }, - "meta-llama/Llama-3.2-3B": { - "id": "meta-llama/Llama-3.2-3B", - "name": "Llama-3.2-3B", - "description": "Small open Llama base model for lightweight text generation and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.1, - "cache_write": 0.1 - } - }, - "meta-llama/Llama-3.2-1B": { - "id": "meta-llama/Llama-3.2-1B", - "name": "Llama-3.2-1B", - "description": "Compact open Llama base model for lightweight and on-device use", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.1, - "cache_write": 0.1 - } - }, - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 16384, - "output": 16384 - }, - "cost": { - "input": 0.9, - "output": 0.9, - "cache_read": 0.9, - "cache_write": 0.9 - } - }, - "XiaomiMiMo/MiMo-V2.5-Pro": { - "id": "XiaomiMiMo/MiMo-V2.5-Pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1050000, - "output": 131000 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036, - "cache_write": 0.435 - } - }, - "XiaomiMiMo/MiMo-V2.5": { - "id": "XiaomiMiMo/MiMo-V2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1050000, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028, - "cache_write": 0.14 - } - }, - "thinkingmachines/inkling-small": { - "id": "thinkingmachines/inkling-small", - "name": "Inkling Small", - "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.5, - "output": 1.2, - "cache_read": 0.1, - "cache_write": 0.5 - } - }, - "meta/muse-spark-1.1": { - "id": "meta/muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15, - "cache_write": 1.25 - } - }, - "HuggingFaceTB/SmolLM3-3B-Base": { - "id": "HuggingFaceTB/SmolLM3-3B-Base", - "name": "SmolLM3 3B Base", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "google/diffusiongemma-26B-A4B-it": { - "id": "google/diffusiongemma-26B-A4B-it", - "name": "DiffusionGemma 26B-A4B IT", - "description": "Gemini model for general assistance, reasoning, and multimodal workflows", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-05-31", - "last_updated": "2026-05-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.5, - "output": 0.5, - "cache_read": 0.5, - "cache_write": 0.5 - } - }, - "google/gemma-4-12B-it": { - "id": "google/gemma-4-12B-it", - "name": "Gemma 4 12B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-05-31", - "last_updated": "2026-05-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.25, - "output": 0.25, - "cache_read": 0.25, - "cache_write": 0.25 - } - }, - "google/gemma-4-E2B-it": { - "id": "google/gemma-4-E2B-it", - "name": "Gemma 4 E2B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.1, - "cache_write": 0.1 - } - }, - "google/gemma-3-4b-pt": { - "id": "google/gemma-3-4b-pt", - "name": "Gemma 3 4B (Pretrained)", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-02-28", - "last_updated": "2025-02-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 0.5, - "cache_read": 0.5, - "cache_write": 0.5 - } - }, - "google/gemma-4-E4B-it": { - "id": "google/gemma-4-E4B-it", - "name": "Gemma 4 E4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.2, - "cache_write": 0.2 - } - }, - "Qwen/Qwen3-4B-Instruct-2507": { - "id": "Qwen/Qwen3-4B-Instruct-2507", - "name": "Qwen3 4B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-07-31", - "last_updated": "2025-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.2, - "cache_write": 0.2 - } - }, - "Qwen/Qwen2.5-Coder-0.5B": { - "id": "Qwen/Qwen2.5-Coder-0.5B", - "name": "Qwen2.5-Coder-0.5B", - "description": "Tiny open Qwen code model for lightweight completion and on-device coding", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-11-12", - "last_updated": "2024-11-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.1, - "cache_write": 0.1 - } - }, - "Qwen/Qwen3-1.7B-Base": { - "id": "Qwen/Qwen3-1.7B-Base", - "name": "Qwen3 1.7B Base", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-03-31", - "last_updated": "2025-03-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.1, - "cache_write": 0.1 - } - }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 1, - "cache_read": 0.028, - "cache_write": 0.175 - } - }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B-A22B Instruct 2507", - "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 1.2, - "cache_read": 1.2, - "cache_write": 1.2 - } - }, - "Qwen/Qwen3-8B": { - "id": "Qwen/Qwen3-8B", - "name": "Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-03-31", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "output": 40960 - }, - "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.2, - "cache_write": 0.2 - } - }, - "Qwen/Qwen3-4B-Base": { - "id": "Qwen/Qwen3-4B-Base", - "name": "Qwen3 4B Base", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-03-31", - "last_updated": "2025-03-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "Qwen/Qwen3-32B": { - "id": "Qwen/Qwen3-32B", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.9, - "output": 0.9, - "cache_read": 0.9, - "cache_write": 0.9 - } - }, - "Qwen/Qwen3.5-9B": { - "id": "Qwen/Qwen3.5-9B", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 0.3, - "cache_read": 0.3, - "cache_write": 0.3 - } - }, - "Qwen/Qwen3.6-27B": { - "id": "Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.6, - "output": 0.6, - "cache_read": 0.6, - "cache_write": 0.6 - } - }, - "LiquidAI/LFM2-24B-A2B": { - "id": "LiquidAI/LFM2-24B-A2B", - "name": "LFM2 24B A2B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "liquid", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-01-31", - "last_updated": "2026-02-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.03, - "output": 0.12, - "cache_read": 0.03, - "cache_write": 0.03 - } - }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek-V3.1", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 131072 - }, - "cost": { - "input": 0.56, - "output": 1.68, - "cache_read": 0.56, - "cache_write": 0.56 - } - }, - "deepseek-ai/DeepSeek-V4-Flash": { - "id": "deepseek-ai/DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.0197, - "cache_write": 0.1 - } - }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625, - "cache_write": 0.435 - } - }, - "deepseek-ai/DeepSeek-V3": { - "id": "deepseek-ai/DeepSeek-V3", - "name": "DeepSeek-V3", - "description": "Open DeepSeek MoE chat model for coding, math, and general reasoning", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2024-12-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 8192 - }, - "cost": { - "input": 0.27, - "output": 1.12, - "cache_read": 0.135, - "cache_write": 0.27 - } - }, - "MiniMaxAI/MiniMax-M3": { - "id": "MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.3 - } - }, - "MiniMaxAI/MiniMax-M2.7": { - "id": "MiniMaxAI/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131000 - }, - "cost": { - "input": 0.279, - "output": 1.2, - "cache_read": 0.279, - "cache_write": 0.279 - } - }, - "mistralai/Mistral-Small-4-119B-2603": { - "id": "mistralai/Mistral-Small-4-119B-2603", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015, - "cache_write": 0.15 - } - }, - "mistralai/Pixtral-12B-2409": { - "id": "mistralai/Pixtral-12B-2409", - "name": "Pixtral 12B", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-09-01", - "last_updated": "2024-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "mistralai/Mistral-Nemo-Instruct-2407": { - "id": "mistralai/Mistral-Nemo-Instruct-2407", - "name": "Mistral Nemo", - "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.02, - "output": 0.03, - "cache_read": 0.02, - "cache_write": 0.02 - } - }, - "mistralai/Magistral-Small-2506": { - "id": "mistralai/Magistral-Small-2506", - "name": "Magistral Small", - "description": "Open Mistral reasoning model for transparent step-by-step problem solving", - "family": "magistral", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-06-10", - "last_updated": "2025-06-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 64000 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.5, - "cache_write": 0.5 - } - }, - "mistralai/Ministral-8B-Instruct-2410": { - "id": "mistralai/Ministral-8B-Instruct-2410", - "name": "Ministral 8B Instruct", - "description": "Efficient open Mistral edge model for on-device chat and function calling", - "family": "ministral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-10-16", - "last_updated": "2024-10-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "mistralai/Codestral-22B-v0.1": { - "id": "mistralai/Codestral-22B-v0.1", - "name": "Codestral-22B-v0.1", - "description": "Open Mistral code model for fill-in-the-middle and 80+ programming languages", - "family": "codestral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-05-29", - "last_updated": "2024-05-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.3, - "cache_write": 0.3 - } - }, - "mistralai/Mistral-7B-Instruct-v0.3": { - "id": "mistralai/Mistral-7B-Instruct-v0.3", - "name": "Mistral 7B Instruct v0.3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2023-04-30", - "last_updated": "2023-04-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.2, - "output": 0.2, - "cache_read": 0.2, - "cache_write": 0.2 - } - }, - "sakana/fugu-ultra": { - "id": "sakana/fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 5 - } - }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 131072 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.34, - "cache_write": 0.95 - } - }, - "moonshotai/Kimi-K3": { - "id": "moonshotai/Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3 - } - }, - "moonshotai/Kimi-K3-Fast": { - "id": "moonshotai/Kimi-K3-Fast", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 4.5, - "output": 22.5, - "cache_read": 0.45, - "cache_write": 4.5 - } - }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 32768 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19, - "cache_write": 0.95 - } - }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202000, - "output": 131072 - }, - "cost": { - "input": 0.98, - "output": 3.08, - "cache_read": 0.182, - "cache_write": 0.98 - } - }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1040000, - "output": 128000 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 1.4 - } - }, - "zai-org/GLM-5.2-Fast": { - "id": "zai-org/GLM-5.2-Fast", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2.1, - "output": 6.6, - "cache_read": 0.21, - "cache_write": 2.1 - } - }, - "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16": { - "id": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.05, - "cache_write": 0.05 - } - }, - "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16": { - "id": "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16", - "name": "Nemotron 3.5 Lightning 30B A3B", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 4096 - }, - "cost": { - "input": 0.5, - "output": 0.5, - "cache_read": 0.5, - "cache_write": 0.5 - } - }, - "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16": { - "id": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 65000 - }, - "cost": { - "input": 0.5, - "output": 2.5, - "cache_read": 0.15, - "cache_write": 0.5 - } - }, - "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8": { - "id": "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8", - "name": "Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 32000 - }, - "cost": { - "input": 0.09, - "output": 0.45, - "cache_read": 0.09, - "cache_write": 0.09 - } - }, - "pioneer/auto": { - "id": "pioneer/auto", - "name": "Pioneer Auto", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2025-06-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 4096 - } - }, - "fastino/gliner2-multi-v1": { - "id": "fastino/gliner2-multi-v1", - "name": "GLiNER2 Multi", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-11-30", - "last_updated": "2025-11-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "fastino/gliner2-privacy-filter-PII-multi": { - "id": "fastino/gliner2-privacy-filter-PII-multi", - "name": "GLiNER2 Privacy Filter PII (Multi)", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "fastino/gliner2.5-multi-v1": { - "id": "fastino/gliner2.5-multi-v1", - "name": "GLiNER 2.5 Multi", - "description": "Multilingual boundary NER and span extraction; non-trainable encoder.", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-08-24", - "last_updated": "2026-08-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 4096, - "output": 4096 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "fastino/gliguard-LLMGuardrails-300M": { - "id": "fastino/gliguard-LLMGuardrails-300M", - "name": "GLiGuard LLM Guardrails 300M", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "fastino/gliner2-large-v1": { - "id": "fastino/gliner2-large-v1", - "name": "GLiNER2 Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "fastino/gliner2-base-v1": { - "id": "fastino/gliner2-base-v1", - "name": "GLiNER2 Base", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "fastino/gliguard-PII-multi": { - "id": "fastino/gliguard-PII-multi", - "name": "GLiNER2-Guardrails-PII-Multi", - "description": "A 300M-parameter multilingual model that runs LLM safety moderation and PII detection in a single forward pass.", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "fastino/gliner2-multi-large-v1": { - "id": "fastino/gliner2-multi-large-v1", - "name": "GLiNER2 Multi Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-11-30", - "last_updated": "2025-11-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.15, - "cache_read": 0.15, - "cache_write": 0.15 - } - }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.07, - "output": 0.3, - "cache_read": 0.035, - "cache_write": 0.07 - } - }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015, - "cache_write": 0.15 - } - } - } - }, - "xiaomi": { - "id": "xiaomi", - "env": [ - "XIAOMI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.xiaomimimo.com/v1", - "name": "Xiaomi", - "doc": "https://platform.xiaomimimo.com/#/docs", - "models": { - "mimo-v2.6-pro": { - "id": "mimo-v2.6-pro", - "name": "MiMo-V2.6-Pro", - "description": "MiMo Pro model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 - } - }, - "mimo-v2-flash": { - "id": "mimo-v2-flash", - "name": "MiMo-V2-Flash", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12-01", - "release_date": "2025-12-16", - "last_updated": "2026-06-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "status": "deprecated", - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 - } - }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-06-24", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 - } - }, - "mimo-v2-omni": { - "id": "mimo-v2-omni", - "name": "MiMo-V2-Omni", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-06-24", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 - } - }, - "mimo-v2.6-pro-ultraspeed": { - "id": "mimo-v2.6-pro-ultraspeed", - "name": "MiMo-V2.6-Pro-UltraSpeed", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-09-21", - "last_updated": "2026-09-21", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 4.35, - "output": 8.7, - "cache_read": 0.036 - } - }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-06-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 - } - }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-06-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 - } - }, - "mimo-v2.6-flash": { - "id": "mimo-v2.6-flash", - "name": "MiMo-V2.6-Flash", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 - } - }, - "mimo-v2.5-pro-ultraspeed": { - "id": "mimo-v2.5-pro-ultraspeed", - "name": "MiMo-V2.5-Pro-UltraSpeed", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-06-08", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "status": "beta", - "cost": { - "input": 1.305, - "output": 2.61, - "cache_read": 0.0108 - } - } - } - }, - "xiaomi-token-plan-sgp": { - "id": "xiaomi-token-plan-sgp", - "env": [ - "XIAOMI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://token-plan-sgp.xiaomimimo.com/v1", - "name": "Xiaomi Token Plan (Singapore)", - "doc": "https://platform.xiaomimimo.com/#/docs", - "models": { - "mimo-v2.6-pro": { - "id": "mimo-v2.6-pro", - "name": "MiMo-V2.6-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "mimo-v2.6-flash": { - "id": "mimo-v2.6-flash", - "name": "MiMo-V2.6-Flash", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "mimo-v2.5-tts-voiceclone": { - "id": "mimo-v2.5-tts-voiceclone", - "name": "MiMo-V2.5-TTS-VoiceClone", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "mimo-v2-tts": { - "id": "mimo-v2-tts", - "name": "MiMo-V2-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo-V2-Pro", - "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "mimo-v2.5-tts": { - "id": "mimo-v2.5-tts", - "name": "MiMo-V2.5-TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mimo-v2.5-tts-voicedesign": { - "id": "mimo-v2.5-tts-voicedesign", - "name": "MiMo-V2.5-TTS-VoiceDesign", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "mimo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - } - } - }, - "minimax": { - "id": "minimax", - "env": [ - "MINIMAX_API_KEY" - ], - "npm": "@ai-sdk/anthropic", - "api": "https://api.minimax.io/anthropic/v1", - "name": "MiniMax (minimax.io)", - "doc": "https://platform.minimax.io/docs/guides/quickstart", - "models": { - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-25", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 512000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "tiers": [ - { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12, - "tier": { - "type": "context", - "size": 512000 - } - } - ], - "context_over_200k": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 - } - } - }, - "MiniMax-M2.1": { - "id": "MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "MiniMax-M2.7": { - "id": "MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "MiniMax-M2.7-highspeed": { - "id": "MiniMax-M2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "MiniMax-M2": { - "id": "MiniMax-M2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "MiniMax-M2.5-highspeed": { - "id": "MiniMax-M2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 - } - } - } - }, - "github-copilot": { - "id": "github-copilot", - "env": [ - "GITHUB_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.githubcopilot.com", - "name": "GitHub Copilot", - "doc": "https://docs.github.com/en/copilot", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } - } - }, - "grok-4.7": { - "id": "grok-4.7", - "name": "Grok 4.7", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "input": 372000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } - } - }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "claude-opus-4.7": { - "id": "claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "input": 168000, - "output": 32000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 30, - "output": 150, - "cache_read": 3, - "cache_write": 37.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 264000, - "input": 128000, - "output": 64000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "claude-opus-4.8": { - "id": "claude-opus-4.8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "input": 168000, - "output": 64000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - }, - "provider": { - "body": { - "speed": "fast" - }, - "headers": { - "anthropic-beta": "fast-mode-2026-02-01" - } - } - } - } - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "gemini-3.6-flash": { - "id": "gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 256, - "max": 32000 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 936000, - "output": 64000 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "mai-code-1-flash-picker": { - "id": "mai-code-1-flash-picker", - "name": "MAI-Code-1-Flash", - "description": "Microsoft coding model built for fast, efficient assistance in everyday developer workflows", - "family": "mai", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-06-02", - "last_updated": "2026-06-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "input": 128000, - "output": 128000 - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "mai-code-1.1-flash": { - "id": "mai-code-1.1-flash", - "name": "MAI-Code-1.1-Flash", - "description": "Microsoft coding model with native vision support, optimized for fast and efficient software development", - "family": "mai", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "input": 128000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02 - } - }, - "claude-haiku-4.5": { - "id": "claude-haiku-4.5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 32000 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "input": 136000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } - } - }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "input": 372000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } - } - }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 936000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 256, - "max": 24000 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "input": 128000, - "output": 64000 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "claude-fable-5.1": { - "id": "claude-fable-5.1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } - } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 936000, - "output": 64000 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "claude-sonnet-4.6": { - "id": "claude-sonnet-4.6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 32000 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "input": 168000, - "output": 32000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "gemini-3.8-flash": { - "id": "gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 936000, - "output": 64000 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "gpt-6-luna": { - "id": "gpt-6-luna", - "name": "GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25 - } - } - }, - "claude-opus-5.5": { - "id": "claude-opus-5.5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } - } - }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 224000, - "output": 32000 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - }, - "grok-4.6": { - "id": "grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "input": 372000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10 - } - } - }, - "gpt-6-sol": { - "id": "gpt-6-sol", - "name": "GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } - } - } - } - }, - "inferx": { - "id": "inferx", - "env": [ - "INFERX_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://model.inferx.net/endpoints/v1", - "name": "InferX", - "doc": "https://model.inferx.net/endpoints", - "models": { - "gemma-4-31B-it-fp8": { - "id": "gemma-4-31B-it-fp8", - "name": "Gemma 4 31B IT FP8", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "Qwen3.6-35B-A3B-fp8-no-thinking": { - "id": "Qwen3.6-35B-A3B-fp8-no-thinking", - "name": "Qwen3.6-35B-A3B-fp8-no-thinking", - "description": "Qwen3.6-35B-A3B-fp8 disable thinking", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "Devstral-2-123B-Instruct-2512-int4-AutoRound": { - "id": "Devstral-2-123B-Instruct-2512-int4-AutoRound", - "name": "Devstral-2-123B-Instruct-2512-int4-AutoRound", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "Qwen3.6-27B-FP8": { - "id": "Qwen3.6-27B-FP8", - "name": "Qwen3.6 27B FP8", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "Qwen3-Coder-Next-FP8": { - "id": "Qwen3-Coder-Next-FP8", - "name": "Qwen3 Coder Next FP8", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256144, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "Qwen3-Coder-Next-FP8-no-thinking": { - "id": "Qwen3-Coder-Next-FP8-no-thinking", - "name": "Qwen3-Coder-Next-FP8-no-thinking", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 260000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mimo-v25": { - "id": "mimo-v25", - "name": "mimo-v25", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 100000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "Qwen3-Embedding-8B": { - "id": "Qwen3-Embedding-8B", - "name": "Qwen3-Embedding-8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-06-05", - "last_updated": "2025-06-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 0 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "Qwen3.6-35B-A3B-FP8": { - "id": "Qwen3.6-35B-A3B-FP8", - "name": "Qwen3.6 35B A3B FP8", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "Ornith-1.0-35B-FP8": { - "id": "Ornith-1.0-35B-FP8", - "name": "Ornith-1.0-35B-FP8", - "description": "Large coding-reasoning model for agentic software tasks and RL search", - "family": "ornith", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-25", - "last_updated": "2026-06-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 100000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "Agents-A1": { - "id": "Agents-A1", - "name": "Agents-A1", - "description": "35B MoE agentic model built for long-horizon search, engineering, and scientific reasoning tasks", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "release_date": "2026-06-26", - "last_updated": "2026-06-26", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 100000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "deepseek-v4-flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 100000 - }, - "cost": { - "input": 0, - "output": 0 - } - } - } - }, - "opencode-go": { - "id": "opencode-go", - "env": [ - "OPENCODE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://opencode.ai/zen/go/v1", - "name": "OpenCode Go", - "doc": "https://opencode.ai/docs/go", - "models": { - "mimo-v2.6-pro": { - "id": "mimo-v2.6-pro", - "name": "MiMo-V2.6-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 - } - }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "qwen3.7-max", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 - } - }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo V2.5", - "description": "MiMo omni model for text, image, video, audio, and agents", - "family": "mimo-v2.5", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 - } - }, - "grok-4.7": { - "id": "grok-4.7", - "name": "Grok 4.7", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } - } - }, - "mimo-v2-omni": { - "id": "mimo-v2-omni", - "name": "MiMo V2 Omni", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo-v2-omni", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 128000 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.08 - } - }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 - } - }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter multimodal flagship for coding, professional work, and long-horizon agentic workflows", - "family": "qwen3.8-max", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Legacy model retained for compatibility with older integrations", - "family": "qwen3.5", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "status": "deprecated", - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 - } - }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-11", - "last_updated": "2026-02-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 32768 - }, - "status": "deprecated", - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 - } - }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 - } - }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax-m2.5", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 65536 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "deepseek-v4-flash-vision-exp": { - "id": "deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 - } - }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 - } - }, - "longcat-2.0": { - "id": "longcat-2.0", - "name": "LongCat-2.0", - "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", - "family": "longcat", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.006 - } - }, - "ox-alpha-free": { - "id": "ox-alpha-free", - "name": "Ox Alpha Free (Unlimited)", - "description": "Stealth reasoning model for coding, agentic tasks, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-08-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 0.6 - } - } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "status": "deprecated", - "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 - } - }, - "mimo-v2-pro": { - "id": "mimo-v2-pro", - "name": "MiMo V2 Pro", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo-v2-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 128000 - }, - "status": "deprecated", - "cost": { - "input": 1, - "output": 3, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } - } - }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax-m2.7", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "space-bunny-free": { - "id": "space-bunny-free", - "name": "Space Bunny Free", - "description": "Anonymous preview reasoning model for coding, agentic tasks, tool use, and multimodal input", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "input": 524288, - "output": 524288 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo V2.5 Pro", - "description": "MiMo pro model for strong multimodal reasoning and agent execution", - "family": "mimo-v2.5-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 128000 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 - } - }, - "mimo-v2.6-flash": { - "id": "mimo-v2.6-flash", - "name": "MiMo-V2.6-Flash", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 - } - }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax-m3", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-31", - "last_updated": "2026-05-31", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "tiers": [ - { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12, - "tier": { - "type": "context", - "size": 512000 - } - } - ], - "context_over_200k": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 - } - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } - } - }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016, - "cache_write": 0.2 - } - }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "hy3": { - "id": "hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 192000, - "output": 128000 - }, - "cost": { - "input": 0.14, - "output": 0.58, - "cache_read": 0.035 - } - }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 32768 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "omen-alpha": { - "id": "omen-alpha", - "name": "Omen Alpha", - "description": "oH man anothEr aLPha ModEl", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 128000 - }, - "status": "deprecated", - "cost": { - "input": 0.2, - "output": 0.66, - "cache_read": 0.04 - } - }, - "muse-spark-1.2-contributor": { - "id": "muse-spark-1.2-contributor", - "name": "Muse Spark 1.2 Contributor", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 - } - }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro (New)", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.66, - "output": 1.98, - "cache_read": 0.022 - } - }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } - } - }, - "hy4-preview": { - "id": "hy4-preview", - "name": "Hy4 preview", - "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1024000, - "output": 64000 - }, - "cost": { - "input": 0.834, - "output": 2.501, - "cache_read": 0.042 - } - }, - "muse-spark-1.3-contributor": { - "id": "muse-spark-1.3-contributor", - "name": "Muse Spark 1.3 Contributor", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for coding and agentic workflows.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 - } - }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - }, - "grok-4.6": { - "id": "grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } - } - }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.7-plus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.04, - "cache_write": 0.5, - "tiers": [ - { - "input": 1.2, - "output": 4.8, - "cache_read": 0.12, - "cache_write": 1.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.2, - "output": 4.8, - "cache_read": 0.12, - "cache_write": 1.5 - } - } - }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 - } - } - } - }, - "openreason": { - "id": "openreason", - "env": [ - "OPENREASON_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.openreason.app/v1", - "name": "OpenReason", - "doc": "https://openreason.app/docs", - "models": { - "deepseek-ai/deepseek-v4-flash-0731": { - "id": "deepseek-ai/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.1371, - "output": 0.2743 - } - }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.0022, - "output": 4.22 - } - }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.1055, - "output": 0.422 - } - } - } - }, - "lmstudio": { - "id": "lmstudio", - "env": [ - "LMSTUDIO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "http://127.0.0.1:1234/v1", - "name": "LMStudio", - "doc": "https://lmstudio.ai/models", - "models": { - "qwen/qwen3-30b-a3b-2507": { - "id": "qwen/qwen3-30b-a3b-2507", - "name": "Qwen3 30B A3B 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-30", - "last_updated": "2025-07-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen/qwen3-coder-30b": { - "id": "qwen/qwen3-coder-30b", - "name": "Qwen3 Coder 30B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - } - } - }, - "aki-io": { - "id": "aki-io", - "env": [ - "AKI_IO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://aki.io/v1", - "name": "AKI.IO", - "doc": "https://aki.io/docs/", - "models": { - "qwen3.8-27b": { - "id": "qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 2.2, - "cache_read": 0.1 - } - }, - "deepseek-v4-flash-0731-284b": { - "id": "deepseek-v4-flash-0731-284b", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 81920 - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.1 - } - }, - "qwen3.6-35b": { - "id": "qwen3.6-35b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.5 - } - }, - "glm5.3-754b": { - "id": "glm5.3-754b", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 81920 - }, - "cost": { - "input": 1, - "output": 3.5, - "cache_read": 0.25 - } - }, - "mistral4-119b": { - "id": "mistral4-119b", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 81920 - }, - "cost": { - "input": 0.2, - "output": 0.6 - } - }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.55 - } - }, - "gemma4-26b": { - "id": "gemma4-26b", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.5 - } - } - } - }, - "tensorx": { - "id": "tensorx", - "env": [ - "TENSORX_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.tensorx.ai/v1", - "name": "TensorX", - "doc": "https://docs.tensorx.ai/", - "models": { - "deepseek/deepseek-v4-flash-0731": { - "id": "deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.25, - "output": 0.3, - "cache_read": 0.06 - } - }, - "deepseek/deepseek-v4.1-flash": { - "id": "deepseek/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.13 - } - }, - "deepseek/deepseek-v4-pro-0813": { - "id": "deepseek/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 64000 - }, - "cost": { - "input": 2, - "output": 4, - "cache_read": 0.5 - } - }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 1.75, - "output": 3.5, - "cache_read": 0.4375, - "cache_write": 2.185 - } - }, - "deepseek/deepseek-r1-0528": { - "id": "deepseek/deepseek-r1-0528", - "name": "DeepSeek R1-0528", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 164000, - "output": 8192 - }, - "cost": { - "input": 0.66, - "output": 2.6, - "cache_read": 0.165, - "cache_write": 0.825 - } - }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 163840 - }, - "cost": { - "input": 0.3, - "output": 0.5, - "cache_read": 0.075, - "cache_write": 0.375 - } - }, - "z-ai/glm-5v-turbo": { - "id": "z-ai/glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.3, - "cache_write": 1.5 - } - }, - "z-ai/glm-5.3-flash": { - "id": "z-ai/glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 64000 - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } - }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 202752 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.25, - "cache_write": 1.25 - } - }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.5, - "output": 4.5, - "cache_read": 0.375 - } - }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 202752 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.35, - "cache_write": 1.75 - } - }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.3, - "cache_write": 1.5 - } - }, - "z-ai/glm-5.3": { - "id": "z-ai/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 64000 - }, - "cost": { - "input": 1.75, - "output": 4.5, - "cache_read": 0.44 - } - }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 - } - }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1, - "output": 4, - "cache_read": 0.25, - "cache_write": 1.25 - } - }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 2.8, - "cache_read": 0.125, - "cache_write": 0.625 - } - }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.25, - "output": 4.5, - "cache_read": 0.3125 - } - }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.075, - "cache_write": 0.375 - } - }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.1 - } - }, - "qwen/qwen3-235b-a22b-2507": { - "id": "qwen/qwen3-235b-a22b-2507", - "name": "Qwen3 235B-A22B-2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-06-30", - "release_date": "2025-07-21", - "last_updated": "2025-07-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131000, - "output": 262144 - }, - "cost": { - "input": 0.072, - "output": 0.464, - "cache_read": 0.018, - "cache_write": 0.09 - } - }, - "qwen/qwen3.8-27b": { - "id": "qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.1 - } - }, - "qwen/qwen3.8-2.4t-a95b": { - "id": "qwen/qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 64000 - }, - "cost": { - "input": 2.5, - "output": 6, - "cache_read": 0.63 - } - }, - "qwen/qwen3.5-122b-a10b": { - "id": "qwen/qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 3.5, - "cache_read": 0.125, - "cache_write": 0.625 - } - }, - "qwen/qwen3.5-9b": { - "id": "qwen/qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.15, - "output": 0.2, - "cache_read": 0.0375, - "cache_write": 0.1875 - } - }, - "qwen/qwen3.8-flash-next": { - "id": "qwen/qwen3.8-flash-next", - "name": "Qwen3.8 Flash Next", - "description": "Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 64000 - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } - } - } - }, - "longcat": { - "id": "longcat", - "env": [ - "LONGCAT_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.longcat.chat/openai", - "name": "LongCat", - "doc": "https://longcat.chat/platform/docs/", - "models": { - "LongCat-2.0": { - "id": "LongCat-2.0", - "name": "LongCat-2.0", - "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", - "family": "longcat", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.75, - "output": 2.95, - "cache_read": 0.015 - } - } - } - }, - "chutes": { - "id": "chutes", - "env": [ - "CHUTES_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://llm.chutes.ai/v1", - "name": "Chutes", - "doc": "https://llm.chutes.ai/v1/models", - "models": { - "Nemotron-3-Nano-Omni-30B-TEE": { - "id": "Nemotron-3-Nano-Omni-30B-TEE", - "name": "Nemotron 3 Nano Omni 30B TEE", - "description": "Omni-modal model for text, vision, audio, and multimodal agent tasks", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-07-23", - "last_updated": "2026-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 0 - }, - "cost": { - "input": 0.0245, - "output": 0.0978, - "cache_read": 0.0024499999999999995 - } - }, - "unsloth/Mistral-Nemo-Instruct-2407-TEE": { - "id": "unsloth/Mistral-Nemo-Instruct-2407-TEE", - "name": "Mistral Nemo Instruct 2407 TEE", - "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.0245, - "output": 0.0978, - "cache_read": 0.0024499999999999995 - } - }, - "google/gemma-4-31B-turbo-TEE": { - "id": "google/gemma-4-31B-turbo-TEE", - "name": "gemma 4 31B turbo TEE", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 65536 - }, - "cost": { - "input": 0.12, - "output": 0.37, - "cache_read": 0.011999999999999997 - } - }, - "Qwen/Qwen3-32B-TEE": { - "id": "Qwen/Qwen3-32B-TEE", - "name": "Qwen3 32B TEE", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "output": 40960 - }, - "cost": { - "input": 0.104, - "output": 0.416, - "cache_read": 0.010399999999999998 - } - }, - "Qwen/Qwen3.6-27B-TEE": { - "id": "Qwen/Qwen3.6-27B-TEE", - "name": "Qwen3.6 27B TEE", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2, - "cache_read": 0.029999999999999992 - } - }, - "Qwen/Qwen3-235B-A22B-Thinking-2507-TEE": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507-TEE", - "name": "Qwen3 235B A22B Thinking 2507 TEE", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07", - "last_updated": "2026-06-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.2989, - "output": 1.1957, - "cache_read": 0.029889999999999993 - } - }, - "Qwen/Qwen3.8-27B-TEE": { - "id": "Qwen/Qwen3.8-27B-TEE", - "name": "Qwen3.8 27B TEE", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-16", - "last_updated": "2026-08-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.24, - "output": 2.2, - "cache_read": 0.023999999999999994 - } - }, - "Qwen/Qwen3.5-397B-A17B-TEE": { - "id": "Qwen/Qwen3.5-397B-A17B-TEE", - "name": "Qwen3.5 397B A17B TEE", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.45, - "output": 3, - "cache_read": 0.04499999999999999 - } - }, - "deepseek-ai/DeepSeek-V4-Flash-0731-TEE": { - "id": "deepseek-ai/DeepSeek-V4-Flash-0731-TEE", - "name": "DeepSeek V4 Flash 0731 TEE", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-08-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.04399999999999999 - } - }, - "deepseek-ai/DeepSeek-V3.2-TEE": { - "id": "deepseek-ai/DeepSeek-V3.2-TEE", - "name": "DeepSeek V3.2 TEE", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12", - "last_updated": "2026-06-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 65536 - }, - "cost": { - "input": 1, - "output": 1, - "cache_read": 0.09999999999999998 - } - }, - "moonshotai/Kimi-K3-TEE": { - "id": "moonshotai/Kimi-K3-TEE", - "name": "Kimi K3 TEE", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-29", - "last_updated": "2026-07-29", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.29999999999999993 - } - }, - "moonshotai/Kimi-K2.6-TEE": { - "id": "moonshotai/Kimi-K2.6-TEE", - "name": "Kimi K2.6 TEE", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65535 - }, - "cost": { - "input": 0.5, - "output": 2.85, - "cache_read": 0.04999999999999999 - } - }, - "zai-org/GLM-5.2-TEE": { - "id": "zai-org/GLM-5.2-TEE", - "name": "GLM 5.2 TEE", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 1.25, - "output": 3.95, - "cache_read": 0.12499999999999997 - } - }, - "zai-org/GLM-5.1-TEE": { - "id": "zai-org/GLM-5.1-TEE", - "name": "GLM 5.1 TEE", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 65535 - }, - "cost": { - "input": 0.98, - "output": 3.08, - "cache_read": 0.09799999999999998 - } - } - } - }, - "edenai": { - "id": "edenai", - "env": [ - "EDENAI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.edenai.run/v3", - "name": "Eden AI", - "doc": "https://docs.edenai.co", - "models": { - "deepinfra/nemotron-3-ultra-550b-a55b": { - "id": "deepinfra/nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B (Deep Infra)", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 128000 - }, - "cost": { - "input": 0.5, - "output": 2.2, - "cache_read": 0.1 - } - }, - "deepinfra/meta-models/Muse-Glimmer-30B": { - "id": "deepinfra/meta-models/Muse-Glimmer-30B", - "name": "Muse Glimmer 30B (Deep Infra)", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.04 - } - }, - "deepinfra/tencent/Hy3": { - "id": "deepinfra/tencent/Hy3", - "name": "Hy3 (Deep Infra)", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 128000 - }, - "cost": { - "input": 0.13, - "output": 0.53, - "cache_read": 0.033 - } - }, - "deepinfra/meta-llama/Llama-Guard-3-8B": { - "id": "deepinfra/meta-llama/Llama-Guard-3-8B", - "name": "Llama-Guard-3-8B (Deep Infra)", - "description": "Llama 3.1-based safety classifier for moderating prompts and model responses", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 4096 - }, - "cost": { - "input": 0.055, - "output": 0.055 - } - }, - "deepinfra/meta-llama/Llama-3.2-11B-Vision-Instruct": { - "id": "deepinfra/meta-llama/Llama-3.2-11B-Vision-Instruct", - "name": "Llama-3.2-11B-Vision-Instruct (Deep Infra)", - "description": "Open multimodal Llama model for image understanding, captioning, and visual QA", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 4096 - }, - "cost": { - "input": 0.345, - "output": 0.345 - } - }, - "deepinfra/meta-llama/Llama-3.3-70B-Instruct": { - "id": "deepinfra/meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama-3.3-70B-Instruct (Deep Infra)", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 4096 - }, - "cost": { - "input": 0.1, - "output": 0.32 - } - }, - "deepinfra/thinkingmachines/Inkling-Small": { - "id": "deepinfra/thinkingmachines/Inkling-Small", - "name": "Inkling Small (Deep Infra)", - "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 1048576 - }, - "cost": { - "input": 0.45, - "output": 1.2, - "cache_read": 0.1 - } - }, - "deepinfra/thinkingmachines/Inkling": { - "id": "deepinfra/thinkingmachines/Inkling", - "name": "Inkling (Deep Infra)", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 1048576 - }, - "cost": { - "input": 0.95, - "output": 4.05, - "cache_read": 0.16 - } - }, - "deepinfra/google/gemma-3-27b-it": { - "id": "deepinfra/google/gemma-3-27b-it", - "name": "Gemma 3 27B IT (Deep Infra)", - "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.08, - "output": 0.16 - } - }, - "deepinfra/google/gemma-3-12b-it": { - "id": "deepinfra/google/gemma-3-12b-it", - "name": "Gemma 3 12B IT (Deep Infra)", - "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.05, - "output": 0.15 - } - }, - "deepinfra/google/gemma-3-4b-it": { - "id": "deepinfra/google/gemma-3-4b-it", - "name": "Gemma 3 4B IT (Deep Infra)", - "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.05, - "output": 0.1 - } - }, - "deepinfra/deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "deepinfra/deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731 (Deep Infra)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.015 - } - }, - "deepinfra/deepseek-ai/DeepSeek-R1": { - "id": "deepinfra/deepseek-ai/DeepSeek-R1", - "name": "DeepSeek-R1 (Deep Infra)", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 32768 - }, - "cost": { - "input": 0.7, - "output": 2.4 - } - }, - "deepinfra/deepseek-ai/DeepSeek-V4-Pro-0813": { - "id": "deepinfra/deepseek-ai/DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813 (Deep Infra)", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 1.3, - "output": 2.6, - "cache_read": 0.1 - } - }, - "deepinfra/deepseek-ai/DeepSeek-V4.1-Flash": { - "id": "deepinfra/deepseek-ai/DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash (Deep Infra)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.2, - "output": 0.6, - "cache_read": 0.006 - } - }, - "deepinfra/deepseek-ai/DeepSeek-V3": { - "id": "deepinfra/deepseek-ai/DeepSeek-V3", - "name": "DeepSeek-V3 (Deep Infra)", - "description": "Open DeepSeek MoE chat model for coding, math, and general reasoning", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2024-12-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 8192 - }, - "cost": { - "input": 0.32, - "output": 0.89 - } - }, - "deepinfra/deepseek-ai/DeepSeek-V3-0324": { - "id": "deepinfra/deepseek-ai/DeepSeek-V3-0324", - "name": "DeepSeek V3 0324 (Deep Infra)", - "description": "March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-03-24", - "last_updated": "2025-03-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 163840 - }, - "cost": { - "input": 0.24, - "output": 0.9, - "cache_read": 0.135 - } - }, - "deepinfra/stepfun-ai/Step-3.7-Flash": { - "id": "deepinfra/stepfun-ai/Step-3.7-Flash", - "name": "Step 3.7 Flash (Deep Infra)", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 256000 - }, - "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 - } - }, - "deepinfra/stepfun-ai/Step-3.5-Flash": { - "id": "deepinfra/stepfun-ai/Step-3.5-Flash", - "name": "Step 3.5 Flash (Deep Infra)", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-02-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 256000 - }, - "cost": { - "input": 0.09, - "output": 0.3, - "cache_read": 0.02 - } - }, - "deepinfra/moonshotai/Kimi-K2.5": { - "id": "deepinfra/moonshotai/Kimi-K2.5", - "name": "Kimi K2.5 (Deep Infra)", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.45, - "output": 2.25, - "cache_read": 0.07 - } - }, - "deepinfra/zai-org/GLM-4.7-Flash": { - "id": "deepinfra/zai-org/GLM-4.7-Flash", - "name": "GLM-4.7-Flash (Deep Infra)", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 0.06, - "output": 0.4, - "cache_read": 0.01 - } - }, - "deepinfra/nvidia/Nemotron-3-Nano-30B-A3B": { - "id": "deepinfra/nvidia/Nemotron-3-Nano-30B-A3B", - "name": "Nemotron 3 Nano 30B A3B (Deep Infra)", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.025 - } - }, - "deepinfra/nvidia/Llama-3.1-Nemotron-70B-Instruct": { - "id": "deepinfra/nvidia/Llama-3.1-Nemotron-70B-Instruct", - "name": "Llama 3.1 Nemotron 70B Instruct (Deep Infra)", - "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-04-15", - "last_updated": "2025-04-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.6, - "output": 0.6 - } - }, - "deepinfra/ByteDance/Seed-2.0-code": { - "id": "deepinfra/ByteDance/Seed-2.0-code", - "name": "Seed 2.0 Code (Deep Infra)", - "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 131072 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1 - } - }, - "deepinfra/ByteDance/Seed-2.0-mini": { - "id": "deepinfra/ByteDance/Seed-2.0-mini", - "name": "Seed 2.0 Mini (Deep Infra)", - "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 32000 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.02 - } - }, - "deepinfra/openai/gpt-oss-20b": { - "id": "deepinfra/openai/gpt-oss-20b", - "name": "GPT OSS 20B (Deep Infra)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.03, - "output": 0.14 - } - }, - "deepinfra/openai/gpt-oss-120b": { - "id": "deepinfra/openai/gpt-oss-120b", - "name": "GPT OSS 120B (Deep Infra)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.037, - "output": 0.17 - } - }, - "cerebras/gpt-oss-120b": { - "id": "cerebras/gpt-oss-120b", - "name": "GPT OSS 120B (Cerebras)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.35, - "output": 0.75, - "cache_read": 0.35 - } - }, - "groq/openai/gpt-oss-20b": { - "id": "groq/openai/gpt-oss-20b", - "name": "GPT OSS 20B (Groq)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.075, - "output": 0.3, - "cache_read": 0.0375 - } - }, - "groq/openai/gpt-oss-safeguard-20b": { - "id": "groq/openai/gpt-oss-safeguard-20b", - "name": "GPT OSS Safeguard 20B (Groq)", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.075, - "output": 0.3, - "cache_read": 0.0375 - } - }, - "groq/openai/gpt-oss-120b": { - "id": "groq/openai/gpt-oss-120b", - "name": "GPT OSS 120B (Groq)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 - } - }, - "zai/glm-4.6v": { - "id": "zai/glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.05 - } - }, - "zai/glm-5v-turbo": { - "id": "zai/glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", - "modalities": { - "input": [ - "image", - "text", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 - } - }, - "zai/glm-5.3-flash": { - "id": "zai/glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 - } - }, - "zai/glm-4.6": { - "id": "zai/glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 - } - }, - "zai/glm-5": { - "id": "zai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 - } - }, - "zai/glm-4.7": { - "id": "zai/glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11 - } - }, - "zai/glm-5.2": { - "id": "zai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "zai/glm-5.1": { - "id": "zai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "zai/glm-5-turbo": { - "id": "zai/glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 - } - }, - "zai/glm-5.3": { - "id": "zai/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "anthropic/claude-opus-latest": { - "id": "anthropic/claude-opus-latest", - "name": "Claude Opus Latest (Claude Opus 5)", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-opus-4-5": { - "id": "anthropic/claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-opus-5-5": { - "id": "anthropic/claude-opus-5-5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 - } - }, - "anthropic/claude-fable-5-1": { - "id": "anthropic/claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "anthropic/claude-sonnet-latest": { - "id": "anthropic/claude-sonnet-latest", - "name": "Claude Sonnet Latest (Claude Sonnet 5)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "anthropic/claude-opus-4-5-20251101": { - "id": "anthropic/claude-opus-4-5-20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-fable-latest": { - "id": "anthropic/claude-fable-latest", - "name": "Claude Fable Latest (Claude Fable 5.1)", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "anthropic/claude-opus-4-8": { - "id": "anthropic/claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "cohere/command-r-08-2024": { - "id": "cohere/command-r-08-2024", - "name": "Command R", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "cohere/command-a-03-2025": { - "id": "cohere/command-a-03-2025", - "name": "Command A", - "description": "Cohere command model for multilingual enterprise agents, tools, and chat", - "family": "command-a", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-03-13", - "last_updated": "2025-03-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 288000, - "output": 8000 - }, - "cost": { - "input": 2.5, - "output": 10 - } - }, - "cohere/command-r7b-12-2024": { - "id": "cohere/command-r7b-12-2024", - "name": "Command R7B", - "description": "Cohere retrieval model for long-context chat and enterprise RAG workflows", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-12-02", - "last_updated": "2024-12-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 132000, - "output": 4000 - }, - "cost": { - "input": 0.0375, - "output": 0.15 - } - }, - "cohere/command-r-plus-08-2024": { - "id": "cohere/command-r-plus-08-2024", - "name": "Command R+", - "description": "Cohere's RAG workhorse for long-context enterprise search and tool use", - "family": "command-r", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2024-08-30", - "last_updated": "2024-08-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4000 - }, - "cost": { - "input": 2.5, - "output": 10 - } - }, - "deepseek/deepseek-v4-flash-vision-exp": { - "id": "deepseek/deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 - } - }, - "deepseek/deepseek-chat": { - "id": "deepseek/deepseek-chat", - "name": "DeepSeek Chat", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-12-01", - "last_updated": "2026-02-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 384000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 - } - }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.66, - "output": 1.98, - "cache_read": 0.022 - } - }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 - } - }, - "databricks/databricks-inkling": { - "id": "databricks/databricks-inkling", - "name": "Inkling (Databricks)", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 1048576 - }, - "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.1, - "cache_write": 1.00002 - } - }, - "databricks/databricks-gpt-oss-20b": { - "id": "databricks/databricks-gpt-oss-20b", - "name": "GPT OSS 20B (Databricks)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.07, - "output": 0.30002, - "cache_read": 0.007, - "cache_write": 0.07 - } - }, - "databricks/databricks-deepseek-v4-flash-0731": { - "id": "databricks/databricks-deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731 (Databricks)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.014, - "cache_write": 0.14 - } - }, - "databricks/databricks-gpt-oss-120b@eu": { - "id": "databricks/databricks-gpt-oss-120b@eu", - "name": "GPT OSS 120B (Databricks, EU)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15001, - "output": 0.59997, - "cache_read": 0.015001, - "cache_write": 0.15001 - } - }, - "databricks/databricks-gpt-oss-120b": { - "id": "databricks/databricks-gpt-oss-120b", - "name": "GPT OSS 120B (Databricks)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15001, - "output": 0.59997, - "cache_read": 0.015001, - "cache_write": 0.15001 - } - }, - "databricks/databricks-deepseek-v4-pro-0813": { - "id": "databricks/databricks-deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813 (Databricks)", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.32, - "output": 3.959999, - "cache_read": 0.132, - "cache_write": 1.31999 - } - }, - "databricks/databricks-gpt-oss-20b@eu": { - "id": "databricks/databricks-gpt-oss-20b@eu", - "name": "GPT OSS 20B (Databricks, EU)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.07, - "output": 0.30002, - "cache_read": 0.007, - "cache_write": 0.07 - } - }, - "together_ai/meta-models/Muse-Glimmer-30B": { - "id": "together_ai/meta-models/Muse-Glimmer-30B", - "name": "Muse Glimmer 30B (Together AI)", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.35, - "output": 1.5, - "cache_read": 0.04 - } - }, - "together_ai/thinkingmachines/Inkling": { - "id": "together_ai/thinkingmachines/Inkling", - "name": "Inkling (Together AI)", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 1048576 - }, - "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.17 - } - }, - "together_ai/deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "together_ai/deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731 (Together AI)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.03 - } - }, - "together_ai/deepseek-ai/DeepSeek-V4-Pro-0813": { - "id": "together_ai/deepseek-ai/DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813 (Together AI)", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.13 - } - }, - "together_ai/deepseek-ai/DeepSeek-V4.1-Flash": { - "id": "together_ai/deepseek-ai/DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash (Together AI)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.006 - } - }, - "together_ai/openai/gpt-oss-120b": { - "id": "together_ai/openai/gpt-oss-120b", - "name": "GPT OSS 120B (Together AI)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "azure/gpt-5.2-codex": { - "id": "azure/gpt-5.2-codex", - "name": "GPT-5.2 Codex (Azure)", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "azure/gpt-5.1-codex": { - "id": "azure/gpt-5.1-codex", - "name": "GPT-5.1 Codex (Azure)", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "azure/gpt-5.1-codex-max": { - "id": "azure/gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max (Azure)", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "azure/gpt-5.1-codex-mini": { - "id": "azure/gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini (Azure)", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "flexai/Step-3.7-Flash": { - "id": "flexai/Step-3.7-Flash", - "name": "Step 3.7 Flash (FlexAI)", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 256000 - }, - "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.03 - } - }, - "flexai/DeepSeek-V4-Flash-0731": { - "id": "flexai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731 (FlexAI)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.065, - "output": 0.18, - "cache_read": 0.0097 - } - }, - "flexai/Muse-Glimmer-30B": { - "id": "flexai/Muse-Glimmer-30B", - "name": "Muse Glimmer 30B (FlexAI)", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.1, - "cache_read": 0.045 - } - }, - "flexai/gpt-oss-20b": { - "id": "flexai/gpt-oss-20b", - "name": "GPT OSS 20B (FlexAI)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.018, - "output": 0.09, - "cache_read": 0.0027 - } - }, - "flexai/gpt-oss-120b": { - "id": "flexai/gpt-oss-120b", - "name": "GPT OSS 120B (FlexAI)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.037, - "output": 0.17, - "cache_read": 0.0055 - } - }, - "google/gemini-flash-latest": { - "id": "google/gemini-flash-latest", - "name": "Gemini Flash Latest (Gemini 3.8 Flash)", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "google/gemini-3.1-flash-image-preview": { - "id": "google/gemini-3.1-flash-image-preview", - "name": "Nano Banana 2 Preview", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-26", - "last_updated": "2026-02-26", - "modalities": { - "input": [ - "image", - "text" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3 - } - }, - "google/gemini-2.5-flash-image": { - "id": "google/gemini-2.5-flash-image", - "name": "Nano Banana", - "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", - "modalities": { - "input": [ - "audio", - "image", - "text", - "video" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333, - "input_audio": 1 - } - }, - "google/gemini-3.6-flash": { - "id": "google/gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "google/gemini-3-pro-image-preview": { - "id": "google/gemini-3-pro-image-preview", - "name": "Nano Banana Pro Preview", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-20", - "last_updated": "2025-11-20", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "input_audio": 2 - } - }, - "google/gemini-3.5-flash-lite": { - "id": "google/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333, - "input_audio": 0.3 - } - }, - "google/gemini-3.1-flash-image": { - "id": "google/gemini-3.1-flash-image", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "image", - "text" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 3 - } - }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "input_audio": 2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "reasoning": 9, - "cache_read": 0.15, - "cache_write": 0.083333, - "input_audio": 3 - } - }, - "google/gemini-3-pro-image": { - "id": "google/gemini-3-pro-image", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "input_audio": 2 - } - }, - "google/gemini-3.1-flash-lite-image": { - "id": "google/gemini-3.1-flash-lite-image", - "name": "Nano Banana 2 Lite", - "description": "Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 4096 - }, - "cost": { - "input": 0.25, - "output": 1.5 - } - }, - "google/gemini-3.7-flash": { - "id": "google/gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "input_audio": 2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "reasoning": 3, - "cache_read": 0.05, - "cache_write": 0.083333, - "input_audio": 1 - } - }, - "google/gemini-pro-latest": { - "id": "google/gemini-pro-latest", - "name": "Gemini Pro Latest (Gemini 3.1 Pro Preview)", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "input_audio": 2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "google/gemini-3.8-flash": { - "id": "google/gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "google/gemini-3.1-flash-lite-preview": { - "id": "google/gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.083333, - "input_audio": 0.5 - } - }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.083333, - "input_audio": 0.5 - } - }, - "xai/grok-latest": { - "id": "xai/grok-latest", - "name": "Grok Latest (Grok 4.6)", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } - } - }, - "xai/grok-4.7": { - "id": "xai/grok-4.7", - "name": "Grok 4.7", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 3.2, - "output": 9.6, - "cache_read": 0.8, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 3.2, - "output": 9.6, - "cache_read": 0.8 - } - } - }, - "xai/grok-4.3": { - "id": "xai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 30000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "xai/grok-4.20-0309-reasoning": { - "id": "xai/grok-4.20-0309-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 30000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "xai/grok-4.5": { - "id": "xai/grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 0.6 - } - } - }, - "xai/grok-4.20-0309-non-reasoning": { - "id": "xai/grok-4.20-0309-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 30000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "xai/grok-build-0.1": { - "id": "xai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 4, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 4, - "cache_read": 0.4 - } - } - }, - "xai/grok-4.6": { - "id": "xai/grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } - } - }, - "nebius/google/gemma-3-27b-it": { - "id": "nebius/google/gemma-3-27b-it", - "name": "Gemma 3 27B IT (Nebius)", - "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 110000, - "output": 131072 - }, - "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.1 - } - }, - "nebius/deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "nebius/deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731 (Nebius)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1024000, - "output": 384000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.14 - } - }, - "nebius/deepseek-ai/DeepSeek-V4-Pro-0813": { - "id": "nebius/deepseek-ai/DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813 (Nebius)", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 979000, - "output": 384000 - }, - "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 1.32 - } - }, - "nebius/deepseek-ai/DeepSeek-V4.1-Flash": { - "id": "nebius/deepseek-ai/DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash (Nebius)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048000, - "output": 384000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.3 - } - }, - "nebius/nvidia/nemotron-3-super-120b-a12b": { - "id": "nebius/nvidia/nemotron-3-super-120b-a12b", - "name": "Nemotron 3 Super 120B A12B (Nebius)", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.3 - } - }, - "nebius/nvidia/Nemotron-3-Ultra-550b-a55b": { - "id": "nebius/nvidia/Nemotron-3-Ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B (Nebius)", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 128000 - }, - "cost": { - "input": 1, - "output": 3, - "cache_read": 1 - } - }, - "nebius/openai/gpt-oss-120b": { - "id": "nebius/openai/gpt-oss-120b", - "name": "GPT OSS 120B (Nebius)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.15 - } - }, - "scaleway/gemma-3-27b-it": { - "id": "scaleway/gemma-3-27b-it", - "name": "Gemma 3 27B IT (Scaleway)", - "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 40000, - "output": 131072 - }, - "cost": { - "input": 0.287125, - "output": 0.57425 - } - }, - "scaleway/deepseek-v4-flash-0731": { - "id": "scaleway/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731 (Scaleway)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 384000 - }, - "cost": { - "input": 0.45644, - "output": 0.91288, - "cache_read": 0.091288 - } - }, - "scaleway/llama-3.3-70b-instruct": { - "id": "scaleway/llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct (Scaleway)", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 1.02699, - "output": 1.02699 - } - }, - "scaleway/gpt-oss-120b": { - "id": "scaleway/gpt-oss-120b", - "name": "GPT OSS 120B (Scaleway)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.171165, - "output": 0.68466 - } - }, - "ionos/meta-llama/Llama-3.3-70B-Instruct": { - "id": "ionos/meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama-3.3-70B-Instruct (IONOS)", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.741715, - "output": 0.741715 - } - }, - "ionos/openai/gpt-oss-120b": { - "id": "ionos/openai/gpt-oss-120b", - "name": "GPT OSS 120B (IONOS)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.171165, - "output": 0.741715 - } - }, - "vertex/gemini-flash-latest": { - "id": "vertex/gemini-flash-latest", - "name": "Gemini Flash Latest (Gemini 3.8 Flash, Vertex AI)", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "vertex/gemini-2.5-flash-image": { - "id": "vertex/gemini-2.5-flash-image", - "name": "Nano Banana (Vertex AI)", - "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333, - "input_audio": 1 - } - }, - "vertex/gemini-3.1-flash-lite@eu": { - "id": "vertex/gemini-3.1-flash-lite@eu", - "name": "Gemini 3.1 Flash Lite (Vertex AI, EU)", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.083333, - "input_audio": 0.5 - } - }, - "vertex/gemini-3.5-flash-lite@eu": { - "id": "vertex/gemini-3.5-flash-lite@eu", - "name": "Gemini 3.5 Flash Lite (Vertex AI, EU)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333, - "input_audio": 0.3 - } - }, - "vertex/gemini-3.6-flash@eu": { - "id": "vertex/gemini-3.6-flash@eu", - "name": "Gemini 3.6 Flash (Vertex AI, EU)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "vertex/gemini-3.6-flash": { - "id": "vertex/gemini-3.6-flash", - "name": "Gemini 3.6 Flash (Vertex AI)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "vertex/gemini-3.5-flash-lite": { - "id": "vertex/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite (Vertex AI)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333, - "input_audio": 0.3 - } - }, - "vertex/gemini-3.1-flash-image": { - "id": "vertex/gemini-3.1-flash-image", - "name": "Nano Banana 2 (Vertex AI)", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "image", - "text" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 3 - } - }, - "vertex/gemini-3.1-pro-preview": { - "id": "vertex/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview (Vertex AI)", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "input_audio": 2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 0.25 - } - } - }, - "vertex/gemini-3.5-flash": { - "id": "vertex/gemini-3.5-flash", - "name": "Gemini 3.5 Flash (Vertex AI)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "reasoning": 9, - "cache_read": 0.15, - "cache_write": 0.083333, - "input_audio": 3 - } - }, - "vertex/gemini-3-pro-image": { - "id": "vertex/gemini-3-pro-image", - "name": "Nano Banana Pro (Vertex AI)", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "input_audio": 2 - } - }, - "vertex/gemini-3.7-flash@us": { - "id": "vertex/gemini-3.7-flash@us", - "name": "Gemini 3.7 Flash (Vertex AI, US)", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "vertex/gemini-3.7-flash@eu": { - "id": "vertex/gemini-3.7-flash@eu", - "name": "Gemini 3.7 Flash (Vertex AI, EU)", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "vertex/gemini-3.1-flash-lite-image": { - "id": "vertex/gemini-3.1-flash-lite-image", - "name": "Nano Banana 2 Lite (Vertex AI)", - "description": "Fastest, most cost-efficient Gemini image model for high-volume 1K generation and editing", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 4096 - }, - "cost": { - "input": 0.25, - "output": 1.5 - } - }, - "vertex/gemini-3.8-flash@eu": { - "id": "vertex/gemini-3.8-flash@eu", - "name": "Gemini 3.8 Flash (Vertex AI, EU)", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "vertex/gemini-3.7-flash": { - "id": "vertex/gemini-3.7-flash", - "name": "Gemini 3.7 Flash (Vertex AI)", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "vertex/gemini-3.6-flash@us": { - "id": "vertex/gemini-3.6-flash@us", - "name": "Gemini 3.6 Flash (Vertex AI, US)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "vertex/gemini-3.5-flash-lite@us": { - "id": "vertex/gemini-3.5-flash-lite@us", - "name": "Gemini 3.5 Flash Lite (Vertex AI, US)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "reasoning": 2.5, - "cache_read": 0.03, - "cache_write": 0.083333, - "input_audio": 0.3 - } - }, - "vertex/gemini-3-flash-preview": { - "id": "vertex/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview (Vertex AI)", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "reasoning": 3, - "cache_read": 0.05, - "cache_write": 0.083333, - "input_audio": 1 - } - }, - "vertex/gemini-3.1-flash-lite@us": { - "id": "vertex/gemini-3.1-flash-lite@us", - "name": "Gemini 3.1 Flash Lite (Vertex AI, US)", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.083333, - "input_audio": 0.5 - } - }, - "vertex/gemini-pro-latest": { - "id": "vertex/gemini-pro-latest", - "name": "Gemini Pro Latest (Gemini 3.1 Pro Preview, Vertex AI)", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "reasoning": 12, - "cache_read": 0.2, - "cache_write": 0.375, - "input_audio": 2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 0.25 - } - } - }, - "vertex/gemini-3.8-flash": { - "id": "vertex/gemini-3.8-flash", - "name": "Gemini 3.8 Flash (Vertex AI)", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "vertex/gemini-3.5-flash@eu": { - "id": "vertex/gemini-3.5-flash@eu", - "name": "Gemini 3.5 Flash (Vertex AI, EU)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "reasoning": 9, - "cache_read": 0.15, - "cache_write": 0.083333, - "input_audio": 3 - } - }, - "vertex/gemini-3.8-flash@us": { - "id": "vertex/gemini-3.8-flash@us", - "name": "Gemini 3.8 Flash (Vertex AI, US)", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "reasoning": 3.75, - "cache_read": 0.075, - "cache_write": 0.041667, - "input_audio": 0.75 - } - }, - "vertex/gemini-3.5-flash@us": { - "id": "vertex/gemini-3.5-flash@us", - "name": "Gemini 3.5 Flash (Vertex AI, US)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "reasoning": 9, - "cache_read": 0.15, - "cache_write": 0.083333, - "input_audio": 3 - } - }, - "vertex/gemini-3.1-flash-lite": { - "id": "vertex/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite (Vertex AI)", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "reasoning": 1.5, - "cache_read": 0.025, - "cache_write": 0.083333, - "input_audio": 0.5 - } - }, - "perplexityai/sonar-pro": { - "id": "perplexityai/sonar-pro", - "name": "Sonar Pro", - "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", - "family": "sonar-pro", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8192 - }, - "cost": { - "input": 3, - "output": 15 - } - }, - "perplexityai/sonar-deep-research": { - "id": "perplexityai/sonar-deep-research", - "name": "Sonar Deep Research", - "description": "Sonar search model for autonomous research and citation-backed long-form reports", - "family": "sonar", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-02-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 8, - "reasoning": 3 - } - }, - "perplexityai/sonar": { - "id": "perplexityai/sonar", - "name": "Sonar", - "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", - "family": "sonar", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 127072, - "output": 4096 - }, - "cost": { - "input": 1, - "output": 1 - } - }, - "perplexityai/sonar-reasoning-pro": { - "id": "perplexityai/sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", - "family": "sonar-reasoning", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 2, - "output": 8 - } - }, - "cloudflare/@cf/meta/llama-guard-3-8b": { - "id": "cloudflare/@cf/meta/llama-guard-3-8b", - "name": "Llama-Guard-3-8B (Cloudflare)", - "description": "Llama 3.1-based safety classifier for moderating prompts and model responses", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 4096 - }, - "cost": { - "input": 0.484, - "output": 0.03 - } - }, - "cloudflare/@cf/aisingapore/gemma-sea-lion-v4-27b-it": { - "id": "cloudflare/@cf/aisingapore/gemma-sea-lion-v4-27b-it", - "name": "Gemma-SEA-LION-v4-27B-IT (Cloudflare)", - "description": "Gemma 3 27B tuned by AI Singapore for Southeast Asian languages and instruction following", - "family": "gemma", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.351, - "output": 0.555 - } - }, - "cloudflare/@cf/deepseek-ai/deepseek-v4-flash-0731": { - "id": "cloudflare/@cf/deepseek-ai/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731 (Cloudflare)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1310720, - "output": 384000 - }, - "cost": { - "input": 0.44, - "output": 1.32, - "cache_read": 0.014 - } - }, - "cloudflare/@cf/deepseek-ai/deepseek-v4-pro-0813": { - "id": "cloudflare/@cf/deepseek-ai/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813 (Cloudflare)", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 - } - }, - "cloudflare/@cf/zai-org/glm-4.7-flash": { - "id": "cloudflare/@cf/zai-org/glm-4.7-flash", - "name": "GLM-4.7-Flash (Cloudflare)", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.0605, - "output": 0.4 - } - }, - "cloudflare/@cf/qwen/qwen2.5-coder-32b-instruct": { - "id": "cloudflare/@cf/qwen/qwen2.5-coder-32b-instruct", - "name": "Qwen2.5-Coder-32B-Instruct (Cloudflare)", - "description": "Open coding-focused Qwen model for code generation, repair, and repository reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2024-11-12", - "last_updated": "2024-11-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 0.66, - "output": 1 - } - }, - "cloudflare/@cf/openai/gpt-oss-20b": { - "id": "cloudflare/@cf/openai/gpt-oss-20b", - "name": "GPT OSS 20B (Cloudflare)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.2, - "output": 0.3 - } - }, - "cloudflare/@cf/openai/gpt-oss-120b": { - "id": "cloudflare/@cf/openai/gpt-oss-120b", - "name": "GPT OSS 120B (Cloudflare)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.35, - "output": 0.75 - } - }, - "minimax/MiniMax-M3": { - "id": "minimax/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 512000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "minimax/MiniMax-M2.1": { - "id": "minimax/MiniMax-M2.1", - "name": "MiniMax-M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 - } - }, - "minimax/MiniMax-M2.5": { - "id": "minimax/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 - } - }, - "minimax/MiniMax-M2.7": { - "id": "minimax/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "minimax/MiniMax-M2": { - "id": "minimax/MiniMax-M2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "tensorx/deepseek/deepseek-v4-flash-0731": { - "id": "tensorx/deepseek/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731 (TensorX)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.25, - "output": 0.3, - "cache_read": 0.0625 - } - }, - "tensorx/deepseek/deepseek-v4.1-flash": { - "id": "tensorx/deepseek/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (TensorX)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.125 - } - }, - "tensorx/deepseek/deepseek-v4-pro-0813": { - "id": "tensorx/deepseek/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813 (TensorX)", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 2, - "output": 4, - "cache_read": 0.5 - } - }, - "tensorx/moonshotai/kimi-k2.5": { - "id": "tensorx/moonshotai/kimi-k2.5", - "name": "Kimi K2.5 (TensorX)", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 2.8, - "cache_read": 0.125 - } - }, - "amazon/google.gemma-3-12b-it": { - "id": "amazon/google.gemma-3-12b-it", - "name": "Gemma 3 12B IT (Amazon Bedrock)", - "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 131072 - }, - "cost": { - "input": 0.09, - "output": 0.29 - } - }, - "amazon/google.gemma-3-4b-it": { - "id": "amazon/google.gemma-3-4b-it", - "name": "Gemma 3 4B IT (Amazon Bedrock)", - "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 131072 - }, - "cost": { - "input": 0.04, - "output": 0.08 - } - }, - "amazon/openai.gpt-oss-safeguard-20b@us": { - "id": "amazon/openai.gpt-oss-safeguard-20b@us", - "name": "GPT OSS Safeguard 20B (Amazon Bedrock, US)", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 131072 - }, - "cost": { - "input": 0.07, - "output": 0.2 - } - }, - "amazon/openai.gpt-oss-safeguard-20b": { - "id": "amazon/openai.gpt-oss-safeguard-20b", - "name": "GPT OSS Safeguard 20B (Amazon Bedrock)", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 131072 - }, - "cost": { - "input": 0.07, - "output": 0.2 - } - }, - "amazon/amazon.nova-lite-v1:0@us": { - "id": "amazon/amazon.nova-lite-v1:0@us", - "name": "Nova Lite (US)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.015 - } - }, - "amazon/mistral.voxtral-mini-3b-2507@us": { - "id": "amazon/mistral.voxtral-mini-3b-2507@us", - "name": "Voxtral Mini 3B 2507 (Amazon Bedrock, US)", - "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", - "family": "voxtral", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.04, - "output": 0.04 - } - }, - "amazon/amazon.nova-micro-v1:0": { - "id": "amazon/amazon.nova-micro-v1:0", - "name": "Nova Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 10000 - }, - "cost": { - "input": 0.035, - "output": 0.14, - "cache_read": 0.00875 - } - }, - "amazon/amazon.nova-pro-v1:0": { - "id": "amazon/amazon.nova-pro-v1:0", - "name": "Nova Pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.8, - "output": 3.2, - "cache_read": 0.2 - } - }, - "amazon/mistral.voxtral-small-24b-2507": { - "id": "amazon/mistral.voxtral-small-24b-2507", - "name": "Voxtral Small 24B 2507 (Amazon Bedrock)", - "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", - "family": "voxtral", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "amazon/amazon.nova-lite-v1:0": { - "id": "amazon/amazon.nova-lite-v1:0", - "name": "Nova Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.015 - } - }, - "amazon/mistral.voxtral-small-24b-2507@us": { - "id": "amazon/mistral.voxtral-small-24b-2507@us", - "name": "Voxtral Small 24B 2507 (Amazon Bedrock, US)", - "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", - "family": "voxtral", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "amazon/google.gemma-3-12b-it@us": { - "id": "amazon/google.gemma-3-12b-it@us", - "name": "Gemma 3 12B IT (Amazon Bedrock, US)", - "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 131072 - }, - "cost": { - "input": 0.09, - "output": 0.29 - } - }, - "amazon/google.gemma-3-27b-it": { - "id": "amazon/google.gemma-3-27b-it", - "name": "Gemma 3 27B IT (Amazon Bedrock)", - "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 131072 - }, - "cost": { - "input": 0.23, - "output": 0.38 - } - }, - "amazon/google.gemma-3-27b-it@us": { - "id": "amazon/google.gemma-3-27b-it@us", - "name": "Gemma 3 27B IT (Amazon Bedrock, US)", - "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 131072 - }, - "cost": { - "input": 0.23, - "output": 0.38 - } - }, - "amazon/mistral.pixtral-large-2502-v1:0": { - "id": "amazon/mistral.pixtral-large-2502-v1:0", - "name": "Pixtral Large (25.02) (Amazon Bedrock)", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-04-08", - "last_updated": "2025-04-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 2, - "output": 6 - } - }, - "amazon/amazon.nova-micro-v1:0@us": { - "id": "amazon/amazon.nova-micro-v1:0@us", - "name": "Nova Micro (US)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 10000 - }, - "cost": { - "input": 0.035, - "output": 0.14, - "cache_read": 0.00875 - } - }, - "amazon/zai.glm-4.7-flash@us": { - "id": "amazon/zai.glm-4.7-flash@us", - "name": "GLM-4.7-Flash (Amazon Bedrock, US)", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 203000, - "output": 131072 - }, - "cost": { - "input": 0.07, - "output": 0.4 - } - }, - "amazon/moonshotai.kimi-k2.5": { - "id": "amazon/moonshotai.kimi-k2.5", - "name": "Kimi K2.5 (Amazon Bedrock)", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 3 - } - }, - "amazon/amazon.nova-pro-v1:0@us": { - "id": "amazon/amazon.nova-pro-v1:0@us", - "name": "Nova Pro (US)", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.8, - "output": 3.2, - "cache_read": 0.2 - } - }, - "amazon/mistral.pixtral-large-2502-v1:0@us": { - "id": "amazon/mistral.pixtral-large-2502-v1:0@us", - "name": "Pixtral Large (25.02) (Amazon Bedrock, US)", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-04-08", - "last_updated": "2025-04-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 2, - "output": 6 - } - }, - "amazon/zai.glm-4.7-flash": { - "id": "amazon/zai.glm-4.7-flash", - "name": "GLM-4.7-Flash (Amazon Bedrock)", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 203000, - "output": 131072 - }, - "cost": { - "input": 0.07, - "output": 0.4 - } - }, - "amazon/moonshot.kimi-k2-thinking": { - "id": "amazon/moonshot.kimi-k2-thinking", - "name": "Kimi K2 Thinking (Amazon Bedrock)", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 2.5 - } - }, - "amazon/google.gemma-3-4b-it@us": { - "id": "amazon/google.gemma-3-4b-it@us", - "name": "Gemma 3 4B IT (Amazon Bedrock, US)", - "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 131072 - }, - "cost": { - "input": 0.04, - "output": 0.08 - } - }, - "amazon/mistral.voxtral-mini-3b-2507": { - "id": "amazon/mistral.voxtral-mini-3b-2507", - "name": "Voxtral Mini 3B 2507 (Amazon Bedrock)", - "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", - "family": "voxtral", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.04, - "output": 0.04 - } - }, - "ovhcloud/gpt-oss-20b": { - "id": "ovhcloud/gpt-oss-20b", - "name": "GPT OSS 20B (OVHcloud)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.05, - "output": 0.18 - } - }, - "ovhcloud/gpt-oss-120b": { - "id": "ovhcloud/gpt-oss-120b", - "name": "GPT OSS 120B (OVHcloud)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.09, - "output": 0.47 - } - }, - "fireworks_ai/gpt-oss-120b": { - "id": "fireworks_ai/gpt-oss-120b", - "name": "GPT OSS 120B (Fireworks AI)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.014 - } - }, - "fireworks_ai/accounts/fireworks/models/deepseek-v4-flash-0731": { - "id": "fireworks_ai/accounts/fireworks/models/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731 (Fireworks AI)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.007 - } - }, - "fireworks_ai/accounts/fireworks/models/muse-glimmer-30b": { - "id": "fireworks_ai/accounts/fireworks/models/muse-glimmer-30b", - "name": "Muse Glimmer 30B (Fireworks AI)", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.35, - "output": 1.5, - "cache_read": 0.04 - } - }, - "fireworks_ai/accounts/fireworks/models/deepseek-v4-pro-0813": { - "id": "fireworks_ai/accounts/fireworks/models/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813 (Fireworks AI)", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 - } - }, - "fireworks_ai/accounts/fireworks/models/inkling": { - "id": "fireworks_ai/accounts/fireworks/models/inkling", - "name": "Inkling (Fireworks AI)", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.17 - } - }, - "mistral/codestral-latest": { - "id": "mistral/codestral-latest", - "name": "Codestral (latest)", - "description": "Mistral code model for completions, refactors, and developer IDE workflows", - "family": "codestral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-05-29", - "last_updated": "2025-01-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 4096 - }, - "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.03 - } - }, - "mistral/mistral-large-latest": { - "id": "mistral/mistral-large-latest", - "name": "Mistral Large (latest)", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 - } - }, - "mistral/mistral-small-latest": { - "id": "mistral/mistral-small-latest", - "name": "Mistral Small (latest)", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 256000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 - } - }, - "mistral/devstral-medium-latest": { - "id": "mistral/devstral-medium-latest", - "name": "Devstral 2 (latest)", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.4, - "output": 2 - } - }, - "mistral/devstral-2512": { - "id": "mistral/devstral-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.04 - } - }, - "mistral/mistral-medium-2505": { - "id": "mistral/mistral-medium-2505", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.4, - "output": 2 - } - }, - "mistral/magistral-medium-latest": { - "id": "mistral/magistral-medium-latest", - "name": "Magistral Medium (latest)", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-20", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 16384 - }, - "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15 - } - }, - "mistral/mistral-small-2603": { - "id": "mistral/mistral-small-2603", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 256000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 - } - }, - "mistral/mistral-large-2512": { - "id": "mistral/mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.05 - } - }, - "mistral/voxtral-small-latest": { - "id": "mistral/voxtral-small-latest", - "name": "Voxtral Small (latest)", - "description": "Instruct model with native audio input for speech understanding and tool use", - "family": "voxtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32000 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 - } - }, - "mistral/mistral-medium-2604": { - "id": "mistral/mistral-medium-2604", - "name": "Mistral Medium 3.5", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15 - } - }, - "mistral/mistral-medium-latest": { - "id": "mistral/mistral-medium-latest", - "name": "Mistral Medium (latest)", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15 - } - }, - "qwen/qwen3-vl-235b-a22b-instruct": { - "id": "qwen/qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 1.6 - } - }, - "qwen/qwq-plus": { - "id": "qwen/qwq-plus", - "name": "QwQ Plus", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-03-05", - "last_updated": "2025-03-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.8, - "output": 2.4 - } - }, - "qwen/qwen3.8-27b": { - "id": "qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1, - "cache_write": 0.625 - } - }, - "qwen/qwen3-vl-235b-a22b-thinking": { - "id": "qwen/qwen3-vl-235b-a22b-thinking", - "name": "Qwen3 VL 235B A22B Thinking", - "description": "Qwen vision-language thinking model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 4 - } - }, - "qwen/qwen3.8-2.4t-a95b": { - "id": "qwen/qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 - } - }, - "qwen/deepseek-v4-flash-0731": { - "id": "qwen/deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731 (Alibaba)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.22, - "output": 0.66, - "cache_read": 0.022 - } - }, - "qwen/qwen3-235b-a22b-instruct-2507": { - "id": "qwen/qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B-A22B Instruct 2507", - "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.23, - "output": 0.92 - } - }, - "qwen/qwen-max": { - "id": "qwen/qwen-max", - "name": "Qwen Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 1.6, - "output": 6.4, - "cache_read": 0.32 - } - }, - "qwen/qwen3-next-80b-a3b-thinking": { - "id": "qwen/qwen3-next-80b-a3b-thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 1.2 - } - }, - "qwen/qwen3.8-max": { - "id": "qwen/qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 - } - }, - "qwen/qwen3-coder-next": { - "id": "qwen/qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 1.5 - } - }, - "qwen/deepseek-v4.1-flash": { - "id": "qwen/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash (Alibaba)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.015 - } - }, - "qwen/qwen3-coder-next@eu": { - "id": "qwen/qwen3-coder-next@eu", - "name": "Qwen3 Coder Next (EU)", - "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 1.5 - } - }, - "qwen/qwen3-max": { - "id": "qwen/qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24, - "cache_write": 1.5 - } - }, - "qwen/qwen3-coder-480b-a35b-instruct": { - "id": "qwen/qwen3-coder-480b-a35b-instruct", - "name": "Qwen3-Coder 480B-A35B Instruct", - "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 7.5 - } - }, - "qwen/qwen3-coder-30b-a3b-instruct": { - "id": "qwen/qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.45, - "output": 2.25 - } - }, - "qwen/qwen-vl-max": { - "id": "qwen/qwen-vl-max", - "name": "Qwen-VL Max", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-08", - "last_updated": "2025-08-13", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.8, - "output": 3.2, - "cache_read": 0.16 - } - }, - "qwen/qwen3-max@eu": { - "id": "qwen/qwen3-max@eu", - "name": "Qwen3 Max (EU)", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 1.2, - "output": 6, - "cache_read": 0.24, - "cache_write": 1.5 - } - }, - "qwen/qwen3-next-80b-a3b-instruct": { - "id": "qwen/qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 1.2 - } - }, - "qwen/qwen-vl-plus": { - "id": "qwen/qwen-vl-plus", - "name": "Qwen-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-08-15", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.21, - "output": 0.63, - "cache_read": 0.042 - } - }, - "qwen/qwen3-coder-flash": { - "id": "qwen/qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 1.5, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "qwen/deepseek-v4-pro-0813": { - "id": "qwen/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813 (Alibaba)", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.66, - "output": 1.98, - "cache_read": 0.066 - } - }, - "qwen/qwen3-coder-plus": { - "id": "qwen/qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.2, - "cache_write": 1.25 - } - }, - "qwen/qwen3.8-flash": { - "id": "qwen/qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016, - "cache_write": 0.2 - } - }, - "qwen/qwen3.8-max-0902": { - "id": "qwen/qwen3.8-max-0902", - "name": "Qwen3.8 Max 0902", - "description": "2026-09-02 upgraded snapshot of Qwen3.8 Max with stronger coding, collaborative agents, and multimodal document understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 - } - }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } - } - }, - "openai/gpt-5.4-pro": { - "id": "openai/gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } - } - }, - "openai/gpt-3.5-turbo": { - "id": "openai/gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16385, - "output": 4096 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } - } - }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "pdf", - "image", - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "openai/gpt-4o": { - "id": "openai/gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "openai/gpt-5-mini": { - "id": "openai/gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "openai/gpt-5.2-pro": { - "id": "openai/gpt-5.2-pro", - "name": "GPT-5.2 Pro", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "image", - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 21, - "output": 168 - } - }, - "openai/o4-mini": { - "id": "openai/o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", - "modalities": { - "input": [ - "image", - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 - } - }, - "openai/o3-mini": { - "id": "openai/o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", - "modalities": { - "input": [ - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 - } - }, - "openai/gpt-4": { - "id": "openai/gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8191, - "output": 8192 - }, - "cost": { - "input": 30, - "output": 60 - } - }, - "openai/gpt-5.3-codex": { - "id": "openai/gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "openai/gpt-4.1-nano": { - "id": "openai/gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "image", - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 - } - }, - "openai/gpt-5-nano": { - "id": "openai/gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 - } - }, - "openai/o1": { - "id": "openai/o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 - } - }, - "openai/gpt-latest": { - "id": "openai/gpt-latest", - "name": "GPT Latest (GPT-6 Astra)", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } - } - }, - "openai/gpt-5-pro": { - "id": "openai/gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", - "modalities": { - "input": [ - "image", - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 272000 - }, - "cost": { - "input": 15, - "output": 120 - } - }, - "openai/gpt-4o-2024-08-06": { - "id": "openai/gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-08-06", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "openai/gpt-6-astra": { - "id": "openai/gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } - } - }, - "openai/gpt-5.1": { - "id": "openai/gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "image", - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 - } - }, - "openai/o3-pro": { - "id": "openai/o3-pro", - "name": "o3-pro", - "description": "High-effort o3 tier for difficult technical reasoning and careful answers", - "family": "o-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-06-10", - "last_updated": "2025-06-10", - "modalities": { - "input": [ - "text", - "pdf", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 20, - "output": 80 - } - }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "pdf", - "image", - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } - } - }, - "openai/gpt-5.2": { - "id": "openai/gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "pdf", - "image", - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "openai/gpt-4.1": { - "id": "openai/gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "openai/gpt-4o-2024-11-20": { - "id": "openai/gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "openai/gpt-mini-latest": { - "id": "openai/gpt-mini-latest", - "name": "GPT Mini Latest (GPT-5.4 mini)", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "pdf", - "image", - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 - } - }, - "openai/gpt-6-luna": { - "id": "openai/gpt-6-luna", - "name": "GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25 - } - } - }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } - } - }, - "openai/gpt-4-turbo": { - "id": "openai/gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 10, - "output": 30 - } - }, - "openai/o3": { - "id": "openai/o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "openai/gpt-5": { - "id": "openai/gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10 - } - } - }, - "openai/gpt-6-sol": { - "id": "openai/gpt-6-sol", - "name": "GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } - } - }, - "openai/o1-pro": { - "id": "openai/o1-pro", - "name": "o1-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2025-03-19", - "last_updated": "2025-03-19", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 150, - "output": 600 - } - }, - "openai/gpt-pro-latest": { - "id": "openai/gpt-pro-latest", - "name": "GPT Pro Latest (GPT-5.5 Pro)", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } - } - }, - "infomaniak/mistralai/Ministral-3-14B-Instruct-2512": { - "id": "infomaniak/mistralai/Ministral-3-14B-Instruct-2512", - "name": "Ministral 3 14B (Infomaniak)", - "description": "Open vision-language model for efficient local deployment, instruction following, and tool use", - "family": "ministral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 100000, - "output": 262144 - }, - "cost": { - "input": 0.34233, - "output": 0.45644 - } - }, - "moonshot/kimi-k3": { - "id": "moonshot/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "moonshot/kimi-k2.7-code-highspeed": { - "id": "moonshot/kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 - } - }, - "moonshot/kimi-k2.6": { - "id": "moonshot/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 - } - }, - "moonshot/kimi-k2.7-code": { - "id": "moonshot/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - } - } - }, - "stepfun": { - "id": "stepfun", - "env": [ - "STEPFUN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.stepfun.com/v1", - "name": "StepFun (China)", - "doc": "https://platform.stepfun.com/docs/zh/overview/concept", - "models": { - "step-1-32k": { - "id": "step-1-32k", - "name": "Step 1 (32K)", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-01-01", - "last_updated": "2026-02-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "input": 32768, - "output": 32768 - }, - "cost": { - "input": 2.05, - "output": 9.59, - "cache_read": 0.41 - } - }, - "step-tts-2": { - "id": "step-tts-2", - "name": "Step TTS 2", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "step", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-03-01", - "last_updated": "2026-07-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "step-5-preview": { - "id": "step-5-preview", - "name": "Step 5 Preview", - "description": "StepFun's next-generation flagship base model for coding and professional knowledge work, with native text, image, and video input and a 1M-token context window", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "release_date": "2026-09-16", - "last_updated": "2026-09-20", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "input": 1000000, - "output": 1000000 - }, - "cost": { - "input": 0.959, - "output": 2.741, - "cache_read": 0.048 - } - }, - "step-3.5-flash-2603": { - "id": "step-3.5-flash-2603", - "name": "Step 3.5 Flash 2603", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 256000, - "output": 256000 - }, - "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 - } - }, - "step-2-16k": { - "id": "step-2-16k", - "name": "Step 2 (16K)", - "description": "StepFun flash model for efficient multimodal reasoning, coding, and tool use", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-01-01", - "last_updated": "2026-02-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16384, - "input": 16384, - "output": 8192 - }, - "cost": { - "input": 5.21, - "output": 16.44, - "cache_read": 1.04 - } - }, - "stepaudio-2.5-tts": { - "id": "stepaudio-2.5-tts", - "name": "StepAudio 2.5 TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "step", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-16", - "last_updated": "2026-07-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "stepaudio-2.5-asr": { - "id": "stepaudio-2.5-asr", - "name": "StepAudio 2.5 ASR", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "step", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-24", - "last_updated": "2026-07-02", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "step-3.5-flash": { - "id": "step-3.5-flash", - "name": "Step 3.5 Flash", - "description": "StepFun flash lane for quick multimodal reasoning and coding assistance", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-29", - "last_updated": "2026-06-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 256000, - "output": 256000 - }, - "cost": { - "input": 0.1, - "output": 0.3, - "cache_read": 0.02 - } - }, - "step-3.7-flash": { - "id": "step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-06-29", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 256000, - "output": 256000 - }, - "cost": { - "input": 0.185, - "output": 1.11, - "cache_read": 0.037 - } - } - } - }, - "hpc-ai": { - "id": "hpc-ai", - "env": [ - "HPC_AI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.hpc-ai.com/inference/v1", - "name": "HPC-AI", - "doc": "https://www.hpc-ai.com/doc/docs/quickstart/", - "models": { - "anthropic/claude-opus-4.7": { - "id": "anthropic/claude-opus-4.7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5 - } - }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1002000, - "output": 128000 - }, - "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.145 - } - }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 128000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 - } - }, - "moonshotai/kimi-k2.5": { - "id": "moonshotai/kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 - } - }, - "moonshotai/kimi-k2.7-code": { - "id": "moonshotai/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - }, - "zai-org/glm-5.2": { - "id": "zai-org/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "zai-org/glm-5.1": { - "id": "zai-org/glm-5.1", - "name": "GLM 5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202000, - "output": 202000 - }, - "cost": { - "input": 0.615, - "output": 2.46, - "cache_read": 0.133 - } - }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196000, - "output": 195000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 - } - }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - } - } - }, - "v0": { - "id": "v0", - "env": [ - "V0_API_KEY" - ], - "npm": "@ai-sdk/vercel", - "name": "v0", - "doc": "https://sdk.vercel.ai/providers/ai-sdk-providers/vercel", - "models": { - "v0-1.5-lg": { - "id": "v0-1.5-lg", - "name": "v0-1.5-lg", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "v0", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-06-09", - "last_updated": "2025-06-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 512000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75 - } - }, - "v0-1.5-md": { - "id": "v0-1.5-md", - "name": "v0-1.5-md", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "v0", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-06-09", - "last_updated": "2025-06-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 32000 - }, - "cost": { - "input": 3, - "output": 15 - } - }, - "v0-1.0-md": { - "id": "v0-1.0-md", - "name": "v0-1.0-md", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "v0", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 32000 - }, - "cost": { - "input": 3, - "output": 15 - } - } - } - }, - "tencent-coding-plan": { - "id": "tencent-coding-plan", - "env": [ - "TENCENT_CODING_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.lkeap.cloud.tencent.com/coding/v3", - "name": "Tencent Coding Plan (China)", - "doc": "https://cloud.tencent.com/document/product/1772/128947", - "models": { - "hunyuan-2.0-thinking": { - "id": "hunyuan-2.0-thinking", - "name": "Tencent HY 2.0 Think", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-03-08", - "last_updated": "2026-03-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 202752, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "hunyuan-2.0-instruct": { - "id": "hunyuan-2.0-instruct", - "name": "Tencent HY 2.0 Instruct", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-03-08", - "last_updated": "2026-03-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi-K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "tc-code-latest": { - "id": "tc-code-latest", - "name": "Auto", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-03-08", - "last_updated": "2026-03-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "hunyuan-turbos": { - "id": "hunyuan-turbos", - "name": "Hunyuan-TurboS", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-03-08", - "last_updated": "2026-03-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "hunyuan-t1": { - "id": "hunyuan-t1", - "name": "Hunyuan-T1", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hunyuan", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-03-08", - "last_updated": "2026-03-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - } - } - }, - "tempr": { - "id": "tempr", - "env": [ - "TEMPR_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.temprhq.io/v1", - "name": "Tempr", - "doc": "https://temprhq.io/docs/gateway-reference.html", - "models": { - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "anthropic/claude-opus-4-5": { - "id": "anthropic/claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-sonnet-4-5": { - "id": "anthropic/claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic/claude-fable-5-1": { - "id": "anthropic/claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "anthropic/claude-opus-4-5-20251101": { - "id": "anthropic/claude-opus-4-5-20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "anthropic/claude-opus-4-8": { - "id": "anthropic/claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-sonnet-4-5-20250929": { - "id": "anthropic/claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "anthropic/claude-opus-4-6": { - "id": "anthropic/claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-haiku-4-5-20251001": { - "id": "anthropic/claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "google/gemini-flash-latest": { - "id": "google/gemini-flash-latest", - "name": "Gemini Flash Latest", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "input_audio": 0.75 - } - }, - "google/gemma-4-31b-it": { - "id": "google/gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - } - }, - "google/gemini-flash-lite-latest": { - "id": "google/gemini-flash-lite-latest", - "name": "Gemini Flash-Lite Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 - } - }, - "google/gemini-embedding-2": { - "id": "google/gemini-embedding-2", - "name": "Gemini Embedding 2", - "description": "Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space", - "family": "gemini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-11", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 1 - }, - "cost": { - "input": 0.2, - "output": 0 - } - }, - "google/gemini-3.6-flash": { - "id": "google/gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "input_audio": 0.75 - } - }, - "google/gemma-4-26b-a4b-it": { - "id": "google/gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - } - }, - "google/gemini-3.5-flash-lite": { - "id": "google/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 - } - }, - "google/gemini-3.1-pro-preview": { - "id": "google/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "google/gemini-3.5-flash": { - "id": "google/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "google/gemini-3.7-flash": { - "id": "google/gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "input_audio": 0.75 - } - }, - "google/gemini-3.1-pro-preview-customtools": { - "id": "google/gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "google/gemini-3-flash-preview": { - "id": "google/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 - } - }, - "google/gemini-3.8-flash": { - "id": "google/gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "input_audio": 0.75 - } - }, - "google/gemini-embedding-001": { - "id": "google/gemini-embedding-001", - "name": "Gemini Embedding 001", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "gemini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-05", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2048, - "output": 1 - }, - "cost": { - "input": 0.15, - "output": 0 - } - }, - "google/gemini-3.1-flash-lite": { - "id": "google/gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 - } - } - } - }, - "inception": { - "id": "inception", - "env": [ - "INCEPTION_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.inceptionlabs.ai/v1/", - "name": "Inception", - "doc": "https://docs.inceptionlabs.ai/get-started/models", - "models": { - "mercury-edit-2": { - "id": "mercury-edit-2", - "name": "Mercury Edit 2", - "description": "Code editing dLLM for autocomplete (FIM) and next-edit suggestions", - "family": "mercury", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-03-30", - "last_updated": "2026-03-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 8192 - }, - "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.025 - } - }, - "mercury-2.5": { - "id": "mercury-2.5", - "name": "Mercury 2.5", - "description": "Mercury 2.5 is the fastest reasoning LLM, and the latest diffusion LLM (dLLM) from Inception", - "family": "mercury", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-11-01", - "release_date": "2026-09-08", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 260000, - "output": 65536 - }, - "cost": { - "input": 0.04, - "output": 0.15, - "cache_read": 0.004 - } - }, - "mercury-2": { - "id": "mercury-2", - "name": "Mercury 2", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "mercury", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01-01", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 50000 - }, - "cost": { - "input": 0.25, - "output": 0.75, - "cache_read": 0.025 - } - } - } - }, - "modelis": { - "id": "modelis", - "env": [ - "MODELIS_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://modelishub.com/v1", - "name": "Modelis", - "doc": "https://modelishub.com/pricing", - "models": { - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 10 - } - }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5 - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50 - } - }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - }, - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25 - } - }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.435, - "output": 0.87 - } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15 - } - }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.0983, - "output": 0.1966 - } - }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 3, - "output": 9 - } - }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.768, - "output": 3.072 - } - } - } - }, - "opencode": { - "id": "opencode", - "env": [ - "OPENCODE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://opencode.ai/zen/v1", - "name": "OpenCode Zen", - "doc": "https://opencode.ai/docs/zen", - "models": { - "ling-3.0-flash-fin-free": { - "id": "ling-3.0-flash-fin-free", - "name": "Ling 3.0 Flash Fin Free", - "description": "Finance-enhanced model for financial research, multi-step investment workflows, and long-horizon planning and execution", - "family": "ling", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } - } - }, - "qwen3.6-plus-free": { - "id": "qwen3.6-plus-free", - "name": "Qwen3.6 Plus Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "qwen-free", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 30, - "output": 180, - "cache_read": 30 - } - }, - "mimo-v2-pro-free": { - "id": "mimo-v2-pro-free", - "name": "MiMo V2 Pro Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo-pro-free", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 64000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "muse-spark-1.2-contributor-free": { - "id": "muse-spark-1.2-contributor-free", - "name": "Muse Spark 1.2 Free", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse-free", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "muse-spark-1.3": { - "id": "muse-spark-1.3", - "name": "Muse Spark 1.3", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 - } - }, - "glm-5-free": { - "id": "glm-5-free", - "name": "GLM-5 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "glm-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-11", - "last_updated": "2026-02-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "trinity-large-preview-free": { - "id": "trinity-large-preview-free", - "name": "Trinity Large Preview", - "description": "Legacy model retained for compatibility with older integrations", - "family": "trinity", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-01-27", - "last_updated": "2026-01-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-5.5-pro": { - "id": "gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 30, - "output": 180, - "cache_read": 30 - } - }, - "grok-4.7": { - "id": "grok-4.7", - "name": "Grok 4.7 (30% Off)", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.4, - "output": 4.2, - "cache_read": 0.35, - "tiers": [ - { - "input": 2.8, - "output": 8.4, - "cache_read": 0.7, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.8, - "output": 8.4, - "cache_read": 0.7 - } - } - }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-14", - "last_updated": "2026-01-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "ling-3.0-tiny-free": { - "id": "ling-3.0-tiny-free", - "name": "Ling-3.0-tiny Free", - "description": "Compact MoE model for responsive agents, instruction following, and multi-turn conversations", - "family": "ling", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-06", - "last_updated": "2026-08-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.07, - "output": 8.5, - "cache_read": 0.107 - } - }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 - } - }, - "kimi-k2.5-free": { - "id": "kimi-k2.5-free", - "name": "Kimi K2.5 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-free", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "glm-4.7-free": { - "id": "glm-4.7-free", - "name": "GLM-4.7 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "glm-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "glm-4.6": { - "id": "glm-4.6", - "name": "GLM-4.6", - "description": "Legacy model retained for compatibility with older integrations", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.1 - } - }, - "laguna-s-2.1-free": { - "id": "laguna-s-2.1-free", - "name": "Laguna S 2.1 Free", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 32000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "GPT-5 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.07, - "output": 8.5, - "cache_read": 0.107 - } - }, - "qwen3-coder": { - "id": "qwen3-coder", - "name": "Qwen3 Coder", - "description": "Legacy model retained for compatibility with older integrations", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "status": "deprecated", - "cost": { - "input": 0.45, - "output": 1.8 - } - }, - "minimax-m3-free": { - "id": "minimax-m3-free", - "name": "MiniMax-M3 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax-m3-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-31", - "last_updated": "2026-05-31", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.5", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 - } - }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-11", - "last_updated": "2026-02-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 - } - }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.006 - } - }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 - } - }, - "deepseek-v4-flash-vision-exp": { - "id": "deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 - } - }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 - } - }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } - } - }, - "claude-opus-5-5": { - "id": "claude-opus-5-5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 - } - }, - "claude-fable-5-1": { - "id": "claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "gemini-3.6-flash": { - "id": "gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google" - }, - "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2.5, - "cache_read": 0.4 - } - }, - "minimax-m2.5-free": { - "id": "minimax-m2.5-free", - "name": "MiniMax-M2.5 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "gpt-5.3-codex-spark": { - "id": "gpt-5.3-codex-spark", - "name": "GPT-5.3 Codex Spark", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex-spark", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "input": 128000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google" - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 - } - }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } - } - }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 0.6, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 0.6 - } - } - }, - "ring-2.6-1t-free": { - "id": "ring-2.6-1t-free", - "name": "Ring 2.6 1T Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "ring-1t-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-05-08", - "last_updated": "2026-05-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 66000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0 - } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.08 - } - }, - "longcat-2.0-free": { - "id": "longcat-2.0-free", - "name": "LongCat-2.0 Free", - "description": "Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window", - "family": "longcat", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.07, - "output": 8.5, - "cache_read": 0.107 - } - }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "claude-3-5-haiku": { - "id": "claude-3-5-haiku", - "name": "Claude Haiku 3.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-haiku", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07-31", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8192 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 - } - }, - "hy3-preview-free": { - "id": "hy3-preview-free", - "name": "Hy3 preview Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "hy3-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 64000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "gemini-3-flash": { - "id": "gemini-3-flash", - "name": "Gemini 3 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google" - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 - } - }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google" - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "space-bunny-free": { - "id": "space-bunny-free", - "name": "Space Bunny Free", - "description": "Anonymous preview reasoning model for coding, agentic tasks, tool use, and multimodal input", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "release_date": "2026-09-23", - "last_updated": "2026-09-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "input": 524288, - "output": 524288 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "nemotron-3-super-free": { - "id": "nemotron-3-super-free", - "name": "Nemotron 3 Super Free", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-03-11", - "last_updated": "2026-03-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 128000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "minimax-m2.1-free": { - "id": "minimax-m2.1-free", - "name": "MiniMax-M2.1 Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "mimo-v2.6-flash-free": { - "id": "mimo-v2.6-flash-free", - "name": "MiMo-V2.6-Flash Free", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "gemini-3-pro": { - "id": "gemini-3-pro", - "name": "Gemini 3 Pro", - "description": "Legacy model retained for compatibility with older integrations", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/google" - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "nemotron-3-ultra-free": { - "id": "nemotron-3-ultra-free", - "name": "Nemotron 3 Ultra Free", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2026-02", - "release_date": "2026-06-04", - "last_updated": "2026-06-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "claude-sonnet-4": { - "id": "claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } - } - }, - "muse-spark-1.2": { - "id": "muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 - } - }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Legacy model retained for compatibility with older integrations", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.1 - } - }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 512000, - "output": 128000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } - } - }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016, - "cache_write": 0.2 - } - }, - "gpt-5.1-codex-max": { - "id": "gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "mimo-v2.5-free": { - "id": "mimo-v2.5-free", - "name": "MiMo V2.5 Free", - "description": "MiMo omni model for text, image, video, audio, and agents", - "family": "mimo-v2.5-free", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "grok-code": { - "id": "grok-code", - "name": "Grok Code Fast 1", - "description": "Legacy model retained for compatibility with older integrations", - "family": "grok", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-20", - "last_updated": "2025-08-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "ling-2.6-flash-free": { - "id": "ling-2.6-flash-free", - "name": "Ling 2.6 Flash Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "ling-flash-free", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262100, - "output": 32800 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0 - } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "deepseek-v4-flash-free": { - "id": "deepseek-v4-flash-free", - "name": "DeepSeek V4 Flash Free", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "hy3-free": { - "id": "hy3-free", - "name": "Hy3 Free", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "hy3-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 190000, - "input": 192000, - "output": 64000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "nemotron-3.5-lightning-free": { - "id": "nemotron-3.5-lightning-free", - "name": "Nemotron 3.5 Lightning Free", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google" - }, - "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "kimi-k2": { - "id": "kimi-k2", - "name": "Kimi K2", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2.5, - "cache_read": 0.4 - } - }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "mimo-v2-omni-free": { - "id": "mimo-v2-omni-free", - "name": "MiMo V2 Omni Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo-omni-free", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 64000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "grok-build-0.1": { - "id": "grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2 - } - }, - "ling-3.0-flash-free": { - "id": "ling-3.0-flash-free", - "name": "Ling-3.0-flash Free", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-23", - "last_updated": "2026-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "gemini-3.8-flash": { - "id": "gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google" - }, - "cost": { - "input": 1.5, - "output": 7.5, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "north-mini-code-free": { - "id": "north-mini-code-free", - "name": "North Mini Code Free", - "description": "Cohere coding model for practical software engineering and agentic edits", - "family": "north-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09-23", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 64000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-6-luna": { - "id": "gpt-6-luna", - "name": "GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25 - } - } - }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.74, - "output": 3.84, - "cache_read": 0.145 - } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-02-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "big-pickle": { - "id": "big-pickle", - "name": "Big Pickle", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "big-pickle", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-10-17", - "last_updated": "2025-10-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "input": 160000, - "output": 32000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625 - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 6.25 - } - } - }, - "mimo-v2-flash-free": { - "id": "mimo-v2-flash-free", - "name": "MiMo V2 Flash Free", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mimo-flash-free", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-16", - "last_updated": "2025-12-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex Mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 1.07, - "output": 8.5, - "cache_read": 0.107 - } - }, - "grok-4.6": { - "id": "grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } - } - }, - "gemini-3.1-pro": { - "id": "gemini-3.1-pro", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/google" - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "minimax-m2.1": { - "id": "minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.1 - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10 - } - } - }, - "x-preview-f-free": { - "id": "x-preview-f-free", - "name": "Ox Alpha Free (Unlimited)", - "description": "Stealth reasoning model for coding, agentic tasks, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-08-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "muse-spark-1.3-contributor-free": { - "id": "muse-spark-1.3-contributor-free", - "name": "Muse Spark 1.3 Free", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for coding and agentic workflows.", - "family": "muse-free", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0 - } - }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 - } - }, - "gpt-6-sol": { - "id": "gpt-6-sol", - "name": "GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai" - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } - } - } - } - }, - "kenari": { - "id": "kenari", - "env": [ - "KENARI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://kenari.id/v1", - "name": "Kenari", - "doc": "https://kenari.id/docs", - "models": { - "gpt-5-4-mini": { - "id": "gpt-5-4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-5-6-luna": { - "id": "gpt-5-6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gemma-4-31b-it": { - "id": "gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "grok-4-5": { - "id": "grok-4-5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gemini-2-5-flash": { - "id": "gemini-2-5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nemotron-3-nano-30b-a3b": { - "id": "nemotron-3-nano-30b-a3b", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "glm-5-3-flash": { - "id": "glm-5-3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "kimi-k2-7-code": { - "id": "kimi-k2-7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-5-6-terra": { - "id": "gpt-5-6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gemini-2-5-flash-lite": { - "id": "gemini-2-5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mimo-v2-5": { - "id": "mimo-v2-5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "grok-build-0-1": { - "id": "grok-build-0-1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "glm-4-7-flash:free": { - "id": "glm-4-7-flash:free", - "name": "GLM-4.7-Flash (Free)", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-image-2": { - "id": "gpt-image-2", - "name": "GPT-Image-2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 272000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen3-8-max": { - "id": "qwen3-8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "minimax-m2-7-highspeed": { - "id": "minimax-m2-7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mimo-v2-5-pro": { - "id": "mimo-v2-5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nemotron-3-super-120b-a12b": { - "id": "nemotron-3-super-120b-a12b", - "name": "Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gemini-3-7-flash": { - "id": "gemini-3-7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "step-3-7-flash:free": { - "id": "step-3-7-flash:free", - "name": "Step 3.7 Flash (Free)", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 256000, - "output": 256000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "kimi-k2-6:free": { - "id": "kimi-k2-6:free", - "name": "Kimi K2.6 (Free)", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-5-6-sol": { - "id": "gpt-5-6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nemotron-3-ultra-550b-a55b": { - "id": "nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gemini-3-1-pro": { - "id": "gemini-3-1-pro", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "glm-5-2": { - "id": "glm-5-2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistral-large:free": { - "id": "mistral-large:free", - "name": "Mistral Large (Free)", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mimo-v2-5:free": { - "id": "mimo-v2-5:free", - "name": "MiMo-V2.5 (Free)", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "minimax-m2-7": { - "id": "minimax-m2-7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "hy3:free": { - "id": "hy3:free", - "name": "Hy3 (Free)", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 192000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "kimi-k2-6": { - "id": "kimi-k2-6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistral-medium-3-5:free": { - "id": "mistral-medium-3-5:free", - "name": "Mistral Medium 3.5 (Free)", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 512000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "glm-5-3": { - "id": "glm-5-3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "deepseek-v4-1-flash": { - "id": "deepseek-v4-1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen3-7-plus": { - "id": "qwen3-7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gemini-3-1-flash-lite": { - "id": "gemini-3-1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "hy3": { - "id": "hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 192000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "glm-5-1": { - "id": "glm-5-1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nemotron-3-super-120b-a12b:free": { - "id": "nemotron-3-super-120b-a12b:free", - "name": "Nemotron 3 Super 120B A12B (Free)", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-5-5": { - "id": "gpt-5-5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "grok-imagine-image-2-0": { - "id": "grok-imagine-image-2-0", - "name": "Grok Imagine Image 2.0", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-08-07", - "last_updated": "2026-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 8000, - "output": 0 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "deepseek-v4-flash:free": { - "id": "deepseek-v4-flash:free", - "name": "DeepSeek V4 Flash (Free)", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "grok-4-6": { - "id": "grok-4-6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "whisper-large-v3-turbo": { - "id": "whisper-large-v3-turbo", - "name": "Whisper Large v3 Turbo", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 448, - "output": 448 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gemini-3-6-flash": { - "id": "gemini-3-6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gemini-3-1-flash-tts": { - "id": "gemini-3-1-flash-tts", - "name": "Gemini 3.1 Flash TTS Preview", - "description": "Low-latency speech generation with steerable prompts and expressive audio tags", - "family": "gemini-flash", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-15", - "last_updated": "2026-04-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "kimi-k2-7-code:free": { - "id": "kimi-k2-7-code:free", - "name": "Kimi K2.7 Code (Free)", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gemini-3-5-flash": { - "id": "gemini-3-5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "step-3-7-flash": { - "id": "step-3-7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 256000, - "output": 256000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0, - "output": 0 - } - } - } - }, - "kimi-code-plan-global": { - "id": "kimi-code-plan-global", - "env": [ - "KIMI_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.kimi.ai/coding/v1", - "name": "Kimi For Coding (kimi.ai)", - "doc": "https://www.kimi.ai/code/docs/en/kimi-code/models.html", - "models": { - "kimi-for-coding-highspeed": { - "id": "kimi-for-coding-highspeed", - "name": "Kimi For Coding HighSpeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "kimi-for-coding": { - "id": "kimi-for-coding", - "name": "kimi-for-coding", - "description": "Kimi coding model with more efficient thinking and up to 1M context, available through Kimi Code", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "k3-256k": { - "id": "k3-256k", - "name": "Kimi K3-256K", - "description": "256K-context version of Kimi K3, reducing token consumption for shorter coding sessions", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "k3": { - "id": "k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - } - } - }, - "trustedrouter": { - "id": "trustedrouter", - "env": [ - "TRUSTEDROUTER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.trustedrouter.com/v1", - "name": "TrustedRouter", - "doc": "https://trustedrouter.com/docs", - "models": { - "trustedrouter/cheap": { - "id": "trustedrouter/cheap", - "name": "Cheap", - "description": "TrustedRouter low-cost routing alias that prefers inexpensive healthy model endpoints.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-01", - "last_updated": "2026-06-27", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - } - }, - "trustedrouter/synth-code": { - "id": "trustedrouter/synth-code", - "name": "Synth Code", - "description": "TrustedRouter code synthesis orchestration alias that combines multiple model responses into one answer.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-20", - "last_updated": "2026-06-27", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - } - }, - "trustedrouter/e2e": { - "id": "trustedrouter/e2e", - "name": "End-to-End Encrypted", - "description": "TrustedRouter privacy routing alias for end-to-end encrypted provider routes where available.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-27", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - } - }, - "trustedrouter/zdr": { - "id": "trustedrouter/zdr", - "name": "Zero Data Retention", - "description": "TrustedRouter privacy routing alias that prefers zero data retention model endpoints.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-27", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - } - }, - "trustedrouter/fast": { - "id": "trustedrouter/fast", - "name": "Fast", - "description": "TrustedRouter speed routing alias that prefers low-latency healthy model endpoints.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-27", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - } - }, - "trustedrouter/synth": { - "id": "trustedrouter/synth", - "name": "Synth", - "description": "TrustedRouter synthesis orchestration alias that combines multiple model responses into one answer.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-20", - "last_updated": "2026-06-27", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - } - }, - "trustedrouter/auto": { - "id": "trustedrouter/auto", - "name": "Auto", - "description": "TrustedRouter automatic routing alias that chooses a healthy supported model endpoint for the request.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-01", - "last_updated": "2026-06-27", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - } - } - } - }, - "wafer.ai": { - "id": "wafer.ai", - "env": [ - "WAFER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://pass.wafer.ai/v1", - "name": "Wafer", - "doc": "https://docs.wafer.ai/wafer-pass", - "models": { - "Kimi-K2.6": { - "id": "Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Kimi K2.6 sparse MoE model with a 262K context window. Available serverless and not included in standard Wafer Pass. Non-ZDR only: requests with `Wafer-ZDR: required` are rejected.", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 1.14, - "output": 4.8, - "cache_read": 0.19, - "cache_write": 0 - } - }, - "MiniMax-M3": { - "id": "MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 512000 - }, - "cost": { - "input": 0.33, - "output": 1.32, - "cache_read": 0.07, - "cache_write": 0, - "tiers": [ - { - "input": 0.66, - "output": 2.64, - "cache_read": 0.13, - "cache_write": 0, - "tier": { - "type": "context", - "size": 512000 - } - } - ], - "context_over_200k": { - "input": 0.66, - "output": 2.64, - "cache_read": 0.13, - "cache_write": 0 - } - } - }, - "GLM-5.1": { - "id": "GLM-5.1", - "name": "GLM-5.1", - "description": "General Language Model 5.1 — high-quality bilingual (EN/ZH) generation with strong coding and reasoning capabilities.", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-07", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.1, - "cache_write": 0 - } - }, - "GLM-5.2": { - "id": "GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4.1, - "cache_read": 0.2, - "cache_write": 0 - } - }, - "glm5.2-fast": { - "id": "glm5.2-fast", - "name": "GLM5.2-Fast", - "description": "The same model served for high TPS.", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 10.25, - "cache_read": 0.5, - "cache_write": 0 - } - } - } - }, - "zhipuai": { - "id": "zhipuai", - "env": [ - "ZHIPU_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://open.bigmodel.cn/api/paas/v4", - "name": "Zhipu AI", - "doc": "https://docs.z.ai/guides/overview/pricing", - "models": { - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 5, - "output": 22, - "cache_read": 1.2, - "cache_write": 0 - } - }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03, - "cache_write": 0 - } - }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0 - } - }, - "glm-5.3-flashx": { - "id": "glm-5.3-flashx", - "name": "GLM-5.3-FlashX", - "description": "High-speed GLM-5.3-Flash serving option for coding and agent workflows", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-18", - "last_updated": "2026-09-18", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.37, - "output": 1.25, - "cache_read": 0.075, - "cache_write": 0 - } - }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 - } - }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 - } - }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26, - "cache_write": 0 - } - }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 98304 - }, - "cost": { - "input": 0.2, - "output": 1.1, - "cache_read": 0.03, - "cache_write": 0 - } - }, - "glm-4.7-flashx": { - "id": "glm-4.7-flashx", - "name": "GLM-4.7-FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0 - } - }, - "glm-4.5-flash": { - "id": "glm-4.5-flash", - "name": "GLM-4.5-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 98304 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 - } - }, - "glm-4.7-flash": { - "id": "glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "glm-4.5v": { - "id": "glm-4.5v", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 64000, - "output": 16384 - }, - "cost": { - "input": 0.6, - "output": 1.8 - } - }, - "glm-4.6v-flash": { - "id": "glm-4.6v-flash", - "name": "GLM-4.6V-Flash", - "description": "Lightweight GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "glm-4.6": { - "id": "glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 - } - }, - "glm-4.5": { - "id": "glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 98304 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 - } - }, - "glm-4.6v": { - "id": "glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 0.9 - } - } - } - }, - "lynkr": { - "id": "lynkr", - "env": [ - "LYNKR_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "http://127.0.0.1:8081/v1", - "name": "Lynkr", - "doc": "https://github.com/Fast-Editor/Lynkr", - "models": { - "lynkr-auto": { - "id": "lynkr-auto", - "name": "Lynkr Auto (complexity routing)", - "description": "Virtual model: Lynkr scores each request on complexity and routes it to the tier model the user configured (local Ollama/llama.cpp for simple requests, configured cloud providers for complex ones).", - "family": "auto", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2026-07-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - } - } - }, - "meganova": { - "id": "meganova", - "env": [ - "MEGANOVA_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.meganova.ai/v1", - "name": "Meganova", - "doc": "https://docs.meganova.ai", - "models": { - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "XiaomiMiMo/MiMo-V2-Flash": { - "id": "XiaomiMiMo/MiMo-V2-Flash", - "name": "MiMo V2 Flash", - "description": "MiMo flash model for fast multimodal assistance and agent workflows", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32000 - }, - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "Qwen/Qwen3.5-Plus": { - "id": "Qwen/Qwen3.5-Plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02", - "last_updated": "2026-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.4, - "output": 2.4, - "reasoning": 2.4 - } - }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.09, - "output": 0.6 - } - }, - "Qwen/Qwen2.5-VL-32B-Instruct": { - "id": "Qwen/Qwen2.5-VL-32B-Instruct", - "name": "Qwen2.5 VL 32B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-03-24", - "last_updated": "2025-03-24", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 16384, - "output": 16384 - }, - "cost": { - "input": 0.2, - "output": 0.6 - } - }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-25", - "last_updated": "2025-08-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 164000, - "output": 164000 - }, - "cost": { - "input": 0.27, - "output": 1 - } - }, - "deepseek-ai/DeepSeek-V3.2": { - "id": "deepseek-ai/DeepSeek-V3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 164000, - "output": 164000 - }, - "cost": { - "input": 0.26, - "output": 0.38 - } - }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 64000 - }, - "cost": { - "input": 0.5, - "output": 2.15 - } - }, - "deepseek-ai/DeepSeek-V3-0324": { - "id": "deepseek-ai/DeepSeek-V3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-03-24", - "last_updated": "2025-03-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 163840 - }, - "cost": { - "input": 0.25, - "output": 0.88 - } - }, - "deepseek-ai/DeepSeek-V3.2-Exp": { - "id": "deepseek-ai/DeepSeek-V3.2-Exp", - "name": "DeepSeek V3.2 Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-10", - "last_updated": "2025-10-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 164000, - "output": 164000 - }, - "cost": { - "input": 0.27, - "output": 0.4 - } - }, - "MiniMaxAI/MiniMax-M2.1": { - "id": "MiniMaxAI/MiniMax-M2.1", - "name": "MiniMax M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 131072 - }, - "cost": { - "input": 0.28, - "output": 1.2 - } - }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "mistralai/Mistral-Nemo-Instruct-2407": { - "id": "mistralai/Mistral-Nemo-Instruct-2407", - "name": "Mistral Nemo Instruct 2407", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 65536 - }, - "cost": { - "input": 0.02, - "output": 0.04 - } - }, - "mistralai/Mistral-Small-3.2-24B-Instruct-2506": { - "id": "mistralai/Mistral-Small-3.2-24B-Instruct-2506", - "name": "Mistral Small 3.2 24B Instruct", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.45, - "output": 2.8 - } - }, - "moonshotai/Kimi-K2-Thinking": { - "id": "moonshotai/Kimi-K2-Thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 2.6 - } - }, - "zai-org/GLM-4.7": { - "id": "zai-org/GLM-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 0.2, - "output": 0.8 - } - }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 0.8, - "output": 2.56 - } - }, - "zai-org/GLM-4.6": { - "id": "zai-org/GLM-4.6", - "name": "GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 0.45, - "output": 1.9 - } - } - } - }, - "ovhcloud": { - "id": "ovhcloud", - "env": [ - "OVHCLOUD_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://oai.endpoints.kepler.ai.cloud.ovh.net/v1", - "name": "OVHcloud AI Endpoints", - "doc": "https://www.ovhcloud.com/en/public-cloud/ai-endpoints/catalog//", - "models": { - "qwen3.8-27b": { - "id": "qwen3.8-27b", - "name": "Qwen3.8-27B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.47, - "output": 3.19 - } - }, - "qwen3guard-gen-8b": { - "id": "qwen3guard-gen-8b", - "name": "Qwen3Guard-Gen-8B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-22", - "last_updated": "2026-01-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mistral-7b-instruct-v0.3": { - "id": "mistral-7b-instruct-v0.3", - "name": "Mistral-7B-Instruct-v0.3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-01", - "last_updated": "2025-04-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 65536, - "output": 65536 - }, - "cost": { - "input": 0.11, - "output": 0.11 - } - }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder-30B-A3B-Instruct", - "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.07, - "output": 0.26 - } - }, - "qwen3guard-gen-0.6b": { - "id": "qwen3guard-gen-0.6b", - "name": "Qwen3Guard-Gen-0.6B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-01-22", - "last_updated": "2026-01-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "gpt-oss-20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2025-08-28", - "last_updated": "2025-08-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.05, - "output": 0.18 - } - }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5-9B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.12, - "output": 0.18 - } - }, - "meta-llama-3_3-70b-instruct": { - "id": "meta-llama-3_3-70b-instruct", - "name": "Meta-Llama-3_3-70B-Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-01", - "last_updated": "2025-04-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.74, - "output": 0.74 - } - }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5-397B-A17B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-18", - "last_updated": "2026-05-18", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.71, - "output": 4.25 - } - }, - "qwen2.5-vl-72b-instruct": { - "id": "qwen2.5-vl-72b-instruct", - "name": "Qwen2.5-VL-72B-Instruct", - "description": "Multimodal model for analyzing text, images, documents, and rich media", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-03-31", - "last_updated": "2025-03-31", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 1.01, - "output": 1.01 - } - }, - "qwen3.6-27b": { - "id": "qwen3.6-27b", - "name": "Qwen3.6-27B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.47, - "output": 3.19 - } - }, - "mistral-small-3.2-24b-instruct-2506": { - "id": "mistral-small-3.2-24b-instruct-2506", - "name": "Mistral-Small-3.2-24B-Instruct-2506", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-16", - "last_updated": "2025-07-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.1, - "output": 0.31 - } - }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2025-08-28", - "last_updated": "2025-08-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.09, - "output": 0.47 - } - }, - "mistral-nemo-instruct-2407": { - "id": "mistral-nemo-instruct-2407", - "name": "Mistral-Nemo-Instruct-2407", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-11-20", - "last_updated": "2024-11-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 65536, - "output": 65536 - }, - "cost": { - "input": 0.14, - "output": 0.14 - } - } - } - }, - "requesty": { - "id": "requesty", - "env": [ - "REQUESTY_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://router.requesty.ai/v1", - "name": "Requesty", - "doc": "https://requesty.ai/solution/llm-routing/models", - "models": { - "ring-2.6-1t": { - "id": "ring-2.6-1t", - "name": "ring-2.6-1t", - "description": "Inclusion AI ring-2.6-1t", - "family": "ring", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "release_date": "2026-05-08", - "last_updated": "2026-05-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.3, - "output": 2.5 - } - }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.75, - "output": 16.5, - "cache_read": 0.275 - } - }, - "nemotron-3-ultra-nvfp4": { - "id": "nemotron-3-ultra-nvfp4", - "name": "nemotron-3-ultra-nvfp4", - "description": "Nemotron-3-Ultra-550B-A55B-NVFP4 is a frontier-scale large language model (LLM) trained by NVIDIA, designed to deliver strong agentic, reasoning, and conversational capabilities. It is optimized for the most demanding workloads, including complex multi-step agents, long-context analysis, and high-accuracy reasoning over code, math, and science. The model employs a hybrid Latent Mixture-of-Experts (LatentMoE) architecture, utilizing interleaved Mamba-2 and MoE layers, along with select Attention layers. Like the Super model, the Ultra model incorporates Multi-Token Prediction (MTP) layers for faster text generation and improved quality, and it is trained using an NVFP4 pre-training recipe to maximize compute efficiency. The model has 55B active parameters and 550B parameters in total.", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.12 - } - }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 180, - "cache_read": 30 - } - }, - "claude-opus-5@eu": { - "id": "claude-opus-5@eu", - "name": "Claude Opus 5 (EU)", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "gemma-4-31b-it": { - "id": "gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "ling-2.6-1t": { - "id": "ling-2.6-1t", - "name": "ling-2.6-1t", - "description": "Inclusion AI ling-2.6-1t", - "family": "ling", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.3, - "output": 2.5 - } - }, - "gpt-4o-mini@eu": { - "id": "gpt-4o-mini@eu", - "name": "GPT-4o mini (EU)", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16000 - }, - "cost": { - "input": 0.165, - "output": 0.66, - "cache_read": 0.0825 - } - }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.25, - "cache_write": 3.125 - } - }, - "nemotron-3.5-lightning-30b-a3b": { - "id": "nemotron-3.5-lightning-30b-a3b", - "name": "nemotron-3.5-lightning-30b-a3b", - "description": "NVIDIA Nemotron 3.5 Lightning 30B-A3B is a hybrid Mamba-2 + MoE + Attention model with 30B total and 3B active parameters, pre-trained on over 20T tokens with an NVFP4 recipe and Multi-Token Prediction for fast generation. Up to 1M token context for long-running autonomous agents, sub-agent workhorse deployments, and agentic workflows. Supports reasoning and tool calling. English and coding languages plus Spanish, French, German, Italian, and Japanese. Open weights under the OpenMDW License Agreement v1.1. Part of the NVIDIA Nemotron family.", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 - } - }, - "gpt-5.5-pro": { - "id": "gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 180 - } - }, - "mistral-medium-3-5@eu": { - "id": "mistral-medium-3-5@eu", - "name": "mistral-medium-3-5@eu", - "description": "Mistral Medium 3.5 is a dense 128B instruction following model from Mistral AI. It supports text and image inputs with text output, and is designed for agentic workflows, coding, and complex multi step reasoning. It is particularly strong at reliable multi tool calling and long horizon tasks, with a 256K context window, configurable reasoning effort per request, and a custom vision encoder that handles variable image sizes and aspect ratios. Self hostable on as few as four GPUs and available under open weights.", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.65, - "output": 8.25, - "cache_read": 1.65 - } - }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "gpt-5@eu": { - "id": "gpt-5@eu", - "name": "GPT-5 (EU)", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.375, - "output": 11, - "cache_read": 0.1375 - } - }, - "deepseek-v4-flash-0731@eu": { - "id": "deepseek-v4-flash-0731@eu", - "name": "DeepSeek V4 Flash 0731 (EU)", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.28, - "output": 0.56, - "cache_read": 0.07 - } - }, - "seed-2.0-pro": { - "id": "seed-2.0-pro", - "name": "Seed 2.0 Pro", - "description": "Flagship ByteDance Seed 2.0 model for complex multimodal reasoning and long-horizon agent workflows", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1 - } - }, - "kimi-k3@eu": { - "id": "kimi-k3@eu", - "name": "Kimi K3 (EU)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 262144 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.45 - } - }, - "seed-1.8": { - "id": "seed-1.8", - "name": "seed-1.8", - "description": "Optimized specifically for multimodal agent scenarios. It features enhanced agent capabilities, upgraded multimodal comprehension, and more flexible context management.", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "release_date": "2026-05-27", - "last_updated": "2026-05-27", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05 - } - }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 262144 - }, - "cost": { - "input": 0.2, - "output": 0.6, - "cache_read": 0.07 - } - }, - "qwen3.5-27b": { - "id": "qwen3.5-27b", - "name": "Qwen3.5 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.26, - "output": 2.6 - } - }, - "mistral-medium-latest@eu": { - "id": "mistral-medium-latest@eu", - "name": "Mistral Medium (latest) (EU)", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.44, - "output": 2.2, - "cache_read": 0.44 - } - }, - "glm-5.2-fast": { - "id": "glm-5.2-fast", - "name": "glm-5.2-fast", - "description": "GLM-5.2 introduces a robust 1M-token context and advanced, multi-effort coding capabilities to significantly enhance performance on long-horizon tasks. Its new IndexShare architecture and improved MTP layer simultaneously boost efficiency by reducing per-token FLOPs and increasing speculative decoding lengths. A 743B-parameter model in Zhipu AI's GLM series, designed to plan, execute, and iterate autonomously on extended, engineering-grade tasks.", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "release_date": "2026-07-13", - "last_updated": "2026-07-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2.1, - "output": 6.6, - "cache_read": 0.21 - } - }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.28, - "output": 0.56, - "cache_read": 0.07 - } - }, - "deepseek-v4-pro-0813@eu": { - "id": "deepseek-v4-pro-0813@eu", - "name": "DeepSeek V4 Pro 0813 (EU)", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.75, - "output": 3.5, - "cache_read": 0.44 - } - }, - "inkling-256k": { - "id": "inkling-256k", - "name": "inkling-256k", - "description": "Inkling 256K is the extended context variant of Inkling, a large MoE hybrid reasoning model from Thinking Machines with audio and vision input support and a 256K context window.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 1.87, - "output": 4.68, - "cache_read": 0.374 - } - }, - "muse-glimmer-30b": { - "id": "muse-glimmer-30b", - "name": "Muse Glimmer 30B", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 20480 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "claude-opus-4-8@eu": { - "id": "claude-opus-4-8@eu", - "name": "Claude Opus 4.8 (EU)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 262144 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.45 - } - }, - "claude-fable-5@eu": { - "id": "claude-fable-5@eu", - "name": "Claude Fable 5 (EU)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 11, - "output": 55, - "cache_read": 1.1, - "cache_write": 13.75 - } - }, - "nemotron-3-nano-omni@eu": { - "id": "nemotron-3-nano-omni@eu", - "name": "nemotron-3-nano-omni@eu", - "description": "The most open, efficient, and accurate omni modal reasoning model for agentic AI.", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-20", - "last_updated": "2026-05-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 300000 - }, - "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.06 - } - }, - "gpt-4.1@eu": { - "id": "gpt-4.1@eu", - "name": "GPT-4.1 (EU)", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 2.2, - "output": 8.8, - "cache_read": 0.55 - } - }, - "gemini-3.1-flash-lite@eu": { - "id": "gemini-3.1-flash-lite@eu", - "name": "Gemini 3.1 Flash Lite (EU)", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.275, - "output": 1.65, - "cache_read": 0.0275, - "cache_write": 0.091663 - } - }, - "nvidia-nemotron-3-super-120b-a12b": { - "id": "nvidia-nemotron-3-super-120b-a12b", - "name": "nvidia-nemotron-3-super-120b-a12b", - "description": "NVIDIA Nemotron 3 Super is a hybrid Mixture-of-Experts (MoE) model engineered for highest compute efficiency and accuracy in multi-agent applications and specialized agentic systems. It is optimized to run many collaborating agents per application on a single GPU, delivering high accuracy for reasoning, tool use, and instruction following.", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.1, - "output": 0.5 - } - }, - "gemini-3.5-flash-lite@eu": { - "id": "gemini-3.5-flash-lite@eu", - "name": "Gemini 3.5 Flash Lite (EU)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.33, - "output": 2.75, - "cache_read": 0.033 - } - }, - "claude-opus-4-5": { - "id": "claude-opus-4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "kimi-k2.7-code@eu": { - "id": "kimi-k2.7-code@eu", - "name": "Kimi K2.7 Code (EU)", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.25, - "output": 4.5, - "cache_read": 0.31 - } - }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 393216 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.05 - } - }, - "glm-5.1@eu": { - "id": "glm-5.1@eu", - "name": "GLM-5.1 (EU)", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 200000 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 1.4 - } - }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "claude-haiku-4-5@eu": { - "id": "claude-haiku-4-5@eu", - "name": "Claude Haiku 4.5 (latest) (EU)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1.1, - "output": 5.5, - "cache_read": 0.11, - "cache_write": 1.375 - } - }, - "claude-sonnet-4-5@eu": { - "id": "claude-sonnet-4-5@eu", - "name": "Claude Sonnet 4.5 (latest) (EU)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.3, - "cache_write": 4.125, - "tiers": [ - { - "input": 6.6, - "output": 24.75, - "cache_read": 0.6, - "cache_write": 8.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6.6, - "output": 24.75, - "cache_read": 0.6, - "cache_write": 8.25 - } - } - }, - "laguna-xs.2": { - "id": "laguna-xs.2", - "name": "Laguna XS.2", - "description": "Agentic coding model from Poolside in the XS size class for local deployment", - "family": "laguna", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "leanstral-1-5@eu": { - "id": "leanstral-1-5@eu", - "name": "leanstral-1-5@eu", - "description": "Leanstral 1.5 is an updated Lean 4 formal proof engineering model from Mistral AI, optimized for automated theorem proving and autoformalization. It has 119B total parameters with 6.5B active and supports a 256K token context window. It supports native function calling and structured output.", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-27", - "last_updated": "2026-05-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 - } - }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } - } - }, - "nemotron-3-super-120b-a12b": { - "id": "nemotron-3-super-120b-a12b", - "name": "Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "claude-opus-5-5": { - "id": "claude-opus-5-5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "claude-sonnet-4-6@eu": { - "id": "claude-sonnet-4-6@eu", - "name": "Claude Sonnet 4.6 (EU)", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.3, - "cache_write": 4.125 - } - }, - "gemini-3.6-flash": { - "id": "gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 1.5, - "output": 7, - "cache_read": 0.15 - } - }, - "gpt-5.5@eu": { - "id": "gpt-5.5@eu", - "name": "GPT-5.5 (EU)", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "ling-3.0-tiny": { - "id": "ling-3.0-tiny", - "name": "ling-3.0-tiny", - "description": "Ling-3.0-tiny is an efficient 7.9B parameter MoE model from inclusionAI with only 1.3B active parameters per token. Built for responsive agents, reliable instruction following and multi turn conversation, with a 256K context window, native function calling, prompt caching and switchable Thinking and Instant modes.", - "family": "ling", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "claude-opus-5-5@eu": { - "id": "claude-opus-5-5@eu", - "name": "Claude Opus 5.5 (EU)", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "gpt-5-mini@eu": { - "id": "gpt-5-mini@eu", - "name": "GPT-5 Mini (EU)", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 0.275, - "output": 2.2, - "cache_read": 0.0275 - } - }, - "mistral-medium-3-5": { - "id": "mistral-medium-3-5", - "name": "mistral-medium-3-5", - "description": "Mistral Medium 3.5 is a dense 128B instruction following model from Mistral AI. It supports text and image inputs with text output, and is designed for agentic workflows, coding, and complex multi step reasoning. It is particularly strong at reliable multi tool calling and long horizon tasks, with a 256K context window, configurable reasoning effort per request, and a custom vision encoder that handles variable image sizes and aspect ratios. Self hostable on as few as four GPUs and available under open weights.", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.65, - "output": 8.25, - "cache_read": 1.65 - } - }, - "grok-4.3": { - "id": "grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "cache_write": 1.25, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "cache_write": 2.5 - } - } - }, - "nemotron-3-ultra-550b-a55b": { - "id": "nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "nemotron-3-nano-omni-30b-a3b-reasoning": { - "id": "nemotron-3-nano-omni-30b-a3b-reasoning", - "name": "Nemotron 3 Nano Omni 30B A3B Reasoning", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 20480 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gemma-4-26b-a4b-it": { - "id": "gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.07, - "output": 0.34, - "cache_read": 0.07 - } - }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 - } - }, - "gpt-5.4@eu": { - "id": "gpt-5.4@eu", - "name": "GPT-5.4 (EU)", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } - } - }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Claude Opus 4.1 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "grok-4.2-beta": { - "id": "grok-4.2-beta", - "name": "grok-4.2-beta", - "description": "Grok 4.20 Beta is xAI's newest flagship model with industry-leading speed and agentic tool calling capabilities. It combines the lowest hallucination rate on the market with strict prompt adherance, delivering consistently precise and truthful responses.", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-03-19", - "last_updated": "2026-03-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 2000000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 0.4, - "cache_write": 4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 0.4, - "cache_write": 4 - } - } - }, - "gemini-3.1-flash-image": { - "id": "gemini-3.1-flash-image", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 2 - } - }, - "nemotron-lightning-3.5-30b-a3b": { - "id": "nemotron-lightning-3.5-30b-a3b", - "name": "nemotron-lightning-3.5-30b-a3b", - "description": "Nemotron-Lightning-3.5-30B-A3B is a 30B-parameter Mixture-of-Experts language model (3B active) from NVIDIA's Nemotron-H family, built on a hybrid Mamba-Transformer architecture for efficient long-context inference. Like other models in the family, it responds to queries by first generating a reasoning trace and then concluding with a final response, with reasoning behavior configurable through a flag in the chat template. It includes a multi-token prediction (MTP) speculative decoding head for low-latency serving.", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-08-15", - "last_updated": "2026-08-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.05, - "output": 0.2, - "cache_read": 0.01 - } - }, - "gpt-5-nano@eu": { - "id": "gpt-5-nano@eu", - "name": "GPT-5 Nano (EU)", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 0.055, - "output": 0.44, - "cache_read": 0.0055 - } - }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2 - } - } - }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 4.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 9, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 9 - } - } - }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "cache_write": 2, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "cache_write": 4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1, - "cache_write": 4 - } - } - }, - "seed-2.0-mini": { - "id": "seed-2.0-mini", - "name": "Seed 2.0 Mini", - "description": "Lightweight ByteDance Seed 2.0 model for low-latency multimodal reasoning and high-volume tasks", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.02 - } - }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "kimi-k2.6@eu": { - "id": "kimi-k2.6@eu", - "name": "Kimi K2.6 (EU)", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 128000 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.95 - } - }, - "qwen3.5-2b": { - "id": "qwen3.5-2b", - "name": "qwen3.5-2b", - "description": "Qwen3.5-2B is a compact yet capable model from Alibaba's Qwen3.5 series. It features a 262K token context window, support for 201 languages, thinking/reasoning mode, and tool calling for agentic workflows. A strong choice for prototyping, fine-tuning, and efficient multilingual deployments.", - "family": "qwen3.5", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-03-10", - "last_updated": "2026-03-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.02, - "output": 0.1 - } - }, - "fugu-ultra": { - "id": "fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "release_date": "2026-06-15", - "last_updated": "2026-06-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 1.2 - } - }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 1.583 - } - }, - "ling-2.6-flash": { - "id": "ling-2.6-flash", - "name": "ling-2.6-flash", - "description": "Inclusion AI ling-2.6-flash", - "family": "ling", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "gemini-3-pro-image": { - "id": "gemini-3-pro-image", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 4.5 - } - }, - "gemini-3.7-flash@eu": { - "id": "gemini-3.7-flash@eu", - "name": "Gemini 3.7 Flash (EU)", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.825, - "output": 4.125, - "cache_read": 0.0825 - } - }, - "claude-fable-5.1": { - "id": "claude-fable-5.1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 - } - }, - "gpt-4.1-nano@eu": { - "id": "gpt-4.1-nano@eu", - "name": "GPT-4.1 nano (EU)", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.11, - "output": 0.44, - "cache_read": 0.0275 - } - }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 - } - }, - "gpt-5.1@eu": { - "id": "gpt-5.1@eu", - "name": "GPT-5.1 (EU)", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.375, - "output": 11, - "cache_read": 0.1375 - } - }, - "gemini-2.5-flash-lite@eu": { - "id": "gemini-2.5-flash-lite@eu", - "name": "Gemini 2.5 Flash-Lite (EU)", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0.18333 - } - }, - "nemotron-3-nano-omni": { - "id": "nemotron-3-nano-omni", - "name": "nemotron-3-nano-omni", - "description": "The most open, efficient, and accurate omni modal reasoning model for agentic AI.", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-20", - "last_updated": "2026-05-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 300000 - }, - "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.06 - } - }, - "claude-sonnet-4@eu": { - "id": "claude-sonnet-4@eu", - "name": "Claude Sonnet 4 (latest) (EU)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } - } - }, - "inkling": { - "id": "inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 65536, - "output": 32768 - }, - "cost": { - "input": 1.87, - "output": 4.68, - "cache_read": 0.374 - } - }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "deepseek-v4.1-flash@eu": { - "id": "deepseek-v4.1-flash@eu", - "name": "DeepSeek V4.1 Flash (EU)", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 393216 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.05 - } - }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04 - } - } - }, - "qwen3.8-2.4T-A95B@eu": { - "id": "qwen3.8-2.4T-A95B@eu", - "name": "Qwen3.8 2.4T A95B (EU)", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 262144 - }, - "cost": { - "input": 2.5, - "output": 6, - "cache_read": 0.63 - } - }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.16, - "output": 0.47, - "cache_read": 0.016, - "cache_write": 0.2 - } - }, - "qwen3.8-flash-next": { - "id": "qwen3.8-flash-next", - "name": "Qwen3.8 Flash Next", - "description": "Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } - }, - "qwen3.8-flash-next@eu": { - "id": "qwen3.8-flash-next@eu", - "name": "Qwen3.8 Flash Next (EU)", - "description": "Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } - }, - "mistral-small-2603": { - "id": "mistral-small-2603", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.165, - "output": 0.66, - "cache_read": 0.165 - } - }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 33, - "cache_read": 0.55 - } - }, - "claude-sonnet-5@eu": { - "id": "claude-sonnet-5@eu", - "name": "Claude Sonnet 5 (EU)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2.2, - "output": 11, - "cache_read": 0.22, - "cache_write": 2.75 - } - }, - "gemini-3.8-flash@eu": { - "id": "gemini-3.8-flash@eu", - "name": "Gemini 3.8 Flash (EU)", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.825, - "output": 4.125, - "cache_read": 0.0825 - } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "qwen3.5-35b-a3b": { - "id": "qwen3.5-35b-a3b", - "name": "Qwen3.5 35B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.14, - "output": 1, - "cache_read": 0.05 - } - }, - "glm-5.3@eu": { - "id": "glm-5.3@eu", - "name": "GLM-5.3 (EU)", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 1.2, - "output": 4.2, - "cache_read": 0.26 - } - }, - "gemini-2.5-flash@eu": { - "id": "gemini-2.5-flash@eu", - "name": "Gemini 2.5 Flash (EU)", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075, - "cache_write": 0.55 - } - }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.8, - "output": 2.55, - "cache_read": 0.16 - } - }, - "hy3": { - "id": "hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.14, - "output": 0.58, - "cache_read": 0.035 - } - }, - "glm-5.2@eu": { - "id": "glm-5.2@eu", - "name": "GLM-5.2 (EU)", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4.2, - "cache_read": 0.26 - } - }, - "gpt-5.6-sol@eu": { - "id": "gpt-5.6-sol@eu", - "name": "GPT-5.6 Sol (EU)", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 4.4, - "output": 22, - "cache_read": 0.44 - } - }, - "seed-2.0-code": { - "id": "seed-2.0-code", - "name": "Seed 2.0 Code", - "description": "ByteDance Seed coding model for multimodal software engineering and long-running agents", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-02-14", - "last_updated": "2026-02-14", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.1 - } - }, - "devstral-latest@eu": { - "id": "devstral-latest@eu", - "name": "devstral-latest@eu", - "description": "An enterprise grade text model, that excels at using tools to explore codebases, editing multiple files and power software engineering agents.", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-27", - "last_updated": "2026-05-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.44, - "output": 2.2, - "cache_read": 0.44 - } - }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "gemini-2.5-pro@eu": { - "id": "gemini-2.5-pro@eu", - "name": "Gemini 2.5 Pro (EU)", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.31, - "cache_write": 2.375, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.62, - "cache_write": 4.75, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.62, - "cache_write": 4.75 - } - } - }, - "gpt-5.6-terra@eu": { - "id": "gpt-5.6-terra@eu", - "name": "GPT-5.6 Terra (EU)", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.2, - "output": 13.2, - "cache_read": 0.22 - } - }, - "grok-build-0.1": { - "id": "grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 1, - "output": 2, - "cache_read": 0.1 - } - }, - "glm-5.3-flash@eu": { - "id": "glm-5.3-flash@eu", - "name": "GLM-5.3-Flash (EU)", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 262144 - }, - "cost": { - "input": 0.2, - "output": 0.6, - "cache_read": 0.07 - } - }, - "minimax-m2.7-highspeed": { - "id": "minimax-m2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 1.2 - } - }, - "deepseek-v4-pro@eu": { - "id": "deepseek-v4-pro@eu", - "name": "DeepSeek V4 Pro (EU)", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 1.75, - "output": 3.5, - "cache_read": 0.44 - } - }, - "devstral-latest": { - "id": "devstral-latest", - "name": "devstral-latest", - "description": "An enterprise grade text model, that excels at using tools to explore codebases, editing multiple files and power software engineering agents.", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-27", - "last_updated": "2026-05-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.44, - "output": 2.2, - "cache_read": 0.44 - } - }, - "o4-mini@eu": { - "id": "o4-mini@eu", - "name": "o4-mini (EU)", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.21, - "output": 4.84, - "cache_read": 0.3025 - } - }, - "mistral-small-2603@eu": { - "id": "mistral-small-2603@eu", - "name": "Mistral Small 4 (EU)", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.165, - "output": 0.66, - "cache_read": 0.165 - } - }, - "qwen3.8-2.4T-A95B": { - "id": "qwen3.8-2.4T-A95B", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 - } - }, - "gemini-3.8-flash": { - "id": "gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "gemini-3.5-flash@eu": { - "id": "gemini-3.5-flash@eu", - "name": "Gemini 3.5 Flash (EU)", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 1.65, - "output": 9.9, - "cache_read": 0.165, - "cache_write": 1.7413 - } - }, - "gpt-6-luna": { - "id": "gpt-6-luna", - "name": "GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02 - } - } - }, - "nemotron-3.5-content-safety": { - "id": "nemotron-3.5-content-safety", - "name": "Nemotron 3.5 Content Safety", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "nemotron", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 - } - }, - "nvidia-nemotron-3-ultra": { - "id": "nvidia-nemotron-3-ultra", - "name": "nvidia-nemotron-3-ultra", - "description": "NVIDIA Nemotron 3 Ultra is NVIDIA's strongest open-weights reasoning model, positioned near GPT-5.4 Mini (xhigh) and ahead of DeepSeek V4-Flash and Qwen3.5-397B-A17B.", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": false, - "structured_output": false, - "release_date": "2026-06-23", - "last_updated": "2026-06-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.5, - "output": 2.5 - } - }, - "claude-fable-5.1@eu": { - "id": "claude-fable-5.1@eu", - "name": "Claude Fable 5.1 (EU)", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 11, - "output": 55, - "cache_read": 0.275, - "cache_write": 13.75 - } - }, - "step-3.7-flash": { - "id": "step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 256000 - }, - "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 - } - }, - "gpt-6-luna@eu": { - "id": "gpt-6-luna@eu", - "name": "GPT-6 Luna (EU)", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.12, - "output": 0.6, - "cache_read": 0.012, - "tiers": [ - { - "input": 0.24, - "output": 0.9, - "cache_read": 0.024, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.24, - "output": 0.9, - "cache_read": 0.024 - } - } - }, - "claude-opus-4-6@eu": { - "id": "claude-opus-4-6@eu", - "name": "Claude Opus 4.6 (EU)", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "leanstral-1-5": { - "id": "leanstral-1-5", - "name": "leanstral-1-5", - "description": "Leanstral 1.5 is an updated Lean 4 formal proof engineering model from Mistral AI, optimized for automated theorem proving and autoformalization. It has 119B total parameters with 6.5B active and supports a 256K token context window. It supports native function calling and structured output.", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "release_date": "2026-05-27", - "last_updated": "2026-05-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625 - } - }, - "gpt-5.6-luna@eu": { - "id": "gpt-5.6-luna@eu", - "name": "GPT-5.6 Luna (EU)", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.22, - "output": 1.32, - "cache_read": 0.022 - } - }, - "claude-opus-4-7@eu": { - "id": "claude-opus-4-7@eu", - "name": "Claude Opus 4.7 (EU)", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "kat-coder-pro": { - "id": "kat-coder-pro", - "name": "kat-coder-pro", - "description": "KAT-Coder-Pro V2 by KwaiKAT is a non-reasoning model optimized for agentic coding. It delivers strong performance on reasoning-style tasks while requiring significantly fewer output tokens than peer models. With the 1210 release, it achieved a score of 64 on the Artificial Analysis Intelligence Index, placing it in the global Top 10 and ranking first among all non-reasoning models.", - "family": "kat-coder", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "release_date": "2026-03-27", - "last_updated": "2026-03-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "laguna-m.1": { - "id": "laguna-m.1", - "name": "Laguna M.1", - "description": "Poolside's open-weight model for agentic coding and long-horizon work", - "family": "laguna", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-6-sol@eu": { - "id": "gpt-6-sol@eu", - "name": "GPT-6 Sol (EU)", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.4, - "output": 12, - "cache_read": 0.24, - "tiers": [ - { - "input": 4.8, - "output": 18, - "cache_read": 0.48, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4.8, - "output": 18, - "cache_read": 0.48 - } - } - }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 1.2, - "output": 4.2, - "cache_read": 0.26 - } - }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - }, - "grok-4.6": { - "id": "grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "cache_write": 2, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "cache_write": 4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1, - "cache_write": 4 - } - } - }, - "minimax-m3@eu": { - "id": "minimax-m3@eu", - "name": "MiniMax-M3 (EU)", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 0.4, - "output": 2, - "cache_read": 0.1 - } - }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.32, - "output": 1.28, - "cache_read": 0.032, - "cache_write": 0.4 - } - }, - "gpt-4.1-mini@eu": { - "id": "gpt-4.1-mini@eu", - "name": "GPT-4.1 mini (EU)", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.44, - "output": 1.76, - "cache_read": 0.11 - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10 - } - } - }, - "claude-opus-4-5@eu": { - "id": "claude-opus-4-5@eu", - "name": "Claude Opus 4.5 (latest) (EU)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.28, - "output": 0.56, - "cache_read": 0.07 - } - }, - "mistral-medium-latest": { - "id": "mistral-medium-latest", - "name": "Mistral Medium (latest)", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.44, - "output": 2.2, - "cache_read": 0.44 - } - }, - "gpt-6-sol": { - "id": "gpt-6-sol", - "name": "GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4 - } - } - }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 0.08333 - } - } - } - }, - "mistral": { - "id": "mistral", - "env": [ - "MISTRAL_API_KEY" - ], - "npm": "@ai-sdk/mistral", - "name": "Mistral", - "doc": "https://docs.mistral.ai/getting-started/models/", - "models": { - "open-mistral-nemo": { - "id": "open-mistral-nemo", - "name": "Open Mistral Nemo", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "status": "deprecated", - "cost": { - "input": 0.15, - "output": 0.15 - } - }, - "codestral-latest": { - "id": "codestral-latest", - "name": "Codestral (latest)", - "description": "Mistral code model for completions, refactors, and developer IDE workflows", - "family": "codestral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-05-29", - "last_updated": "2025-01-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 4096 - }, - "cost": { - "input": 0.3, - "output": 0.9 - } - }, - "mistral-large-2411": { - "id": "mistral-large-2411", - "name": "Mistral Large 2.1", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-18", - "last_updated": "2024-11-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 2, - "output": 6 - } - }, - "mistral-nemo": { - "id": "mistral-nemo", - "name": "Mistral Nemo", - "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.15, - "output": 0.15 - } - }, - "voxtral-mini-tts-latest": { - "id": "voxtral-mini-tts-latest", - "name": "Voxtral Mini TTS (latest)", - "description": "Multilingual text-to-speech model with zero-shot voice cloning", - "family": "voxtral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-03-01", - "last_updated": "2026-03-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "mistral-medium-2508": { - "id": "mistral-medium-2508", - "name": "Mistral Medium 3.1", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-08-12", - "last_updated": "2025-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.4, - "output": 2 - } - }, - "mistral-large-latest": { - "id": "mistral-large-latest", - "name": "Mistral Large (latest)", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "mistral-small-latest": { - "id": "mistral-small-latest", - "name": "Mistral Small (latest)", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "zai-glm-5-2": { - "id": "zai-glm-5-2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "status": "beta", - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.14 - } - }, - "ministral-8b-latest": { - "id": "ministral-8b-latest", - "name": "Ministral 8B (latest)", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.1, - "output": 0.1 - } - }, - "devstral-medium-latest": { - "id": "devstral-medium-latest", - "name": "Devstral 2 (latest)", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2 - } - }, - "open-mixtral-8x22b": { - "id": "open-mixtral-8x22b", - "name": "Mixtral 8x22B", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mixtral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-17", - "last_updated": "2024-04-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 64000, - "output": 64000 - }, - "cost": { - "input": 2, - "output": 6 - } - }, - "devstral-2512": { - "id": "devstral-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2 - } - }, - "mistral-medium-2505": { - "id": "mistral-medium-2505", - "name": "Mistral Medium 3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.4, - "output": 2 - } - }, - "magistral-medium-latest": { - "id": "magistral-medium-latest", - "name": "Magistral Medium (latest)", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-medium", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2, - "output": 5 - } - }, - "pixtral-12b": { - "id": "pixtral-12b", - "name": "Pixtral 12B", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-09-01", - "last_updated": "2024-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.15, - "output": 0.15 - } - }, - "mistral-embed": { - "id": "mistral-embed", - "name": "Mistral Embed", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "mistral-embed", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2023-12-11", - "last_updated": "2023-12-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8000, - "output": 3072 - }, - "cost": { - "input": 0.1, - "output": 0 - } - }, - "devstral-small-2505": { - "id": "devstral-small-2505", - "name": "Devstral Small 2505", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "status": "deprecated", - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "mistral-small-2603": { - "id": "mistral-small-2603", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "devstral-medium-2507": { - "id": "devstral-medium-2507", - "name": "Devstral Medium", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-07-10", - "last_updated": "2025-07-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2 - } - }, - "magistral-small": { - "id": "magistral-small", - "name": "Magistral Small", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-small", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-03-17", - "last_updated": "2025-03-17", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "voxtral-mini-latest": { - "id": "voxtral-mini-latest", - "name": "Voxtral Mini (latest)", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "voxtral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-02-01", - "last_updated": "2026-02-01", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 0 - } - }, - "labs-devstral-small-2512": { - "id": "labs-devstral-small-2512", - "name": "Devstral Small 2", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "status": "deprecated", - "cost": { - "input": 0, - "output": 0 - } - }, - "mistral-large-2512": { - "id": "mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "devstral-latest": { - "id": "devstral-latest", - "name": "Devstral 2", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2 - } - }, - "devstral-small-2507": { - "id": "devstral-small-2507", - "name": "Devstral Small", - "description": "Legacy model retained for compatibility with older integrations", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-07-10", - "last_updated": "2025-07-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "status": "deprecated", - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "pixtral-large-latest": { - "id": "pixtral-large-latest", - "name": "Pixtral Large (latest)", - "description": "Mistral's larger vision model for document-heavy image understanding and chat", - "family": "pixtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2024-11-04", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 6 - } - }, - "voxtral-small-latest": { - "id": "voxtral-small-latest", - "name": "Voxtral Small (latest)", - "description": "Instruct model with native audio input for speech understanding and tool use", - "family": "voxtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "open-mixtral-8x7b": { - "id": "open-mixtral-8x7b", - "name": "Mixtral 8x7B", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mixtral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-01", - "release_date": "2023-12-11", - "last_updated": "2023-12-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0.7, - "output": 0.7 - } - }, - "mistral-small-2506": { - "id": "mistral-small-2506", - "name": "Mistral Small 3.2", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "zai-glm-5-3": { - "id": "zai-glm-5-3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.14 - } - }, - "mistral-medium-2604": { - "id": "mistral-medium-2604", - "name": "Mistral Medium 3.5", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.5, - "output": 7.5 - } - }, - "open-mistral-7b": { - "id": "open-mistral-7b", - "name": "Mistral 7B", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-09-27", - "last_updated": "2023-09-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 8000, - "output": 8000 - }, - "cost": { - "input": 0.25, - "output": 0.25 - } - }, - "ministral-3b-latest": { - "id": "ministral-3b-latest", - "name": "Ministral 3B (latest)", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.04, - "output": 0.04 - } - }, - "mistral-medium-latest": { - "id": "mistral-medium-latest", - "name": "Mistral Medium (latest)", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.5, - "output": 7.5 - } - } - } - }, - "amazon-bedrock": { - "id": "amazon-bedrock", - "env": [ - "AWS_ACCESS_KEY_ID", - "AWS_SECRET_ACCESS_KEY", - "AWS_REGION", - "AWS_BEARER_TOKEN_BEDROCK" - ], - "npm": "@ai-sdk/amazon-bedrock", - "name": "Amazon Bedrock", - "doc": "https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html", - "models": { - "google.gemma-3-12b-it": { - "id": "google.gemma-3-12b-it", - "name": "Gemma 3 12B IT", - "description": "Open multimodal Gemma instruction model for multilingual text generation and image understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.09, - "output": 0.29 - } - }, - "google.gemma-3-4b-it": { - "id": "google.gemma-3-4b-it", - "name": "Gemma 3 4B IT", - "description": "Open multimodal Gemma instruction model for efficient text generation and image understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 4096 - }, - "cost": { - "input": 0.04, - "output": 0.08 - } - }, - "eu.anthropic.claude-fable-5": { - "id": "eu.anthropic.claude-fable-5", - "name": "Claude Fable 5 (EU)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 11, - "output": 55, - "cache_read": 1.1, - "cache_write": 13.75 - } - }, - "qwen.qwen3-coder-480b-a35b-v1:0": { - "id": "qwen.qwen3-coder-480b-a35b-v1:0", - "name": "Qwen3-Coder 480B-A35B Instruct", - "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-09-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 65536 - }, - "cost": { - "input": 0.45, - "output": 1.8 - } - }, - "google.gemma-4-31b": { - "id": "google.gemma-4-31b", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 0.14, - "output": 0.4 - } - }, - "us.writer.palmyra-x5-v1:0": { - "id": "us.writer.palmyra-x5-v1:0", - "name": "Palmyra X5 (US)", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "palmyra", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1040000, - "input": 1040000, - "output": 8192 - }, - "cost": { - "input": 0.6, - "output": 6 - } - }, - "google.gemma-4-e2b": { - "id": "google.gemma-4-e2b", - "name": "Gemma 4 E2B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 0.04, - "output": 0.08 - } - }, - "us.anthropic.claude-opus-4-7": { - "id": "us.anthropic.claude-opus-4-7", - "name": "Claude Opus 4.7 (US)", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "global.openai.gpt-6-astra": { - "id": "global.openai.gpt-6-astra", - "name": "GPT-6 Astra (Global)", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } - } - }, - "eu.amazon.nova-lite-v1:0": { - "id": "eu.amazon.nova-lite-v1:0", - "name": "Nova Lite (EU)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.069, - "output": 0.276, - "cache_read": 0.01725, - "cache_write": 0.069 - } - }, - "global.anthropic.claude-opus-5-5": { - "id": "global.anthropic.claude-opus-5-5", - "name": "Claude Opus 5.5 (Global)", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 - } - }, - "deepseek.r1-v1:0": { - "id": "deepseek.r1-v1:0", - "name": "DeepSeek-R1", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 1.35, - "output": 5.4 - } - }, - "openai.gpt-oss-safeguard-20b": { - "id": "openai.gpt-oss-safeguard-20b", - "name": "GPT OSS Safeguard 20B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.07, - "output": 0.2 - } - }, - "anthropic.claude-opus-5-5": { - "id": "anthropic.claude-opus-5-5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 - } - }, - "global.anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "global.anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5 (Global)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "eu.amazon.nova-2-lite-v1:0": { - "id": "eu.amazon.nova-2-lite-v1:0", - "name": "Nova 2 Lite (EU)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-10", - "release_date": "2025-12-02", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65535 - }, - "cost": { - "input": 0.374, - "output": 3.157, - "cache_read": 0.0935, - "cache_write": 0.374 - } - }, - "us.amazon.nova-pro-v1:0": { - "id": "us.amazon.nova-pro-v1:0", - "name": "Nova Pro (US)", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.8, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0.8 - } - }, - "jp.anthropic.claude-opus-5-5": { - "id": "jp.anthropic.claude-opus-5-5", - "name": "Claude Opus 5.5 (JP)", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4.4, - "output": 22, - "cache_read": 0.22, - "cache_write": 5.5 - } - }, - "openai.gpt-5.6-luna": { - "id": "openai.gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 0.22, - "output": 1.32, - "cache_read": 0.022, - "cache_write": 0.275, - "tiers": [ - { - "input": 0.44, - "output": 1.98, - "cache_read": 0.044, - "cache_write": 0.55, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.44, - "output": 1.98, - "cache_read": 0.044, - "cache_write": 0.55 - } - } - }, - "us.meta.llama4-maverick-17b-instruct-v1:0": { - "id": "us.meta.llama4-maverick-17b-instruct-v1:0", - "name": "Llama 4 Maverick 17B Instruct (US)", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 8192 - }, - "cost": { - "input": 0.24, - "output": 0.97 - } - }, - "apac.amazon.nova-micro-v1:0": { - "id": "apac.amazon.nova-micro-v1:0", - "name": "Nova Micro (APAC)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 10000 - }, - "cost": { - "input": 0.037, - "output": 0.148, - "cache_read": 0.00925, - "cache_write": 0.037 - } - }, - "eu.anthropic.claude-sonnet-4-20250514-v1:0": { - "id": "eu.anthropic.claude-sonnet-4-20250514-v1:0", - "name": "Claude Sonnet 4 (EU)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "status": "deprecated", - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "eu.anthropic.claude-opus-4-5-20251101-v1:0": { - "id": "eu.anthropic.claude-opus-4-5-20251101-v1:0", - "name": "Claude Opus 4.5 (EU)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "global.openai.gpt-6-sol": { - "id": "global.openai.gpt-6-sol", - "name": "GPT-6 Sol (Global)", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } - } - }, - "qwen.qwen3-coder-30b-a3b-v1:0": { - "id": "qwen.qwen3-coder-30b-a3b-v1:0", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-31", - "last_updated": "2025-09-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "us.amazon.nova-premier-v1:0": { - "id": "us.amazon.nova-premier-v1:0", - "name": "Nova Premier (US)", - "description": "Multimodal model for complex analysis, long-context understanding, tool use, and model distillation", - "family": "nova", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-04-30", - "last_updated": "2025-04-30", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 10000 - }, - "status": "deprecated", - "cost": { - "input": 2.5, - "output": 12.5, - "cache_read": 0.625, - "cache_write": 2.5 - } - }, - "anthropic.claude-sonnet-4-6": { - "id": "anthropic.claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 - } - }, - "meta.llama3-1-70b-instruct-v1:0": { - "id": "meta.llama3-1-70b-instruct-v1:0", - "name": "Llama 3.1 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.72, - "output": 0.72 - } - }, - "us.writer.palmyra-x4-v1:0": { - "id": "us.writer.palmyra-x4-v1:0", - "name": "Palmyra X4 (US)", - "description": "Enterprise language model for workflow automation, coding, data analysis, and tool use", - "family": "palmyra", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2024-10-09", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 122880, - "output": 8192 - }, - "cost": { - "input": 2.5, - "output": 10 - } - }, - "apac.amazon.nova-pro-v1:0": { - "id": "apac.amazon.nova-pro-v1:0", - "name": "Nova Pro (APAC)", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.84, - "output": 3.36, - "cache_read": 0.21, - "cache_write": 0.84 - } - }, - "global.anthropic.claude-fable-5-1": { - "id": "global.anthropic.claude-fable-5-1", - "name": "Claude Fable 5.1 (Global)", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "us.anthropic.claude-sonnet-5": { - "id": "us.anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5 (US)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2.2, - "output": 11, - "cache_read": 0.22, - "cache_write": 2.75 - } - }, - "amazon.nova-micro-v1:0": { - "id": "amazon.nova-micro-v1:0", - "name": "Nova Micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 10000 - }, - "cost": { - "input": 0.035, - "output": 0.14, - "cache_read": 0.00875, - "cache_write": 0.035 - } - }, - "minimax.minimax-m2.5": { - "id": "minimax.minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 98304 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "xai.grok-4.3": { - "id": "xai.grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-06-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 - } - }, - "us.amazon.nova-lite-v1:0": { - "id": "us.amazon.nova-lite-v1:0", - "name": "Nova Lite (US)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.015, - "cache_write": 0.06 - } - }, - "amazon.nova-pro-v1:0": { - "id": "amazon.nova-pro-v1:0", - "name": "Nova Pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.8, - "output": 3.2, - "cache_read": 0.2, - "cache_write": 0.8 - } - }, - "qwen.qwen3-32b-v1:0": { - "id": "qwen.qwen3-32b-v1:0", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-09-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "openai.gpt-5.5": { - "id": "openai.gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 5.5, - "output": 33, - "cache_read": 0.55 - } - }, - "mistral.voxtral-small-24b-2507": { - "id": "mistral.voxtral-small-24b-2507", - "name": "Voxtral Small 24B 2507", - "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", - "family": "voxtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "global.amazon.nova-2-lite-v1:0": { - "id": "global.amazon.nova-2-lite-v1:0", - "name": "Nova 2 Lite (Global)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-10", - "release_date": "2025-12-02", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65535 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075, - "cache_write": 0.3 - } - }, - "us.openai.gpt-5.6-terra": { - "id": "us.openai.gpt-5.6-terra", - "name": "GPT-5.6 Terra (US)", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.2, - "output": 13.2, - "cache_read": 0.22, - "cache_write": 2.75, - "tiers": [ - { - "input": 4.4, - "output": 19.8, - "cache_read": 0.44, - "cache_write": 5.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4.4, - "output": 19.8, - "cache_read": 0.44, - "cache_write": 5.5 - } - } - }, - "mistral.devstral-2-123b": { - "id": "mistral.devstral-2-123b", - "name": "Devstral 2 123B", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.4, - "output": 2 - } - }, - "us.xai.grok-4.6": { - "id": "us.xai.grok-4.6", - "name": "Grok 4.6 (US)", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-18", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2.2, - "output": 6.6, - "cache_read": 0.55 - } - }, - "jp.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "jp.anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5 (JP)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 - } - }, - "anthropic.claude-fable-5": { - "id": "anthropic.claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "jp.anthropic.claude-opus-4-8": { - "id": "jp.anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8 (JP)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "us.openai.gpt-5.6-luna": { - "id": "us.openai.gpt-5.6-luna", - "name": "GPT-5.6 Luna (US)", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.22, - "output": 1.32, - "cache_read": 0.022, - "cache_write": 0.275, - "tiers": [ - { - "input": 0.44, - "output": 1.98, - "cache_read": 0.044, - "cache_write": 0.55, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.44, - "output": 1.98, - "cache_read": 0.044, - "cache_write": 0.55 - } - } - }, - "us.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "us.anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5 (US)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 - } - }, - "au.anthropic.claude-opus-5": { - "id": "au.anthropic.claude-opus-5", - "name": "Claude Opus 5 (AU)", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "nvidia.nemotron-nano-12b-v2": { - "id": "nvidia.nemotron-nano-12b-v2", - "name": "NVIDIA Nemotron Nano 12B v2 VL BF16", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.2, - "output": 0.6 - } - }, - "us.amazon.nova-2-lite-v1:0": { - "id": "us.amazon.nova-2-lite-v1:0", - "name": "Nova 2 Lite (US)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-10", - "release_date": "2025-12-02", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65535 - }, - "cost": { - "input": 0.33, - "output": 2.75, - "cache_read": 0.0825, - "cache_write": 0.33 - } - }, - "us.openai.gpt-6-astra": { - "id": "us.openai.gpt-6-astra", - "name": "GPT-6 Astra (US)", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 11, - "output": 55, - "cache_read": 1.1, - "cache_write": 13.75, - "tiers": [ - { - "input": 22, - "output": 82.5, - "cache_read": 2.2, - "cache_write": 27.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 22, - "output": 82.5, - "cache_read": 2.2, - "cache_write": 27.5 - } - } - }, - "openai.gpt-5.6-terra": { - "id": "openai.gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 2.2, - "output": 13.2, - "cache_read": 0.22, - "cache_write": 2.75, - "tiers": [ - { - "input": 4.4, - "output": 19.8, - "cache_read": 0.44, - "cache_write": 5.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4.4, - "output": 19.8, - "cache_read": 0.44, - "cache_write": 5.5 - } - } - }, - "anthropic.claude-opus-4-6-v1": { - "id": "anthropic.claude-opus-4-6-v1", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "us.anthropic.claude-opus-4-1-20250805-v1:0": { - "id": "us.anthropic.claude-opus-4-1-20250805-v1:0", - "name": "Claude Opus 4.1 (US)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "us.meta.llama3-1-8b-instruct-v1:0": { - "id": "us.meta.llama3-1-8b-instruct-v1:0", - "name": "Llama 3.1 8B Instruct (US)", - "description": "Compact open Llama model for lightweight chat, drafting, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.22, - "output": 0.22 - } - }, - "eu.anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "eu.anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5 (EU)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1.1, - "output": 5.5, - "cache_read": 0.11, - "cache_write": 1.375 - } - }, - "global.anthropic.claude-opus-4-7": { - "id": "global.anthropic.claude-opus-4-7", - "name": "Claude Opus 4.7 (Global)", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "apac.anthropic.claude-sonnet-4-20250514-v1:0": { - "id": "apac.anthropic.claude-sonnet-4-20250514-v1:0", - "name": "Claude Sonnet 4 (APAC)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "status": "deprecated", - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "deepseek.v3.2": { - "id": "deepseek.v3.2", - "name": "DeepSeek V3.2", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2026-02-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 81920 - }, - "cost": { - "input": 0.62, - "output": 1.85 - } - }, - "qwen.qwen3-235b-a22b-2507-v1:0": { - "id": "qwen.qwen3-235b-a22b-2507-v1:0", - "name": "Qwen3 235B-A22B Instruct 2507", - "description": "Updated large open Qwen3 MoE instruct model for multilingual chat, coding, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-09-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.22, - "output": 0.88 - } - }, - "amazon.nova-lite-v1:0": { - "id": "amazon.nova-lite-v1:0", - "name": "Nova Lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.06, - "output": 0.24, - "cache_read": 0.015, - "cache_write": 0.06 - } - }, - "global.anthropic.claude-opus-5": { - "id": "global.anthropic.claude-opus-5", - "name": "Claude Opus 5 (Global)", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "openai.gpt-oss-20b-1:0": { - "id": "openai.gpt-oss-20b-1:0", - "name": "gpt-oss-20b", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.07, - "output": 0.3 - } - }, - "anthropic.claude-opus-5": { - "id": "anthropic.claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "meta.llama3-3-70b-instruct-v1:0": { - "id": "meta.llama3-3-70b-instruct-v1:0", - "name": "Llama 3.3 70B Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.72, - "output": 0.72 - } - }, - "minimax.minimax-m2": { - "id": "minimax.minimax-m2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204608, - "output": 128000 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "mistral.mistral-large-3-675b-instruct": { - "id": "mistral.mistral-large-3-675b-instruct", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "in.openai.gpt-5.6-luna": { - "id": "in.openai.gpt-5.6-luna", - "name": "GPT-5.6 Luna (India)", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.22, - "output": 1.32, - "cache_read": 0.022, - "cache_write": 0.275, - "tiers": [ - { - "input": 0.44, - "output": 1.98, - "cache_read": 0.044, - "cache_write": 0.55, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.44, - "output": 1.98, - "cache_read": 0.044, - "cache_write": 0.55 - } - } - }, - "eu.anthropic.claude-opus-4-8": { - "id": "eu.anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8 (EU)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "global.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "global.anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5 (Global)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "zai.glm-4.7": { - "id": "zai.glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.2 - } - }, - "au.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "au.anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5 (AU)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 - } - }, - "eu.anthropic.claude-opus-5-5": { - "id": "eu.anthropic.claude-opus-5-5", - "name": "Claude Opus 5.5 (EU)", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4.4, - "output": 22, - "cache_read": 0.22, - "cache_write": 5.5 - } - }, - "global.anthropic.claude-sonnet-4-6": { - "id": "global.anthropic.claude-sonnet-4-6", - "name": "Claude Sonnet 4.6 (Global)", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "jp.anthropic.claude-sonnet-4-6": { - "id": "jp.anthropic.claude-sonnet-4-6", - "name": "Claude Sonnet 4.6 (JP)", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 - } - }, - "google.gemma-3-27b-it": { - "id": "google.gemma-3-27b-it", - "name": "Gemma 3 27B IT", - "description": "Largest open Gemma 3 instruction model for multilingual text generation and visual understanding", - "family": "gemma", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 8192 - }, - "cost": { - "input": 0.23, - "output": 0.38 - } - }, - "eu.anthropic.claude-opus-4-6-v1": { - "id": "eu.anthropic.claude-opus-4-6-v1", - "name": "Claude Opus 4.6 (EU)", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "us.anthropic.claude-opus-5-5": { - "id": "us.anthropic.claude-opus-5-5", - "name": "Claude Opus 5.5 (US)", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4.4, - "output": 22, - "cache_read": 0.22, - "cache_write": 5.5 - } - }, - "us.anthropic.claude-opus-5": { - "id": "us.anthropic.claude-opus-5", - "name": "Claude Opus 5 (US)", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "us.openai.gpt-5.6-sol": { - "id": "us.openai.gpt-5.6-sol", - "name": "GPT-5.6 Sol (US)", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 4.4, - "output": 22, - "cache_read": 0.44, - "cache_write": 5.5, - "tiers": [ - { - "input": 8.8, - "output": 33, - "cache_read": 0.88, - "cache_write": 11, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8.8, - "output": 33, - "cache_read": 0.88, - "cache_write": 11 - } - } - }, - "us.moonshotai.kimi-k3": { - "id": "us.moonshotai.kimi-k3", - "name": "Kimi K3 (US)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 - } - }, - "anthropic.claude-sonnet-5": { - "id": "anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "mistral.magistral-small-2509": { - "id": "mistral.magistral-small-2509", - "name": "Magistral Small 1.2", - "description": "Open multimodal reasoning model for transparent analysis of text and images", - "family": "magistral", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-18", - "last_updated": "2025-09-18", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 40000 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "mistral.pixtral-large-2502-v1:0": { - "id": "mistral.pixtral-large-2502-v1:0", - "name": "Pixtral Large (25.02)", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-04-08", - "last_updated": "2025-04-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 2, - "output": 6 - } - }, - "anthropic.claude-opus-4-5-20251101-v1:0": { - "id": "anthropic.claude-opus-4-5-20251101-v1:0", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "nvidia.nemotron-nano-9b-v2": { - "id": "nvidia.nemotron-nano-9b-v2", - "name": "NVIDIA Nemotron Nano 9B v2", - "description": "Compact Nemotron model for efficient reasoning and deployable AI agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-18", - "last_updated": "2025-08-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.06, - "output": 0.23 - } - }, - "eu.amazon.nova-micro-v1:0": { - "id": "eu.amazon.nova-micro-v1:0", - "name": "Nova Micro (EU)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 10000 - }, - "cost": { - "input": 0.04, - "output": 0.16, - "cache_read": 0.01, - "cache_write": 0.04 - } - }, - "zai.glm-5": { - "id": "zai.glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 3.2 - } - }, - "us.meta.llama3-1-70b-instruct-v1:0": { - "id": "us.meta.llama3-1-70b-instruct-v1:0", - "name": "Llama 3.1 70B Instruct (US)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.72, - "output": 0.72 - } - }, - "jp.anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "jp.anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5 (JP)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1.1, - "output": 5.5, - "cache_read": 0.11, - "cache_write": 1.375 - } - }, - "nvidia.nemotron-nano-3-30b": { - "id": "nvidia.nemotron-nano-3-30b", - "name": "NVIDIA Nemotron Nano 3 30B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 8192 - }, - "cost": { - "input": 0.06, - "output": 0.24 - } - }, - "mistral.ministral-3-3b-instruct": { - "id": "mistral.ministral-3-3b-instruct", - "name": "Ministral 3 3B", - "description": "Compact open vision-language model for edge deployment, instruction following, and tool use", - "family": "ministral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.1, - "output": 0.1 - } - }, - "us.openai.gpt-6-luna": { - "id": "us.openai.gpt-6-luna", - "name": "GPT-6 Luna (US)", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.11, - "output": 0.55, - "cache_read": 0.011, - "cache_write": 0.1375, - "tiers": [ - { - "input": 0.22, - "output": 0.825, - "cache_read": 0.022, - "cache_write": 0.275, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.22, - "output": 0.825, - "cache_read": 0.022, - "cache_write": 0.275 - } - } - }, - "global.openai.gpt-5.6-sol": { - "id": "global.openai.gpt-5.6-sol", - "name": "GPT-5.6 Sol (Global)", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10 - } - } - }, - "google.gemma-4-26b-a4b": { - "id": "google.gemma-4-26b-a4b", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 0.13, - "output": 0.4 - } - }, - "xai.grok-4.6": { - "id": "xai.grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-18", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 2.2, - "output": 6.6, - "cache_read": 0.55 - } - }, - "global.xai.grok-4.6": { - "id": "global.xai.grok-4.6", - "name": "Grok 4.6 (Global)", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-18", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "au.anthropic.claude-opus-4-6-v1": { - "id": "au.anthropic.claude-opus-4-6-v1", - "name": "AU Anthropic Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "qwen.qwen3-vl-235b-a22b": { - "id": "qwen.qwen3-vl-235b-a22b", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language instruct model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-09-23", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262000 - }, - "cost": { - "input": 0.53, - "output": 2.66 - } - }, - "anthropic.claude-fable-5-1": { - "id": "anthropic.claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "jp.anthropic.claude-sonnet-5": { - "id": "jp.anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5 (JP)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2.2, - "output": 11, - "cache_read": 0.22, - "cache_write": 2.75 - } - }, - "qwen.qwen3-coder-next": { - "id": "qwen.qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Open-weight Qwen coding model for agents, repository edits, and multi-turn tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 1.2 - } - }, - "anthropic.claude-opus-4-8": { - "id": "anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "au.anthropic.claude-sonnet-5": { - "id": "au.anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5 (AU)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2.2, - "output": 11, - "cache_read": 0.22, - "cache_write": 2.75 - } - }, - "meta.llama4-scout-17b-instruct-v1:0": { - "id": "meta.llama4-scout-17b-instruct-v1:0", - "name": "Llama 4 Scout 17B Instruct", - "description": "Open Llama with long-context vision for efficient multimodal agents", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 10000000, - "output": 8192 - }, - "cost": { - "input": 0.17, - "output": 0.66 - } - }, - "openai.gpt-oss-safeguard-120b": { - "id": "openai.gpt-oss-safeguard-120b", - "name": "GPT OSS Safeguard 120B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "us.meta.llama3-3-70b-instruct-v1:0": { - "id": "us.meta.llama3-3-70b-instruct-v1:0", - "name": "Llama 3.3 70B Instruct (US)", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.72, - "output": 0.72 - } - }, - "global.openai.gpt-5.6-luna": { - "id": "global.openai.gpt-5.6-luna", - "name": "GPT-5.6 Luna (Global)", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } - } - }, - "au.anthropic.claude-sonnet-4-6": { - "id": "au.anthropic.claude-sonnet-4-6", - "name": "AU Anthropic Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 - } - }, - "openai.gpt-oss-120b": { - "id": "openai.gpt-oss-120b", - "name": "gpt-oss-120b", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/v1", - "shape": "responses" - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "deepseek.v3-v1:0": { - "id": "deepseek.v3-v1:0", - "name": "DeepSeek-V3.1", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-09-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 81920 - }, - "cost": { - "input": 0.58, - "output": 1.68 - } - }, - "moonshotai.kimi-k2.5": { - "id": "moonshotai.kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-02-06", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262143, - "output": 16384 - }, - "cost": { - "input": 0.6, - "output": 3 - } - }, - "eu.anthropic.claude-sonnet-4-6": { - "id": "eu.anthropic.claude-sonnet-4-6", - "name": "Claude Sonnet 4.6 (EU)", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 - } - }, - "us-gov.openai.gpt-oss-120b-1:0": { - "id": "us-gov.openai.gpt-oss-120b-1:0", - "name": "gpt-oss-120b (GovCloud)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.18, - "output": 0.72 - } - }, - "us.anthropic.claude-fable-5": { - "id": "us.anthropic.claude-fable-5", - "name": "Claude Fable 5 (US)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 11, - "output": 55, - "cache_read": 1.1, - "cache_write": 13.75 - } - }, - "us.deepseek.r1-v1:0": { - "id": "us.deepseek.r1-v1:0", - "name": "DeepSeek-R1 (US)", - "description": "Classic open reasoning model for transparent math, coding, and deliberate problem solving", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 1.35, - "output": 5.4 - } - }, - "eu.anthropic.claude-opus-5": { - "id": "eu.anthropic.claude-opus-5", - "name": "Claude Opus 5 (EU)", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "ca.amazon.nova-lite-v1:0": { - "id": "ca.amazon.nova-lite-v1:0", - "name": "Nova Lite (CA)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.064, - "output": 0.256, - "cache_read": 0.016, - "cache_write": 0.064 - } - }, - "us.openai.gpt-6-sol": { - "id": "us.openai.gpt-6-sol", - "name": "GPT-6 Sol (US)", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.2, - "output": 11, - "cache_read": 0.22, - "cache_write": 2.75, - "tiers": [ - { - "input": 4.4, - "output": 16.5, - "cache_read": 0.44, - "cache_write": 5.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4.4, - "output": 16.5, - "cache_read": 0.44, - "cache_write": 5.5 - } - } - }, - "au.anthropic.claude-opus-4-7": { - "id": "au.anthropic.claude-opus-4-7", - "name": "Claude Opus 4.7 (AU)", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "us.meta.llama4-scout-17b-instruct-v1:0": { - "id": "us.meta.llama4-scout-17b-instruct-v1:0", - "name": "Llama 4 Scout 17B Instruct (US)", - "description": "Open Llama with long-context vision for efficient multimodal agents", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 10000000, - "output": 8192 - }, - "cost": { - "input": 0.17, - "output": 0.66 - } - }, - "global.anthropic.claude-sonnet-5": { - "id": "global.anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5 (Global)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "global.anthropic.claude-fable-5": { - "id": "global.anthropic.claude-fable-5", - "name": "Claude Fable 5 (Global)", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "apac.amazon.nova-lite-v1:0": { - "id": "apac.amazon.nova-lite-v1:0", - "name": "Nova Lite (APAC)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.063, - "output": 0.252, - "cache_read": 0.01575, - "cache_write": 0.063 - } - }, - "us.anthropic.claude-opus-4-8": { - "id": "us.anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8 (US)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "nvidia.nemotron-super-3-120b": { - "id": "nvidia.nemotron-super-3-120b", - "name": "NVIDIA Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.65 - } - }, - "global.openai.gpt-6-luna": { - "id": "global.openai.gpt-6-luna", - "name": "GPT-6 Luna (Global)", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25 - } - } - }, - "us-gov.openai.gpt-oss-20b-1:0": { - "id": "us-gov.openai.gpt-oss-20b-1:0", - "name": "gpt-oss-20b (GovCloud)", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.084, - "output": 0.36 - } - }, - "us.anthropic.claude-opus-4-6-v1": { - "id": "us.anthropic.claude-opus-4-6-v1", - "name": "Claude Opus 4.6 (US)", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "meta.llama3-1-8b-instruct-v1:0": { - "id": "meta.llama3-1-8b-instruct-v1:0", - "name": "Llama 3.1 8B Instruct", - "description": "Compact open Llama model for lightweight chat, drafting, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.22, - "output": 0.22 - } - }, - "writer.palmyra-x5-v1:0": { - "id": "writer.palmyra-x5-v1:0", - "name": "Palmyra X5", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "palmyra", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1040000, - "input": 1040000, - "output": 8192 - }, - "cost": { - "input": 0.6, - "output": 6 - } - }, - "anthropic.claude-opus-4-1-20250805-v1:0": { - "id": "anthropic.claude-opus-4-1-20250805-v1:0", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "us.anthropic.claude-sonnet-4-6": { - "id": "us.anthropic.claude-sonnet-4-6", - "name": "Claude Sonnet 4.6 (US)", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 - } - }, - "global.anthropic.claude-opus-4-6-v1": { - "id": "global.anthropic.claude-opus-4-6-v1", - "name": "Claude Opus 4.6 (Global)", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "eu.anthropic.claude-sonnet-4-5-20250929-v1:0": { - "id": "eu.anthropic.claude-sonnet-4-5-20250929-v1:0", - "name": "Claude Sonnet 4.5 (EU)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3.3, - "output": 16.5, - "cache_read": 0.33, - "cache_write": 4.125 - } - }, - "jp.anthropic.claude-opus-4-7": { - "id": "jp.anthropic.claude-opus-4-7", - "name": "Claude Opus 4.7 (JP)", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "au.anthropic.claude-opus-5-5": { - "id": "au.anthropic.claude-opus-5-5", - "name": "Claude Opus 5.5 (AU)", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4.4, - "output": 22, - "cache_read": 0.22, - "cache_write": 5.5 - } - }, - "writer.palmyra-x4-v1:0": { - "id": "writer.palmyra-x4-v1:0", - "name": "Palmyra X4", - "description": "Enterprise language model for workflow automation, coding, data analysis, and tool use", - "family": "palmyra", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2024-10-09", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 122880, - "output": 8192 - }, - "cost": { - "input": 2.5, - "output": 10 - } - }, - "global.anthropic.claude-opus-4-5-20251101-v1:0": { - "id": "global.anthropic.claude-opus-4-5-20251101-v1:0", - "name": "Claude Opus 4.5 (Global)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic.claude-opus-4-7": { - "id": "anthropic.claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "us.anthropic.claude-sonnet-4-20250514-v1:0": { - "id": "us.anthropic.claude-sonnet-4-20250514-v1:0", - "name": "Claude Sonnet 4 (US)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "status": "deprecated", - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "eu.anthropic.claude-sonnet-5": { - "id": "eu.anthropic.claude-sonnet-5", - "name": "Claude Sonnet 5 (EU)", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2.2, - "output": 11, - "cache_read": 0.22, - "cache_write": 2.75 - } - }, - "zai.glm-4.7-flash": { - "id": "zai.glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0.07, - "output": 0.4 - } - }, - "moonshot.kimi-k2-thinking": { - "id": "moonshot.kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262143, - "output": 16000 - }, - "cost": { - "input": 0.6, - "output": 2.5 - } - }, - "eu.amazon.nova-pro-v1:0": { - "id": "eu.amazon.nova-pro-v1:0", - "name": "Nova Pro (EU)", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.92, - "output": 3.68, - "cache_read": 0.23, - "cache_write": 0.92 - } - }, - "global.openai.gpt-5.6-terra": { - "id": "global.openai.gpt-5.6-terra", - "name": "GPT-5.6 Terra (Global)", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } - } - }, - "openai.gpt-6-astra": { - "id": "openai.gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 11, - "output": 55, - "cache_read": 1.1, - "cache_write": 13.75, - "tiers": [ - { - "input": 22, - "output": 82.5, - "cache_read": 2.2, - "cache_write": 27.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 22, - "output": 82.5, - "cache_read": 2.2, - "cache_write": 27.5 - } - } - }, - "us.anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "us.anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5 (US)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1.1, - "output": 5.5, - "cache_read": 0.11, - "cache_write": 1.375 - } - }, - "meta.llama4-maverick-17b-instruct-v1:0": { - "id": "meta.llama4-maverick-17b-instruct-v1:0", - "name": "Llama 4 Maverick 17B Instruct", - "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2025-04-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 8192 - }, - "cost": { - "input": 0.24, - "output": 0.97 - } - }, - "qwen.qwen3-next-80b-a3b": { - "id": "qwen.qwen3-next-80b-a3b", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-11", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262000 - }, - "cost": { - "input": 0.15, - "output": 1.2 - } - }, - "eu.mistral.pixtral-large-2502-v1:0": { - "id": "eu.mistral.pixtral-large-2502-v1:0", - "name": "Pixtral Large (25.02) (EU)", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-04-08", - "last_updated": "2025-04-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 2, - "output": 6 - } - }, - "au.anthropic.claude-opus-4-8": { - "id": "au.anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8 (AU)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "us.anthropic.claude-fable-5-1": { - "id": "us.anthropic.claude-fable-5-1", - "name": "Claude Fable 5.1 (US)", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 11, - "output": 55, - "cache_read": 0.275, - "cache_write": 13.75 - } - }, - "global.moonshotai.kimi-k3": { - "id": "global.moonshotai.kimi-k3", - "name": "Kimi K3 (Global)", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "in.openai.gpt-5.6-terra": { - "id": "in.openai.gpt-5.6-terra", - "name": "GPT-5.6 Terra (India)", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.2, - "output": 13.2, - "cache_read": 0.22, - "cache_write": 2.75, - "tiers": [ - { - "input": 4.4, - "output": 19.8, - "cache_read": 0.44, - "cache_write": 5.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4.4, - "output": 19.8, - "cache_read": 0.44, - "cache_write": 5.5 - } - } - }, - "openai.gpt-oss-20b": { - "id": "openai.gpt-oss-20b", - "name": "gpt-oss-20b", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/v1", - "shape": "responses" - }, - "cost": { - "input": 0.07, - "output": 0.3 - } - }, - "mistral.ministral-3-14b-instruct": { - "id": "mistral.ministral-3-14b-instruct", - "name": "Ministral 14B 3.0", - "description": "Open vision-language model for efficient local deployment, instruction following, and tool use", - "family": "ministral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.2, - "output": 0.2 - } - }, - "au.anthropic.claude-haiku-4-5-20251001-v1:0": { - "id": "au.anthropic.claude-haiku-4-5-20251001-v1:0", - "name": "Claude Haiku 4.5 (AU)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1.1, - "output": 5.5, - "cache_read": 0.11, - "cache_write": 1.375 - } - }, - "openai.gpt-5.6-sol": { - "id": "openai.gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 4.4, - "output": 22, - "cache_read": 0.44, - "cache_write": 5.5, - "tiers": [ - { - "input": 8.8, - "output": 33, - "cache_read": 0.88, - "cache_write": 11, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8.8, - "output": 33, - "cache_read": 0.88, - "cache_write": 11 - } - } - }, - "minimax.minimax-m2.1": { - "id": "minimax.minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "us.anthropic.claude-opus-4-5-20251101-v1:0": { - "id": "us.anthropic.claude-opus-4-5-20251101-v1:0", - "name": "Claude Opus 4.5 (US)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "eu.anthropic.claude-opus-4-7": { - "id": "eu.anthropic.claude-opus-4-7", - "name": "Claude Opus 4.7 (EU)", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "jp.amazon.nova-2-lite-v1:0": { - "id": "jp.amazon.nova-2-lite-v1:0", - "name": "Nova 2 Lite (JP)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-10", - "release_date": "2025-12-02", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65535 - }, - "cost": { - "input": 0.396, - "output": 3.311, - "cache_read": 0.099, - "cache_write": 0.396 - } - }, - "us.mistral.pixtral-large-2502-v1:0": { - "id": "us.mistral.pixtral-large-2502-v1:0", - "name": "Pixtral Large (25.02) (US)", - "description": "Mistral vision-language model for image understanding and multimodal chat", - "family": "pixtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-04-08", - "last_updated": "2025-04-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 2, - "output": 6 - } - }, - "jp.anthropic.claude-opus-5": { - "id": "jp.anthropic.claude-opus-5", - "name": "Claude Opus 5 (JP)", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.5, - "cache_read": 0.55, - "cache_write": 6.875 - } - }, - "global.anthropic.claude-sonnet-4-20250514-v1:0": { - "id": "global.anthropic.claude-sonnet-4-20250514-v1:0", - "name": "Claude Sonnet 4 (Global)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "status": "deprecated", - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "global.anthropic.claude-opus-4-8": { - "id": "global.anthropic.claude-opus-4-8", - "name": "Claude Opus 4.8 (Global)", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "openai.gpt-5.4": { - "id": "openai.gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/amazon-bedrock/mantle", - "api": "https://bedrock-mantle.${AWS_REGION}.api.aws/openai/v1", - "shape": "responses" - }, - "cost": { - "input": 2.75, - "output": 16.5, - "cache_read": 0.275 - } - }, - "mistral.voxtral-mini-3b-2507": { - "id": "mistral.voxtral-mini-3b-2507", - "name": "Voxtral Mini 3B 2507", - "description": "Open audio-language model for speech transcription, audio understanding, and voice-driven tool use", - "family": "voxtral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-15", - "last_updated": "2025-07-15", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 4096 - }, - "cost": { - "input": 0.04, - "output": 0.04 - } - }, - "openai.gpt-oss-120b-1:0": { - "id": "openai.gpt-oss-120b-1:0", - "name": "gpt-oss-120b", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "amazon.nova-2-lite-v1:0": { - "id": "amazon.nova-2-lite-v1:0", - "name": "Nova 2 Lite", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "nova", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-10", - "release_date": "2025-12-02", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65535 - }, - "cost": { - "input": 0.33, - "output": 2.75, - "cache_read": 0.0825, - "cache_write": 0.33 - } - }, - "us.amazon.nova-micro-v1:0": { - "id": "us.amazon.nova-micro-v1:0", - "name": "Nova Micro (US)", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 10000 - }, - "cost": { - "input": 0.035, - "output": 0.14, - "cache_read": 0.00875, - "cache_write": 0.035 - } - }, - "mistral.ministral-3-8b-instruct": { - "id": "mistral.ministral-3-8b-instruct", - "name": "Ministral 3 8B", - "description": "Compact open vision-language model for edge deployment, instruction following, and tool use", - "family": "ministral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.15, - "output": 0.15 - } - } - } - }, - "synthetic": { - "id": "synthetic", - "env": [ - "SYNTHETIC_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.synthetic.new/openai/v1", - "name": "Synthetic", - "doc": "https://synthetic.new/pricing", - "models": { - "hf:moonshotai/Kimi-K3": { - "id": "hf:moonshotai/Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-16", - "last_updated": "2026-07-27", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 65536 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.45 - } - }, - "hf:moonshotai/Kimi-K2.7-Code": { - "id": "hf:moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.95 - } - }, - "hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4": { - "id": "hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4", - "name": "Nemotron 3 Super 120B A12B", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 1, - "cache_read": 0.3 - } - }, - "hf:zai-org/GLM-5.2": { - "id": "hf:zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 65536 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 1.4 - } - }, - "hf:zai-org/GLM-5.3-Flash": { - "id": "hf:zai-org/GLM-5.3-Flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 65536 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.04 - } - }, - "hf:zai-org/GLM-4.7-Flash": { - "id": "hf:zai-org/GLM-4.7-Flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.1 - } - }, - "hf:deepseek-ai/DeepSeek-V4.1-Flash": { - "id": "hf:deepseek-ai/DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 65536 - }, - "cost": { - "input": 0.6, - "output": 1.2, - "cache_read": 0.03 - } - }, - "hf:Qwen/Qwen3.6-27B": { - "id": "hf:Qwen/Qwen3.6-27B", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.45, - "output": 3.6, - "cache_read": 0.45 - } - }, - "hf:openai/gpt-oss-120b": { - "id": "hf:openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.1, - "cache_read": 0.1 - } - }, - "hf:MiniMaxAI/MiniMax-M3": { - "id": "hf:MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 65536 - }, - "cost": { - "input": 0.6, - "output": 1.2, - "cache_read": 0.6 - } - } - } - }, - "llmgateway": { - "id": "llmgateway", - "env": [ - "LLMGATEWAY_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.llmgateway.io/v1", - "name": "DevPass (LLM Gateway)", - "doc": "https://llmgateway.io/docs", - "models": { - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "Grok 4.1 Fast Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 30000 - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } - }, - "glm-4.6v": { - "id": "glm-4.6v", - "name": "GLM-4.6V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-08", - "last_updated": "2025-12-08", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 0.9, - "cache_read": 0.05 - } - }, - "qwen-flash": { - "id": "qwen-flash", - "name": "Qwen Flash", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0.0625 - } - }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } - }, - "glm-4.5": { - "id": "glm-4.5", - "name": "GLM-4.5", - "description": "Hybrid-reasoning GLM release that made the 4.5 line broadly useful", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131000, - "output": 98304 - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 - } - }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 180 - } - }, - "mimo-v2.6-pro": { - "id": "mimo-v2.6-pro", - "name": "MiMo-V2.6-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036 - } - }, - "gemma-4-31b-it": { - "id": "gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.25, - "cache_read": 0.01 - } - }, - "grok-4-5": { - "id": "grok-4-5", - "name": "Grok 4.5", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3 - } - }, - "gpt-3.5-turbo": { - "id": "gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16385, - "output": 4096 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0 - } - }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "Grok 4.1 Fast Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 2000000 - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } - }, - "qwen3-vl-235b-a22b-instruct": { - "id": "qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-15", - "last_updated": "2025-09-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 8192 - }, - "cost": { - "input": 0.2, - "output": 0.88, - "cache_read": 0.11 - } - }, - "seed-1-6-flash-250715": { - "id": "seed-1-6-flash-250715", - "name": "Seed 1.6 Flash (250715)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-26", - "last_updated": "2025-07-26", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.07, - "output": 0.3, - "cache_read": 0.015 - } - }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 3.75, - "cache_read": 0.25, - "cache_write": 3.125 - } - }, - "muse-spark-1.3": { - "id": "muse-spark-1.3", - "name": "Muse Spark 1.3", - "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It improves long-horizon agent collaboration, instruction following, and coding efficiency relative to Muse Spark 1.2.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 - } - }, - "Qwen3.8-27B": { - "id": "Qwen3.8-27B", - "name": "Qwen3.8 27B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.08, - "output": 0.35, - "cache_read": 0.05 - } - }, - "gpt-4o-transcribe": { - "id": "gpt-4o-transcribe", - "name": "GPT-4o Transcribe", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-03-20", - "last_updated": "2025-03-20", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16000, - "output": 16000 - }, - "cost": { - "input": 2.5, - "output": 10 - } - }, - "mimo-v2.5": { - "id": "mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028, - "tiers": [ - { - "input": 0.8, - "output": 4, - "cache_read": 0.16, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.8, - "output": 4, - "cache_read": 0.16 - } - } - }, - "gpt-5.5-pro": { - "id": "gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } - } - }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "mistral-large-latest": { - "id": "mistral-large-latest", - "name": "Mistral Large (latest)", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2024-11-01", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 262144 - }, - "cost": { - "input": 4, - "output": 12 - } - }, - "nemotron-3.5-lightning": { - "id": "nemotron-3.5-lightning", - "name": "Nemotron 3.5 Lightning 30B A3B", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.08, - "output": 0.2 - } - }, - "qwen3-vl-235b-a22b-thinking": { - "id": "qwen3-vl-235b-a22b-thinking", - "name": "Qwen3 VL 235B A22B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-15", - "last_updated": "2025-09-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.98, - "output": 3.95 - } - }, - "qwen3.8-2.4t-a95b": { - "id": "qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1010000, - "output": 131072 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 - } - }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Codex GPT for repository edits, code review, and practical software agents", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "qwen-coder-plus": { - "id": "qwen-coder-plus", - "name": "Qwen Coder Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2024-09-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.502, - "output": 1.004 - } - }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 - } - }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "qwen3-vl-plus": { - "id": "qwen3-vl-plus", - "name": "Qwen3-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.2, - "output": 1.6, - "reasoning": 4.8, - "cache_read": 0.04, - "cache_write": 0.25 - } - }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.088, - "output": 0.25, - "cache_read": 0.025 - } - }, - "fugu-max": { - "id": "fugu-max", - "name": "Fugu Max", - "description": "Multi-agent model for routing expert agents across complex analytical tasks", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25 - } - }, - "ministral-8b-2512": { - "id": "ministral-8b-2512", - "name": "Ministral 8B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "mistral", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.15 - } - }, - "glm-4.6": { - "id": "glm-4.6", - "name": "GLM-4.6", - "description": "Late GLM-4 workhorse for coding agents, reasoning, and structured tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-30", - "last_updated": "2025-09-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.55, - "output": 2.2, - "cache_read": 0.11, - "cache_write": 0 - } - }, - "glm-5.2-fast": { - "id": "glm-5.2-fast", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 2.2, - "output": 6.5, - "cache_read": 0.45 - } - }, - "muse-glimmer-30b": { - "id": "muse-glimmer-30b", - "name": "Muse Glimmer 30B", - "description": "Muse Glimmer is a 30-billion-parameter open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark for always-on local agents, tool use, coding, and image understanding.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-01-04", - "release_date": "2026-08-10", - "last_updated": "2026-08-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.04 - } - }, - "grok-build-0-1": { - "id": "grok-build-0-1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2, - "tiers": [ - { - "input": 2, - "output": 4, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 4, - "cache_read": 0.4 - } - } - }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B A22B Instruct (2507)", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-08", - "last_updated": "2025-07-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 8192 - }, - "cost": { - "input": 0.09, - "output": 0.58 - } - }, - "qwen-max": { - "id": "qwen-max", - "name": "Qwen Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 8192 - }, - "cost": { - "input": 1.6, - "output": 6.4 - } - }, - "qwen3-next-80b-a3b-thinking": { - "id": "qwen3-next-80b-a3b-thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Efficient Qwen thinking model for local reasoning, math, and coding agents", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 1.2 - } - }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max Preview", - "description": "Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-19", - "last_updated": "2026-07-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.25, - "cache_write": 2.5 - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.108, - "output": 0.675, - "cache_read": 0.06 - } - }, - "glm-4.5-x": { - "id": "glm-4.5-x", - "name": "GLM-4.5 X", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "status": "beta", - "cost": { - "input": 2.2, - "output": 8.9, - "cache_read": 0.45 - } - }, - "grok-4": { - "id": "grok-4", - "name": "Grok 4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-07-09", - "last_updated": "2025-07-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 - } - }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "llama-4-maverick-17b-instruct": { - "id": "llama-4-maverick-17b-instruct", - "name": "Llama 4 Maverick 17B Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 2048 - }, - "cost": { - "input": 0.27, - "output": 0.85 - } - }, - "gpt-5.2-pro": { - "id": "gpt-5.2-pro", - "name": "GPT-5.2 Pro", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 21, - "output": 168 - } - }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 - } - }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 - } - }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 203000, - "output": 131072 - }, - "cost": { - "input": 0.72, - "output": 2.3, - "cache_read": 0.144, - "cache_write": 0 - } - }, - "fugu-ultra-v2.0": { - "id": "fugu-ultra-v2.0", - "name": "Fugu Ultra v2.0", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "family": "fugu", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-11", - "last_updated": "2026-09-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - } - }, - "gpt-4": { - "id": "gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 30, - "output": 60 - } - }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1050000, - "output": 384000 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.003 - } - }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 - } - }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 228700, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "kimi-k3-fast": { - "id": "kimi-k3-fast", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1040384, - "output": 131072 - }, - "cost": { - "input": 4.5, - "output": 22.5, - "cache_read": 0.45 - } - }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "output": 16384 - }, - "cost": { - "input": 0.36, - "output": 0.87, - "reasoning": 8.4 - } - }, - "muse-spark-1.1": { - "id": "muse-spark-1.1", - "name": "Muse Spark 1.1", - "description": "Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-08", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 - } - }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 - } - }, - "qwen-plus": { - "id": "qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 1.2, - "reasoning": 4, - "cache_read": 0.08, - "cache_write": 0.5 - } - }, - "kimi-k2.7-code-highspeed": { - "id": "kimi-k2.7-code-highspeed", - "name": "Kimi K2.7 Code Highspeed", - "description": "Lower-latency Kimi Code variant for interactive edits and coding-agent loops", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.9, - "output": 8, - "cache_read": 0.38 - } - }, - "qwen3.7-flash": { - "id": "qwen3.7-flash", - "name": "Qwen3.7 Flash", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-07-27", - "last_updated": "2026-07-27", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 0.03, - "output": 0.13, - "cache_read": 0.006, - "cache_write": 0.0375 - } - }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 3.05, - "cache_read": 0.13 - } - }, - "minimax-m2.1-lightning": { - "id": "minimax-m2.1-lightning", - "name": "MiniMax M2.1 Lightning", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 131072 - }, - "cost": { - "input": 0.12, - "output": 0.48 - } - }, - "claude-sonnet-4-5": { - "id": "claude-sonnet-4-5", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen3 model for coding agents, complex reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.845, - "output": 3.38, - "cache_read": 0.6, - "cache_write": 3.75 - } - }, - "claude-opus-5-5": { - "id": "claude-opus-5-5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 - } - }, - "glm-4.5-airx": { - "id": "glm-4.5-airx", - "name": "GLM-4.5 AirX", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.1, - "output": 4.5, - "cache_read": 0.22 - } - }, - "glm-4.5v": { - "id": "glm-4.5v", - "name": "GLM-4.5V", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-08-11", - "last_updated": "2025-08-11", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.6, - "output": 1.8, - "cache_read": 0.11 - } - }, - "claude-fable-5-1": { - "id": "claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "devstral-2512": { - "id": "devstral-2512", - "name": "Devstral 2", - "description": "Mistral's coding-agent model for repository work, terminal tasks, and software fixes", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.4, - "output": 2 - } - }, - "llama-4-scout-17b-instruct": { - "id": "llama-4-scout-17b-instruct", - "name": "Llama 4 Scout 17B Instruct", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-04-05", - "last_updated": "2025-04-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 2048 - }, - "cost": { - "input": 0.18, - "output": 0.59 - } - }, - "gemini-3.6-flash": { - "id": "gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.08333 - } - }, - "qwen3-235b-a22b-fp8": { - "id": "qwen3-235b-a22b-fp8", - "name": "Qwen3 235B A22B FP8", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-28", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "output": 8192 - }, - "cost": { - "input": 0.2, - "output": 0.8 - } - }, - "o1": { - "id": "o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 - } - }, - "glm-4-32b-0414-128k": { - "id": "glm-4-32b-0414-128k", - "name": "GLM-4 32B (0414-128k)", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.1, - "output": 0.1 - } - }, - "qwen3-coder-480b-a35b-instruct": { - "id": "qwen3-coder-480b-a35b-instruct", - "name": "Qwen3-Coder 480B-A35B Instruct", - "description": "Open Qwen coding heavyweight for repository reasoning and agentic engineering", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.38, - "output": 1.55 - } - }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 65536 - }, - "cost": { - "input": 0.07, - "output": 0.27 - } - }, - "sonar-pro": { - "id": "sonar-pro", - "name": "Sonar Pro", - "description": "Deeper Sonar search model with broader retrieval and stronger synthesis", - "family": "sonar-pro", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8192 - }, - "cost": { - "input": 3, - "output": 15 - } - }, - "ling-3.0-flash": { - "id": "ling-3.0-flash", - "name": "InclusionAI Ling 3.0 Flash", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "ling", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-02", - "last_updated": "2026-08-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.06, - "output": 0.18, - "cache_read": 0.012 - } - }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Thinking Kimi model for slower research passes, planning, and hard technical questions", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.06 - } - }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 272000 - }, - "cost": { - "input": 15, - "output": 120 - } - }, - "qwen3-vl-30b-a3b-instruct": { - "id": "qwen3-vl-30b-a3b-instruct", - "name": "Qwen3 VL 30B A3B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-02", - "last_updated": "2025-10-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "seed-1-6-250915": { - "id": "seed-1-6-250915", - "name": "Seed 1.6 (250915)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-15", - "last_updated": "2025-09-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05 - } - }, - "gemma-4-26b-a4b-it": { - "id": "gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.07, - "output": 0.34 - } - }, - "claude-opus-4-5-20251101": { - "id": "claude-opus-4-5-20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.08333 - } - }, - "llama-3-70b-instruct": { - "id": "llama-3-70b-instruct", - "name": "Llama 3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2024-04-18", - "last_updated": "2024-04-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 8000 - }, - "cost": { - "input": 0.51, - "output": 0.74 - } - }, - "qwen3-vl-flash": { - "id": "qwen3-vl-flash", - "name": "Qwen3 VL Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-09", - "last_updated": "2025-10-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.01 - } - }, - "qwen3-235b-a22b-thinking-2507": { - "id": "qwen3-235b-a22b-thinking-2507", - "name": "Qwen3 235B A22B Thinking (2507)", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-08", - "last_updated": "2025-07-08", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 8192 - }, - "cost": { - "input": 0.3, - "output": 3 - } - }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-09-03", - "last_updated": "2026-09-03", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "output": 1050000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "minimax-text-01": { - "id": "minimax-text-01", - "name": "MiniMax Text 01", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-01-15", - "last_updated": "2025-01-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.2, - "output": 1.1 - } - }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "glm-4.6v-flashx": { - "id": "glm-4.6v-flashx", - "name": "GLM-4.6V FlashX", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-08", - "last_updated": "2025-12-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16000 - }, - "cost": { - "input": 0.04, - "output": 0.4, - "cache_read": 0.004 - } - }, - "grok-4-20-beta-0309-non-reasoning": { - "id": "grok-4-20-beta-0309-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 30000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.405, - "output": 1.98, - "cache_read": 0.225 - } - }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 4096 - }, - "cost": { - "input": 0.135, - "output": 0.4 - } - }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 - } - }, - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.248, - "output": 1.485 - } - }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.15, - "output": 1.2 - } - }, - "fugu-ultra": { - "id": "fugu-ultra", - "name": "Fugu Ultra", - "description": "Quality-first multi-agent model for hard research, analysis, and competitions", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-06-22", - "last_updated": "2026-06-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - } - }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32766 - }, - "cost": { - "input": 0.04, - "output": 0.19, - "cache_read": 0.01 - } - }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.08, - "output": 0.32, - "cache_read": 0.017, - "cache_write": 0.375 - } - }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 0.08333 - } - }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } - } - }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.05, - "cache_write": 0.3125 - } - }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1, - "max": 24576 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 - } - }, - "qwen3-coder-flash": { - "id": "qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 1.5, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "nemotron-3-ultra-550b": { - "id": "nemotron-3-ultra-550b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 128000 - }, - "cost": { - "input": 0.5, - "output": 2.2, - "cache_read": 0.1 - } - }, - "llama-3.2-11b-instruct": { - "id": "llama-3.2-11b-instruct", - "name": "Llama 3.2 11B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2024-09-25", - "last_updated": "2024-09-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.07, - "output": 0.33 - } - }, - "inkling-small": { - "id": "inkling-small", - "name": "Inkling Small", - "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 1048576 - }, - "cost": { - "input": 0.45, - "output": 1.2, - "cache_read": 0.1 - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.0036, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.4 - } - } - }, - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.2, - "cache_write": 1.25 - } - }, - "glm-4.7-flash": { - "id": "glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Budget GLM lane for fast coding help, routing, and everyday automation", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0.06, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0 - } - }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.15 - } - }, - "inkling": { - "id": "inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 1048576 - }, - "cost": { - "input": 0.95, - "output": 4.05, - "cache_read": 0.16 - } - }, - "muse-spark-1.2": { - "id": "muse-spark-1.2", - "name": "Muse Spark 1.2", - "description": "Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1 with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows.", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-05", - "last_updated": "2026-08-05", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.25, - "output": 4.25, - "cache_read": 0.15 - } - }, - "mimo-v2.6-flash": { - "id": "mimo-v2.6-flash", - "name": "MiMo-V2.6-Flash", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 - } - }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "seed-1-8-251228": { - "id": "seed-1-8-251228", - "name": "Seed 1.8 (251228)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-18", - "last_updated": "2025-12-18", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05 - } - }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.38, - "output": 1.98, - "cache_read": 0.19, - "cache_write": 0 - } - }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 512000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25 - } - }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016, - "cache_write": 0.2 - } - }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "seed-1-6-250615": { - "id": "seed-1-6-250615", - "name": "Seed 1.6 (250615)", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "seed", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-06-25", - "last_updated": "2025-06-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 8192 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.05 - } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "qwen35-397b-a17b": { - "id": "qwen35-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.6, - "output": 3.6 - } - }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": false, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "qwen3.6-max-preview": { - "id": "qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-20", - "last_updated": "2026-04-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 1.3, - "output": 7.8, - "cache_read": 0.13, - "cache_write": 1.625 - } - }, - "claude-opus-4-6": { - "id": "claude-opus-4-6", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.08333 - } - }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 0.8, - "output": 2.55, - "cache_read": 0.16, - "cache_write": 0 - } - }, - "ernie-4.5-vl-424b-a47b": { - "id": "ernie-4.5-vl-424b-a47b", - "name": "ERNIE 4.5 VL 424B A47B", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "ernie", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2025-06-30", - "last_updated": "2025-06-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 123000, - "output": 123000 - }, - "cost": { - "input": 0.42, - "output": 1.25 - } - }, - "hy3": { - "id": "hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "input": 192000, - "output": 128000 - }, - "cost": { - "input": 0.132, - "output": 0.528, - "cache_read": 0.033 - } - }, - "kimi-k2": { - "id": "kimi-k2", - "name": "Kimi K2", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-11", - "last_updated": "2025-07-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 16384 - }, - "cost": { - "input": 0.57, - "output": 2.3, - "cache_read": 0.5 - } - }, - "qwen-omni-turbo": { - "id": "qwen-omni-turbo", - "name": "Qwen-Omni Turbo", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01-19", - "last_updated": "2025-03-26", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 2048 - }, - "cost": { - "input": 0.2, - "output": 0.8 - } - }, - "hy-mt2-plus": { - "id": "hy-mt2-plus", - "name": "Hy-MT2 Plus", - "description": "Tool-capable chat model for instruction following and agentic application workflows", - "family": "Hy", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.074, - "output": 0.295 - } - }, - "grok-4-3": { - "id": "grok-4-3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 30000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.931, - "output": 2.93, - "cache_read": 0.173, - "cache_write": 0 - } - }, - "grok-4-20-non-reasoning": { - "id": "grok-4-20-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 30000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05 - } - }, - "grok-4-20-beta-0309-reasoning": { - "id": "grok-4-20-beta-0309-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 30000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "minimax-m2.7-highspeed": { - "id": "minimax-m2.7-highspeed", - "name": "MiniMax-M2.7-highspeed", - "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.06, - "cache_write": 0.375 - } - }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "glm-4.7-flashx": { - "id": "glm-4.7-flashx", - "name": "GLM-4.7-FlashX", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 0.07, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0 - } - }, - "muse-spark-1.2-contributor": { - "id": "muse-spark-1.2-contributor", - "name": "Muse Spark 1.2 Contributor", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-06", - "last_updated": "2026-08-06", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 - } - }, - "sonar": { - "id": "sonar", - "name": "Sonar", - "description": "Fast web-grounded Sonar for current answers, citations, and lightweight retrieval", - "family": "sonar", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 130000, - "output": 4096 - }, - "cost": { - "input": 1, - "output": 1 - } - }, - "mistral-large-2512": { - "id": "mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.5, - "output": 1.5 - } - }, - "grok-4-6": { - "id": "grok-4-6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "gemini-pro-latest": { - "id": "gemini-pro-latest", - "name": "Gemini Pro Latest", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-27", - "last_updated": "2026-02-27", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 - } - }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 - } - }, - "minimax-m2.5-highspeed": { - "id": "minimax-m2.5-highspeed", - "name": "MiniMax-M2.5-highspeed", - "description": "High-speed MiniMax model for low-latency coding and agent workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-13", - "last_updated": "2026-02-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.6, - "output": 2.4, - "cache_read": 0.03, - "cache_write": 0.375 - } - }, - "gemini-3.8-flash": { - "id": "gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.08333 - } - }, - "ministral-14b-2512": { - "id": "ministral-14b-2512", - "name": "Ministral 14B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "mistral", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 8192 - }, - "cost": { - "input": 0.2, - "output": 0.2 - } - }, - "glm-5-turbo": { - "id": "glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.24 - } - }, - "gpt-6-luna": { - "id": "gpt-6-luna", - "name": "GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125 - } - }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1050000, - "output": 384000 - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 - } - }, - "step-3.7-flash": { - "id": "step-3.7-flash", - "name": "Step 3.7 Flash", - "description": "Newer StepFun flash model for faster agents, coding, and multimodal prompts", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2026-03-01", - "release_date": "2026-05-29", - "last_updated": "2026-05-29", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 256000 - }, - "cost": { - "input": 0.2, - "output": 1.15, - "cache_read": 0.04 - } - }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 - } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "gpt-4o-mini-transcribe": { - "id": "gpt-4o-mini-transcribe", - "name": "GPT-4o Mini Transcribe", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2025-03-20", - "last_updated": "2025-03-20", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16000, - "output": 16000 - }, - "cost": { - "input": 1.25, - "output": 5 - } - }, - "sonar-reasoning-pro": { - "id": "sonar-reasoning-pro", - "name": "Sonar Reasoning Pro", - "description": "Web-grounded Sonar for multi-step research questions that need cited reasoning", - "family": "sonar-reasoning", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 2, - "output": 8 - } - }, - "llama-3.2-3b-instruct": { - "id": "llama-3.2-3b-instruct", - "name": "Llama 3.2 3B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2024-09-18", - "last_updated": "2024-09-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32000 - }, - "cost": { - "input": 0.03, - "output": 0.05 - } - }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } - } - }, - "mistral-small-2506": { - "id": "mistral-small-2506", - "name": "Mistral Small 3.2", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.1, - "output": 0.3 - } - }, - "grok-4-20-reasoning": { - "id": "grok-4-20-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 30000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "grok-4-7": { - "id": "grok-4-7", - "name": "Grok 4.7", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-09-21", - "last_updated": "2026-09-21", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 16384 - }, - "cost": { - "input": 0.26, - "output": 0.38, - "cache_read": 0.13 - } - }, - "gpt-4-turbo": { - "id": "gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 10, - "output": 30 - } - }, - "atria-dawn-preview": { - "id": "atria-dawn-preview", - "name": "Atria Dawn Preview", - "description": "Preview model for early access evaluation, prototyping, and compatibility testing", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-09-12", - "last_updated": "2026-09-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "hy4-preview": { - "id": "hy4-preview", - "name": "Hy4 preview", - "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.834, - "output": 2.501, - "cache_read": 0.042 - } - }, - "muse-spark-1.3-contributor": { - "id": "muse-spark-1.3-contributor", - "name": "Muse Spark 1.3 Contributor", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "muse", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 0.1, - "output": 0.2, - "cache_read": 0.002 - } - }, - "glm-4.5-air": { - "id": "glm-4.5-air", - "name": "GLM-4.5-Air", - "description": "Lighter GLM-4.5 variant for fast coding assistance and cheaper agents", - "family": "glm-air", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131000, - "output": 98304 - }, - "cost": { - "input": 0.13, - "output": 0.85, - "cache_read": 0.025, - "cache_write": 0 - } - }, - "qwen-plus-latest": { - "id": "qwen-plus-latest", - "name": "Qwen Plus Latest", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-01-25", - "last_updated": "2025-01-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 8192 - }, - "cost": { - "input": 0.4, - "output": 1.2, - "cache_read": 0.08, - "cache_write": 0.5 - } - }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32766 - }, - "cost": { - "input": 0.032, - "output": 0.14, - "cache_read": 0.032 - } - }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "GPT-5.1 Codex mini", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "ministral-3b-2512": { - "id": "ministral-3b-2512", - "name": "Ministral 3B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "mistral", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.1, - "output": 0.1 - } - }, - "o3": { - "id": "o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4, - "cache_read": 0.2 - } - }, - "claude-opus-4-7": { - "id": "claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.08, - "cache_write": 0.5 - } - }, - "minimax-m2.1": { - "id": "minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "Earlier MiniMax agent model for practical coding and productivity tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.27, - "output": 1.1 - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5 - } - }, - "auto": { - "id": "auto", - "name": "Auto Route", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "llama-3.1-70b-instruct": { - "id": "llama-3.1-70b-instruct", - "name": "Llama 3.1 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 2048 - }, - "status": "beta", - "cost": { - "input": 0.72, - "output": 0.72 - } - }, - "minimax-m2": { - "id": "minimax-m2", - "name": "MiniMax-M2", - "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 131072 - }, - "cost": { - "input": 0.2, - "output": 1, - "cache_read": 0.03 - } - }, - "custom": { - "id": "custom", - "name": "Custom Model", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-01-01", - "last_updated": "2024-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "codestral-2508": { - "id": "codestral-2508", - "name": "Codestral", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "mistral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-30", - "last_updated": "2025-07-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 16384 - }, - "cost": { - "input": 0.3, - "output": 0.9 - } - }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1050000, - "output": 384000 - }, - "cost": { - "input": 0.05, - "output": 0.1, - "cache_read": 0.01 - } - }, - "gpt-6-sol": { - "id": "gpt-6-sol", - "name": "GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 0.08333 - } - } - } - }, - "sap-ai-core": { - "id": "sap-ai-core", - "env": [ - "AICORE_SERVICE_KEY" - ], - "npm": "@jerome-benoit/sap-ai-provider-v2", - "name": "SAP AI Core", - "doc": "https://help.sap.com/docs/sap-ai-core", - "models": { - "anthropic--claude-4-opus": { - "id": "anthropic--claude-4-opus", - "name": "anthropic--claude-4-opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "gpt-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } - }, - "nvidia--llama-3.2-nv-embedqa-1b": { - "id": "nvidia--llama-3.2-nv-embedqa-1b", - "name": "nvidia--llama-3.2-nv-embedqa-1b", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-09-25", - "last_updated": "2024-09-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 4096 - }, - "cost": { - "input": 0.07, - "output": 0 - } - }, - "text-embedding-3-small": { - "id": "text-embedding-3-small", - "name": "text-embedding-3-small", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8191, - "output": 1536 - }, - "cost": { - "input": 0.02, - "output": 0 - } - }, - "sap-abap-1": { - "id": "sap-abap-1", - "name": "sap-abap-1", - "description": "SAP-hosted model for ABAP code generation and enterprise development tasks", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-11-26", - "last_updated": "2025-11-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 4096 - }, - "cost": { - "input": 0.48, - "output": 1.7 - } - }, - "anthropic--claude-4.8-opus": { - "id": "anthropic--claude-4.8-opus", - "name": "anthropic--claude-4.8-opus", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "gpt-5-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "anthropic--claude-3-haiku": { - "id": "anthropic--claude-3-haiku", - "name": "anthropic--claude-3-haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-03-13", - "last_updated": "2024-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 4096 - }, - "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.03, - "cache_write": 0.3 - } - }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "gpt-4.1-nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.08, - "output": 0.26 - } - }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "gpt-5-nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 - } - }, - "anthropic--claude-4.5-haiku": { - "id": "anthropic--claude-4.5-haiku", - "name": "anthropic--claude-4.5-haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "anthropic--claude-4.7-opus": { - "id": "anthropic--claude-4.7-opus", - "name": "anthropic--claude-4.7-opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gemini-embedding": { - "id": "gemini-embedding", - "name": "Gemini Embedding 001", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "gemini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-05", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2048, - "output": 1 - } - }, - "gemini-embedding-2": { - "id": "gemini-embedding-2", - "name": "Gemini Embedding 2", - "description": "Multimodal embedding model mapping text, images, video, audio, and PDFs into a unified embedding space", - "family": "gemini", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-11", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 3072 - } - }, - "amazon--nova-pro": { - "id": "amazon--nova-pro", - "name": "amazon--nova-pro", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 8192 - }, - "cost": { - "input": 0.56, - "output": 2.13 - } - }, - "amazon--nova-lite": { - "id": "amazon--nova-lite", - "name": "amazon--nova-lite", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0.3, - "output": 2.37 - } - }, - "sonar-pro": { - "id": "sonar-pro", - "name": "sonar-pro", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8192 - }, - "cost": { - "input": 3, - "output": 15 - } - }, - "anthropic--claude-3.5-sonnet": { - "id": "anthropic--claude-3.5-sonnet", - "name": "anthropic--claude-3.5-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04-30", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8192 - }, - "status": "deprecated", - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "text-embedding-3-large": { - "id": "text-embedding-3-large", - "name": "text-embedding-3-large", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-01-25", - "last_updated": "2024-01-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8191, - "output": 3072 - }, - "cost": { - "input": 0.09, - "output": 0 - } - }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - } - }, - "amazon--nova-micro": { - "id": "amazon--nova-micro", - "name": "amazon--nova-micro", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "nova-micro", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.03, - "output": 0.1 - } - }, - "anthropic--claude-4.5-opus": { - "id": "anthropic--claude-4.5-opus", - "name": "anthropic--claude-4.5-opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic--claude-4-sonnet": { - "id": "anthropic--claude-4-sonnet", - "name": "anthropic--claude-4-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "gemini-3.5-flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "gemini-2.5-pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-03-25", - "last_updated": "2025-06-05", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "anthropic--claude-3-sonnet": { - "id": "anthropic--claude-3-sonnet", - "name": "anthropic--claude-3-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-03-04", - "last_updated": "2024-03-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 4096 - }, - "status": "deprecated", - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "sonar-deep-research": { - "id": "sonar-deep-research", - "name": "sonar-deep-research", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar-deep-research", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-02-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 8, - "reasoning": 3 - } - }, - "mistralai--mistral-small": { - "id": "mistralai--mistral-small", - "name": "mistralai--mistral-small", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.07, - "output": 0.28 - } - }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "gemini-2.5-flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-04-17", - "last_updated": "2025-06-05", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 - } - }, - "anthropic--claude-3.7-sonnet": { - "id": "anthropic--claude-3.7-sonnet", - "name": "anthropic--claude-3.7-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-10-31", - "release_date": "2025-02-24", - "last_updated": "2025-02-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "status": "deprecated", - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "gpt-5.6-luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1, - "tiers": [ - { - "input": 2, - "output": 9, - "cache_read": 0.2, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 9, - "cache_read": 0.2 - } - } - }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "gpt-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 9.44, - "cache_read": 0.12 - } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "gpt-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - } - }, - "anthropic--claude-4.6-opus": { - "id": "anthropic--claude-4.6-opus", - "name": "anthropic--claude-4.6-opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "mistralai--mistral-medium-instruct": { - "id": "mistralai--mistral-medium-instruct", - "name": "mistralai--mistral-medium-instruct", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-medium", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-07", - "last_updated": "2025-05-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.36, - "output": 1.22 - } - }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "gpt-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.32 - } - }, - "anthropic--claude-3-opus": { - "id": "anthropic--claude-3-opus", - "name": "anthropic--claude-3-opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-08-31", - "release_date": "2024-02-29", - "last_updated": "2024-02-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 4096 - }, - "status": "deprecated", - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "anthropic--claude-4.6-sonnet": { - "id": "anthropic--claude-4.6-sonnet", - "name": "anthropic--claude-4.6-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "sonar": { - "id": "sonar", - "name": "sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-09-01", - "release_date": "2024-01-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 1, - "output": 1 - } - }, - "amazon--titan-embed-text": { - "id": "amazon--titan-embed-text", - "name": "amazon--titan-embed-text", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-04-30", - "last_updated": "2024-04-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 1536 - }, - "cost": { - "input": 0.14, - "output": 0 - } - }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "gpt-4.1-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 - } - }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "gemini-2.5-flash-lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "gpt-5.6-terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } - } - }, - "mistralai--mistral-medium": { - "id": "mistralai--mistral-medium", - "name": "Mistral Medium 3.5", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - } - }, - "gpt-5": { - "id": "gpt-5", - "name": "gpt-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "anthropic--claude-4.5-sonnet": { - "id": "anthropic--claude-4.5-sonnet", - "name": "anthropic--claude-4.5-sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "gpt-5.6-sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "cohere--command-a-reasoning": { - "id": "cohere--command-a-reasoning", - "name": "cohere--command-a-reasoning", - "description": "Cohere reasoning model for multilingual enterprise agents, tools, and complex workflows", - "family": "command-a", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-06-01", - "release_date": "2025-08-21", - "last_updated": "2025-08-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 32000 - }, - "cost": { - "input": 0.63, - "output": 5.05 - } - }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "gemini-3.1-flash-lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 - } - } - } - }, - "vivgrid": { - "id": "vivgrid", - "env": [ - "VIVGRID_API_KEY" - ], - "npm": "@ai-sdk/openai", - "api": "https://api.vivgrid.com/v1", - "name": "Vivgrid", - "doc": "https://docs.vivgrid.com/models", - "models": { - "viv-fast": { - "id": "viv-fast", - "name": "Viv Fast", - "description": "Fast coding model", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-09", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 256000 - }, - "cost": { - "input": 0.13, - "output": 0.4, - "cache_read": 0.05 - } - }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } - }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "gpt-5.2-codex": { - "id": "gpt-5.2-codex", - "name": "GPT-5.2 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-01-14", - "last_updated": "2026-01-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "GPT-5.1 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.04 - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.03 - } - }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.31, - "output": 1.23, - "cache_read": 0.01 - } - }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-24", - "last_updated": "2026-02-24", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "claude-fable-5-1": { - "id": "claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.5, - "cache_write": 12.5 - } - }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } - } - }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "claude-fable-5": { - "id": "claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1.25, - "cache_write": 12.5 - } - }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.35, - "output": 3, - "reasoning": 3, - "cache_read": 0.05 - } - }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT 5.6 Luna", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 1, - "output": 6, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "gpt-5.1-codex-max": { - "id": "gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/anthropic" - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075 - } - }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 1.2, - "output": 4.2, - "cache_read": 0.3 - } - }, - "gemini-3.8-flash": { - "id": "gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.15 - } - }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT 5.6 Terra", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 3.125 - } - }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek-V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 0.28, - "output": 0.42 - } - }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.2, - "output": 4.2, - "cache_read": 0.26 - } - }, - "gemini-3.1-flash-lite-preview": { - "id": "gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible" - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "cache_write": 1 - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT 5.6 Sol", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.15, - "output": 0.3, - "reasoning": 0.3, - "cache_read": 0.03 - } - } - } - }, - "klokintegration": { - "id": "klokintegration", - "env": [ - "KLOKINTEGRATION_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api-gw.klok.ipaas.se/proxy/kloker-key/v1", - "name": "klokintegration.se", - "doc": "https://klokintegration.se/docs/ai-api", - "models": { - "Kloker-Integration-Developer": { - "id": "Kloker-Integration-Developer", - "name": "Kloker Integration Developer", - "description": "Knows the customer integration environment and Klok best practices. Opinionated about implementation. The gateway runs ecosystem lookup tools server-side and appends a system-prompt injection. Client system prompts and OpenAI tool calls are preserved.", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-29", - "last_updated": "2026-08-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 50000 - }, - "status": "beta", - "cost": { - "input": 0.23, - "output": 1.16 - } - }, - "Kloker": { - "id": "Kloker", - "name": "Kloker", - "description": "Cheap general model with a clean context. Nothing from the customer environment is packed in. It tracks the current best open source model. The Klok team verifies it and upgrades it periodically.", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-29", - "last_updated": "2026-08-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 50000 - }, - "status": "beta", - "cost": { - "input": 0.23, - "output": 1.16 - } - }, - "Kloker-Integration-Architect": { - "id": "Kloker-Integration-Architect", - "name": "Kloker Integration Architect", - "description": "Knows the customer integration environment and Klok best practices. Opinionated about structure. The gateway runs ecosystem lookup tools server-side and appends a system-prompt injection (data contracts, CloudEvents, event-driven flows). Client system prompts and OpenAI tool calls are preserved.", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "release_date": "2026-08-29", - "last_updated": "2026-08-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 50000 - }, - "status": "beta", - "cost": { - "input": 0.23, - "output": 1.16 - } - } - } - }, - "google-vertex": { - "id": "google-vertex", - "env": [ - "GOOGLE_VERTEX_PROJECT", - "GOOGLE_VERTEX_LOCATION", - "GOOGLE_APPLICATION_CREDENTIALS" - ], - "npm": "@ai-sdk/google-vertex", - "name": "Vertex", - "doc": "https://cloud.google.com/vertex-ai/generative-ai/docs/models", - "models": { - "claude-opus-4-8@default": { - "id": "claude-opus-4-8@default", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "gemini-2.5-flash-tts": { - "id": "gemini-2.5-flash-tts", - "name": "Gemini 2.5 Flash TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gemini-flash", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-09-30", - "last_updated": "2025-12-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 0.5, - "output": 10 - } - }, - "gemini-flash-latest": { - "id": "gemini-flash-latest", - "name": "Gemini Flash Latest", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "claude-sonnet-4-5@20250929": { - "id": "claude-sonnet-4-5@20250929", - "name": "Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "gemini-2.5-flash-image": { - "id": "gemini-2.5-flash-image", - "name": "Nano Banana", - "description": "Nano Banana image model for fast generation, edits, and character-consistent assets", - "family": "gemini-flash", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-26", - "last_updated": "2025-08-26", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 30 - } - }, - "claude-opus-4-6@default": { - "id": "claude-opus-4-6@default", - "name": "Claude Opus 4.6", - "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "claude-opus-4@20250514": { - "id": "claude-opus-4@20250514", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "gemini-flash-lite-latest": { - "id": "gemini-flash-lite-latest", - "name": "Gemini Flash-Lite Latest", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 - } - }, - "claude-opus-5-5@default": { - "id": "claude-opus-5-5@default", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 - } - }, - "gemini-3.6-flash": { - "id": "gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "input_audio": 0.75 - } - }, - "claude-opus-5@default": { - "id": "claude-opus-5@default", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "claude-sonnet-4-6@default": { - "id": "claude-sonnet-4-6@default", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } - } - }, - "claude-haiku-4-5@20251001": { - "id": "claude-haiku-4-5@20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03 - } - }, - "gemini-3.1-flash-image": { - "id": "gemini-3.1-flash-image", - "name": "Nano Banana 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "high" - ] - } - ], - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.5, - "output": 60, - "cache_read": 0.05 - } - }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "claude-opus-4-1@20250805": { - "id": "claude-opus-4-1@20250805", - "name": "Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "claude-opus-4-5@20251101": { - "id": "claude-opus-4-5@20251101", - "name": "Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-01", - "last_updated": "2025-11-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "claude-opus-4-7@default": { - "id": "claude-opus-4-7@default", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } - } - }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "input_audio": 1.5 - } - }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } - } - }, - "gemini-3-pro-image": { - "id": "gemini-3-pro-image", - "name": "Nano Banana Pro", - "description": "Nano Banana Pro for higher-fidelity image generation and design-heavy edits", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 120, - "cache_read": 0.2 - } - }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "input_audio": 1 - } - }, - "claude-fable-5-1@default": { - "id": "claude-fable-5-1@default", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "input_audio": 0.75 - } - }, - "gemini-3.1-pro-preview-customtools": { - "id": "gemini-3.1-pro-preview-customtools", - "name": "Gemini 3.1 Pro Preview Custom Tools", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } - } - }, - "gemini-3-flash-preview": { - "id": "gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "input_audio": 1 - } - }, - "claude-fable-5@default": { - "id": "claude-fable-5@default", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "gemini-3.8-flash": { - "id": "gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "input_audio": 0.75 - } - }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "input_audio": 0.3 - } - }, - "gemini-2.5-pro-tts": { - "id": "gemini-2.5-pro-tts", - "name": "Gemini 2.5 Pro TTS", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "gemini-pro", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-09-30", - "last_updated": "2025-12-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 1, - "output": 20 - } - }, - "gemini-embedding-001": { - "id": "gemini-embedding-001", - "name": "Gemini Embedding 001", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "gemini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-05", - "release_date": "2025-05-20", - "last_updated": "2025-05-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2048, - "output": 1 - }, - "cost": { - "input": 0.15, - "output": 0 - } - }, - "claude-sonnet-4@20250514": { - "id": "claude-sonnet-4@20250514", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "gemini-3.1-flash-lite-preview": { - "id": "gemini-3.1-flash-lite-preview", - "name": "Gemini 3.1 Flash Lite Preview", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "status": "deprecated", - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 - } - }, - "claude-sonnet-5@default": { - "id": "claude-sonnet-5@default", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "provider": { - "npm": "@ai-sdk/google-vertex/anthropic" - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.25, - "output": 1.5, - "cache_read": 0.025, - "input_audio": 0.5 - } - }, - "meta/llama-4-maverick-17b-128e-instruct-maas": { - "id": "meta/llama-4-maverick-17b-128e-instruct-maas", - "name": "Llama 4 Maverick 17B 128E Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-29", - "last_updated": "2025-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 8192 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.35, - "output": 1.15 - } - }, - "meta/llama-3.3-70b-instruct-maas": { - "id": "meta/llama-3.3-70b-instruct-maas", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2025-04-29", - "last_updated": "2025-04-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 8192 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.72, - "output": 0.72 - } - }, - "xai/grok-4.20-reasoning": { - "id": "xai/grok-4.20-reasoning", - "name": "Grok 4.20 (Reasoning)", - "description": "Reasoning Grok for document-heavy analysis and long-horizon tool use", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 30000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "xai/grok-4.20-non-reasoning": { - "id": "xai/grok-4.20-non-reasoning", - "name": "Grok 4.20 (Non-Reasoning)", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-09", - "last_updated": "2026-03-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 30000 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "xai/grok-4.1-fast-reasoning": { - "id": "xai/grok-4.1-fast-reasoning", - "name": "Grok 4.1 Fast (Reasoning)", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 30000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } - }, - "xai/grok-4.3": { - "id": "xai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 30000 - }, - "status": "beta", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4 - } - } - }, - "xai/grok-4.1-fast-non-reasoning": { - "id": "xai/grok-4.1-fast-non-reasoning", - "name": "Grok 4.1 Fast", - "description": "Fast Grok model for responsive chat, tool-assisted work, and low-latency responses", - "family": "grok", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-11-19", - "last_updated": "2025-11-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 30000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.2, - "output": 0.5, - "cache_read": 0.05 - } - }, - "xai/grok-4.6": { - "id": "xai/grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 524288, - "output": 500000 - }, - "status": "beta", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1 - } - } - }, - "deepseek-ai/deepseek-v3.1-maas": { - "id": "deepseek-ai/deepseek-v3.1-maas", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-28", - "last_updated": "2025-08-28", - "modalities": { - "input": [ - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 32768 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.6, - "output": 1.7, - "cache_read": 0.06 - } - }, - "deepseek-ai/deepseek-v3.2-maas": { - "id": "deepseek-ai/deepseek-v3.2-maas", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-17", - "last_updated": "2026-04-04", - "modalities": { - "input": [ - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 65536 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.56, - "output": 1.68, - "cache_read": 0.056 - } - }, - "moonshotai/kimi-k2-thinking-maas": { - "id": "moonshotai/kimi-k2-thinking-maas", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.6, - "output": 2.5, - "cache_read": 0.06 - } - }, - "zai-org/glm-5-maas": { - "id": "zai-org/glm-5-maas", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.1 - } - }, - "zai-org/glm-4.7-maas": { - "id": "zai-org/glm-4.7-maas", - "name": "GLM-4.7", - "description": "GLM vision model for visual reasoning, documents, and multimodal agents", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-06", - "last_updated": "2026-01-06", - "modalities": { - "input": [ - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.6, - "output": 2.2, - "cache_read": 0.06 - } - }, - "zai-org/glm-5.2-maas": { - "id": "zai-org/glm-5.2-maas", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 64000 - }, - "status": "beta", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.14 - } - }, - "qwen/qwen3-235b-a22b-instruct-2507-maas": { - "id": "qwen/qwen3-235b-a22b-instruct-2507-maas", - "name": "Qwen3 235B A22B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-13", - "last_updated": "2025-08-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 16384 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.22, - "output": 0.88 - } - }, - "openai/gpt-oss-20b-maas": { - "id": "openai/gpt-oss-20b-maas", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "status": "deprecated", - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.07, - "output": 0.25, - "cache_read": 0.007 - } - }, - "openai/gpt-oss-120b-maas": { - "id": "openai/gpt-oss-120b-maas", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "provider": { - "npm": "@ai-sdk/openai-compatible", - "api": "https://${GOOGLE_VERTEX_ENDPOINT}/v1/projects/${GOOGLE_VERTEX_PROJECT}/locations/${GOOGLE_VERTEX_LOCATION}/endpoints/openapi" - }, - "cost": { - "input": 0.09, - "output": 0.36 - } - } - } - }, - "evroc": { - "id": "evroc", - "env": [ - "EVROC_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://models.think.evroc.com/v1", - "name": "evroc", - "doc": "https://docs.evroc.com/products/think/overview.html", - "models": { - "google/gemma-4-26B-A4B-it": { - "id": "google/gemma-4-26B-A4B-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.144, - "output": 0.575 - } - }, - "intfloat/multilingual-e5-large-instruct": { - "id": "intfloat/multilingual-e5-large-instruct", - "name": "E5 Multi-Lingual Large Embeddings 0.6B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2024-06-01", - "last_updated": "2024-06-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 512, - "output": 512 - }, - "cost": { - "input": 0.114, - "output": 0.114 - } - }, - "Qwen/Qwen3.8-27B": { - "id": "Qwen/Qwen3.8-27B", - "name": "Qwen3.8-27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.87, - "output": 3.5 - } - }, - "Qwen/Qwen3-Reranker-4B": { - "id": "Qwen/Qwen3-Reranker-4B", - "name": "Qwen3 Reranker 4B", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2025-07-30", - "last_updated": "2025-07-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32000, - "output": 4096 - }, - "cost": { - "input": 0.0575, - "output": 0 - } - }, - "Qwen/Qwen3.6-35B-A3B": { - "id": "Qwen/Qwen3.6-35B-A3B", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.345, - "output": 1.38 - } - }, - "Qwen/Qwen3-Embedding-8B": { - "id": "Qwen/Qwen3-Embedding-8B", - "name": "Qwen3 Embedding 8B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2025-07-30", - "last_updated": "2025-07-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 40960, - "output": 4096 - }, - "cost": { - "input": 0.115, - "output": 0.115 - } - }, - "mistralai/Mistral-Medium-3.5-128B": { - "id": "mistralai/Mistral-Medium-3.5-128B", - "name": "Mistral Medium 3.5", - "description": "Balanced Mistral model for enterprise assistants, multilingual work, and tools", - "family": "mistral-medium", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-29", - "last_updated": "2026-04-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.725, - "output": 6.9 - } - }, - "mistralai/Voxtral-Small-24B-2507": { - "id": "mistralai/Voxtral-Small-24B-2507", - "name": "Voxtral Small 24B", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "voxtral", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2025-03-01", - "last_updated": "2025-03-01", - "modalities": { - "input": [ - "audio", - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0.0023, - "output": 0.0023, - "output_audio": 2.3 - } - }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 1.4375, - "output": 5.75 - } - }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 131072 - }, - "cost": { - "input": 1.4375, - "output": 5.75 - } - }, - "nvidia/Llama-3.3-70B-Instruct-FP8": { - "id": "nvidia/Llama-3.3-70B-Instruct-FP8", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 1.15, - "output": 1.15 - } - }, - "KBLab/kb-whisper-large": { - "id": "KBLab/kb-whisper-large", - "name": "KB Whisper", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 448, - "output": 448 - }, - "cost": { - "input": 0.0023, - "output": 0.0023, - "output_audio": 2.3 - } - }, - "evroc/roc": { - "id": "evroc/roc", - "name": "roc", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-06-06", - "last_updated": "2026-06-06", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 2.875, - "output": 11.516 - } - }, - "openai/whisper-large-v3": { - "id": "openai/whisper-large-v3", - "name": "Whisper 3 Large", - "description": "Open Whisper checkpoint for robust multilingual transcription and captioning", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 448, - "output": 4096 - }, - "cost": { - "input": 0.0023, - "output": 0.0023, - "output_audio": 2.3 - } - }, - "openai/whisper-large-v3-turbo": { - "id": "openai/whisper-large-v3-turbo", - "name": "Whisper Large v3 Turbo", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", - "family": "whisper", - "attachment": false, - "reasoning": false, - "tool_call": false, - "release_date": "2024-10-01", - "last_updated": "2024-10-01", - "modalities": { - "input": [ - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 448, - "output": 448 - }, - "cost": { - "input": 0.0023, - "output": 0.0023, - "output_audio": 2.3 - } - }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 65536, - "output": 65536 - }, - "cost": { - "input": 0.23, - "output": 0.92 - } - } - } - }, - "tokengo": { - "id": "tokengo", - "env": [ - "TOKENGO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.tokengo.com/v1", - "name": "TokenGo", - "doc": "https://www.tokengo.com/docs", - "models": { - "deepseek/deepseek-v3.1": { - "id": "deepseek/deepseek-v3.1", - "name": "DeepSeek-V3.1", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-21", - "last_updated": "2025-08-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.19, - "output": 0.71, - "cache_read": 0.06 - } - }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.435, - "output": 0.87 - } - }, - "deepseek/deepseek-v3.2": { - "id": "deepseek/deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 64000 - }, - "cost": { - "input": 0.2174, - "output": 0.326, - "cache_read": 0.06 - } - }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.098, - "output": 0.196, - "cache_read": 0.028 - } - }, - "z-ai/glm-5.3-flash": { - "id": "z-ai/glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.075, - "output": 0.025, - "cache_read": 0.015 - } - }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.89, - "output": 3.2647, - "cache_read": 0.2226 - } - }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "z-ai/glm-5.3": { - "id": "z-ai/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "moonshotai/kimi-k2.6": { - "id": "moonshotai/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.16 - } - }, - "minimax/minimax-m2.5": { - "id": "minimax/minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.03 - } - }, - "qwen/qwen3.5-397b-a17b": { - "id": "qwen/qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.4, - "output": 2.65, - "cache_read": 0.2 - } - } - } - }, - "submodel": { - "id": "submodel", - "env": [ - "SUBMODEL_INSTAGEN_ACCESS_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://llm.submodel.ai/v1", - "name": "submodel", - "doc": "https://submodel.gitbook.io", - "models": { - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen3 235B A22B Thinking 2507", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.2, - "output": 0.6 - } - }, - "Qwen/Qwen3-235B-A22B-Instruct-2507": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507", - "name": "Qwen3 235B A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.2, - "output": 0.3 - } - }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.2, - "output": 0.8 - } - }, - "deepseek-ai/DeepSeek-V3.1": { - "id": "deepseek-ai/DeepSeek-V3.1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 75000, - "output": 163840 - }, - "cost": { - "input": 0.2, - "output": 0.8 - } - }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 75000, - "output": 163840 - }, - "cost": { - "input": 0.5, - "output": 2.15 - } - }, - "deepseek-ai/DeepSeek-V3-0324": { - "id": "deepseek-ai/DeepSeek-V3-0324", - "name": "DeepSeek V3 0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 75000, - "output": 163840 - }, - "cost": { - "input": 0.2, - "output": 0.8 - } - }, - "zai-org/GLM-4.5-FP8": { - "id": "zai-org/GLM-4.5-FP8", - "name": "GLM 4.5 FP8", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.2, - "output": 0.8 - } - }, - "zai-org/GLM-4.5-Air": { - "id": "zai-org/GLM-4.5-Air", - "name": "GLM 4.5 Air", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-air", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.1, - "output": 0.5 - } - }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2025-08-23", - "last_updated": "2025-08-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.5 - } - } - } - }, - "kosmik": { - "id": "kosmik", - "env": [ - "KOSMIK_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.koscompute.com/v1", - "name": "Kosmik Compute", - "doc": "https://api.koscompute.com/docs/", - "models": { - "qwen/qwen3.8-27b": { - "id": "qwen/qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.35, - "output": 2.2, - "cache_read": 0.09 - } - } - } - }, - "tencent-token-plan": { - "id": "tencent-token-plan", - "env": [ - "TENCENT_TOKEN_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.lkeap.cloud.tencent.com/plan/v3", - "name": "Tencent Token Plan", - "doc": "https://cloud.tencent.com/document/product/1823/130060", - "models": { - "hy3": { - "id": "hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 192000, - "output": 128000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "hy4-preview": { - "id": "hy4-preview", - "name": "Hy4 preview", - "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1024000, - "output": 64000 - }, - "cost": { - "input": 0.834, - "output": 2.501, - "cache_read": 0.042 - } - } - } - }, - "togetherai": { - "id": "togetherai", - "env": [ - "TOGETHER_API_KEY" - ], - "npm": "@ai-sdk/togetherai", - "name": "Together AI", - "doc": "https://docs.together.ai/docs/serverless-models", - "models": { - "meta-llama/Llama-3.3-70B-Instruct-Turbo": { - "id": "meta-llama/Llama-3.3-70B-Instruct-Turbo", - "name": "Llama 3.3 70B", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2026-07-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 1.04, - "output": 1.04 - } - }, - "meta-llama/Meta-Llama-3-8B-Instruct-Lite": { - "id": "meta-llama/Meta-Llama-3-8B-Instruct-Lite", - "name": "Meta Llama 3 8B Instruct Lite", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2024-04-18", - "last_updated": "2024-04-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.14, - "output": 0.14 - } - }, - "pearl-ai/gemma-4-31b-it": { - "id": "pearl-ai/gemma-4-31b-it", - "name": "Pearl AI Gemma 4 31B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0.28, - "output": 0.86 - } - }, - "deepcogito/cogito-v2-1-671b": { - "id": "deepcogito/cogito-v2-1-671b", - "name": "Cogito v2.1 671B", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "cogito", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": false, - "temperature": true, - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 163840, - "output": 163840 - }, - "cost": { - "input": 1.25, - "output": 1.25 - } - }, - "thinkingmachines/Inkling": { - "id": "thinkingmachines/Inkling", - "name": "Inkling", - "description": "Multimodal MoE reasoning model (975B total, 41B active) for text, image, and audio", - "family": "ling", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "max", - "xhigh", - "high", - "medium", - "low", - "none" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 131072 - }, - "cost": { - "input": 1, - "output": 4.05, - "cache_read": 0.17 - } - }, - "essentialai/Rnj-1-Instruct": { - "id": "essentialai/Rnj-1-Instruct", - "name": "Rnj-1 Instruct", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "rnj", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2025-12-05", - "last_updated": "2025-12-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "status": "deprecated", - "cost": { - "input": 0.15, - "output": 0.15 - } - }, - "google/gemma-3n-E4B-it": { - "id": "google/gemma-3n-E4B-it", - "name": "Gemma 3N E4B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": true, - "temperature": true, - "release_date": "2025-05-20", - "last_updated": "2025-05-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.06, - "output": 0.12 - } - }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B Instruct", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.39, - "output": 0.97 - } - }, - "Qwen/Qwen3.7-Max": { - "id": "Qwen/Qwen3.7-Max", - "name": "Qwen3.7 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-07-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 500000 - }, - "cost": { - "input": 1.25, - "output": 3.75, - "cache_read": 0.125 - } - }, - "Qwen/Qwen2.5-7B-Instruct-Turbo": { - "id": "Qwen/Qwen2.5-7B-Instruct-Turbo", - "name": "Qwen 2.5 7B Instruct Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-09-19", - "last_updated": "2024-09-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.3, - "output": 0.3 - } - }, - "Qwen/Qwen3-Coder-Next-FP8": { - "id": "Qwen/Qwen3-Coder-Next-FP8", - "name": "Qwen3 Coder Next FP8", - "description": "Legacy model retained for compatibility with older integrations", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2026-02-03", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.5, - "output": 1.2 - } - }, - "Qwen/Qwen3-235B-A22B-Instruct-2507-tput": { - "id": "Qwen/Qwen3-235B-A22B-Instruct-2507-tput", - "name": "Qwen3 235B A22B Instruct 2507 FP8", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.2, - "output": 0.6 - } - }, - "Qwen/Qwen3.5-9B": { - "id": "Qwen/Qwen3.5-9B", - "name": "Qwen3.5 9B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-03", - "last_updated": "2026-03-03", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0.17, - "output": 0.25 - } - }, - "Qwen/Qwen3.5-397B-A17B": { - "id": "Qwen/Qwen3.5-397B-A17B", - "name": "Qwen3.5 397B A17B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-16", - "last_updated": "2026-06-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 130000 - }, - "status": "deprecated", - "cost": { - "input": 0.6, - "output": 3.6, - "cache_read": 0.35 - } - }, - "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8": { - "id": "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8", - "name": "Qwen3 Coder 480B A35B Instruct", - "description": "Legacy model retained for compatibility with older integrations", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 2, - "output": 2 - } - }, - "Qwen/Qwen3.6-Plus": { - "id": "Qwen/Qwen3.6-Plus", - "name": "Qwen3.6 Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 500000 - }, - "cost": { - "input": 0.5, - "output": 3 - } - }, - "LiquidAI/LFM2-24B-A2B": { - "id": "LiquidAI/LFM2-24B-A2B", - "name": "LFM2-24B-A2B", - "description": "Open-weight instruction model for adaptable chat and self-hosted production workloads", - "family": "liquid", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "release_date": "2026-02-25", - "last_updated": "2026-02-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32768, - "output": 32768 - }, - "cost": { - "input": 0.03, - "output": 0.12 - } - }, - "deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.03 - } - }, - "deepseek-ai/DeepSeek-R1": { - "id": "deepseek-ai/DeepSeek-R1", - "name": "DeepSeek-R1", - "description": "Legacy model retained for compatibility with older integrations", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-03-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163839, - "output": 163839 - }, - "status": "deprecated", - "cost": { - "input": 3, - "output": 7 - } - }, - "deepseek-ai/DeepSeek-V4-Pro-0813": { - "id": "deepseek-ai/DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.13 - } - }, - "deepseek-ai/DeepSeek-V4.1-Flash": { - "id": "deepseek-ai/DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 384000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.006 - } - }, - "deepseek-ai/DeepSeek-V3-1": { - "id": "deepseek-ai/DeepSeek-V3-1", - "name": "DeepSeek V3.1", - "description": "Legacy model retained for compatibility with older integrations", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-21", - "last_updated": "2025-08-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0.6, - "output": 1.7 - } - }, - "deepseek-ai/DeepSeek-V4-Pro": { - "id": "deepseek-ai/DeepSeek-V4-Pro", - "name": "DeepSeek V4 Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 512000, - "output": 384000 - }, - "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.2 - } - }, - "deepseek-ai/DeepSeek-V3": { - "id": "deepseek-ai/DeepSeek-V3", - "name": "DeepSeek-V3", - "description": "Legacy model retained for compatibility with older integrations", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-12-26", - "last_updated": "2025-05-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 1.25, - "output": 1.25 - } - }, - "MiniMaxAI/MiniMax-M3": { - "id": "MiniMaxAI/MiniMax-M3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 524288, - "output": 250000 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "MiniMaxAI/MiniMax-M2.5": { - "id": "MiniMaxAI/MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 204800, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "MiniMaxAI/MiniMax-M2.7": { - "id": "MiniMaxAI/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 - } - }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131000 - }, - "cost": { - "input": 1.2, - "output": 4.5, - "cache_read": 0.2 - } - }, - "moonshotai/Kimi-K3": { - "id": "moonshotai/Kimi-K3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 - } - }, - "moonshotai/Kimi-K2.7-Code": { - "id": "moonshotai/Kimi-K2.7-Code", - "name": "Kimi K2.7 Code", - "description": "Kimi coding model for software agents, refactors, and repository reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-14", - "last_updated": "2026-06-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 131072 - }, - "cost": { - "input": 0.95, - "output": 4, - "cache_read": 0.19 - } - }, - "moonshotai/Kimi-K2.5": { - "id": "moonshotai/Kimi-K2.5", - "name": "Kimi K2.5", - "description": "Legacy model retained for compatibility with older integrations", - "family": "kimi-k2", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2026-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "status": "deprecated", - "cost": { - "input": 0.5, - "output": 2.8 - } - }, - "zai-org/GLM-5.1": { - "id": "zai-org/GLM-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-11", - "release_date": "2026-04-07", - "last_updated": "2026-07-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "zai-org/GLM-5.2": { - "id": "zai-org/GLM-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-16", - "last_updated": "2026-06-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 512000, - "output": 164000 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "zai-org/GLM-5": { - "id": "zai-org/GLM-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 131072 - }, - "status": "deprecated", - "cost": { - "input": 1, - "output": 3.2 - } - }, - "zai-org/GLM-5.3-Flash": { - "id": "zai-org/GLM-5.3-Flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048575, - "output": 400000 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 - } - }, - "zai-org/GLM-5.3": { - "id": "zai-org/GLM-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 262144 - }, - "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 - } - }, - "nvidia/nemotron-3-ultra-550b-a55b": { - "id": "nvidia/nemotron-3-ultra-550b-a55b", - "name": "Nemotron 3 Ultra 550B A55B", - "description": "Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 512300, - "output": 512300 - }, - "cost": { - "input": 0.6, - "output": 3.6, - "cache_read": 0.2 - } - }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.05, - "output": 0.2 - } - }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - } - } - }, - "helicone": { - "id": "helicone", - "env": [ - "HELICONE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://ai-gateway.helicone.ai/v1", - "name": "Helicone", - "doc": "https://helicone.ai/models", - "models": { - "grok-4-1-fast-reasoning": { - "id": "grok-4-1-fast-reasoning", - "name": "xAI Grok 4.1 Fast Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-17", - "last_updated": "2025-11-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 2000000 - }, - "cost": { - "input": 0.19999999999999998, - "output": 0.5, - "cache_read": 0.049999999999999996 - } - }, - "grok-3": { - "id": "grok-3", - "name": "xAI Grok 3", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 - } - }, - "llama-4-scout": { - "id": "llama-4-scout", - "name": "Meta Llama 4 Scout 17B 16E", - "description": "Open multimodal Llama model for long-context analysis and efficient agents", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.08, - "output": 0.3 - } - }, - "llama-guard-4": { - "id": "llama-guard-4", - "name": "Meta Llama Guard 4 12B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 1024 - }, - "cost": { - "input": 0.21, - "output": 0.21 - } - }, - "mistral-large-2411": { - "id": "mistral-large-2411", - "name": "Mistral-Large", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-24", - "last_updated": "2024-07-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 6 - } - }, - "grok-4-1-fast-non-reasoning": { - "id": "grok-4-1-fast-non-reasoning", - "name": "xAI Grok 4.1 Fast Non-Reasoning", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-17", - "last_updated": "2025-11-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 30000 - }, - "cost": { - "input": 0.19999999999999998, - "output": 0.5, - "cache_read": 0.049999999999999996 - } - }, - "qwen3-vl-235b-a22b-instruct": { - "id": "qwen3-vl-235b-a22b-instruct", - "name": "Qwen3 VL 235B A22B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 16384 - }, - "cost": { - "input": 0.3, - "output": 1.5 - } - }, - "mistral-nemo": { - "id": "mistral-nemo", - "name": "Mistral Nemo", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16400 - }, - "cost": { - "input": 20, - "output": 40 - } - }, - "gpt-5.1-codex": { - "id": "gpt-5.1-codex", - "name": "OpenAI: GPT-5.1 Codex", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-codex", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 - } - }, - "gemini-3-pro-preview": { - "id": "gemini-3-pro-preview", - "name": "Google Gemini 3 Pro Preview", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-18", - "last_updated": "2025-11-18", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.19999999999999998 - } - }, - "gpt-4o": { - "id": "gpt-4o", - "name": "OpenAI GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-05", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "gpt-4.1-mini-2025-04-14": { - "id": "gpt-4.1-mini-2025-04-14", - "name": "OpenAI GPT-4.1 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.39999999999999997, - "output": 1.5999999999999999, - "cache_read": 0.09999999999999999 - } - }, - "glm-4.6": { - "id": "glm-4.6", - "name": "Zai GLM-4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 204800, - "output": 131072 - }, - "cost": { - "input": 0.44999999999999996, - "output": 1.5 - } - }, - "claude-3.5-haiku": { - "id": "claude-3.5-haiku", - "name": "Anthropic: Claude 3.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8192 - }, - "cost": { - "input": 0.7999999999999999, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 - } - }, - "o1-mini": { - "id": "o1-mini", - "name": "OpenAI: o1-mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 65536 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 - } - }, - "grok-4": { - "id": "grok-4", - "name": "xAI Grok 4", - "description": "Grok model for agentic tool use, reasoning, coding, and live assistance", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-09", - "last_updated": "2024-07-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.75 - } - }, - "gpt-5-codex": { - "id": "gpt-5-codex", - "name": "OpenAI: GPT-5 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 - } - }, - "llama-4-maverick": { - "id": "llama-4-maverick", - "name": "Meta Llama 4 Maverick 17B 128E", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.15, - "output": 0.6 - } - }, - "qwen3-coder": { - "id": "qwen3-coder", - "name": "Qwen3 Coder 480B A35B Instruct Turbo", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 16384 - }, - "cost": { - "input": 0.22, - "output": 0.95 - } - }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "OpenAI GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.024999999999999998 - } - }, - "o4-mini": { - "id": "o4-mini", - "name": "OpenAI o4 Mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 - } - }, - "o3-mini": { - "id": "o3-mini", - "name": "OpenAI o3 Mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2023-10", - "release_date": "2023-10-01", - "last_updated": "2023-10-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 - } - }, - "kimi-k2-0711": { - "id": "kimi-k2-0711", - "name": "Kimi K2 (07/11)", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.5700000000000001, - "output": 2.3 - } - }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "OpenAI GPT-4.1 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.09999999999999999, - "output": 0.39999999999999997, - "cache_read": 0.024999999999999998 - } - }, - "grok-4-fast-reasoning": { - "id": "grok-4-fast-reasoning", - "name": "xAI: Grok 4 Fast Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-01", - "last_updated": "2025-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 2000000 - }, - "cost": { - "input": 0.19999999999999998, - "output": 0.5, - "cache_read": 0.049999999999999996 - } - }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-28", - "last_updated": "2025-04-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 40960 - }, - "cost": { - "input": 0.29, - "output": 0.59 - } - }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "OpenAI GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 0.049999999999999996, - "output": 0.39999999999999997, - "cache_read": 0.005 - } - }, - "claude-3.5-sonnet-v2": { - "id": "claude-3.5-sonnet-v2", - "name": "Anthropic: Claude 3.5 Sonnet v2", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-22", - "last_updated": "2024-10-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8192 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.30000000000000004, - "cache_write": 3.75 - } - }, - "gpt-5.1-chat-latest": { - "id": "gpt-5.1-chat-latest", - "name": "OpenAI GPT-5.1 Chat", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-codex", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 - } - }, - "claude-4.5-sonnet": { - "id": "claude-4.5-sonnet", - "name": "Anthropic: Claude Sonnet 4.5", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.30000000000000004, - "cache_write": 3.75 - } - }, - "deepseek-tng-r1t2-chimera": { - "id": "deepseek-tng-r1t2-chimera", - "name": "DeepSeek TNG R1T2 Chimera", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-02", - "last_updated": "2025-07-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 130000, - "output": 163840 - }, - "cost": { - "input": 0.3, - "output": 1.2 - } - }, - "deepseek-v3": { - "id": "deepseek-v3", - "name": "DeepSeek V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-26", - "last_updated": "2024-12-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 8192 - }, - "cost": { - "input": 0.56, - "output": 1.68, - "cache_read": 0.07 - } - }, - "deepseek-r1-distill-llama-70b": { - "id": "deepseek-r1-distill-llama-70b", - "name": "DeepSeek R1 Distill Llama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.03, - "output": 0.13 - } - }, - "o1": { - "id": "o1", - "name": "OpenAI: o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 - } - }, - "mistral-small": { - "id": "mistral-small", - "name": "Mistral Small 3.2", - "description": "Efficient Mistral model for fast chat, extraction, and production assistants", - "family": "mistral-small", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-06-20", - "last_updated": "2025-06-20", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.075, - "output": 0.2 - } - }, - "chatgpt-4o-latest": { - "id": "chatgpt-4o-latest", - "name": "OpenAI ChatGPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2024-08-14", - "last_updated": "2024-08-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 5, - "output": 20, - "cache_read": 2.5 - } - }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3 Coder 30B A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-31", - "last_updated": "2025-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.09999999999999999, - "output": 0.3 - } - }, - "sonar-pro": { - "id": "sonar-pro", - "name": "Perplexity Sonar Pro", - "description": "Advanced Sonar search model for deeper research and cited synthesis", - "family": "sonar-pro", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-27", - "last_updated": "2025-01-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 4096 - }, - "cost": { - "input": 3, - "output": 15 - } - }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-06", - "last_updated": "2025-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 262144 - }, - "cost": { - "input": 0.48, - "output": 2 - } - }, - "grok-4-fast-non-reasoning": { - "id": "grok-4-fast-non-reasoning", - "name": "xAI Grok 4 Fast Non-Reasoning", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-19", - "last_updated": "2025-09-19", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 2000000, - "output": 2000000 - }, - "cost": { - "input": 0.19999999999999998, - "output": 0.5, - "cache_read": 0.049999999999999996 - } - }, - "claude-opus-4-1-20250805": { - "id": "claude-opus-4-1-20250805", - "name": "Anthropic: Claude Opus 4.1 (20250805)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "OpenAI: GPT-5 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 32768 - }, - "cost": { - "input": 15, - "output": 120 - } - }, - "gpt-5-chat-latest": { - "id": "gpt-5-chat-latest", - "name": "OpenAI GPT-5 Chat Latest", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2024-09", - "release_date": "2024-09-30", - "last_updated": "2024-09-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 - } - }, - "deepseek-v3.1-terminus": { - "id": "deepseek-v3.1-terminus", - "name": "DeepSeek V3.1 Terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-22", - "last_updated": "2025-09-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.27, - "output": 1, - "cache_read": 0.21600000000000003 - } - }, - "claude-opus-4-1": { - "id": "claude-opus-4-1", - "name": "Anthropic: Claude Opus 4.1", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "OpenAI GPT-5.1", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 - } - }, - "llama-3.3-70b-versatile": { - "id": "llama-3.3-70b-versatile", - "name": "Meta Llama 3.3 70B Versatile", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32678 - }, - "cost": { - "input": 0.59, - "output": 0.7899999999999999 - } - }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Meta Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16400 - }, - "cost": { - "input": 0.13, - "output": 0.39 - } - }, - "grok-3-mini": { - "id": "grok-3-mini", - "name": "xAI Grok 3 Mini", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.3, - "output": 0.5, - "cache_read": 0.075 - } - }, - "gemma-3-12b-it": { - "id": "gemma-3-12b-it", - "name": "Google Gemma 3 12B", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.049999999999999996, - "output": 0.09999999999999999 - } - }, - "qwen2.5-coder-7b-fast": { - "id": "qwen2.5-coder-7b-fast", - "name": "Qwen2.5 Coder 7B fast", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-09-15", - "last_updated": "2024-09-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 8192 - }, - "cost": { - "input": 0.03, - "output": 0.09 - } - }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "OpenAI GPT-4o-mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 - } - }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3 Next 80B A3B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262000, - "output": 16384 - }, - "cost": { - "input": 0.14, - "output": 1.4 - } - }, - "o3-pro": { - "id": "o3-pro", - "name": "OpenAI o3 Pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 20, - "output": 80 - } - }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "OpenAI GPT-OSS 20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.049999999999999996, - "output": 0.19999999999999998 - } - }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Google Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "budget_tokens", - "min": 128, - "max": 32768 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.3125, - "cache_write": 1.25 - } - }, - "sonar-deep-research": { - "id": "sonar-deep-research", - "name": "Perplexity Sonar Deep Research", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar-deep-research", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-27", - "last_updated": "2025-01-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 127000, - "output": 4096 - }, - "cost": { - "input": 2, - "output": 8 - } - }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Google Gemini 2.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 0, - "max": 24576 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.075, - "cache_write": 0.3 - } - }, - "claude-4.5-haiku": { - "id": "claude-4.5-haiku", - "name": "Anthropic: Claude 4.5 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-10", - "release_date": "2025-10-01", - "last_updated": "2025-10-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8192 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.09999999999999999, - "cache_write": 1.25 - } - }, - "gemma2-9b-it": { - "id": "gemma2-9b-it", - "name": "Google Gemma 2", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-06-25", - "last_updated": "2024-06-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "cost": { - "input": 0.01, - "output": 0.03 - } - }, - "claude-3.7-sonnet": { - "id": "claude-3.7-sonnet", - "name": "Anthropic: Claude 3.7 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-02", - "release_date": "2025-02-19", - "last_updated": "2025-02-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.30000000000000004, - "cache_write": 3.75 - } - }, - "hermes-2-pro-llama-3-8b": { - "id": "hermes-2-pro-llama-3-8b", - "name": "Hermes 2 Pro Llama 3 8B", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-05", - "release_date": "2024-05-27", - "last_updated": "2024-05-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.14, - "output": 0.14 - } - }, - "claude-sonnet-4": { - "id": "claude-sonnet-4", - "name": "Anthropic: Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-14", - "last_updated": "2025-05-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.30000000000000004, - "cache_write": 3.75 - } - }, - "qwen3-235b-a22b-thinking": { - "id": "qwen3-235b-a22b-thinking", - "name": "Qwen3 235B A22B Thinking", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-25", - "last_updated": "2025-07-25", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 81920 - }, - "cost": { - "input": 0.3, - "output": 2.9000000000000004 - } - }, - "grok-code-fast-1": { - "id": "grok-code-fast-1", - "name": "xAI Grok Code Fast 1", - "description": "Fast Grok model for responsive chat, reasoning, and tool-assisted work", - "family": "grok", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2024-08-25", - "last_updated": "2024-08-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 10000 - }, - "cost": { - "input": 0.19999999999999998, - "output": 1.5, - "cache_read": 0.02 - } - }, - "claude-sonnet-4-5-20250929": { - "id": "claude-sonnet-4-5-20250929", - "name": "Anthropic: Claude Sonnet 4.5 (20250929)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.30000000000000004, - "cache_write": 3.75 - } - }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "OpenAI GPT-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "qwen3-30b-a3b": { - "id": "qwen3-30b-a3b", - "name": "Qwen3 30B A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2025-06-01", - "last_updated": "2025-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 41000, - "output": 41000 - }, - "cost": { - "input": 0.08, - "output": 0.29 - } - }, - "kimi-k2-0905": { - "id": "kimi-k2-0905", - "name": "Kimi K2 (09/05)", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-05", - "last_updated": "2025-09-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 16384 - }, - "cost": { - "input": 0.5, - "output": 2, - "cache_read": 0.39999999999999997 - } - }, - "llama-3.1-8b-instruct": { - "id": "llama-3.1-8b-instruct", - "name": "Meta Llama 3.1 8B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16384, - "output": 16384 - }, - "cost": { - "input": 0.02, - "output": 0.049999999999999996 - } - }, - "llama-3.1-8b-instruct-turbo": { - "id": "llama-3.1-8b-instruct-turbo", - "name": "Meta Llama 3.1 8B Instruct Turbo", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.02, - "output": 0.03 - } - }, - "llama-prompt-guard-2-86m": { - "id": "llama-prompt-guard-2-86m", - "name": "Meta Llama Prompt Guard 2 86M", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 512, - "output": 2 - }, - "cost": { - "input": 0.01, - "output": 0.01 - } - }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Anthropic: Claude 4.5 Haiku (20251001)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-10", - "release_date": "2025-10-01", - "last_updated": "2025-10-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8192 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.09999999999999999, - "cache_write": 1.25 - } - }, - "sonar": { - "id": "sonar", - "name": "Perplexity Sonar", - "description": "Sonar search model for current answers, retrieval, and citation-backed chat", - "family": "sonar", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-27", - "last_updated": "2025-01-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 127000, - "output": 4096 - }, - "cost": { - "input": 1, - "output": 1 - } - }, - "deepseek-reasoner": { - "id": "deepseek-reasoner", - "name": "DeepSeek Reasoner", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 64000 - }, - "cost": { - "input": 0.56, - "output": 1.68, - "cache_read": 0.07 - } - }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "OpenAI GPT-4.1 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.39999999999999997, - "output": 1.5999999999999999, - "cache_read": 0.09999999999999999 - } - }, - "ernie-4.5-21b-a3b-thinking": { - "id": "ernie-4.5-21b-a3b-thinking", - "name": "Baidu Ernie 4.5 21B A3B Thinking", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "family": "ernie", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-03-16", - "last_updated": "2025-03-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 8000 - }, - "cost": { - "input": 0.07, - "output": 0.28 - } - }, - "claude-4.5-opus": { - "id": "claude-4.5-opus", - "name": "Anthropic: Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 63999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-11", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gemini-2.5-flash-lite": { - "id": "gemini-2.5-flash-lite", - "name": "Google Gemini 2.5 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 512, - "max": 24576 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-07", - "release_date": "2025-07-22", - "last_updated": "2025-07-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.09999999999999999, - "output": 0.39999999999999997, - "cache_read": 0.024999999999999998, - "cache_write": 0.09999999999999999 - } - }, - "claude-3-haiku-20240307": { - "id": "claude-3-haiku-20240307", - "name": "Anthropic: Claude 3 Haiku", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-03", - "release_date": "2024-03-07", - "last_updated": "2024-03-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 4096 - }, - "cost": { - "input": 0.25, - "output": 1.25, - "cache_read": 0.03, - "cache_write": 0.3 - } - }, - "sonar-reasoning-pro": { - "id": "sonar-reasoning-pro", - "name": "Perplexity Sonar Reasoning Pro", - "description": "Web-grounded reasoning model for multi-step research and cited answers", - "family": "sonar-reasoning", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-27", - "last_updated": "2025-01-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 127000, - "output": 4096 - }, - "cost": { - "input": 2, - "output": 8 - } - }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2025-09-22", - "last_updated": "2025-09-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 163840, - "output": 65536 - }, - "cost": { - "input": 0.27, - "output": 0.41 - } - }, - "llama-prompt-guard-2-22m": { - "id": "llama-prompt-guard-2-22m", - "name": "Meta Llama Prompt Guard 2 22M", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-10-01", - "last_updated": "2024-10-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 512, - "output": 2 - }, - "cost": { - "input": 0.01, - "output": 0.01 - } - }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "OpenAI GPT-OSS 120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 131072 - }, - "cost": { - "input": 0.04, - "output": 0.16 - } - }, - "gpt-5.1-codex-mini": { - "id": "gpt-5.1-codex-mini", - "name": "OpenAI: GPT-5.1 Codex Mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-codex", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.024999999999999998 - } - }, - "o3": { - "id": "o3", - "name": "OpenAI o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2024-06", - "release_date": "2024-06-01", - "last_updated": "2024-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "gpt-5": { - "id": "gpt-5", - "name": "OpenAI GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2025-01-01", - "last_updated": "2025-01-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.12500000000000003 - } - }, - "llama-3.1-8b-instant": { - "id": "llama-3.1-8b-instant", - "name": "Meta Llama 3.1 8B Instant", - "description": "Compact Llama instruction model for fast chat and local deployment", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131072, - "output": 32678 - }, - "cost": { - "input": 0.049999999999999996, - "output": 0.08 - } - }, - "sonar-reasoning": { - "id": "sonar-reasoning", - "name": "Perplexity Sonar Reasoning", - "description": "Web-grounded reasoning model for multi-step research and cited answers", - "family": "sonar-reasoning", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-01-27", - "last_updated": "2025-01-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 127000, - "output": 4096 - }, - "cost": { - "input": 1, - "output": 5 - } - }, - "claude-opus-4": { - "id": "claude-opus-4", - "name": "Anthropic: Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024, - "max": 31999 - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-05-14", - "last_updated": "2025-05-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 32000 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - } - } - }, - "cortecs": { - "id": "cortecs", - "env": [ - "CORTECS_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.cortecs.ai/v1", - "name": "Cortecs", - "doc": "https://api.cortecs.ai/v1/models", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "output": 128000 - }, - "cost": { - "input": 2.898, - "output": 15.453, - "cache_read": 0.242 - } - }, - "claude-haiku-4-5": { - "id": "claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 0.996, - "output": 4.982, - "cache_read": 0.099, - "cache_write": 1.186 - } - }, - "gpt-oss-safeguard-120b": { - "id": "gpt-oss-safeguard-120b", - "name": "GPT OSS Safeguard 120B", - "description": "Safety model for policy screening, moderation, and risk-aware routing workflows", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-29", - "last_updated": "2025-10-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.179, - "output": 0.697 - } - }, - "gemma-4-31b-it": { - "id": "gemma-4-31b-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.223, - "output": 0.39 - } - }, - "qwen3.8-27b": { - "id": "qwen3.8-27b", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.04 - } - }, - "gemma-3-27b-it": { - "id": "gemma-3-27b-it", - "name": "Gemma 3 27B IT", - "description": "Gemma 3 is a family of lightweight, multimodal models from Google, supporting text and image inputs, multilingual capabilities, and a 131K context window.", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": false, - "knowledge": "2024-08", - "release_date": "2025-03-12", - "last_updated": "2025-03-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131000, - "output": 110000 - }, - "cost": { - "input": 0.099, - "output": 0.299 - } - }, - "qwen3.8-2.4t-a95b": { - "id": "qwen3.8-2.4t-a95b", - "name": "Qwen3.8 2.4T A95B", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 2.5, - "output": 6, - "cache_read": 0.625 - } - }, - "mistral-large-2402": { - "id": "mistral-large-2402", - "name": "mistral-large-2402", - "description": "Mistral Large (24.02) is Mistral AI’s most advanced language model, built for complex multilingual reasoning, code generation, and deep text understanding.", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-05-26", - "last_updated": "2025-05-26", - "modalities": { - "input": [ - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 8192 - }, - "cost": { - "input": 4.284, - "output": 12.952 - } - }, - "glm-5v-turbo": { - "id": "glm-5v-turbo", - "name": "GLM-5V-Turbo", - "description": "Fast GLM vision model for screenshots, documents, and multimodal agent tasks", - "family": "glm", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-04-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 202752, - "output": 202752 - }, - "cost": { - "input": 1.186, - "output": 3.955, - "cache_read": 0.296, - "cache_write": 1.544 - } - }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16000 - }, - "cost": { - "input": 2.659, - "output": 10.635, - "cache_read": 1.33 - } - }, - "qwen3guard-gen-8b": { - "id": "qwen3guard-gen-8b", - "name": "qwen3guard-gen-8b", - "description": "Qwen3Guard-Gen-8B is a large-scale multilingual safety moderation model designed for high-accuracy prompt and response classification.", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "release_date": "2026-02-04", - "last_updated": "2026-02-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 0.1, - "output": 0.35, - "cache_read": 0.018 - } - }, - "ministral-8b-2512": { - "id": "ministral-8b-2512", - "name": "ministral-8b-2512", - "description": "Ministral 3 8B is a balanced, efficient multimodal model offering strong text and vision capabilities, optimized for edge and local deployment.", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.179, - "output": 0.179, - "cache_read": 0.017 - } - }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 0.055, - "output": 0.174, - "cache_read": 0.009 - } - }, - "mistral-7b-instruct-v0.2": { - "id": "mistral-7b-instruct-v0.2", - "name": "mistral-7b-instruct-v0.2", - "description": "Mistral 7B Instruct is a compact, 7B parameter model optimized for fast and efficient text and code generation with a 32K token context window.", - "attachment": true, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "release_date": "2025-05-26", - "last_updated": "2025-05-26", - "modalities": { - "input": [ - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 8192 - }, - "cost": { - "input": 0.159, - "output": 0.219 - } - }, - "qwen3-235b-a22b-instruct-2507": { - "id": "qwen3-235b-a22b-instruct-2507", - "name": "Qwen3 235B-A22B Instruct 2507", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-07-21", - "last_updated": "2025-07-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.069, - "output": 0.455, - "cache_read": 0.018 - } - }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 3, - "output": 14.999 - } - }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-09", - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.167, - "output": 0.891 - } - }, - "nova-lite-v1": { - "id": "nova-lite-v1", - "name": "nova-lite-v1", - "description": "Nova Lite is a fast, low-cost multimodal foundation model capable of reasoning over text, images, and video in 200+ languages.", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.069, - "output": 0.275 - } - }, - "apertus-70b": { - "id": "apertus-70b", - "name": "Apertus 70B", - "description": "Apertus 70B is an open, multilingual language model designed for research, long-context reasoning, and sovereignty-focused AI systems.", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-09", - "release_date": "2025-09-02", - "last_updated": "2025-09-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 65536, - "output": 16384 - }, - "cost": { - "input": 1.393, - "output": 2.228 - } - }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.279, - "output": 2.192, - "cache_read": 0.056 - } - }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 202752 - }, - "cost": { - "input": 0.988, - "output": 3.164, - "cache_read": 0.247 - } - }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 64000 - }, - "cost": { - "input": 0.5, - "output": 1.499, - "cache_read": 0.13 - } - }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.111, - "output": 0.434, - "cache_read": 0.056 - } - }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196000, - "output": 196000 - }, - "cost": { - "input": 0.296, - "output": 1.186, - "cache_read": 0.075 - } - }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 16384, - "output": 16384 - }, - "cost": { - "input": 0.179, - "output": 0.697 - } - }, - "mistral-medium-3.5": { - "id": "mistral-medium-3.5", - "name": "mistral-medium-3.5", - "description": "Mistral Medium 3.5 is a frontier multimodal 128B model combining reasoning, coding, and instruction-following with strong agentic performance and efficient deployment.", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-04-30", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 1.532, - "output": 7.843, - "cache_read": 0.154 - } - }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.06, - "output": 0.439, - "cache_read": 0.019 - } - }, - "claude-4-6-sonnet": { - "id": "claude-4-6-sonnet", - "name": "Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3.196, - "output": 15.94, - "cache_read": 0.32, - "cache_write": 3.999 - } - }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Kimi K2.6", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.516, - "output": 2.869, - "cache_read": 0.115 - } - }, - "claude-4-5-sonnet": { - "id": "claude-4-5-sonnet", - "name": "Claude Sonnet 4.5 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 2.989, - "output": 14.945, - "cache_read": 0.326, - "cache_write": 4.078 - } - }, - "devstral-2512": { - "id": "devstral-2512", - "name": "Devstral 2", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-12", - "release_date": "2025-12-09", - "last_updated": "2025-12-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.478, - "output": 2.392, - "cache_read": 0.045 - } - }, - "gemini-3.6-flash": { - "id": "gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.038 - } - }, - "mistral-7b-instruct-v0.3": { - "id": "mistral-7b-instruct-v0.3", - "name": "mistral-7b-instruct-v0.3", - "description": "Mistral-7B-Instruct-v0.3 model is a fine-tuned version of the Mistral 7B base model, optimized for instruction-following tasks. Released in 2023, it is intended for demonstration purposes and does not include built-in guardrails or moderation features.", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-05-26", - "last_updated": "2025-05-26", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 127000, - "output": 127000 - }, - "cost": { - "input": 0.111, - "output": 0.111 - } - }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Smaller Qwen coder for efficient local agents and repo-level fixes", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.067, - "output": 0.245, - "cache_read": 0.014 - } - }, - "mixtral-8x7B-instruct-v0.1": { - "id": "mixtral-8x7B-instruct-v0.1", - "name": "Mixtral 8x7B Instruct v0.1", - "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2023-12-11", - "last_updated": "2023-12-11", - "modalities": { - "input": [ - "text", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32000, - "output": 4096 - }, - "cost": { - "input": 0.488, - "output": 0.758 - } - }, - "gemma-4-26b-a4b-it": { - "id": "gemma-4-26b-a4b-it", - "name": "Gemma 4 26B A4B IT", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 81920 - }, - "cost": { - "input": 0.111, - "output": 0.557 - } - }, - "gemini-3.5-flash-lite": { - "id": "gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.33, - "output": 2.749, - "cache_read": 0.033 - } - }, - "voxtral-small-2507": { - "id": "voxtral-small-2507", - "name": "voxtral-small-2507", - "description": "Voxtral Small is a multimodal model with audio input, combining advanced speech capabilities with strong text performance for transcription, translation, and audio understanding.", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-02-02", - "last_updated": "2026-02-02", - "modalities": { - "input": [ - "text", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0.123, - "output": 0.368, - "cache_read": 0.012 - } - }, - "qwen3.5-122b-a10b": { - "id": "qwen3.5-122b-a10b", - "name": "Qwen3.5 122B-A10B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.495, - "output": 3.46, - "cache_read": 0.124 - } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.495, - "output": 2.768, - "cache_read": 0.124 - } - }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.375, - "output": 10.96, - "cache_read": 0.156 - } - }, - "claude-opus-5": { - "id": "claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 27.498, - "cache_read": 0.55, - "cache_write": 6.874 - } - }, - "llama-3.3-70b-instruct": { - "id": "llama-3.3-70b-instruct", - "name": "Llama-3.3-70B-Instruct", - "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131000, - "output": 131000 - }, - "cost": { - "input": 0.724, - "output": 0.724 - } - }, - "qwen3guard-gen-0.6b": { - "id": "qwen3guard-gen-0.6b", - "name": "qwen3guard-gen-0.6b", - "description": "Qwen3Guard-Gen-0.6B is a lightweight multilingual safety moderation model that classifies prompts and responses into safe, controversial, or unsafe categories.", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": false, - "release_date": "2026-02-04", - "last_updated": "2026-02-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16000 - }, - "cost": { - "input": 0.159, - "output": 0.638, - "cache_read": 0.081 - } - }, - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 32768 - }, - "cost": { - "input": 0.167, - "output": 0.557 - } - }, - "pixtral-12b-2409": { - "id": "pixtral-12b-2409", - "name": "pixtral-12b-2409", - "description": "Pixtral 2409 12B is a state-of-the-art multimodal model with 12B parameters and a 400M vision encoder, natively trained on interleaved text and image data. It excels in tasks spanning vision-language reasoning, instruction following, and pure text understanding, making it highly effective for real-world multimodal applications.", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2024-11-09", - "last_updated": "2024-11-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.223, - "output": 0.223 - } - }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131000, - "output": 131000 - }, - "cost": { - "input": 0.045, - "output": 0.167 - } - }, - "minimax-m2.7": { - "id": "minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 196608 - }, - "cost": { - "input": 0.668, - "output": 2.674 - } - }, - "gemini-3.5-flash": { - "id": "gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 1.649, - "output": 9.899, - "cache_read": 0.165, - "cache_write": 1 - } - }, - "gemini-2.5-pro": { - "id": "gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 1.495, - "output": 9.964, - "cache_read": 0.242, - "cache_write": 0.434 - } - }, - "hermes-4-405b": { - "id": "hermes-4-405b", - "name": "hermes-4-405b", - "description": "Hermes 4 405B is a frontier hybrid-mode reasoning model built on Llama 3.1, optimized for advanced logic, math, coding, and structured output generation.", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2024-08-13", - "last_updated": "2024-08-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.996, - "output": 2.989 - } - }, - "gemini-2.5-flash": { - "id": "gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.299, - "output": 2.491, - "cache_read": 0.029, - "cache_write": 0.097 - } - }, - "qwen3-vl-235b-a22b": { - "id": "qwen3-vl-235b-a22b", - "name": "qwen3-vl-235b-a22b", - "description": "Qwen3 VL 235B A22B is a 235B-parameter MoE vision-language flagship model (≈22B active) designed for frontier-level multimodal understanding across text, images, documents, and long videos.", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-01-13", - "last_updated": "2026-01-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.617, - "output": 3.119, - "cache_read": 0.052 - } - }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 64000 - }, - "cost": { - "input": 2, - "output": 3.999, - "cache_read": 0.5 - } - }, - "glm-4.7-flash": { - "id": "glm-4.7-flash", - "name": "GLM-4.7-Flash", - "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", - "family": "glm-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-19", - "last_updated": "2026-01-19", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 203000, - "output": 203000 - }, - "cost": { - "input": 0.08, - "output": 0.478 - } - }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.111, - "output": 0.167 - } - }, - "claude-opus4-6": { - "id": "claude-opus4-6", - "name": "Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.313, - "output": 26.561, - "cache_read": 0.531, - "cache_write": 6.645 - } - }, - "claude-sonnet-4": { - "id": "claude-sonnet-4", - "name": "Claude Sonnet 4 (latest)", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 65000 - }, - "cost": { - "input": 2.898, - "output": 14.493, - "cache_read": 0.29, - "cache_write": 3.624 - } - }, - "minimax-m3": { - "id": "minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 0.395, - "output": 1.977, - "cache_read": 0.099 - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 0.219, - "output": 1.32, - "cache_read": 0.022, - "cache_write": 0.275 - } - }, - "qwen3.8-flash-next": { - "id": "qwen3.8-flash-next", - "name": "Qwen3.8 Flash Next", - "description": "Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 64000 - }, - "cost": { - "input": 0.201, - "output": 0.5, - "cache_read": 0.05 - } - }, - "mistral-small-2603": { - "id": "mistral-small-2603", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 256000 - }, - "cost": { - "input": 0.156, - "output": 0.625, - "cache_read": 0.016 - } - }, - "nemotron-nano-v2-12b": { - "id": "nemotron-nano-v2-12b", - "name": "nemotron-nano-v2-12b", - "description": "NVIDIA Nemotron Nano v2 12B is a 12-billion-parameter multimodal reasoning model designed for advanced video understanding, document intelligence, and visual reasoning, built with a hybrid Transformer-Mamba architecture for high efficiency and low latency.", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-10-31", - "last_updated": "2025-10-31", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.24, - "output": 0.707 - } - }, - "mistral-small-2503": { - "id": "mistral-small-2503", - "name": "mistral-small-2503", - "description": "Combines advanced text and vision capabilities with 24 billion parameters, supporting multilingual tasks and long contexts up to 131k tokens, making it versatile for various applications without sacrificing performance.", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-03-20", - "last_updated": "2025-03-20", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.111, - "output": 0.334 - } - }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 2.192, - "output": 8.769, - "cache_read": 0.546 - } - }, - "claude-sonnet-5": { - "id": "claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2.2, - "output": 11, - "cache_read": 0.219, - "cache_write": 2.749 - } - }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-02-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.668, - "output": 4.01 - } - }, - "gemini-3.7-flash": { - "id": "gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.038 - } - }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1000000 - }, - "cost": { - "input": 1.1, - "output": 2.99, - "cache_read": 0.18 - } - }, - "pixtral-large-2502": { - "id": "pixtral-large-2502", - "name": "Pixtral Large (25.02)", - "description": "Pixtral Large (25.02) is a 124B open-weight multimodal model built on Mistral Large 2, offering advanced image understanding and strong performance across text and code tasks.", - "family": "pixtral", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-04-08", - "last_updated": "2025-04-08", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 1.993, - "output": 5.978 - } - }, - "claude-opus4-7": { - "id": "claude-opus4-7", - "name": "Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.437, - "output": 27.186, - "cache_read": 0.544, - "cache_write": 6.797 - } - }, - "minicpm-v-4.5": { - "id": "minicpm-v-4.5", - "name": "minicpm-v-4.5", - "description": "MiniCPM-V 4.5 is a compact, high-performance vision-language model excelling in video understanding, OCR, and multimodal reasoning with efficient deployment.", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0.651, - "output": 1.097 - } - }, - "llama-3.1-8b-instruct": { - "id": "llama-3.1-8b-instruct", - "name": "Llama-3.1-8B-Instruct", - "description": "Optimized for dialogue, this LLM by Meta outperforms other open-source chat models in benchmarks while prioritizing helpfulness and safety.", - "family": "llama", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.167, - "output": 0.167 - } - }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 202752 - }, - "cost": { - "input": 1.384, - "output": 4.348, - "cache_read": 0.346 - } - }, - "qwen3-30b-a3b-instruct-2507": { - "id": "qwen3-30b-a3b-instruct-2507", - "name": "qwen3-30b-a3b-instruct-2507", - "description": "Qwen3-30B-A3B-Instruct-2507 is an advanced Mixture-of-Experts model optimized for reasoning, coding, and multilingual instruction following.", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-07-28", - "last_updated": "2025-07-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.099, - "output": 0.299 - } - }, - "qwen2.5-vl-72b-instruct": { - "id": "qwen2.5-vl-72b-instruct", - "name": "qwen2.5-vl-72b-instruct", - "description": "Qwen2.5-VL is a powerful vision-language model with advanced capabilities in visual understanding, long video reasoning, and structured output generation.", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": false, - "structured_output": true, - "temperature": false, - "release_date": "2025-01-27", - "last_updated": "2025-01-27", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 1.014, - "output": 1.014 - } - }, - "mistral-large-2512": { - "id": "mistral-large-2512", - "name": "Mistral Large 3", - "description": "Mistral's largest general model for enterprise agents, coding, and multilingual reasoning", - "family": "mistral-large", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-12-02", - "last_updated": "2025-12-02", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.613, - "output": 1.838, - "cache_read": 0.061 - } - }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.434, - "output": 1.704, - "cache_read": 0.134 - } - }, - "gemini-3.8-flash": { - "id": "gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.825, - "output": 4.125, - "cache_read": 0.082, - "cache_write": 0.084 - } - }, - "ministral-14b-2512": { - "id": "ministral-14b-2512", - "name": "ministral-14b-2512", - "description": "Ministral 3 14B is a frontier-level 14B multimodal model optimized for local deployment, delivering state-of-the-art text and vision reasoning with a 256K context window and strong agentic capabilities.", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.24, - "output": 0.24, - "cache_read": 0.022 - } - }, - "glm-5-turbo": { - "id": "glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 202752, - "output": 202752 - }, - "cost": { - "input": 1.186, - "output": 3.955, - "cache_read": 0.296, - "cache_write": 1.544 - } - }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 1.73, - "output": 3.46, - "cache_read": 0.432 - } - }, - "qwen3.6-27b": { - "id": "qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262000, - "output": 262000 - }, - "cost": { - "input": 0.446, - "output": 3.008 - } - }, - "claude-opus4-8": { - "id": "claude-opus4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5.437, - "output": 27.186, - "cache_read": 0.544, - "cache_write": 6.797 - } - }, - "mistral-small-3.2-24b-instruct-2506": { - "id": "mistral-small-3.2-24b-instruct-2506", - "name": "mistral-small-3.2-24b-instruct-2506", - "description": "Mistral-Small-3.2-24B-Instruct-2506 is a 24B parameter instruction-tuned model with enhanced long-context support (128k) and state-of-the-art vision understanding.", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-05-26", - "last_updated": "2025-05-26", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 131000, - "output": 131000 - }, - "cost": { - "input": 0.1, - "output": 0.312 - } - }, - "deepseek-r1-0528": { - "id": "deepseek-r1-0528", - "name": "DeepSeek R1 0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-05-28", - "last_updated": "2025-05-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 164000, - "output": 164000 - }, - "cost": { - "input": 0.652, - "output": 2.57, - "cache_read": 0.163 - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.2, - "output": 13.199, - "cache_read": 0.219, - "cache_write": 2.749 - } - }, - "deepseek-v3.2": { - "id": "deepseek-v3.2", - "name": "DeepSeek V3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-12-01", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 163840, - "output": 163840 - }, - "cost": { - "input": 0.296, - "output": 0.495, - "cache_read": 0.075 - } - }, - "nova-2-lite": { - "id": "nova-2-lite", - "name": "Nova 2 Lite", - "description": "Nova 2 Lite is an advanced multimodal reasoning model that combines efficiency and performance, delivering reliable AI for agentic workflows and enterprise applications.", - "family": "nova", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-10", - "release_date": "2025-12-02", - "last_updated": "2025-12-01", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65535 - }, - "cost": { - "input": 0.373, - "output": 3.144 - } - }, - "nova-pro-v1": { - "id": "nova-pro-v1", - "name": "Nova Pro 1.0", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "nova-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-12-03", - "last_updated": "2024-12-03", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 300000, - "output": 10000 - }, - "cost": { - "input": 0.918, - "output": 3.671 - } - }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131000, - "output": 131000 - }, - "cost": { - "input": 0.089, - "output": 0.446, - "cache_read": 0.01 - } - }, - "ministral-3b-2512": { - "id": "ministral-3b-2512", - "name": "ministral-3b-2512", - "description": "Ministral 3 3B is a compact, efficient multimodal model with strong language, vision capabilities, and ideal for custom fine-tuning.", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.123, - "output": 0.123, - "cache_read": 0.012 - } - }, - "claude-opus4-5": { - "id": "claude-opus4-5", - "name": "Claude Opus 4.5 (latest)", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 5.313, - "output": 26.568, - "cache_read": 0.531, - "cache_write": 6.645 - } - }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 1048576 - }, - "cost": { - "input": 1.114, - "output": 3.899, - "cache_read": 0.279 - } - }, - "kimi-k2.7-code": { - "id": "kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.706, - "output": 3.208, - "cache_read": 0.18 - } - }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.375, - "output": 10.96, - "cache_read": 0.156 - } - }, - "minimax-m2.1": { - "id": "minimax-m2.1", - "name": "MiniMax-M2.1", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-23", - "last_updated": "2025-12-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196000, - "output": 196000 - }, - "cost": { - "input": 0.359, - "output": 1.435 - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5.5, - "output": 32.998, - "cache_read": 0.55, - "cache_write": 6.879 - } - }, - "mistral-nemo-instruct-2407": { - "id": "mistral-nemo-instruct-2407", - "name": "mistral-nemo-instruct-2407", - "description": "A 12B parameter, instruct-tuned language model by Mistral AI and NVIDIA, designed for advanced instruction following, multi-turn conversations, and generating text and code across multiple languages.", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2024-08-07", - "last_updated": "2024-08-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 128000 - }, - "cost": { - "input": 0.145, - "output": 0.145, - "cache_read": 0.014 - } - }, - "minimax-m2": { - "id": "minimax-m2", - "name": "MiniMax-M2", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2025-10-27", - "last_updated": "2025-10-27", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 400000, - "output": 196000 - }, - "cost": { - "input": 0.349, - "output": 1.405 - } - }, - "nova-micro-v1": { - "id": "nova-micro-v1", - "name": "nova-micro-v1", - "description": "Nova Micro is a multilingual text-to-text foundation model with strong reasoning capabilities and broad language coverage across 200+ languages.", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 10000 - }, - "cost": { - "input": 0.04, - "output": 0.159 - } - }, - "codestral-2508": { - "id": "codestral-2508", - "name": "Codestral 2508", - "description": "Mistral coding model for code completion, generation, and developer workflows", - "family": "mistral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-03", - "release_date": "2025-07-30", - "last_updated": "2025-07-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.368, - "output": 1.103, - "cache_read": 0.037 - } - }, - "nvidia-nemotron-3-nano-30b-a3b": { - "id": "nvidia-nemotron-3-nano-30b-a3b", - "name": "nvidia-nemotron-3-nano-30b-a3b", - "description": "Nemotron-Nano-3-30B-A3B is a compact Mixture-of-Experts model optimized for efficient reasoning, chat, and coding, with strong multilingual support and long-context RAG and agent workflows.", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "release_date": "2026-01-12", - "last_updated": "2026-01-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 0.06, - "output": 0.24 - } - }, - "gemini-3.1-flash-lite": { - "id": "gemini-3.1-flash-lite", - "name": "Gemini 3.1 Flash Lite", - "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-07", - "last_updated": "2026-05-07", - "modalities": { - "input": [ - "text", - "image", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65535 - }, - "cost": { - "input": 0.272, - "output": 1.631, - "cache_read": 0.025, - "cache_write": 0.082 - } - } - } - }, - "agnes": { - "id": "agnes", - "env": [ - "AGNES_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://apihub.agnes-ai.com/v1", - "name": "Agnes AI", - "doc": "https://agnes-ai.com/doc", - "models": { - "agnes-2.5-pro-alpha": { - "id": "agnes-2.5-pro-alpha", - "name": "Agnes 2.5 Pro Alpha", - "description": "Paid reasoning model for advanced coding, scientific reasoning, long-context analysis, agentic workflows, and multimodal understanding.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.45, - "output": 0.9, - "cache_read": 0.0038 - } - }, - "agnes-2.0-flash": { - "id": "agnes-2.0-flash", - "name": "Agnes 2.0 Flash", - "description": "Fast and efficient model for agent workflows, tool calling, coding, and image understanding.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-25", - "last_updated": "2026-05-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 512000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "agnes-2.5-flash": { - "id": "agnes-2.5-flash", - "name": "Agnes 2.5 Flash", - "description": "Upgraded model with improved coding, agent workflows, tool calling, and multimodal understanding.", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "release_date": "2026-07", - "last_updated": "2026-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 512000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0 - } - } - } - }, - "daoxe": { - "id": "daoxe", - "env": [ - "DAOXE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://daoxe.com/v1", - "name": "DaoXE", - "doc": "https://daoxe.com/pricing", - "models": { - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25 - } - }, - "grok-4.3": { - "id": "grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 30000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2 - } - }, - "gemini-3.1-pro-preview": { - "id": "gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", - "modalities": { - "input": [ - "text", - "image", - "video", - "audio", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2 - } - }, - "grok-4.5": { - "id": "grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5 - } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-01", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 - }, - "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 - } - }, - "claude-opus-4-8": { - "id": "claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - } - }, - "claude-haiku-4-5-20251001": { - "id": "claude-haiku-4-5-20251001", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 5 - } - }, - "claude-sonnet-4-6": { - "id": "claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - } - } - }, - "morph": { - "id": "morph", - "env": [ - "MORPH_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.morphllm.com/v1", - "name": "Morph", - "doc": "https://docs.morphllm.com/api-reference/introduction", - "models": { - "morph-v3-large": { - "id": "morph-v3-large", - "name": "Morph v3 Large", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "morph", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0.9, - "output": 1.9 - } - }, - "morph-v3-fast": { - "id": "morph-v3-fast", - "name": "Morph v3 Fast", - "description": "Efficient model for low-latency assistance, extraction, and routine automation", - "family": "morph", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-08-15", - "last_updated": "2024-08-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16000, - "output": 16000 - }, - "cost": { - "input": 0.8, - "output": 1.2 - } - }, - "auto": { - "id": "auto", - "name": "Auto", - "description": "Automatic model router for matching prompts to suitable backends and budgets", - "family": "auto", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-06-01", - "last_updated": "2024-06-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32000, - "output": 32000 - }, - "cost": { - "input": 0.85, - "output": 1.55 - } - } - } - }, - "openai": { - "id": "openai", - "env": [ - "OPENAI_API_KEY" - ], - "npm": "@ai-sdk/openai", - "name": "OpenAI", - "doc": "https://platform.openai.com/docs/models", - "models": { - "chatgpt-image-latest": { - "id": "chatgpt-image-latest", - "name": "chatgpt-image-latest", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-12-16", - "last_updated": "2025-12-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "input": 0, - "output": 0 - } - }, - "gpt-5.4": { - "id": "gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } - } - }, - "gpt-5.4-pro": { - "id": "gpt-5.4-pro", - "name": "GPT-5.4 Pro", - "description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } - } - }, - "gpt-3.5-turbo": { - "id": "gpt-3.5-turbo", - "name": "GPT-3.5-turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": false, - "reasoning": false, - "tool_call": false, - "structured_output": false, - "temperature": true, - "knowledge": "2021-09-01", - "release_date": "2023-03-01", - "last_updated": "2023-11-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 16385, - "output": 4096 - }, - "status": "deprecated", - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0 - } - }, - "gpt-5.5-pro": { - "id": "gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } - } - }, - "text-embedding-3-small": { - "id": "text-embedding-3-small", - "name": "text-embedding-3-small", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2024-01", - "release_date": "2024-01-25", - "last_updated": "2024-01-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8191, - "output": 1536 - }, - "cost": { - "input": 0.02, - "output": 0 - } - }, - "gpt-5.4-nano": { - "id": "gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 - } - }, - "gpt-realtime-2.1": { - "id": "gpt-realtime-2.1", - "name": "GPT-Realtime-2.1", - "description": "Realtime speech-to-speech model with configurable reasoning, tool use, and robust voice-agent behavior", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text", - "audio", - "image" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "input": 96000, - "output": 32000 - }, - "cost": { - "input": 4, - "output": 24, - "cache_read": 0.4, - "input_audio": 32, - "output_audio": 64 - } - }, - "gpt-4o-2024-05-13": { - "id": "gpt-4o-2024-05-13", - "name": "GPT-4o (2024-05-13)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-05-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "status": "deprecated", - "cost": { - "input": 5, - "output": 15 - } - }, - "gpt-4o": { - "id": "gpt-4o", - "name": "GPT-4o", - "description": "Omni-era GPT for multimodal chat, practical coding, and general assistants", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "gpt-5-mini": { - "id": "gpt-5-mini", - "name": "GPT-5 Mini", - "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.25, - "output": 2, - "cache_read": 0.025 - } - }, - "gpt-image-2": { - "id": "gpt-image-2", - "name": "gpt-image-2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-21", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "input": 0, - "output": 0 - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 1.25 - } - }, - "gpt-5.2-pro": { - "id": "gpt-5.2-pro", - "name": "GPT-5.2 Pro", - "description": "Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": false, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 21, - "output": 168 - } - }, - "o4-mini": { - "id": "o4-mini", - "name": "o4-mini", - "description": "Fast o-series model for compact reasoning, coding, and tool use", - "family": "o-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "status": "deprecated", - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.275 - } - }, - "o3-mini": { - "id": "o3-mini", - "name": "o3-mini", - "description": "Smaller o-series reasoner for economical coding, math, and planning tasks", - "family": "o-mini", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "status": "deprecated", - "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 - } - }, - "text-embedding-ada-002": { - "id": "text-embedding-ada-002", - "name": "text-embedding-ada-002", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2022-12", - "release_date": "2022-12-15", - "last_updated": "2022-12-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 1536 - }, - "cost": { - "input": 0.1, - "output": 0 - } - }, - "gpt-4": { - "id": "gpt-4", - "name": "GPT-4", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-11", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8192, - "output": 8192 - }, - "status": "deprecated", - "cost": { - "input": 30, - "output": 60 - } - }, - "gpt-5.3-codex": { - "id": "gpt-5.3-codex", - "name": "GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "gpt-4.1-nano": { - "id": "gpt-4.1-nano", - "name": "GPT-4.1 nano", - "description": "Tiny GPT-4.1 option for classification, routing, and very high-volume tasks", - "family": "gpt-nano", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "status": "deprecated", - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.025 - } - }, - "gpt-5-nano": { - "id": "gpt-5-nano", - "name": "GPT-5 Nano", - "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 - } - }, - "gpt-5.6": { - "id": "gpt-5.6", - "name": "GPT-5.6", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 8, - "output": 40, - "cache_read": 0.8, - "cache_write": 10 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10 - } - } - }, - "gpt-5.2-chat-latest": { - "id": "gpt-5.2-chat-latest", - "name": "GPT-5.2 Chat", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt-codex", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "status": "deprecated", - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "o1": { - "id": "o1", - "name": "o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "status": "deprecated", - "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 - } - }, - "gpt-5-pro": { - "id": "gpt-5-pro", - "name": "GPT-5 Pro", - "description": "Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-10-06", - "last_updated": "2025-10-06", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 272000 - }, - "cost": { - "input": 15, - "output": 120 - } - }, - "gpt-5.3-codex-spark": { - "id": "gpt-5.3-codex-spark", - "name": "GPT-5.3 Codex Spark", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex-spark", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "input": 100000, - "output": 32000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "text-embedding-3-large": { - "id": "text-embedding-3-large", - "name": "text-embedding-3-large", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2024-01", - "release_date": "2024-01-25", - "last_updated": "2024-01-25", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 8191, - "output": 3072 - }, - "cost": { - "input": 0.13, - "output": 0 - } - }, - "gpt-4o-2024-08-06": { - "id": "gpt-4o-2024-08-06", - "name": "GPT-4o (2024-08-06)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-08-06", - "last_updated": "2024-08-06", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "gpt-6-astra": { - "id": "gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 20, - "output": 100, - "cache_read": 2, - "cache_write": 25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } - } - }, - "gpt-5.1": { - "id": "gpt-5.1", - "name": "GPT-5.1", - "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-09-30", - "release_date": "2025-11-13", - "last_updated": "2025-11-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "gpt-4o-mini": { - "id": "gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 - } - }, - "o3-pro": { - "id": "o3-pro", - "name": "o3-pro", - "description": "High-effort o3 tier for difficult technical reasoning and careful answers", - "family": "o-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-06-10", - "last_updated": "2025-06-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 20, - "output": 80 - } - }, - "gpt-image-1": { - "id": "gpt-image-1", - "name": "gpt-image-1", - "description": "OpenAI image model for production generation, edits, and brand-safe visual workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-04-24", - "last_updated": "2025-04-24", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "input": 0, - "output": 0 - }, - "status": "deprecated" - }, - "gpt-5.4-mini": { - "id": "gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 - } - }, - "gpt-5.6-luna": { - "id": "gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 0.4, - "output": 2.4, - "cache_read": 0.04, - "cache_write": 0.5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } - }, - "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } - } - }, - "gpt-5.2": { - "id": "gpt-5.2", - "name": "GPT-5.2", - "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "gpt-5.5": { - "id": "gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 12.5, - "output": 75, - "cache_read": 1.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - } - } - }, - "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } - } - }, - "gpt-4.1": { - "id": "gpt-4.1", - "name": "GPT-4.1", - "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "gpt-4o-2024-11-20": { - "id": "gpt-4o-2024-11-20", - "name": "GPT-4o (2024-11-20)", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-11-20", - "last_updated": "2024-11-20", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 - } - }, - "gpt-4.1-mini": { - "id": "gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1047576, - "output": 32768 - }, - "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 - } - }, - "gpt-6-luna": { - "id": "gpt-6-luna", - "name": "GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 0.2, - "output": 1, - "cache_read": 0.02, - "cache_write": 0.25 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } - }, - "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25 - } - } - }, - "gpt-5.3-chat-latest": { - "id": "gpt-5.3-chat-latest", - "name": "GPT-5.3 Chat (latest)", - "description": "Chat-tuned GPT model for conversational assistance, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-03", - "last_updated": "2026-03-03", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 16384 - }, - "status": "deprecated", - "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 - } - }, - "gpt-image-1-mini": { - "id": "gpt-image-1-mini", - "name": "gpt-image-1-mini", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-09-26", - "last_updated": "2025-09-26", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "input": 0, - "output": 0 - } - }, - "gpt-image-1.5": { - "id": "gpt-image-1.5", - "name": "gpt-image-1.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-11-25", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text", - "image" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "input": 0, - "output": 0 - } - }, - "gpt-5.6-terra": { - "id": "gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 4, - "output": 24, - "cache_read": 0.4, - "cache_write": 5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } - }, - "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } - } - }, - "gpt-4-turbo": { - "id": "gpt-4-turbo", - "name": "GPT-4 Turbo", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": false, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2023-11-06", - "last_updated": "2024-04-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "status": "deprecated", - "cost": { - "input": 10, - "output": 30 - } - }, - "o3": { - "id": "o3", - "name": "o3", - "description": "Deliberate o-series reasoner for hard math, coding, and multi-step analysis", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 - } - }, - "gpt-5": { - "id": "gpt-5", - "name": "GPT-5", - "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-09-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 400000, - "input": 272000, - "output": 128000 - }, - "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 - } - }, - "gpt-5.6-sol": { - "id": "gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 8, - "output": 40, - "cache_read": 0.8, - "cache_write": 10 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10 - } - } - }, - "gpt-6-sol": { - "id": "gpt-6-sol", - "name": "GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "pdf" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 - }, - "experimental": { - "modes": { - "fast": { - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5 - }, - "provider": { - "body": { - "service_tier": "priority" - } - } - }, - "pro": { - "provider": { - "body": { - "reasoning": { - "mode": "pro" - } - } - } - } - } - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } - } - }, - "o1-pro": { - "id": "o1-pro", - "name": "o1-pro", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2025-03-19", - "last_updated": "2025-03-19", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 100000 - }, - "status": "deprecated", - "cost": { - "input": 150, - "output": 600 - } - } - } - }, - "alibaba-coding-plan-cn": { - "id": "alibaba-coding-plan-cn", - "env": [ - "ALIBABA_CODING_PLAN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://coding.dashscope.aliyuncs.com/v1", - "name": "Alibaba Coding Plan (China)", - "doc": "https://help.aliyun.com/zh/model-studio/coding-plan", - "models": { - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 - } - }, - "qwen3-coder-next": { - "id": "qwen3-coder-next", - "name": "Qwen3 Coder Next", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-03", - "last_updated": "2026-02-03", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 202752, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "qwen3-max-2026-01-23": { - "id": "qwen3-max-2026-01-23", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-23", - "last_updated": "2026-01-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0.1875, - "output": 1.125, - "cache_write": 0.234375 - } - }, - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2025-07-23", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 196608, - "output": 24576 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "glm-4.7": { - "id": "glm-4.7", - "name": "GLM-4.7", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 202752, - "output": 16384 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 65536 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 64000 - }, - "cost": { - "input": 0, - "output": 0, - "cache_read": 0, - "cache_write": 0 - } - } - } - }, - "io-net": { - "id": "io-net", - "env": [ - "IOINTELLIGENCE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.intelligence.io.solutions/api/v1", - "name": "IO.NET", - "doc": "https://io.net/docs/guides/intelligence/io-intelligence", - "models": { - "meta-llama/Llama-3.2-90B-Vision-Instruct": { - "id": "meta-llama/Llama-3.2-90B-Vision-Instruct", - "name": "Llama 3.2 90B Vision Instruct", - "description": "Open Llama multimodal model for image understanding and text reasoning", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-09-25", - "last_updated": "2024-09-25", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 16000, - "output": 4096 - }, - "cost": { - "input": 0.35, - "output": 0.4, - "cache_read": 0.175, - "cache_write": 0.7 - } - }, - "meta-llama/Llama-3.3-70B-Instruct": { - "id": "meta-llama/Llama-3.3-70B-Instruct", - "name": "Llama 3.3 70B Instruct", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.13, - "output": 0.38, - "cache_read": 0.065, - "cache_write": 0.26 - } - }, - "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8": { - "id": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", - "name": "Llama 4 Maverick 17B 128E Instruct", - "description": "Open multimodal Llama model for strong reasoning and fast responses", - "family": "llama", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-15", - "last_updated": "2025-01-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 430000, - "output": 4096 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075, - "cache_write": 0.3 - } - }, - "Qwen/Qwen3-Next-80B-A3B-Instruct": { - "id": "Qwen/Qwen3-Next-80B-A3B-Instruct", - "name": "Qwen 3 Next 80B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-10", - "last_updated": "2025-01-10", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 4096 - }, - "cost": { - "input": 0.1, - "output": 0.8, - "cache_read": 0.05, - "cache_write": 0.2 - } - }, - "Qwen/Qwen3-235B-A22B-Thinking-2507": { - "id": "Qwen/Qwen3-235B-A22B-Thinking-2507", - "name": "Qwen 3 235B Thinking", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-07-01", - "last_updated": "2025-07-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 4096 - }, - "cost": { - "input": 0.11, - "output": 0.6, - "cache_read": 0.055, - "cache_write": 0.22 - } - }, - "Qwen/Qwen2.5-VL-32B-Instruct": { - "id": "Qwen/Qwen2.5-VL-32B-Instruct", - "name": "Qwen 2.5 VL 32B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-11-01", - "last_updated": "2024-11-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 32000, - "output": 4096 - }, - "cost": { - "input": 0.05, - "output": 0.22, - "cache_read": 0.025, - "cache_write": 0.1 - } - }, - "deepseek-ai/DeepSeek-R1-0528": { - "id": "deepseek-ai/DeepSeek-R1-0528", - "name": "DeepSeek R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2025-01-20", - "last_updated": "2025-05-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 2, - "output": 8.75, - "cache_read": 1, - "cache_write": 4 - } - }, - "mistralai/Mistral-Nemo-Instruct-2407": { - "id": "mistralai/Mistral-Nemo-Instruct-2407", - "name": "Mistral Nemo Instruct 2407", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral-nemo", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-05", - "release_date": "2024-07-01", - "last_updated": "2024-07-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.02, - "output": 0.04, - "cache_read": 0.01, - "cache_write": 0.04 - } - }, - "mistralai/Magistral-Small-2506": { - "id": "mistralai/Magistral-Small-2506", - "name": "Magistral Small 2506", - "description": "Mistral reasoning model for transparent analysis, math, and complex decisions", - "family": "magistral-small", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-01", - "last_updated": "2025-06-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.5, - "output": 1.5, - "cache_read": 0.25, - "cache_write": 1 - } - }, - "mistralai/Devstral-Small-2505": { - "id": "mistralai/Devstral-Small-2505", - "name": "Devstral Small 2505", - "description": "Mistral coding agent model for repository tasks and software engineering workflows", - "family": "devstral", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-05-01", - "last_updated": "2025-05-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 0.05, - "output": 0.22, - "cache_read": 0.025, - "cache_write": 0.1 - } - }, - "mistralai/Mistral-Large-Instruct-2411": { - "id": "mistralai/Mistral-Large-Instruct-2411", - "name": "Mistral Large Instruct 2411", - "description": "Flagship Mistral model for advanced reasoning, coding, and multilingual work", - "family": "mistral-large", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-11-01", - "last_updated": "2024-11-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 128000, - "output": 4096 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 1, - "cache_write": 4 - } - }, - "Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar": { - "id": "Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar", - "name": "Qwen 3 Coder 480B", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-15", - "last_updated": "2025-01-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 106000, - "output": 4096 - }, - "cost": { - "input": 0.22, - "output": 0.95, - "cache_read": 0.11, - "cache_write": 0.44 - } - }, - "moonshotai/Kimi-K2-Thinking": { - "id": "moonshotai/Kimi-K2-Thinking", - "name": "Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2024-11-01", - "last_updated": "2024-11-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 4096 - }, - "cost": { - "input": 0.55, - "output": 2.25, - "cache_read": 0.275, - "cache_write": 1.1 - } - }, - "moonshotai/Kimi-K2-Instruct-0905": { - "id": "moonshotai/Kimi-K2-Instruct-0905", - "name": "Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2024-09-05", - "last_updated": "2024-09-05", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 32768, - "output": 4096 - }, - "cost": { - "input": 0.39, - "output": 1.9, - "cache_read": 0.195, - "cache_write": 0.78 - } - }, - "zai-org/GLM-4.6": { - "id": "zai-org/GLM-4.6", - "name": "GLM 4.6", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-11-15", - "last_updated": "2024-11-15", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 4096 - }, - "cost": { - "input": 0.4, - "output": 1.75, - "cache_read": 0.2, - "cache_write": 0.8 - } - }, - "openai/gpt-oss-20b": { - "id": "openai/gpt-oss-20b", - "name": "GPT-OSS 20B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 64000, - "output": 4096 - }, - "cost": { - "input": 0.03, - "output": 0.14, - "cache_read": 0.015, - "cache_write": 0.06 - } - }, - "openai/gpt-oss-120b": { - "id": "openai/gpt-oss-120b", - "name": "GPT-OSS 120B", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-10", - "release_date": "2024-12-01", - "last_updated": "2024-12-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 4096 - }, - "cost": { - "input": 0.04, - "output": 0.4, - "cache_read": 0.02, - "cache_write": 0.08 - } - } - } - }, - "infomaniak": { - "id": "infomaniak", - "env": [ - "INFOMANIAK_API_KEY", - "INFOMANIAK_PRODUCT_ID" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.infomaniak.com/2/ai/${INFOMANIAK_PRODUCT_ID}/openai/v1", - "name": "Infomaniak", - "doc": "https://www.infomaniak.com/en/hosting/ai-services/open-source-models", - "models": { - "mini_lm_l12_v2": { - "id": "mini_lm_l12_v2", - "name": "All-MiniLM-L12-v2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2021-08-30", - "last_updated": "2026-08-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 128, - "input": 128, - "output": 384 - }, - "cost": { - "input": 0, - "output": 0 - } - }, - "bge_multilingual_gemma2": { - "id": "bge_multilingual_gemma2", - "name": "BGE Multilingual Gemma2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-07-25", - "last_updated": "2026-08-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 8000, - "input": 8000, - "output": 3584 - }, - "cost": { - "input": 0.08, - "output": 0 - } - }, - "swiss-ai/Apertus-v1.5-70B": { - "id": "swiss-ai/Apertus-v1.5-70B", - "name": "Apertus v1.5 70B", - "description": "Open, ethically-sourced Swiss AI model for multilingual, multimodal chat and instruction following", - "attachment": true, - "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2026-07-24", - "last_updated": "2026-08-01", - "modalities": { - "input": [ - "text", - "image", - "audio" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 100000, - "input": 100000, - "output": 8192 - }, - "status": "beta", - "cost": { - "input": 0.87, - "output": 3.1 - } - }, - "google/gemma-4-31B-it": { - "id": "google/gemma-4-31B-it", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-08-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 100000, - "input": 100000, - "output": 32768 - }, - "cost": { - "input": 0.25, - "output": 0.5 - } - }, - "Qwen/Qwen3.5-122B-A10B-FP8": { - "id": "Qwen/Qwen3.5-122B-A10B-FP8", - "name": "Qwen3.5 122B-A10B FP8", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-08-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "input": 200000, - "output": 65536 - }, - "cost": { - "input": 0.5, - "output": 3.97 - } - }, - "Qwen/Qwen3.5-397B-A17B-FP8": { - "id": "Qwen/Qwen3.5-397B-A17B-FP8", - "name": "Qwen3.5 397B-A17B FP8", - "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-08-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "input": 200000, - "output": 65536 - }, - "status": "beta", - "cost": { - "input": 0.99, - "output": 4.46 - } - }, - "mistralai/Mistral-Small-4-119B-2603": { - "id": "mistralai/Mistral-Small-4-119B-2603", - "name": "Mistral Small 4", - "description": "Fast Mistral production model for chat, extraction, and cost-sensitive agents", - "family": "mistral-small", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "interleaved": true, - "temperature": true, - "knowledge": "2025-06", - "release_date": "2026-03-16", - "last_updated": "2026-08-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 256000, - "output": 256000 - }, - "cost": { - "input": 0.25, - "output": 0.93 - } - }, - "mistralai/Ministral-3-14B-Instruct-2512": { - "id": "mistralai/Ministral-3-14B-Instruct-2512", - "name": "Ministral 3 14B Instruct", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": true, - "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-02", - "last_updated": "2026-08-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 100000, - "input": 100000, - "output": 25600 - }, - "status": "beta", - "cost": { - "input": 0.37, - "output": 0.5 - } - }, - "moonshotai/Kimi-K2.6": { - "id": "moonshotai/Kimi-K2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-08-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 256000, - "input": 256000, - "output": 256000 - }, - "status": "beta", - "cost": { - "input": 0.74, - "output": 3.72 - } - }, - "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8": { - "id": "nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8", - "name": "Nemotron 3 Nano 30B A3B FP8", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": true, - "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2026-08-01", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "input": 1000000, - "output": 262144 - }, - "status": "beta", - "cost": { - "input": 0.06, - "output": 0.25 - } - } - } - }, - "llmtech": { - "id": "llmtech", - "env": [ - "LLMTECH_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.llmtech.eu/v1", - "name": "LLM Tech", - "doc": "https://llmtech.eu/models/qwen3.8-27b", - "models": { - "nvidia/Qwen3.8-27B-NVFP4": { - "id": "nvidia/Qwen3.8-27B-NVFP4", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.25, - "output": 2.09, - "cache_read": 0.04 - } - } - } - }, - "crossmodel": { - "id": "crossmodel", - "env": [ - "CROSSMODEL_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.crossmodel.ai/v1", - "name": "CrossModel", - "doc": "https://www.crossmodel.ai/docs", - "models": { - "anthropic/claude-haiku-4-5": { - "id": "anthropic/claude-haiku-4-5", - "name": "Claude Haiku 4.5 (latest)", - "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 64000 - }, - "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 - } - }, - "anthropic/claude-opus-5-5": { - "id": "anthropic/claude-opus-5-5", - "name": "Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5 - } - }, - "anthropic/claude-fable-5-1": { - "id": "anthropic/claude-fable-5-1", - "name": "Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 - } - }, - "anthropic/claude-opus-5": { - "id": "anthropic/claude-opus-5", - "name": "Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-fable-5": { - "id": "anthropic/claude-fable-5", - "name": "Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-09", - "last_updated": "2026-06-09", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 - } - }, - "anthropic/claude-opus-4-8": { - "id": "anthropic/claude-opus-4-8", - "name": "Claude Opus 4.8", - "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01", - "release_date": "2026-05-28", - "last_updated": "2026-05-28", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "anthropic/claude-sonnet-5": { - "id": "anthropic/claude-sonnet-5", - "name": "Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 - } - }, - "anthropic/claude-sonnet-4-6": { - "id": "anthropic/claude-sonnet-4-6", - "name": "Claude Sonnet 4.6", - "description": "Claude workhorse for coding agents, careful analysis, and production cost control", - "family": "claude-sonnet", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - }, - { - "type": "budget_tokens", - "min": 1024 - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 - } - }, - "anthropic/claude-opus-4-7": { - "id": "anthropic/claude-opus-4-7", - "name": "Claude Opus 4.7", - "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 - } - }, - "deepseek/deepseek-v4.1-flash": { - "id": "deepseek/deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.27, - "output": 1.08, - "cache_read": 0.0054, - "cache_write": 0.27 - } - }, - "deepseek/deepseek-v4-flash-vision-exp": { - "id": "deepseek/deepseek-v4-flash-vision-exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.27, - "output": 1.08, - "cache_read": 0.0054, - "cache_write": 0.27 - } - }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 1.215, - "output": 3.645, - "cache_read": 0.0405, - "cache_write": 1.215 - } - }, - "deepseek/deepseek-v4-flash": { - "id": "deepseek/deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 384000 - }, - "cost": { - "input": 0.27, - "output": 1.08, - "cache_read": 0.0054, - "cache_write": 0.27 - } - }, - "tencent/hy3": { - "id": "tencent/hy3", - "name": "Hy3", - "description": "Tencent Hy reasoning model for coding, instruction following, and agent tasks", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-06", - "last_updated": "2026-07-06", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "input": 192000, - "output": 131072 - }, - "cost": { - "input": 0.16, - "output": 0.64, - "cache_read": 0.04, - "cache_write": 0.16 - } - }, - "tencent/hy4-preview": { - "id": "tencent/hy4-preview", - "name": "Hy4 preview", - "description": "A next-generation productivity model with significantly enhanced Agent and complex task execution capabilities.", - "family": "Hy", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-08-28", - "last_updated": "2026-08-28", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 65536 - }, - "cost": { - "input": 0.96, - "output": 2.88, - "cache_read": 0.048, - "cache_write": 0.96 - } - }, - "z-ai/glm-5.3-flash": { - "id": "z-ai/glm-5.3-flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", - "modalities": { - "input": [ - "text", - "image", - "video" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03, - "cache_write": 0.15 - } - }, - "z-ai/glm-5": { - "id": "z-ai/glm-5", - "name": "GLM-5", - "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.16, - "cache_write": 0.6, - "tiers": [ - { - "input": 0.8, - "output": 3.4, - "cache_read": 0.2, - "cache_write": 0.8, - "tier": { - "type": "context", - "size": 32000 - } - } - ] - } - }, - "z-ai/glm-4.7": { - "id": "z-ai/glm-4.7", - "name": "GLM-4.7", - "description": "Mature GLM model for dependable coding, reasoning, and structured agent tasks", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-12-22", - "last_updated": "2025-12-22", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.47, - "output": 2.16, - "cache_read": 0.1, - "cache_write": 0.47, - "tiers": [ - { - "input": 0.62, - "output": 2.47, - "cache_read": 0.13, - "cache_write": 0.62, - "tier": { - "type": "context", - "size": 32000 - } - } - ] - } - }, - "z-ai/glm-5.2": { - "id": "z-ai/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 1.2, - "output": 4.4, - "cache_read": 0.3, - "cache_write": 1.2 - } - }, - "z-ai/glm-5.1": { - "id": "z-ai/glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 1, - "output": 3.8, - "cache_read": 0.2, - "cache_write": 1, - "tiers": [ - { - "input": 1.2, - "output": 4.4, - "cache_read": 0.3, - "cache_write": 1.2, - "tier": { - "type": "context", - "size": 32000 - } - } - ] - } - }, - "z-ai/glm-5-turbo": { - "id": "z-ai/glm-5-turbo", - "name": "GLM-5-Turbo", - "description": "Faster GLM-5 lane for coding agents that need lower latency", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-16", - "last_updated": "2026-03-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 128000 - }, - "cost": { - "input": 0.9, - "output": 3.7, - "cache_read": 0.18, - "cache_write": 0.9, - "tiers": [ - { - "input": 1.1, - "output": 4.3, - "cache_read": 0.27, - "cache_write": 1.1, - "tier": { - "type": "context", - "size": 32000 - } - } - ] - } - }, - "z-ai/glm-5.3": { - "id": "z-ai/glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 1000000, - "output": 128000 - }, - "cost": { - "input": 1.2, - "output": 4.4, - "cache_read": 0.3, - "cache_write": 1.2 - } - }, - "x-ai/grok-4.7": { - "id": "x-ai/grok-4.7", - "name": "Grok 4.7", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-05", - "release_date": "2026-09-21", - "last_updated": "2026-09-21", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "cache_write": 2, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "cache_write": 4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1, - "cache_write": 4 - } - } - }, - "x-ai/grok-4.3": { - "id": "x-ai/grok-4.3", - "name": "Grok 4.3", - "description": "xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 1000000, - "output": 1000000 - }, - "cost": { - "input": 1.25, - "output": 2.5, - "cache_read": 0.2, - "cache_write": 1.25, - "tiers": [ - { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 5, - "cache_read": 0.4, - "cache_write": 2.5 - } - } - }, - "x-ai/grok-4.5": { - "id": "x-ai/grok-4.5", - "name": "Grok 4.5", - "description": "xAI's Grok model for chat, coding, agentic tools, and lower hallucination risk", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-08", - "last_updated": "2026-07-08", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.3, - "cache_write": 2, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 0.6, - "cache_write": 4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 0.6, - "cache_write": 4 - } - } - }, - "x-ai/grok-build-0.1": { - "id": "x-ai/grok-build-0.1", - "name": "Grok Build 0.1", - "description": "Fast Grok coding model tuned for agentic engineering and iterative edits", - "family": "grok-build", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-16", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 256000, - "output": 256000 - }, - "cost": { - "input": 1, - "output": 2, - "cache_read": 0.2, - "cache_write": 1, - "tiers": [ - { - "input": 2, - "output": 4, - "cache_read": 0.4, - "cache_write": 2, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 4, - "cache_read": 0.4, - "cache_write": 2 - } - } - }, - "x-ai/grok-4.6": { - "id": "x-ai/grok-4.6", - "name": "Grok 4.6", - "description": "xAI's frontier model for long-running agents, coding, knowledge work, and visual projects", - "family": "grok", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-02-01", - "release_date": "2026-08-12", - "last_updated": "2026-08-12", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 500000, - "output": 500000 - }, - "cost": { - "input": 2, - "output": 6, - "cache_read": 0.5, - "cache_write": 2, - "tiers": [ - { - "input": 4, - "output": 12, - "cache_read": 1, - "cache_write": 4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 12, - "cache_read": 1, - "cache_write": 4 - } - } - }, - "xiaomi/mimo-v2.6-pro": { - "id": "xiaomi/mimo-v2.6-pro", - "name": "MiMo-V2.6-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" + "release_date": "2025-09-11", + "last_updated": "2025-09-11", + "modalities": { + "input": [ + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 256000, + "input": 256000, + "output": 262144 }, "cost": { - "input": 0.47, - "output": 0.94, - "cache_read": 0.005, - "cache_write": 0.47 + "input": 0.3, + "output": 0.5 } }, - "xiaomi/mimo-v2.5": { - "id": "xiaomi/mimo-v2.5", - "name": "MiMo-V2.5", - "description": "Open MiMo model for multimodal coding agents and long-context automation", - "family": "mimo", + "qwen/qwen3.5-plus": { + "id": "qwen/qwen3.5-plus", + "name": "Qwen3.5 Plus", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 983616, + "input": 983616, + "output": 65536 }, "cost": { - "input": 0.16, - "output": 0.32, - "cache_read": 0.004, - "cache_write": 0.16 + "input": 0.4, + "output": 2.4, + "cache_read": 0.04 } }, - "xiaomi/mimo-v2.5-pro": { - "id": "xiaomi/mimo-v2.5-pro", - "name": "MiMo-V2.5-Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "qwen/qwen3.5-plus-thinking": { + "id": "qwen/qwen3.5-plus-thinking", + "name": "Qwen3.5 Plus Thinking", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 983616, + "input": 983616, + "output": 65536 }, "cost": { - "input": 0.47, - "output": 0.94, - "cache_read": 0.005, - "cache_write": 0.47 + "input": 0.4, + "output": 2.4, + "cache_read": 0.04 } }, - "xiaomi/mimo-v2.6-flash": { - "id": "xiaomi/mimo-v2.6-flash", - "name": "MiMo-V2.6-Flash", - "description": "MiMo Flash model for multimodal coding agents and long-context automation", - "family": "mimo", + "qwen/qwen3-235b-a22b": { + "id": "qwen/qwen3-235b-a22b", + "name": "Qwen 3 235b A22B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning": false, "tool_call": true, - "temperature": true, - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "structured_output": true, + "release_date": "2025-04-29", + "last_updated": "2025-04-29", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 41000, + "input": 41000, + "output": 32768 }, "cost": { - "input": 0.16, - "output": 0.32, - "cache_read": 0.004, - "cache_write": 0.16 + "input": 0.3, + "output": 0.5 } }, - "minimax/minimax-m2.7": { - "id": "minimax/minimax-m2.7", - "name": "MiniMax-M2.7", - "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", - "family": "minimax", + "qwen/qwen-2.5-72b-instruct": { + "id": "qwen/qwen-2.5-72b-instruct", + "name": "Qwen2.5 72B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-03", + "last_updated": "2025-07-03", "modalities": { "input": [ "text" @@ -316867,80 +188878,50 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 131072, + "input": 131072, + "output": 8192 }, "cost": { - "input": 0.33, - "output": 1.32, - "cache_read": 0.066, - "cache_write": 0.42 + "input": 0.357, + "output": 0.408 } }, - "minimax/minimax-m3": { - "id": "minimax/minimax-m3", - "name": "MiniMax-M3", - "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", - "family": "minimax", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "qwen/qwen3-coder-next": { + "id": "qwen/qwen3-coder-next", + "name": "Qwen3 Coder Next", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-06-01", - "last_updated": "2026-06-01", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1024000, - "output": 512000 + "context": 262144, + "input": 262144, + "output": 65536 }, "cost": { - "input": 0.33, - "output": 1.32, - "cache_read": 0.066, - "cache_write": 0.33, - "tiers": [ - { - "input": 0.66, - "output": 2.63, - "cache_read": 0.132, - "cache_write": 0.66, - "tier": { - "type": "context", - "size": 512000 - } - } - ], - "context_over_200k": { - "input": 0.66, - "output": 2.63, - "cache_read": 0.132, - "cache_write": 0.66 - } + "input": 0.15, + "output": 1.5 } }, - "gemini/gemini-3.6-flash": { - "id": "gemini/gemini-3.6-flash", - "name": "Gemini 3.6 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", + "qwen/qwen3.5-9b": { + "id": "qwen/qwen3.5-9b", + "name": "Qwen3.5 9B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -316948,27 +188929,19 @@ "type": "toggle" }, { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "budget_tokens", + "min": 1024, + "max": 81920 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-10", + "last_updated": "2026-03-10", "modalities": { "input": [ "text", - "image", - "audio", - "video" + "image" ], "output": [ "text" @@ -316976,95 +188949,60 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 256000, + "input": 256000, "output": 65536 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.75 + "input": 0.05, + "output": 0.15 } }, - "gemini/gemini-3.5-flash-lite": { - "id": "gemini/gemini-3.5-flash-lite", - "name": "Gemini 3.5 Flash Lite", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-07-21", - "last_updated": "2026-07-21", + "qwen/qwen3.5-397b-a17b": { + "id": "qwen/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-16", + "last_updated": "2026-02-16", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1048576, + "context": 258048, + "input": 258048, "output": 65536 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.3 + "input": 0.6, + "output": 3.6 } }, - "gemini/gemini-3.1-pro-preview": { - "id": "gemini/gemini-3.1-pro-preview", - "name": "Gemini 3.1 Pro Preview", - "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-02-19", - "last_updated": "2026-02-19", + "qwen/qwen3-coder-flash": { + "id": "qwen/qwen3-coder-flash", + "name": "Qwen3 Coder Flash", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-09-17", + "last_updated": "2025-09-17", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -317072,67 +189010,28 @@ }, "open_weights": false, "limit": { - "context": 1048576, + "context": 128000, + "input": 128000, "output": 65536 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 4, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 4 - } + "input": 0.3, + "output": 1.5 } }, - "gemini/gemini-3.5-flash": { - "id": "gemini/gemini-3.5-flash", - "name": "Gemini 3.5 Flash", - "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-05-19", - "last_updated": "2026-05-19", + "qwen/Qwen2.5-Coder-32B-Instruct": { + "id": "qwen/Qwen2.5-Coder-32B-Instruct", + "name": "Qwen 2.5 Coder 32b", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-03", + "last_updated": "2025-07-03", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -317140,47 +189039,28 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 32000, + "input": 32000, + "output": 8192 }, "cost": { - "input": 1.5, - "output": 9, - "cache_read": 0.15, - "cache_write": 1.5 + "input": 0.2006, + "output": 0.2006 } }, - "gemini/gemini-2.5-pro": { - "id": "gemini/gemini-2.5-pro", - "name": "Gemini 2.5 Pro", - "description": "Google's proven reasoning model for coding, math, and multimodal analysis", - "family": "gemini-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "qwen/Qwen3-8B": { + "id": "qwen/Qwen3-8B", + "name": "Qwen 3 8B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -317188,65 +189068,28 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 41000, + "input": 41000, + "output": 32768 }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125, - "cache_write": 1.25, - "tiers": [ - { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 2.5 - } + "input": 0.47, + "output": 0.47 } }, - "gemini/gemini-2.5-flash": { - "id": "gemini/gemini-2.5-flash", - "name": "Gemini 2.5 Flash", - "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "qwen/qwen3-14b": { + "id": "qwen/qwen3-14b", + "name": "Qwen 3 14b", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-01-01", + "last_updated": "2024-01-01", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -317254,45 +189097,28 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 41000, + "input": 41000, + "output": 32768 }, "cost": { - "input": 0.3, - "output": 2.5, - "cache_read": 0.03, - "cache_write": 0.3 + "input": 0.08, + "output": 0.24 } }, - "gemini/gemini-3.7-flash": { - "id": "gemini/gemini-3.7-flash", - "name": "Gemini 3.7 Flash", - "description": "High-efficiency Gemini model for agentic workflows, coding, and multimodal reasoning", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "qwen/Qwen3-235B-A22B-Instruct-2507": { + "id": "qwen/Qwen3-235B-A22B-Instruct-2507", + "name": "Qwen 3 235b A22B 2507", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2026-03", - "release_date": "2026-08-13", - "last_updated": "2026-08-13", + "release_date": "2025-07-25", + "last_updated": "2025-07-25", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -317300,49 +189126,28 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 256000, + "input": 256000, + "output": 262144 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.75 + "input": 0.13, + "output": 0.5 } }, - "gemini/gemini-3-flash-preview": { - "id": "gemini/gemini-3-flash-preview", - "name": "Gemini 3 Flash Preview", - "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-12-17", - "last_updated": "2025-12-17", + "qwen/qwen3-30b-a3b": { + "id": "qwen/qwen3-30b-a3b", + "name": "Qwen3 30B A3B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-27", + "last_updated": "2025-02-27", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" @@ -317350,44 +189155,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 41000, + "input": 41000, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.5 + "input": 0.1, + "output": 0.3 } }, - "gemini/gemini-3.8-flash": { - "id": "gemini/gemini-3.8-flash", - "name": "Gemini 3.8 Flash", - "description": "Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows", - "family": "gemini-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-09-02", - "last_updated": "2026-09-02", + "amazon/nova-lite-v1": { + "id": "amazon/nova-lite-v1", + "name": "Amazon Nova Lite 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-lite", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -317395,47 +189185,29 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 300000, + "input": 300000, + "output": 5120 }, "cost": { - "input": 0.75, - "output": 3.75, - "cache_read": 0.075, - "cache_write": 0.75 + "input": 0.0595, + "output": 0.238 } }, - "gemini/gemini-2.5-flash-lite": { - "id": "gemini/gemini-2.5-flash-lite", - "name": "Gemini 2.5 Flash-Lite", - "description": "Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents", - "family": "gemini-flash-lite", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2025-06-17", - "last_updated": "2025-06-17", + "amazon/nova-pro-v1": { + "id": "amazon/nova-pro-v1", + "name": "Amazon Nova Pro 1.0", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "nova-pro", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -317443,33 +189215,26 @@ }, "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 300000, + "input": 300000, + "output": 32000 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01, - "cache_write": 0.1 + "input": 0.7989999999999999, + "output": 3.1959999999999997 } }, - "qwen/qwen3.7-max": { - "id": "qwen/qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "amazon/nova-micro-v1": { + "id": "amazon/nova-micro-v1", + "name": "Amazon Nova Micro 1.0", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova-micro", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ "text" @@ -317480,49 +189245,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 128000, + "input": 128000, + "output": 5120 }, "cost": { - "input": 1.88, - "output": 5.63, - "cache_read": 0.375, - "cache_write": 2.35 + "input": 0.0357, + "output": 0.1394 } }, - "qwen/qwen3.8-max": { - "id": "qwen/qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "amazon/nova-2-lite-v1": { + "id": "amazon/nova-2-lite-v1", + "name": "Amazon Nova 2 Lite", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "nova", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-03", + "last_updated": "2024-12-03", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -317531,32 +189276,25 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 131072 + "input": 1000000, + "output": 65535 }, "cost": { - "input": 1.88, - "output": 5.63, - "cache_read": 0.23, - "cache_write": 2.35 + "input": 0.5099999999999999, + "output": 4.25 } }, - "qwen/qwen3.7-flash": { - "id": "qwen/qwen3.7-flash", - "name": "Qwen3.7 Flash", - "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "alibaba/qwen3.6-flash": { + "id": "alibaba/qwen3.6-flash", + "name": "Qwen3.6 Flash", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen3.6", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-17", + "last_updated": "2026-04-17", "modalities": { "input": [ "text", @@ -317569,70 +189307,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "input": 991000, + "context": 991800, "output": 65536 }, "cost": { - "input": 0.04, - "output": 0.13, - "cache_read": 0.01, - "cache_write": 0.04, - "tiers": [ - { - "input": 0.1, - "output": 0.37, - "cache_read": 0.02, - "cache_write": 0.12, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.19, - "output": 0.74, - "cache_read": 0.04, - "cache_write": 0.24, - "tier": { - "type": "context", - "size": 256000 - } - } - ] + "input": 0.19, + "output": 1.16 } }, - "qwen/qwen3.8-omni-flash": { - "id": "qwen/qwen3.8-omni-flash", - "name": "Qwen3.8 Omni Flash", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", + "alibaba/qwen3.6-27b": { + "id": "alibaba/qwen3.6-27b", + "name": "Qwen3.6 27B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, - { - "type": "budget_tokens" - } - ], - "tool_call": true, - "release_date": "2026-09-17", - "last_updated": "2026-09-17", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ @@ -317641,33 +189337,32 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 260096, + "input": 260096, + "output": 65536 }, "cost": { - "input": 0.13, - "output": 0.43, - "cache_read": 0.016, - "cache_write": 0.13 + "input": 0.203, + "output": 2.24 } }, - "qwen/qwen3.6-flash": { - "id": "qwen/qwen3.6-flash", - "name": "Qwen3.6 Flash", + "alibaba/qwen3.6-27b:thinking": { + "id": "alibaba/qwen3.6-27b:thinking", + "name": "Qwen3.6 27B Thinking", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "budget_tokens", + "min": 1024, + "max": 131072 } ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ "text", @@ -317680,63 +189375,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, + "context": 260096, + "input": 260096, "output": 65536 }, "cost": { - "input": 0.19, - "output": 1.13, - "cache_read": 0.019, - "cache_write": 0.24, - "tiers": [ - { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075, - "cache_write": 0.94, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075, - "cache_write": 0.94 - } + "input": 0.203, + "output": 2.24 } }, - "qwen/qwen3.8-flash": { - "id": "qwen/qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "aion-labs/aion-rp-llama-3.1-8b": { + "id": "aion-labs/aion-rp-llama-3.1-8b", + "name": "Llama 3.1 8b (uncensored)", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -317744,39 +189405,59 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 32768, + "input": 32768, + "output": 16384 }, "cost": { - "input": 0.13, - "output": 0.43, - "cache_read": 0.016, - "cache_write": 0.2 + "input": 0.2006, + "output": 0.2006 } }, - "qwen/qwen3.6-plus": { - "id": "qwen/qwen3.6-plus", - "name": "Qwen3.6 Plus", - "description": "Earlier Qwen multimodal workhorse for million-token agent and document tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "aion-labs/aion-2.5": { + "id": "aion-labs/aion-2.5", + "name": "AionLabs: Aion-2.5", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-20", + "last_updated": "2026-03-20", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 131072, + "input": 131072, + "output": 32768 + }, + "cost": { + "input": 1, + "output": 3, + "cache_read": 0.35 + } + }, + "aion-labs/aion-1.0-mini": { + "id": "aion-labs/aion-1.0-mini", + "name": "Aion 1.0 mini (DeepSeek)", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-20", + "last_updated": "2025-02-20", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -317784,57 +189465,28 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 131072, + "input": 131072, + "output": 8192 }, "cost": { - "input": 0.32, - "output": 1.88, - "cache_read": 0.032, - "cache_write": 0.4, - "tiers": [ - { - "input": 1.25, - "output": 7.5, - "cache_read": 0.124, - "cache_write": 1.57, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 1.25, - "output": 7.5, - "cache_read": 0.124, - "cache_write": 1.57 - } + "input": 0.7989999999999999, + "output": 1.394 } }, - "qwen/qwen3.7-plus": { - "id": "qwen/qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "aion-labs/aion-2.0": { + "id": "aion-labs/aion-2.0", + "name": "AionLabs: Aion-2.0", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -317842,63 +189494,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 64000 + "context": 131072, + "input": 131072, + "output": 32768 }, "cost": { - "input": 0.32, - "output": 1.25, - "cache_read": 0.032, - "cache_write": 0.4, - "tiers": [ - { - "input": 0.96, - "output": 3.75, - "cache_read": 0.096, - "cache_write": 1.2, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 0.96, - "output": 3.75, - "cache_read": 0.096, - "cache_write": 1.2 - } + "input": 0.8, + "output": 1.6 } }, - "openai/gpt-5.4": { - "id": "openai/gpt-5.4", - "name": "GPT-5.4", - "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "aion-labs/aion-1.0": { + "id": "aion-labs/aion-1.0", + "name": "Aion 1.0", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-01", + "last_updated": "2025-02-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -317906,62 +189524,30 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 65536, + "input": 65536, + "output": 8192 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "cache_write": 2.5, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "cache_write": 5 - } + "input": 3.995, + "output": 7.99 } }, - "openai/gpt-5.5-pro": { - "id": "openai/gpt-5.5-pro", - "name": "GPT-5.5 Pro", - "description": "Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding", - "family": "gpt-pro", - "attachment": true, + "pamanseau/OpenReasoning-Nemotron-32B": { + "id": "pamanseau/OpenReasoning-Nemotron-32B", + "name": "OpenReasoning Nemotron 32B", + "description": "Nemotron model for efficient reasoning, coding, and specialized AI agents", + "family": "nemotron", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -317969,58 +189555,29 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 32768, + "input": 32768, + "output": 65536 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 0.1, + "output": 0.4 } }, - "openai/gpt-5.4-nano": { - "id": "openai/gpt-5.4-nano", - "name": "GPT-5.4 nano", - "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", - "family": "gpt-nano", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "LatitudeGames/Wayfarer-Large-70B-Llama-3.3": { + "id": "LatitudeGames/Wayfarer-Large-70B-Llama-3.3", + "name": "Llama 3.3 70B Wayfarer", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-20", + "last_updated": "2025-02-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -318028,46 +189585,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02, - "cache_write": 0.2 + "input": 0.700000007, + "output": 0.700000007 } }, - "openai/gpt-6-astra": { - "id": "openai/gpt-6-astra", - "name": "GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "baseten/Kimi-K2-Instruct-FP4": { + "id": "baseten/Kimi-K2-Instruct-FP4", + "name": "Kimi K2 0711 Instruct FP4", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-11", + "last_updated": "2025-07-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -318075,52 +189615,29 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 131072 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2, - "cache_write": 25 - } + "input": 0.1, + "output": 2 } }, - "openai/gpt-4o-mini": { - "id": "openai/gpt-4o-mini", - "name": "GPT-4o mini", - "description": "Small omni GPT for cheap multimodal assistance and production-scale traffic", - "family": "gpt-mini", - "attachment": true, + "inflection/inflection-3-pi": { + "id": "inflection/inflection-3-pi", + "name": "Inflection 3 Pi", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "tool_call": false, + "structured_output": false, + "release_date": "2024-10-11", + "last_updated": "2024-10-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -318128,45 +189645,29 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 + "context": 8000, + "input": 8000, + "output": 4096 }, "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075, - "cache_write": 0.15 + "input": 2.499, + "output": 9.996 } }, - "openai/gpt-5.4-mini": { - "id": "openai/gpt-5.4-mini", - "name": "GPT-5.4 mini", - "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", - "family": "gpt-mini", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "inflection/inflection-3-productivity": { + "id": "inflection/inflection-3-productivity", + "name": "Inflection 3 Productivity", + "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "family": "gpt", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-10-11", + "last_updated": "2024-10-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -318174,47 +189675,29 @@ }, "open_weights": false, "limit": { - "context": 400000, - "input": 272000, - "output": 128000 + "context": 8000, + "input": 8000, + "output": 4096 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075, - "cache_write": 0.75 + "input": 2.499, + "output": 9.996 } }, - "openai/gpt-5.6-luna": { - "id": "openai/gpt-5.6-luna", - "name": "GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0": { + "id": "ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0", + "name": "Omega Directive 24B Unslop v2.0", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -318222,64 +189705,29 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 32768 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "cache_write": 0.25, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "cache_write": 0.5 - } + "input": 0.5, + "output": 0.5 } }, - "openai/gpt-5.5": { - "id": "openai/gpt-5.5", - "name": "GPT-5.5", - "description": "Default frontier GPT for coding, computer use, research, and knowledge work", - "family": "gpt", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-23", + "MiniMaxAI/MiniMax-M1-80k": { + "id": "MiniMaxAI/MiniMax-M1-80k", + "name": "MiniMax M1 80K", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-06-16", + "last_updated": "2025-06-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -318287,65 +189735,28 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "input": 1000000, + "output": 131072 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "cache_write": 5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1, - "cache_write": 10 - } + "input": 0.6052, + "output": 2.4225000000000003 } }, - "openai/gpt-6-luna": { - "id": "openai/gpt-6-luna", - "name": "GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "upstage/solar-pro-3": { + "id": "upstage/solar-pro-3", + "name": "Solar Pro 3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -318353,65 +189764,31 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, + "context": 128000, + "input": 128000, "output": 128000 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "cache_write": 0.125, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "cache_write": 0.25 - } + "input": 0.15, + "output": 0.6, + "cache_read": 0.015 } }, - "openai/gpt-5.6-terra": { - "id": "openai/gpt-5.6-terra", - "name": "GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, + "allenai/olmo-3-32b-think": { + "id": "allenai/olmo-3-32b-think", + "name": "Olmo 3 32B Think", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "family": "allenai", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-11-01", + "last_updated": "2025-11-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -318419,65 +189796,29 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 0.3, + "output": 0.44999999999999996 } }, - "openai/gpt-5.6-sol": { - "id": "openai/gpt-5.6-sol", - "name": "GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "essentialai/rnj-1-instruct": { + "id": "essentialai/rnj-1-instruct", + "name": "RNJ-1 Instruct 8B", + "description": "General-purpose chat model for instruction following, writing, and analysis", + "family": "rnj", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-13", + "last_updated": "2025-12-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -318485,65 +189826,40 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 128000, + "input": 128000, + "output": 8192 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.8, - "cache_write": 10 - } + "input": 0.15, + "output": 0.15 } }, - "openai/gpt-6-sol": { - "id": "openai/gpt-6-sol", - "name": "GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", - "attachment": true, + "deepseek/deepseek-v4-flash": { + "id": "deepseek/deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "none", - "low", - "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -318551,182 +189867,143 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "input": 1048576, + "output": 384000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4, - "cache_write": 5 - } + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "moonshot/kimi-k3": { - "id": "moonshot/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, + "deepseek/deepseek-v4-flash:thinking": { + "id": "deepseek/deepseek-v4-flash:thinking", + "name": "DeepSeek V4 Flash (Thinking)", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, - "output": 1048576 + "input": 1048576, + "output": 384000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3 + "input": 0.14, + "output": 0.28, + "cache_read": 0.028 } }, - "moonshot/kimi-k2.6": { - "id": "moonshot/kimi-k2.6", - "name": "Kimi K2.6", - "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", - "family": "kimi-k2", - "attachment": true, + "deepseek/deepseek-v4-pro-cheaper:thinking": { + "id": "deepseek/deepseek-v4-pro-cheaper:thinking", + "name": "DeepSeek V4 Pro Cheaper (Thinking)", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "high", + "xhigh" + ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "release_date": "2026-04-25", + "last_updated": "2026-04-25", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 1048576, + "input": 1048576, + "output": 384000 }, "cost": { - "input": 1, - "output": 4.16, - "cache_read": 0.18, - "cache_write": 1 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "moonshot/kimi-k2.7-code": { - "id": "moonshot/kimi-k2.7-code", - "name": "Kimi K2.7 Code", - "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", - "family": "kimi-k2", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-06-12", - "last_updated": "2026-06-12", + "deepseek/deepseek-prover-v2-671b": { + "id": "deepseek/deepseek-prover-v2-671b", + "name": "DeepSeek Prover v2 671B", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-04-30", + "last_updated": "2025-04-30", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262000, - "output": 262000 + "context": 160000, + "input": 160000, + "output": 16384 }, "cost": { "input": 1, - "output": 4.16, - "cache_read": 0.18, - "cache_write": 1 + "output": 2.5 } - } - } - }, - "arcee": { - "id": "arcee", - "env": [ - "ARCEE_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.arcee.ai/api/v1", - "name": "Arcee", - "doc": "https://docs.arcee.ai", - "models": { - "trinity-large-thinking": { - "id": "trinity-large-thinking", - "name": "Trinity Large Thinking", - "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", - "family": "trinity", + }, + "deepseek/deepseek-latest": { + "id": "deepseek/deepseek-latest", + "name": "DeepSeek Latest", + "description": "DeepSeek chat model for instruction following, coding, and analysis", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "high", + "xhigh" + ] + } + ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-05-28", + "structured_output": true, + "release_date": "2026-05-03", + "last_updated": "2026-05-03", "modalities": { "input": [ "text" @@ -318735,42 +190012,40 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "input": 1048576, + "output": 384000 }, - "status": "beta", "cost": { - "input": 0.25, - "output": 0.8, - "cache_read": 0.06 + "input": 1.1, + "output": 2.2, + "cache_read": 0.11 } }, - "deepseek/deepseek-v4-pro-0813": { - "id": "deepseek/deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", + "deepseek/deepseek-v4-pro": { + "id": "deepseek/deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" @@ -318779,23 +190054,22 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, + "input": 1048576, "output": 384000 }, - "status": "beta", "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 + "input": 1.1, + "output": 2.2, + "cache_read": 0.11 } }, - "deepseek/deepseek-v4-pro": { - "id": "deepseek/deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "deepseek/deepseek-v4-pro:thinking": { + "id": "deepseek/deepseek-v4-pro:thinking", + "name": "DeepSeek V4 Pro (Thinking)", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -318803,17 +190077,12 @@ "type": "effort", "values": [ "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", "release_date": "2026-04-24", "last_updated": "2026-04-24", "modalities": { @@ -318824,43 +190093,72 @@ "text" ] }, - "open_weights": true, + "open_weights": false, + "limit": { + "context": 1048576, + "input": 1048576, + "output": 384000 + }, + "cost": { + "input": 1.1, + "output": 2.2, + "cache_read": 0.11 + } + }, + "deepseek/deepseek-v3.2-speciale": { + "id": "deepseek/deepseek-v3.2-speciale", + "name": "DeepSeek V3.2 Speciale", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-02", + "last_updated": "2025-12-02", + "modalities": { + "input": [ + "text", + "pdf" + ], + "output": [ + "text" + ] + }, + "open_weights": false, "limit": { - "context": 512000, - "output": 384000 + "context": 163000, + "input": 163000, + "output": 65536 }, - "status": "beta", "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.2 + "input": 0.27999999999999997, + "output": 0.42000000000000004 } }, - "deepseek/deepseek-v4-flash-latest": { - "id": "deepseek/deepseek-v4-flash-latest", - "name": "DeepSeek V4 Flash Latest", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + "deepseek/deepseek-v4-pro-cheaper": { + "id": "deepseek/deepseek-v4-pro-cheaper", + "name": "DeepSeek V4 Pro Cheaper", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "high", - "max" + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "release_date": "2026-04-25", + "last_updated": "2026-04-25", "modalities": { "input": [ "text" @@ -318869,166 +190167,125 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 1048576, + "input": 1048576, "output": 384000 }, - "status": "beta", "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "thinkingmachines/inkling-small": { - "id": "thinkingmachines/inkling-small", - "name": "Inkling Small", - "description": "Multimodal MoE reasoning model (276B total, 12B active) for text, image, and audio", - "family": "ling", - "attachment": false, + "deepseek/deepseek-v3.2:thinking": { + "id": "deepseek/deepseek-v3.2:thinking", + "name": "DeepSeek V3.2 Thinking", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-07-30", - "last_updated": "2026-07-30", + "structured_output": true, + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 163000, + "input": 163000, + "output": 65536 }, - "status": "beta", "cost": { - "input": 0.5, - "output": 1.2, - "cache_read": 0.1 + "input": 0.27999999999999997, + "output": 0.42000000000000004 } }, - "moonshotai/kimi-k3": { - "id": "moonshotai/kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "deepseek/deepseek-v3.2": { + "id": "deepseek/deepseek-v3.2", + "name": "DeepSeek V3.2", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": true, + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 163000, + "input": 163000, + "output": 65536 }, - "status": "beta", "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.27999999999999997, + "output": 0.42000000000000004 } }, - "zai-org/glm-5.2": { - "id": "zai-org/glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", - "attachment": false, + "minimax/minimax-m3:thinking": { + "id": "minimax/minimax-m3:thinking", + "name": "MiniMax M3 Thinking", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 512000, + "input": 512000, + "output": 80000 }, - "status": "beta", "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } - } - } - }, - "drun": { - "id": "drun", - "env": [ - "DRUN_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://chat.d.run/v1", - "name": "D.Run (China)", - "doc": "https://www.d.run", - "models": { - "public/deepseek-v3": { - "id": "public/deepseek-v3", - "name": "DeepSeek V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + }, + "minimax/minimax-m2.5": { + "id": "minimax/minimax-m2.5", + "name": "MiniMax M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-12-26", - "last_updated": "2024-12-26", + "structured_output": true, + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -319037,31 +190294,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 204800, + "input": 204800, + "output": 131072 }, "cost": { - "input": 0.28, - "output": 1.1 + "input": 0.3, + "output": 1.2 } }, - "public/minimax-m25": { - "id": "public/minimax-m25", - "name": "MiniMax M2.5", + "minimax/minimax-m2.1": { + "id": "minimax/minimax-m2.1", + "name": "MiniMax M2.1", "description": "MiniMax model for chat, coding, office work, and agentic tasks", "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_details" - }, - "temperature": true, - "release_date": "2025-03-01", - "last_updated": "2025-03-01", + "structured_output": true, + "release_date": "2025-12-19", + "last_updated": "2025-12-19", "modalities": { "input": [ "text" @@ -319072,30 +190327,26 @@ }, "open_weights": false, "limit": { - "context": 204800, + "context": 200000, + "input": 200000, "output": 131072 }, "cost": { - "input": 0.29, - "output": 1.16 + "input": 0.33, + "output": 1.32 } }, - "public/deepseek-r1": { - "id": "public/deepseek-r1", - "name": "DeepSeek R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "minimax/minimax-m2.7-turbo": { + "id": "minimax/minimax-m2.7-turbo", + "name": "MiniMax M2.7 Turbo", + "description": "Efficient MiniMax model for quick assistance, coding, and routine automation", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-01-20", - "last_updated": "2025-01-20", + "structured_output": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -319104,95 +190355,58 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32000 + "context": 204800, + "input": 204800, + "output": 131072 }, "cost": { - "input": 0.55, - "output": 2.2 + "input": 0.6, + "output": 2.4 } - } - } - }, - "amd": { - "id": "amd", - "env": [ - "AMD_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://developer.amd.com.cn/radeon/api/v1", - "name": "AMD", - "doc": "https://developer.amd.com.cn/radeon/tokenfactory", - "models": { - "Qwen3.8-27B": { - "id": "Qwen3.8-27B", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium" - ] - } - ], - "tool_call": true, - "interleaved": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + }, + "minimax/minimax-01": { + "id": "minimax/minimax-01", + "name": "MiniMax 01", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-01-15", + "last_updated": "2025-01-15", "modalities": { "input": [ - "text" + "text", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 1000192, + "input": 1000192, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 0.1394, + "output": 1.1219999999999999 } }, - "DeepSeek-V4.1-Flash": { - "id": "DeepSeek-V4.1-Flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + "minimax/minimax-m3": { + "id": "minimax/minimax-m3", + "name": "MiniMax M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", @@ -319202,44 +190416,29 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 512000, + "input": 512000, + "output": 80000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "DeepSeek-V4-Flash": { - "id": "DeepSeek-V4-Flash", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + "minimax/minimax-m2-her": { + "id": "minimax/minimax-m2-her", + "name": "MiniMax M2-her", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-01-24", + "last_updated": "2026-01-24", "modalities": { "input": [ "text" @@ -319248,39 +190447,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 384000 + "context": 65532, + "input": 65532, + "output": 2048 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.30200000000000005, + "output": 1.2069999999999999 } }, - "Qwen3.8-Flash-Next": { - "id": "Qwen3.8-Flash-Next", - "name": "Qwen3.8 Flash Next", - "description": "Open-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understanding", - "family": "qwen", + "minimax/minimax-latest": { + "id": "minimax/minimax-latest", + "name": "MiniMax Latest", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-27", - "last_updated": "2026-08-27", + "release_date": "2026-05-03", + "last_updated": "2026-05-03", "modalities": { "input": [ "text", @@ -319290,28 +190482,30 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 131072 + "context": 512000, + "input": 512000, + "output": 80000 }, "cost": { - "input": 0.15, - "output": 0.47, - "cache_read": 0.016 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06 } }, - "MiniCPM5-2B": { - "id": "MiniCPM5-2B", - "name": "MiniCPM5-2B", - "description": "Dense 2B-class open-source model for on-device and resource-constrained use, with native long-context support, tool calling, and agentic tasks", + "minimax/minimax-m2.7": { + "id": "minimax/minimax-m2.7", + "name": "MiniMax M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-09-06", - "last_updated": "2026-09-12", + "structured_output": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -319320,85 +190514,57 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, + "context": 204800, + "input": 204800, "output": 131072 }, "cost": { - "input": 0.124, - "output": 0.7425, - "cache_read": 0.124 + "input": 0.3, + "output": 1.2 } }, - "DeepSeek-V4-Flash-Vision-Exp": { - "id": "DeepSeek-V4-Flash-Vision-Exp", - "name": "DeepSeek V4 Flash Vision Exp", - "description": "Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work", - "family": "deepseek-flash", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "minimal", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-21", - "last_updated": "2026-09-01", + "kwaipilot/kat-coder-pro-v2": { + "id": "kwaipilot/kat-coder-pro-v2", + "name": "KAT Coder Pro V2", + "description": "Coding model for repository understanding, refactors, and agentic engineering tasks", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-28", + "last_updated": "2026-03-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 256000, + "input": 256000, + "output": 80000 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0.3, + "output": 1.2 } - } - } - }, - "qvac": { - "id": "qvac", - "env": [ - "QVAC_API_KEY" - ], - "npm": "@qvac/ai-sdk-provider", - "name": "QVAC", - "doc": "https://www.npmjs.com/package/@qvac/ai-sdk-provider", - "models": { - "qwen3.5-0.8b": { - "id": "qwen3.5-0.8b", - "name": "Qwen3.5 0.8B", - "description": "Qwen instruction model for multilingual chat and tool use", - "family": "qwen", + }, + "mlabonne/NeuralDaredevil-8B-abliterated": { + "id": "mlabonne/NeuralDaredevil-8B-abliterated", + "name": "Neural Daredevil 8B abliterated", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ "text" @@ -319407,93 +190573,88 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, + "context": 8192, + "input": 8192, "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.44, + "output": 0.44 } }, - "gemma4-31b": { - "id": "gemma4-31b", - "name": "Gemma 4 31B IT", - "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", - "family": "gemma", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "VongolaChouko/Starcannon-Unleashed-12B-v1.0": { + "id": "VongolaChouko/Starcannon-Unleashed-12B-v1.0", + "name": "Mistral Nemo Starcannon 12b v1", + "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", + "family": "mistral-nemo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0, - "output": 0 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "qwen3.6-35b-a3b": { - "id": "qwen3.6-35b-a3b", - "name": "Qwen3.6 35B-A3B", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", - "family": "qwen", - "attachment": true, + "TEE/minimax-m2.5": { + "id": "TEE/minimax-m2.5", + "name": "MiniMax M2.5 TEE", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 196608, + "input": 196608, + "output": 131072 }, "cost": { - "input": 0, - "output": 0 + "input": 0.2, + "output": 1.38 } }, - "qwen3.5-2b": { - "id": "qwen3.5-2b", - "name": "Qwen3.5 2B", - "description": "Qwen instruction model for multilingual chat and tool use", - "family": "qwen", + "TEE/glm-4.7": { + "id": "TEE/glm-4.7", + "name": "GLM 4.7 TEE", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "tool_call": false, + "structured_output": false, + "release_date": "2026-01-29", + "last_updated": "2026-01-29", "modalities": { "input": [ "text" @@ -319502,29 +190663,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 131000, + "input": 131000, + "output": 65535 }, "cost": { - "input": 0, - "output": 0 + "input": 0.85, + "output": 3.3 } }, - "gpt-oss-20b": { - "id": "gpt-oss-20b", - "name": "GPT OSS 20B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "TEE/llama3-3-70b": { + "id": "TEE/llama3-3-70b", + "name": "Llama 3.3 70B", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-03", + "last_updated": "2025-07-03", "modalities": { "input": [ "text" @@ -319533,68 +190693,62 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0, - "output": 0 + "input": 2, + "output": 2 } }, - "qwen3.5-9b": { - "id": "qwen3.5-9b", - "name": "Qwen3.5 9B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": true, + "TEE/gemma-4-31b-it": { + "id": "TEE/gemma-4-31b-it", + "name": "Gemma 4 31B IT TEE", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-02-23", - "last_updated": "2026-02-23", + "structured_output": false, + "release_date": "2026-05-26", + "last_updated": "2026-05-26", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 262144, + "input": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.46 } }, - "qwen3.5-4b": { - "id": "qwen3.5-4b", - "name": "Qwen3.5 4B", - "description": "Qwen instruction model for multilingual chat and tool use", + "TEE/qwen3-30b-a3b-instruct-2507": { + "id": "TEE/qwen3-30b-a3b-instruct-2507", + "name": "Qwen3 30B A3B Instruct 2507 TEE", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "release_date": "2025-11-01", - "last_updated": "2025-11-01", + "tool_call": false, + "structured_output": false, + "release_date": "2025-07-29", + "last_updated": "2025-07-29", "modalities": { "input": [ "text" @@ -319603,63 +190757,62 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 8192 + "context": 262000, + "input": 262000, + "output": 32768 }, "cost": { - "input": 0, - "output": 0 + "input": 0.15, + "output": 0.44999999999999996 } }, - "qwen3.6-27b": { - "id": "qwen3.6-27b", - "name": "Qwen3.6 27B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "TEE/glm-5.1": { + "id": "TEE/glm-5.1", + "name": "GLM 5.1 TEE", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text", - "image", - "video", - "audio" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 202752, + "input": 202752, + "output": 65535 }, "cost": { - "input": 0, - "output": 0 + "input": 1.5, + "output": 5.25, + "cache_read": 0.3 } }, - "gpt-oss-120b": { - "id": "gpt-oss-120b", - "name": "GPT OSS 120B", - "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", - "family": "gpt-oss", + "TEE/deepseek-v4-pro": { + "id": "TEE/deepseek-v4-pro", + "name": "DeepSeek V4 Pro TEE", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2025-08-05", - "last_updated": "2025-08-05", + "release_date": "2026-04-25", + "last_updated": "2026-04-25", "modalities": { "input": [ "text" @@ -319668,45 +190821,32 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 800000, + "input": 800000, + "output": 65536 }, "cost": { - "input": 0, - "output": 0 + "input": 1.5, + "output": 5.25, + "cache_read": 0.15 } - } - } - }, - "claudinio": { - "id": "claudinio", - "env": [ - "CLAUDINIO_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.claudin.io/v1", - "name": "Claudinio", - "doc": "https://claudin.io", - "models": { - "claudinio": { - "id": "claudinio", - "name": "Claudinio", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + }, + "TEE/deepseek-v4-pro:thinking": { + "id": "TEE/deepseek-v4-pro:thinking", + "name": "DeepSeek V4 Pro Thinking TEE", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, - "knowledge": "2026-05", - "release_date": "2026-05-12", - "last_updated": "2026-06-02", + "structured_output": true, + "release_date": "2026-04-29", + "last_updated": "2026-04-29", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -319714,32 +190854,30 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 800000, + "input": 800000, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 2, + "input": 1.5, + "output": 5.25, "cache_read": 0.15 } }, - "claudius": { - "id": "claudius", - "name": "Claudius", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "attachment": true, + "TEE/gemma4-31b:thinking": { + "id": "TEE/gemma4-31b:thinking", + "name": "Gemma 4 31B Thinking TEE", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "attachment": false, "reasoning": true, "reasoning_options": [], - "tool_call": true, - "knowledge": "2026-05", - "release_date": "2026-05-12", - "last_updated": "2026-05-12", + "tool_call": false, + "structured_output": true, + "release_date": "2026-05-02", + "last_updated": "2026-05-02", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ "text" @@ -319747,49 +190885,26 @@ }, "open_weights": false, "limit": { - "context": 256000, - "output": 64000 + "context": 262144, + "input": 262144, + "output": 131072 }, "cost": { - "input": 3, - "output": 8, - "cache_read": 0.9 + "input": 0.45, + "output": 1 } - } - } - }, - "runinfra": { - "id": "runinfra", - "env": [ - "RUNINFRA_GATEWAY_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.runinfra.ai/v1", - "name": "RunInfra", - "doc": "https://runinfra.ai/docs", - "models": { - "Inferact/Qwen3.8-2.4T-A95B-NVFP4": { - "id": "Inferact/Qwen3.8-2.4T-A95B-NVFP4", - "name": "Qwen3.8 2.4T A95B (NVFP4)", - "description": "Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows", - "family": "qwen", + }, + "TEE/kimi-k2.5": { + "id": "TEE/kimi-k2.5", + "name": "Kimi K2.5 TEE", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-12", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-01-29", + "last_updated": "2026-01-29", "modalities": { "input": [ "text" @@ -319798,40 +190913,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 65535 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 0.3, + "output": 1.9 } }, - "Qwen/Qwen3.8-27B": { - "id": "Qwen/Qwen3.8-27B", - "name": "Qwen3.8 27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", + "TEE/qwen2.5-vl-72b-instruct": { + "id": "TEE/qwen2.5-vl-72b-instruct", + "name": "Qwen2.5 VL 72B TEE", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-02-01", + "last_updated": "2025-02-01", "modalities": { "input": [ "text", @@ -319841,40 +190944,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 65536, + "input": 65536, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 + "input": 0.7, + "output": 0.7 } }, - "deepseek-ai/DeepSeek-V4-Flash-0731": { - "id": "deepseek-ai/DeepSeek-V4-Flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + "TEE/gpt-oss-120b": { + "id": "TEE/gpt-oss-120b", + "name": "GPT-OSS 120B TEE", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "max" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -319883,76 +190974,58 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 32768 + "context": 131072, + "input": 131072, + "output": 16384 }, "cost": { - "input": 0.13, - "output": 0.27, - "cache_read": 0.01 + "input": 2, + "output": 2 } }, - "deepseek-ai/DeepSeek-V4-Pro-0813": { - "id": "deepseek-ai/DeepSeek-V4-Pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "TEE/qwen3.5-27b": { + "id": "TEE/qwen3.5-27b", + "name": "Qwen3.5 27B TEE", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "attachment": true, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-03-13", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 32768 + "context": 262144, + "input": 262144, + "output": 65536 }, "cost": { - "input": 0.6, - "output": 1.9, - "cache_read": 0.03 + "input": 0.3, + "output": 2.4 } }, - "zai-org/GLM-5.3-Flash": { - "id": "zai-org/GLM-5.3-Flash", - "name": "GLM-5.3-Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", + "TEE/gemma-4-26b-a4b-uncensored": { + "id": "TEE/gemma-4-26b-a4b-uncensored", + "name": "Gemma 4 26B A4B Uncensored TEE", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": true, "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "release_date": "2026-05-23", + "last_updated": "2026-05-23", "modalities": { "input": [ "text", @@ -319962,118 +191035,57 @@ "text" ] }, - "open_weights": true, - "limit": { - "context": 1048576, - "output": 32768 - }, - "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 - } - }, - "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16": { - "id": "nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16", - "name": "Nemotron 3.5 Lightning 30B A3B", - "description": "Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads", - "family": "nemotron", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-11", - "last_updated": "2026-08-11", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 - }, - "cost": { - "input": 0.05, - "output": 0.15, - "cache_read": 0.01 - } - }, - "ornith-ai/Ornith-1.5-35B-A3B": { - "id": "ornith-ai/Ornith-1.5-35B-A3B", - "name": "Ornith 1.5 35B A3B", - "description": "Mixture-of-experts coding-reasoning model for agentic software tasks, tool use, and image understanding", - "family": "ornith", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" + "context": 65536, + "input": 65536, + "output": 65536 }, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-18", - "last_updated": "2026-08-23", + "cost": { + "input": 0.15, + "output": 0.7 + } + }, + "TEE/deepseek-v3.1": { + "id": "TEE/deepseek-v3.1", + "name": "DeepSeek V3.1 TEE", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-08-21", + "last_updated": "2025-08-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 32768 + "context": 164000, + "input": 164000, + "output": 8192 }, "cost": { - "input": 0.1, - "output": 0.4, - "cache_read": 0.01 + "input": 1, + "output": 2.5 } - } - } - }, - "hetzner": { - "id": "hetzner", - "env": [ - "HETZNER_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.hetzner.com/api/v1", - "name": "Hetzner", - "doc": "https://experiments.hetzner.com/docs/inference", - "models": { - "Qwen3.8-27B": { - "id": "Qwen3.8-27B", - "name": "Qwen3.8-27B", - "description": "Dense 27B vision-language model for coding, agent tasks, and image and video understanding", - "family": "qwen", + }, + "TEE/kimi-k2.6": { + "id": "TEE/kimi-k2.6", + "name": "Kimi K2.6 TEE", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", "attachment": true, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "structured_output": false, + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -320083,90 +191095,60 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 262144, - "output": 262144 + "input": 262144, + "output": 65536 }, - "status": "beta", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 1.5, + "output": 5.25, + "cache_read": 0.375 } }, - "Qwen/Qwen3.6-35B-A3B-FP8": { - "id": "Qwen/Qwen3.6-35B-A3B-FP8", - "name": "Qwen3.6 35B A3B FP8", - "description": "Open multimodal Qwen MoE for local agents that need vision, audio, and code", + "TEE/qwen3.5-397b-a17b": { + "id": "TEE/qwen3.5-397b-a17b", + "name": "Qwen3.5 397B A17B TEE", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-04-17", - "last_updated": "2026-04-17", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-28", + "last_updated": "2026-02-28", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 258048, + "input": 258048, + "output": 65536 }, - "status": "beta", "cost": { - "input": 0, - "output": 0, - "cache_read": 0 + "input": 0.6, + "output": 3.6 } - } - } - }, - "digitalocean": { - "id": "digitalocean", - "env": [ - "DIGITALOCEAN_ACCESS_TOKEN" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://inference.do-ai.run/v1", - "name": "DigitalOcean", - "doc": "https://docs.digitalocean.com/products/gradient-ai-platform/details/models/", - "models": { - "deepseek-3.2": { - "id": "deepseek-3.2", - "name": "Deepseek 3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + }, + "TEE/kimi-k2.5-thinking": { + "id": "TEE/kimi-k2.5-thinking", + "name": "Kimi K2.5 Thinking TEE", + "description": "Kimi reasoning model for long-horizon research, planning, and tool use", + "family": "kimi-thinking", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2025-12-02", - "last_updated": "2026-04-30", + "reasoning_options": [], + "tool_call": false, + "structured_output": false, + "release_date": "2026-01-29", + "last_updated": "2026-01-29", "modalities": { "input": [ "text" @@ -320175,42 +191157,37 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 163840, - "output": 163840 + "context": 128000, + "input": 128000, + "output": 65535 }, "cost": { - "input": 0.5, - "output": 1.6, - "cache_read": 0.15 + "input": 0.3, + "output": 1.9 } }, - "openai-gpt-5.5": { - "id": "openai-gpt-5.5", - "name": "OpenAI GPT-5.5", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", + "TEE/qwen3.6-35b-a3b-uncensored": { + "id": "TEE/qwen3.6-35b-a3b-uncensored", + "name": "Qwen3.6 35B A3B Uncensored TEE", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 131072 } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-12-01", - "release_date": "2026-04-23", - "last_updated": "2026-04-30", + "release_date": "2026-05-23", + "last_updated": "2026-05-23", "modalities": { "input": [ "text", @@ -320222,58 +191199,29 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 131072, + "input": 131072, + "output": 131072 }, "cost": { - "input": 5, - "output": 30, - "cache_read": 0.5, - "tiers": [ - { - "input": 10, - "output": 45, - "cache_read": 1, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 45, - "cache_read": 1 - } + "input": 0.3, + "output": 1.5 } }, - "openai-gpt-5.2-pro": { - "id": "openai-gpt-5.2-pro", - "name": "OpenAI GPT-5.2 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "TEE/glm-4.7-flash": { + "id": "TEE/glm-4.7-flash", + "name": "GLM 4.7 Flash TEE", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-flash", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -320281,25 +191229,25 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 + "context": 203000, + "input": 203000, + "output": 65535 }, "cost": { - "input": 21, - "output": 168 + "input": 0.15, + "output": 0.5 } }, - "all-mini-lm-l6-v2": { - "id": "all-mini-lm-l6-v2", - "name": "All-MiniLM-L6-v2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "TEE/gemma4-31b": { + "id": "TEE/gemma4-31b", + "name": "Gemma 4 31B", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2021-08-30", - "last_updated": "2026-04-16", + "structured_output": true, + "release_date": "2026-04-04", + "last_updated": "2026-04-04", "modalities": { "input": [ "text" @@ -320308,27 +191256,28 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256, - "output": 384 + "context": 262144, + "input": 262144, + "output": 131072 }, "cost": { - "input": 0.009, - "output": 0 + "input": 0.45, + "output": 1 } }, - "multi-qa-mpnet-base-dot-v1": { - "id": "multi-qa-mpnet-base-dot-v1", - "name": "Multi-QA-mpnet-base-dot-v1", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "TEE/gemma-3-27b-it": { + "id": "TEE/gemma-3-27b-it", + "name": "Gemma 3 27B TEE", + "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", + "family": "gemma", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2021-08-30", - "last_updated": "2026-04-16", + "structured_output": false, + "release_date": "2025-03-10", + "last_updated": "2025-03-10", "modalities": { "input": [ "text" @@ -320337,63 +191286,61 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 512, - "output": 768 + "context": 131072, + "input": 131072, + "output": 8192 }, "cost": { - "input": 0.009, - "output": 0 + "input": 0.2, + "output": 0.8 } }, - "openai-gpt-image-1": { - "id": "openai-gpt-image-1", - "name": "GPT Image 1", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, + "TEE/gpt-oss-20b": { + "id": "TEE/gpt-oss-20b", + "name": "GPT-OSS 20B TEE", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2025-04-24", - "last_updated": "2025-04-24", + "structured_output": false, + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ - "image" + "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 0 + "context": 131072, + "input": 131072, + "output": 8192 }, "cost": { - "input": 5, - "output": 40, - "cache_read": 1.25 + "input": 0.2, + "output": 0.8 } }, - "anthropic-claude-3-opus": { - "id": "anthropic-claude-3-opus", - "name": "Claude 3 Opus", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-opus", - "attachment": true, + "TEE/glm-5": { + "id": "TEE/glm-5", + "name": "GLM 5 TEE", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", + "attachment": false, "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2023-08", - "release_date": "2024-02-29", - "last_updated": "2024-02-29", + "tool_call": false, + "structured_output": false, + "release_date": "2026-02-11", + "last_updated": "2026-02-11", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -320401,44 +191348,29 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 4096 + "context": 203000, + "input": 203000, + "output": 65535 }, - "status": "deprecated", "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 1.2, + "output": 3.5 } }, - "anthropic-claude-4.6-sonnet": { - "id": "anthropic-claude-4.6-sonnet", - "name": "Anthropic Claude Sonnet 4.6", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "TEE/glm-5.1-thinking": { + "id": "TEE/glm-5.1-thinking", + "name": "GLM 5.1 Thinking TEE", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2025-08-31", - "release_date": "2026-02-17", - "last_updated": "2026-03-13", + "structured_output": true, + "release_date": "2026-04-20", + "last_updated": "2026-04-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -320446,60 +191378,30 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 202752, + "input": 202752, + "output": 65535 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 1.5, + "output": 5.25, + "cache_read": 0.3 } }, - "anthropic-claude-haiku-4.5": { - "id": "anthropic-claude-haiku-4.5", - "name": "Anthropic Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "TEE/deepseek-v3.2": { + "id": "TEE/deepseek-v3.2", + "name": "DeepSeek V3.2 TEE", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2025-12-01", + "last_updated": "2025-12-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -320507,88 +191409,63 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 164000, + "input": 164000, + "output": 65536 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 0.1, - "cache_write": 1.25 + "input": 0.5, + "output": 1 } }, - "glm-5.3-flash": { - "id": "glm-5.3-flash", - "name": "GLM5.3 Flash", - "description": "Native multimodal GLM model for efficient coding and long-horizon agent tasks", - "family": "glm-flash", - "attachment": true, + "TEE/qwen3.5-122b-a10b": { + "id": "TEE/qwen3.5-122b-a10b", + "name": "Qwen3.5 122B A10B TEE", + "description": "Reasoning model for deliberate analysis, multi-step problem solving, and tool use", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "structured_output": false, + "release_date": "2026-05-26", + "last_updated": "2026-05-26", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 262144, + "input": 262144, + "output": 262144 }, "cost": { - "input": 0.15, - "output": 0.5, - "cache_read": 0.03 + "input": 0.46, + "output": 3.68 } }, - "openai-o1": { - "id": "openai-o1", - "name": "OpenAI o1", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2023-09", - "release_date": "2024-12-05", - "last_updated": "2024-12-05", + "Sao10K/L3.1-70B-Hanami-x1": { + "id": "Sao10K/L3.1-70B-Hanami-x1", + "name": "Llama 3.1 70B Hanami", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -320596,38 +191473,26 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 100000 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 15, - "output": 60, - "cache_read": 7.5 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "deepseek-v4-flash-0731": { - "id": "deepseek-v4-flash-0731", - "name": "DeepSeek V4 Flash 0731", - "description": "Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding", - "family": "deepseek-flash", + "Sao10K/L3-8B-Stheno-v3.2": { + "id": "Sao10K/L3-8B-Stheno-v3.2", + "name": "Sao10K Stheno 8b", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-07-31", - "last_updated": "2026-07-31", + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-11-29", + "last_updated": "2024-11-29", "modalities": { "input": [ "text" @@ -320636,44 +191501,31 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 16384, + "input": 16384, + "output": 8192 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 0.2006, + "output": 0.2006 } }, - "anthropic-claude-opus-4.7": { - "id": "anthropic-claude-opus-4.7", - "name": "Anthropic Claude Opus 4.7", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-04-16", - "last_updated": "2026-04-16", + "Sao10K/L3.3-70B-Euryale-v2.3": { + "id": "Sao10K/L3.3-70B-Euryale-v2.3", + "name": "Llama 3.3 70B Euryale", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -320681,74 +191533,59 @@ }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 20480, + "input": 20480, + "output": 16384 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "stable-diffusion-3.5-large": { - "id": "stable-diffusion-3.5-large", - "name": "Stable Diffusion 3.5 Large", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "stable-diffusion", + "Sao10K/L3.1-70B-Euryale-v2.2": { + "id": "Sao10K/L3.1-70B-Euryale-v2.2", + "name": "Llama 3.1 70B Euryale", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2024-10-22", - "last_updated": "2026-04-30", + "structured_output": false, + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 256, - "output": 1 + "context": 20480, + "input": 20480, + "output": 16384 }, "cost": { - "input": 0.08, - "output": 0 + "input": 0.306, + "output": 0.357 } }, - "openai-gpt-6-luna": { - "id": "openai-gpt-6-luna", - "name": "OpenAI GPT-6 Luna", - "description": "OpenAI's most efficient model for focused, high-volume tasks", - "family": "gpt-luna", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], - "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-05-18", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "Steelskull/L3.3-MS-Evalebis-70b": { + "id": "Steelskull/L3.3-MS-Evalebis-70b", + "name": "MS Evalebis 70b", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -320756,60 +191593,29 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 0.1, - "output": 0.5, - "cache_read": 0.01, - "tiers": [ - { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.2, - "output": 0.75, - "cache_read": 0.02 - } + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8-Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], - "tool_call": true, - "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "Steelskull/L3.3-MS-Nevoria-70b": { + "id": "Steelskull/L3.3-MS-Nevoria-70b", + "name": "Steelskull Nevoria 70b", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -320817,88 +191623,59 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 262144 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 2, - "output": 6, - "cache_read": 0.2 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], - "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "Steelskull/L3.3-Cu-Mai-R1-70b": { + "id": "Steelskull/L3.3-Cu-Mai-R1-70b", + "name": "Llama 3.3 70B Cu Mai", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 131072 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "anthropic-claude-opus-5.5": { - "id": "anthropic-claude-opus-5.5", - "name": "Anthropic Claude Opus 5.5", - "description": "Claude model for long-running agentic coding and knowledge work", - "family": "claude-opus", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], - "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "Steelskull/L3.3-MS-Evayale-70B": { + "id": "Steelskull/L3.3-MS-Evayale-70B", + "name": "Evayale 70b ", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -320906,135 +191683,133 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 16384, + "input": 16384, + "output": 16384 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.2, - "cache_write": 5, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.4, - "cache_write": 10, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.4, - "cache_write": 10 - } + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "llama-4-maverick": { - "id": "llama-4-maverick", - "name": "Llama 4 Maverick", - "description": "Open multimodal Llama model for strong reasoning and fast responses", + "Steelskull/L3.3-Nevoria-R1-70b": { + "id": "Steelskull/L3.3-Nevoria-R1-70b", + "name": "Steelskull Nevoria R1 70b", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", "family": "llama", - "attachment": true, + "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2024-08", - "release_date": "2025-04-05", - "last_updated": "2026-04-30", + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, + "context": 16384, + "input": 16384, "output": 16384 }, "cost": { - "input": 0.25, - "output": 0.87 + "input": 0.49299999999999994, + "output": 0.49299999999999994 } }, - "openai-o3-mini": { - "id": "openai-o3-mini", - "name": "OpenAI o3 mini", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o-mini", - "attachment": true, + "Steelskull/L3.3-Electra-R1-70b": { + "id": "Steelskull/L3.3-Electra-R1-70b", + "name": "Steelskull Electra R1 70b", + "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "family": "llama", + "attachment": false, + "reasoning": false, + "tool_call": false, + "structured_output": false, + "release_date": "2024-12-06", + "last_updated": "2024-12-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 16384, + "input": 16384, + "output": 16384 + }, + "cost": { + "input": 0.69989, + "output": 0.69989 + } + } + } + }, + "moark": { + "id": "moark", + "env": [ + "MOARK_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://moark.com/v1", + "name": "Moark", + "doc": "https://moark.com/docs/openapi/v1#tag/%E6%96%87%E6%9C%AC%E7%94%9F%E6%88%90", + "models": { + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2024-12-20", - "last_updated": "2025-01-29", + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 100000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.1, - "output": 4.4, - "cache_read": 0.55 + "input": 2.1, + "output": 8.4 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM 5", + "GLM-4.7": { + "id": "GLM-4.7", + "name": "GLM-4.7", "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-04-16", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -321045,39 +191820,47 @@ }, "open_weights": true, "limit": { - "context": 64000, - "output": 64000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1, - "output": 3.2, - "cache_read": 0.2 + "input": 3.5, + "output": 14 } - }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", + } + } + }, + "lilac": { + "id": "lilac", + "env": [ + "LILAC_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.getlilac.com/v1", + "name": "Lilac", + "doc": "https://docs.getlilac.com/inference/models", + "models": { + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "xhigh", - "max" - ] + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -321089,41 +191872,37 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 1048576 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.006 + "input": 0.7, + "output": 3.5, + "cache_read": 0.2 } }, - "minimax-m2.5": { - "id": "minimax-m2.5", - "name": "MiniMax M2.5 (Public Preview)", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax-m2.5", + "minimaxai/minimax-m3": { + "id": "minimaxai/minimax-m3", + "name": "MiniMax M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax-m3", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-08", - "release_date": "2026-02-12", - "last_updated": "2026-04-16", + "release_date": "2026-06-01", + "last_updated": "2026-06-01", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -321131,231 +191910,164 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 1048576, + "output": 1048576 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06 + "input": 0.28, + "output": 1.1, + "cache_read": 0.05 } }, - "anthropic-claude-4.1-opus": { - "id": "anthropic-claude-4.1-opus", - "name": "Anthropic Claude 4.1 Opus", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "google/gemma-4-31b-it": { + "id": "google/gemma-4-31b-it", + "name": "Gemma 4 31B IT", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-08-05", - "last_updated": "2025-08-05", - "modalities": { - "input": [ - "text", - "image" - ], - "output": [ - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 200000, - "output": 8192 - }, - "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 - } - }, - "wan2-2-t2v-a14b": { - "id": "wan2-2-t2v-a14b", - "name": "Wan2.2-T2V-A14B", - "description": "Video model for prompt-guided generation, editing, and motion workflows", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-07-28", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "video" - ] - }, - "open_weights": true, - "limit": { - "context": 100, - "output": 1 + "interleaved": { + "field": "reasoning_content" }, - "cost": { - "input": 0.6, - "output": 0 - } - }, - "openai-gpt-4o": { - "id": "openai-gpt-4o", - "name": "OpenAI GPT-4o", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", - "attachment": true, - "reasoning": false, - "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-05-13", - "last_updated": "2024-08-06", + "knowledge": "2025-01", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 262100, + "output": 262100 }, "cost": { - "input": 2.5, - "output": 10, - "cache_read": 1.25 + "input": 0.11, + "output": 0.35 } }, - "anthropic-claude-4.5-haiku": { - "id": "anthropic-claude-4.5-haiku", - "name": "Claude Haiku 4.5", - "description": "Fast Claude model for responsive assistance, classification, and lightweight agents", - "family": "claude-haiku", - "attachment": true, + "zai-org/glm-5.2": { + "id": "zai-org/glm-5.2", + "name": "GLM 5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ - "low", - "medium", - "high" + "high", + "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-02-28", - "release_date": "2025-10-15", - "last_updated": "2025-10-15", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 524288, + "output": 524288 }, "cost": { - "input": 1, - "output": 5, - "cache_read": 1, - "cache_write": 1.25 + "input": 0.9, + "output": 3, + "cache_read": 0.27 } - }, - "anthropic-claude-fable-5.1": { - "id": "anthropic-claude-fable-5.1", - "name": "Anthropic Claude Fable 5.1", - "description": "Claude model for demanding reasoning and long-horizon agentic work", - "family": "claude-fable", + } + } + }, + "ambient": { + "id": "ambient", + "env": [ + "AMBIENT_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.ambient.xyz/v1", + "name": "Ambient", + "doc": "https://ambient.xyz", + "models": { + "moonshotai/kimi-k2.7-code": { + "id": "moonshotai/kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "knowledge": "2026-06", - "release_date": "2026-09-01", - "last_updated": "2026-09-01", + "structured_output": true, + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 262144, + "output": 262144 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 0.25, - "cache_write": 12.5 + "input": 0.75, + "output": 3.5, + "cache_read": 0.16, + "cache_write": 0 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", + "moonshotai/kimi-k2.6": { + "id": "moonshotai/kimi-k2.6", "name": "Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-01", "release_date": "2026-04-21", "last_updated": "2026-04-21", @@ -321376,21 +192088,26 @@ "cost": { "input": 0.95, "output": 4, - "cache_read": 0.19 + "cache_read": 0.2, + "cache_write": 0 } }, - "deepseek-v3": { - "id": "deepseek-v3", - "name": "DeepSeek V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "zai-org/GLM-5.1-FP8": { + "id": "zai-org/GLM-5.1-FP8", + "name": "GLM 5.1", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-12-26", - "last_updated": "2025-03-24", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -321401,80 +192118,103 @@ }, "open_weights": true, "limit": { - "context": 163840, + "context": 202752, "output": 131072 + }, + "cost": { + "input": 1.4, + "output": 4.4, + "cache_read": 0, + "cache_write": 0 } - }, - "deepseek-r1-distill-llama-70b": { - "id": "deepseek-r1-distill-llama-70b", - "name": "DeepSeek R1 Distill Llama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + } + } + }, + "neon": { + "id": "neon", + "env": [ + "NEON_AI_GATEWAY_BASE_URL", + "NEON_AI_GATEWAY_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/mlflow/v1", + "name": "Neon", + "doc": "https://neon.com/docs", + "models": { + "gemini-3-flash": { + "id": "gemini-3-flash", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-01-30", - "last_updated": "2025-01-30", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32678, - "output": 8192 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.99, - "output": 0.99 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 } }, - "openai-gpt-5-nano": { - "id": "openai-gpt-5-nano", - "name": "OpenAI GPT-5 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + "claude-sonnet-4": { + "id": "claude-sonnet-4", + "name": "Claude Sonnet 4.5", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -321482,56 +192222,28 @@ }, "open_weights": false, "limit": { - "context": 400000, - "output": 128000 - }, - "cost": { - "input": 0.05, - "output": 0.4, - "cache_read": 0.005 - } - }, - "qwen3-embedding-0.6b": { - "id": "qwen3-embedding-0.6b", - "name": "Qwen3 Embedding 0.6B", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-06-03", - "last_updated": "2026-04-16", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 8000, - "output": 1024 + "context": 200000, + "output": 64000 }, - "status": "beta", "cost": { - "input": 0.04, - "output": 0 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "mistral-7b-instruct-v0.3": { - "id": "mistral-7b-instruct-v0.3", - "name": "Mistral 7B Instruct v0.3", - "description": "Mistral model for multilingual chat, reasoning, and tool-assisted workflows", - "family": "mistral", + "qwen3-next-80b-a3b-instruct": { + "id": "qwen3-next-80b-a3b-instruct", + "name": "Qwen3-Next 80B-A3B Instruct", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "release_date": "2024-05-22", - "last_updated": "2024-05-22", + "knowledge": "2025-04", + "release_date": "2025-09", + "last_updated": "2025-09", "modalities": { "input": [ "text" @@ -321542,25 +192254,30 @@ }, "open_weights": true, "limit": { - "context": 32768, + "context": 131072, "output": 32768 + }, + "cost": { + "input": 0.15, + "output": 1.2 } }, - "qwen-2.5-14b-instruct": { - "id": "qwen-2.5-14b-instruct", - "name": "Qwen 2.5 14B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "llama-4-maverick": { + "id": "llama-4-maverick", + "name": "Llama 4 Maverick 17B Instruct", + "description": "Open multimodal Llama for strong reasoning with efficient everyday serving", + "family": "llama", + "attachment": true, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-09", - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "knowledge": "2024-08", + "release_date": "2025-04-05", + "last_updated": "2025-04-05", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -321568,37 +192285,41 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 1000000, + "output": 16384 + }, + "cost": { + "input": 0.5, + "output": 1.5 } }, - "openai-o3": { - "id": "openai-o3", - "name": "OpenAI o3", - "description": "O-series reasoning model for hard analysis, math, coding, and planning", - "family": "o", + "claude-opus-4-5": { + "id": "claude-opus-4-5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2024-05", - "release_date": "2025-04-16", - "last_updated": "2025-04-16", + "temperature": true, + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -321607,18 +192328,19 @@ "open_weights": false, "limit": { "context": 200000, - "output": 100000 + "output": 64000 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai-gpt-5": { - "id": "openai-gpt-5", - "name": "OpenAI GPT-5", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", + "gpt-5": { + "id": "gpt-5", + "name": "GPT-5", + "description": "Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows", "family": "gpt", "attachment": true, "reasoning": true, @@ -321645,33 +192367,40 @@ "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" + }, "cost": { "input": 1.25, "output": 10, "cache_read": 0.125 } }, - "openai-gpt-4.1": { - "id": "openai-gpt-4.1", - "name": "OpenAI GPT-4.1", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "gemma-3-12b": { + "id": "gemma-3-12b", + "name": "Gemma 3 12B", + "description": "Google's open-weight Gemma 3 vision-language model for text and image understanding", + "family": "gemma", "attachment": true, "reasoning": false, "tool_call": true, "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2024-08-31", + "release_date": "2025-03-13", + "last_updated": "2025-03-13", "modalities": { "input": [ "text", @@ -321681,20 +192410,19 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 131072, + "output": 16384 }, "cost": { - "input": 2, - "output": 8, - "cache_read": 0.5 + "input": 0.15, + "output": 0.5 } }, - "openai-gpt-5.1-codex-max": { - "id": "openai-gpt-5.1-codex-max", - "name": "GPT-5.1 Codex Max", + "gpt-5-1-codex-mini": { + "id": "gpt-5-1-codex-mini", + "name": "GPT-5.1 Codex mini", "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", "family": "gpt-codex", "attachment": true, @@ -321705,8 +192433,7 @@ "values": [ "low", "medium", - "high", - "xhigh" + "high" ] } ], @@ -321722,99 +192449,119 @@ "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" + }, "cost": { - "input": 1.25, - "output": 10, - "cache_read": 0.125 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "anthropic-claude-3.5-haiku": { - "id": "anthropic-claude-3.5-haiku", - "name": "Claude 3.5 Haiku", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-haiku", - "attachment": false, - "reasoning": false, + "gpt-5-1": { + "id": "gpt-5-1", + "name": "GPT-5.1", + "description": "Sharper GPT-5 generation for coding, product work, and tool-assisted tasks", + "family": "gpt", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-07", - "release_date": "2024-11-05", - "last_updated": "2024-11-05", + "structured_output": true, + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" }, - "status": "deprecated", "cost": { - "input": 0.8, - "output": 4, - "cache_read": 0.08, - "cache_write": 1 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "anthropic-claude-3.5-sonnet": { - "id": "anthropic-claude-3.5-sonnet", - "name": "Claude 3.5 Sonnet", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-sonnet", - "attachment": true, + "meta-llama-3-3-70b-instruct": { + "id": "meta-llama-3-3-70b-instruct", + "name": "Llama-3.3-70B-Instruct", + "description": "Popular open Llama workhorse for multilingual chat, coding, and self-hosting", + "family": "llama", + "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-06-20", - "last_updated": "2024-10-22", + "knowledge": "2023-12", + "release_date": "2024-12-06", + "last_updated": "2024-12-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 + "context": 128000, + "output": 4096 }, - "status": "deprecated", "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.5, + "output": 1.5 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Kimi K2.5", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "gemini-3-pro": { + "id": "gemini-3-pro", + "name": "Gemini 3 Pro Preview", + "description": "Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts", + "family": "gemini-pro", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", "high" @@ -321822,83 +192569,107 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, + "temperature": true, "knowledge": "2025-01", - "release_date": "2026-01", - "last_updated": "2026-04-16", + "release_date": "2025-11-18", + "last_updated": "2025-11-18", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.5, - "output": 2.7, - "cache_read": 0.203 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "anthropic-claude-opus-4": { - "id": "anthropic-claude-opus-4", - "name": "Claude Opus 4", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "gpt-5-2": { + "id": "gpt-5-2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image", - "pdf" + "image" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 200000, - "output": 32000 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" }, "cost": { - "input": 15, - "output": 75, - "cache_read": 1.5, - "cache_write": 18.75 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "openai-gpt-6-astra": { - "id": "openai-gpt-6-astra", - "name": "OpenAI GPT-6 Astra", - "description": "GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.", - "family": "gpt-astra", + "gemini-3-1-flash-lite": { + "id": "gemini-3-1-flash-lite", + "name": "Gemini 3.1 Flash Lite Preview", + "description": "Low-latency Gemini model for high-volume multimodal and agent workloads", + "family": "gemini-flash-lite", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -321907,22 +192678,23 @@ "values": [ "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2026-04-30", - "release_date": "2026-09-04", - "last_updated": "2026-09-04", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-03-03", + "last_updated": "2026-03-03", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -321930,122 +192702,92 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "tiers": [ - { - "input": 20, - "output": 75, - "cache_read": 2, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 20, - "output": 75, - "cache_read": 2 - } + "input": 0.25, + "output": 1.5, + "cache_read": 0.025, + "input_audio": 0.5 } }, - "openai-gpt-image-2": { - "id": "openai-gpt-image-2", - "name": "OpenAI GPT Image 2", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", + "claude-sonnet-4-5": { + "id": "claude-sonnet-4-5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-04-24", - "last_updated": "2025-04-24", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", "modalities": { "input": [ "text", - "image" - ], - "output": [ "image", - "text" - ] - }, - "open_weights": false, - "limit": { - "context": 0, - "output": 16384 - }, - "cost": { - "input": 8, - "output": 30 - } - }, - "openai-gpt-image-1.5": { - "id": "openai-gpt-image-1.5", - "name": "OpenAI GPT Image 1.5", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "gpt-image", - "attachment": true, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-11-25", - "last_updated": "2025-11-25", - "modalities": { - "input": [ - "text", - "image" + "pdf" ], "output": [ - "image", "text" ] }, "open_weights": false, "limit": { - "context": 0, - "output": 16384 + "context": 200000, + "output": 64000 }, "cost": { - "input": 5, - "output": 10, - "cache_read": 1 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openai-gpt-4o-mini": { - "id": "openai-gpt-4o-mini", - "name": "OpenAI GPT-4o mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "claude-opus-4-7": { + "id": "claude-opus-4-7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ + { + "type": "toggle" + }, { "type": "effort", "values": [ "low", "medium", - "high" + "high", + "xhigh", + "max" ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "knowledge": "2023-09", - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -322053,60 +192795,43 @@ }, "open_weights": false, "limit": { - "context": 128000, - "output": 16384 - }, - "cost": { - "input": 0.15, - "output": 0.6, - "cache_read": 0.075 - } - }, - "qwen3-coder-flash": { - "id": "qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2026-04-30", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 262144 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.45, - "output": 1.7, - "cache_read": 0.09 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-4-flash": { - "id": "deepseek-4-flash", - "name": "Deepseek V4 Flash", - "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", - "family": "deepseek", - "attachment": false, - "reasoning": false, + "gemini-2-5-pro": { + "id": "gemini-2-5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 128, + "max": 32768 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-27", - "last_updated": "2026-05-29", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" @@ -322115,28 +192840,62 @@ "open_weights": false, "limit": { "context": 1048576, - "output": 384000 + "output": 65536 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.028 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "nemotron-3-ultra-550b": { - "id": "nemotron-3-ultra-550b", - "name": "Nemotron 3 Ultra", - "description": "Flagship Nemotron model for high-throughput reasoning and complex agents", - "family": "nemotron", - "attachment": false, - "reasoning": false, + "claude-opus-4-8": { + "id": "claude-opus-4-8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2026-06-04", - "last_updated": "2026-06-12", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -322144,29 +192903,32 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.9, - "output": 1.7 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "openai-gpt-5.2": { - "id": "openai-gpt-5.2", - "name": "OpenAI GPT-5.2", - "description": "GPT model for general reasoning, writing, coding, and tool-assisted tasks", - "family": "gpt", + "gpt-5-3-codex": { + "id": "gpt-5-3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "minimal", + "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], @@ -322174,34 +192936,42 @@ "structured_output": true, "temperature": false, "knowledge": "2025-08-31", - "release_date": "2025-12-11", - "last_updated": "2025-12-11", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { "context": 400000, + "input": 272000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" + }, "cost": { "input": 1.75, "output": 14, "cache_read": 0.175 } }, - "nvidia-nemotron-3-super-120b": { - "id": "nvidia-nemotron-3-super-120b", - "name": "NVIDIA Nemotron 3 Super 120B (Public Preview)", - "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", - "family": "nemotron", - "attachment": false, + "gpt-5-4-nano": { + "id": "gpt-5-4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -322211,154 +192981,191 @@ "low", "medium", "high", - "max" + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-03-11", - "last_updated": "2026-03-11", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" }, "cost": { - "input": 0.3, - "output": 0.65, - "cache_read": 0.06 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "deepseek-v4-pro-0813": { - "id": "deepseek-v4-pro-0813", - "name": "DeepSeek V4 Pro 0813", - "description": "DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes", - "family": "deepseek-thinking", - "attachment": false, + "claude-opus-4-1": { + "id": "claude-opus-4-1", + "name": "Claude Opus 4.1 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "none", - "low", - "high", - "max" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 31999 } ], "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-08-12", - "last_updated": "2026-08-22", + "knowledge": "2025-03-31", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 200000, + "output": 32000 }, "cost": { - "input": 1.32, - "output": 3.96, - "cache_read": 0.044 + "input": 15, + "output": 75, + "cache_read": 1.5, + "cache_write": 18.75 } }, - "mimo-v2.5-pro": { - "id": "mimo-v2.5-pro", - "name": "MiMo V2.5 Pro", - "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", - "family": "mimo", - "attachment": false, + "gpt-5-4-mini": { + "id": "gpt-5-4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ "none", - "high" + "low", + "medium", + "high", + "xhigh" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "knowledge": "2024-12", - "release_date": "2026-04-22", - "last_updated": "2026-04-22", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" }, "cost": { - "input": 0.8, - "output": 3, - "cache_read": 0.16 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "mistral-3-14B": { - "id": "mistral-3-14B", - "name": "Ministral 3 14B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", - "attachment": false, - "reasoning": false, + "gemini-2-5-flash": { + "id": "gemini-2-5-flash", + "name": "Gemini 2.5 Flash", + "description": "Fast Gemini workhorse for multimodal apps where latency and price matter", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 0, + "max": 24576 + } + ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2026-04-30", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 128000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 0.2 + "input": 0.3, + "output": 2.5, + "cache_read": 0.03, + "input_audio": 1 } }, - "anthropic-claude-opus-5": { - "id": "anthropic-claude-opus-5", - "name": "Anthropic Claude Opus 5", - "description": "Strongest Claude Opus model for coding, agents, and professional work", - "family": "claude-opus", + "gpt-5-2-codex": { + "id": "gpt-5-2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -322368,66 +193175,69 @@ "low", "medium", "high", - "xhigh", - "max" + "xhigh" ] } ], "tool_call": true, + "structured_output": true, "temperature": false, - "knowledge": "2026-05", - "release_date": "2026-07-24", - "last_updated": "2026-07-24", + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 1000000, + "context": 400000, + "input": 272000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" + }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "nemotron-3-nano-omni": { - "id": "nemotron-3-nano-omni", - "name": "Nemotron 3 Nano Omni", - "description": "Open Nemotron omni model combining reasoning with text, vision, and audio", - "family": "nemotron", - "attachment": true, + "gpt-oss-120b": { + "id": "gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open GPT reasoning model for self-hosted agents and controllable deployments", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, "temperature": true, - "release_date": "2026-04-28", - "last_updated": "2026-04-28", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -322435,57 +193245,68 @@ }, "open_weights": true, "limit": { - "context": 65536, - "output": 65536 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.5, - "output": 0.9 + "input": 0.072, + "output": 0.28 } }, - "nemotron-3-nano-30b": { - "id": "nemotron-3-nano-30b", - "name": "Nemotron 3 Nano 30B A3B", - "description": "Small Nemotron 3 MoE for efficient coding, math, and long-context agents", - "family": "nemotron", - "attachment": false, + "gemini-3-5-flash": { + "id": "gemini-3-5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 1048576, + "output": 65536 + }, + "cost": { + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 } }, - "arcee-trinity-large-thinking": { - "id": "arcee-trinity-large-thinking", - "name": "Arcee Trinity Large Thinking (Public Preview)", - "description": "Reasoning-optimized 398B MoE agent model with extended thinking for long-horizon and multi-turn tool use", - "family": "trinity", - "attachment": false, + "gemini-3-1-pro": { + "id": "gemini-3-1-pro", + "name": "Gemini 3.1 Pro Preview Custom Tools", + "description": "Advanced Gemini model for complex reasoning, coding, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -322498,54 +193319,77 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-04-01", - "last_updated": "2026-05-28", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 32000 + "context": 1048576, + "output": 65536 }, "cost": { - "input": 0.25, - "output": 0.9, - "cache_read": 0.06 + "input": 2, + "output": 12, + "cache_read": 0.2, + "tiers": [ + { + "input": 4, + "output": 18, + "cache_read": 0.4, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "anthropic-claude-5-sonnet": { - "id": "anthropic-claude-5-sonnet", - "name": "Anthropic Claude Sonnet 5", - "description": "Everyday Claude agent model for coding, planning, browsing, and general work", - "family": "claude-sonnet", - "attachment": false, + "claude-haiku-4-5": { + "id": "claude-haiku-4-5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 } ], "tool_call": true, - "temperature": false, - "knowledge": "2026-01-31", - "release_date": "2026-06-30", - "last_updated": "2026-06-30", + "temperature": true, + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -322553,42 +193397,43 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 200000, + "output": 64000 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "cache_write": 2.5 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "openai-gpt-5.4-pro": { - "id": "openai-gpt-5.4-pro", - "name": "OpenAI GPT-5.4 Pro", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt-pro", + "claude-opus-4-6": { + "id": "claude-opus-4-6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 127999 } ], "tool_call": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ "text" @@ -322596,40 +193441,39 @@ }, "open_weights": false, "limit": { - "context": 1050000, + "context": 1000000, "output": 128000 }, "cost": { - "input": 30, - "output": 180, - "tiers": [ - { - "input": 60, - "output": 270, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 60, - "output": 270 - } + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "llama3-8b-instruct": { - "id": "llama3-8b-instruct", - "name": "Llama 3.1 Instruct (8B)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", - "family": "llama", + "qwen35-122b-a10b": { + "id": "qwen35-122b-a10b", + "name": "Qwen3.5 122B-A10B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-07-23", - "last_updated": "2024-07-23", + "release_date": "2026-02-23", + "last_updated": "2026-02-23", "modalities": { "input": [ "text" @@ -322640,26 +193484,27 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 131072 + "context": 262144, + "output": 8000 }, "cost": { - "input": 0.198, - "output": 0.198 + "input": 0.22, + "output": 2.2 } }, - "llama3.3-70b-instruct": { - "id": "llama3.3-70b-instruct", - "name": "Llama 3.3 Instruct (70B)", - "description": "Open Llama instruction model for multilingual chat, reasoning, and coding", + "meta-llama-3-1-8b-instruct": { + "id": "meta-llama-3-1-8b-instruct", + "name": "Llama 3.1 8B Instruct", + "description": "Meta's compact open-weight Llama 3.1 model for fast, low-cost text generation", "family": "llama", "attachment": false, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2023-12", - "release_date": "2024-12-06", - "last_updated": "2024-12-06", + "knowledge": "2023-12-31", + "release_date": "2024-07-23", + "last_updated": "2024-07-23", "modalities": { "input": [ "text" @@ -322670,25 +193515,26 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 131072, + "output": 16384 }, "cost": { - "input": 0.65, - "output": 0.65 + "input": 0.15, + "output": 0.45 } }, - "openai-gpt-oss-120b": { - "id": "openai-gpt-oss-120b", - "name": "OpenAI GPT-oss-120b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", - "attachment": false, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", "medium", "high" @@ -322696,80 +193542,138 @@ } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-05", - "last_updated": "2026-04-16", + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 128000, - "output": 4096 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" }, "cost": { - "input": 0.1, - "output": 0.7, - "cache_read": 0.02 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "alibaba-qwen3-32b": { - "id": "alibaba-qwen3-32b", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "gpt-5-nano": { + "id": "gpt-5-nano", + "name": "GPT-5 Nano", + "description": "Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs", + "family": "gpt-nano", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", + "minimal", "low", "medium", - "high", - "max" + "high" ] } ], "tool_call": true, "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", + "modalities": { + "input": [ + "text", + "image" + ], + "output": [ + "text", + "image" + ] + }, + "open_weights": false, + "limit": { + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" + }, + "cost": { + "input": 0.05, + "output": 0.4, + "cache_read": 0.005 + } + }, + "claude-sonnet-4-6": { + "id": "claude-sonnet-4-6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "budget_tokens", + "min": 1024, + "max": 63999 + } + ], + "tool_call": true, "temperature": true, - "release_date": "2025-04-30", - "last_updated": "2026-04-16", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 32768, - "output": 32768 + "context": 1000000, + "output": 64000 }, "cost": { - "input": 0.25, - "output": 0.55 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "openai-gpt-5.3-codex": { - "id": "openai-gpt-5.3-codex", - "name": "OpenAI GPT-5.3 Codex", - "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", - "family": "gpt-codex", + "gpt-5-4": { + "id": "gpt-5-4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", "attachment": true, "reasoning": true, "reasoning_options": [ @@ -322788,78 +193692,109 @@ "structured_output": true, "temperature": false, "knowledge": "2025-08-31", - "release_date": "2026-02-05", - "last_updated": "2026-02-05", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ "text", - "image" + "image", + "pdf" ], "output": [ - "text" + "text", + "image" ] }, "open_weights": false, "limit": { - "context": 400000, + "context": 1050000, + "input": 922000, "output": 128000 }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" + }, "cost": { - "input": 1.75, - "output": 14, - "cache_read": 0.175 + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tiers": [ + { + "input": 5, + "output": 22.5, + "cache_read": 0.5, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen 3.5 397B A17B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen3.5", + "gpt-5-1-codex-max": { + "id": "gpt-5-1-codex-max", + "name": "GPT-5.1 Codex Max", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high" + "high", + "xhigh" ] } ], "tool_call": true, "structured_output": true, - "temperature": true, - "release_date": "2026-02-15", - "last_updated": "2026-04-30", + "temperature": false, + "knowledge": "2024-09-30", + "release_date": "2025-11-13", + "last_updated": "2025-11-13", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ - "text" + "text", + "image" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 131072 + "context": 400000, + "input": 272000, + "output": 128000 + }, + "provider": { + "npm": "@ai-sdk/openai", + "api": "${NEON_AI_GATEWAY_BASE_URL}/ai-gateway/openai/v1", + "shape": "responses" }, "cost": { - "input": 0.55, - "output": 3.5, - "cache_read": 0.111 + "input": 1.25, + "output": 10, + "cache_read": 0.125 } }, - "anthropic-claude-3.7-sonnet": { - "id": "anthropic-claude-3.7-sonnet", - "name": "Claude 3.7 Sonnet", - "description": "Legacy model retained for compatibility with older integrations", - "family": "claude-sonnet", - "attachment": true, + "gpt-oss-20b": { + "id": "gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -322872,58 +193807,64 @@ } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-11", - "release_date": "2025-02-24", - "last_updated": "2025-02-24", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 64000 + "context": 131072, + "output": 32768 }, - "status": "deprecated", "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75 + "input": 0.05, + "output": 0.2 } - }, - "anthropic-claude-opus-4.8": { - "id": "anthropic-claude-opus-4.8", - "name": "Anthropic Claude Opus 4.8", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + } + } + }, + "upstage": { + "id": "upstage", + "env": [ + "UPSTAGE_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.upstage.ai/v1/solar", + "name": "Upstage", + "doc": "https://developers.upstage.ai/docs/apis/chat", + "models": { + "solar-pro2": { + "id": "solar-pro2", + "name": "solar-pro2", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "solar-pro", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "low", - "medium", - "high", - "max" + "minimal", + "high" ] } ], "tool_call": true, "temperature": true, - "release_date": "2026-05-28", - "last_updated": "2026-05-29", + "knowledge": "2025-03", + "release_date": "2025-05-20", + "last_updated": "2025-05-20", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -322931,46 +193872,39 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 128000 + "context": 65536, + "output": 8192 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.25, + "output": 0.25 } }, - "openai-gpt-5.6-luna": { - "id": "openai-gpt-5.6-luna", - "name": "OpenAI GPT-5.6 Luna", - "description": "Cost-efficient GPT-5.6 model for fast, high-volume workloads", - "family": "gpt-luna", - "attachment": true, + "solar-pro3": { + "id": "solar-pro3", + "name": "solar-pro3", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "solar-pro", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", "low", "medium", - "high", - "xhigh", - "max" + "high" ] } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "knowledge": "2025-03", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -322978,72 +193912,66 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 8192 }, "cost": { - "input": 0.2, - "output": 1.2, - "cache_read": 0.02, - "tiers": [ - { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 0.4, - "output": 1.8, - "cache_read": 0.04 - } + "input": 0.25, + "output": 0.25 } }, - "qwen3-tts-voicedesign": { - "id": "qwen3-tts-voicedesign", - "name": "Qwen3 TTS VoiceDesign", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "qwen", + "solar-mini": { + "id": "solar-mini", + "name": "solar-mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "solar-mini", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2026-04-21", - "last_updated": "2026-04-30", + "tool_call": true, + "temperature": true, + "knowledge": "2024-09", + "release_date": "2024-06-12", + "last_updated": "2025-04-22", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 32768, - "output": 1 + "output": 4096 + }, + "cost": { + "input": 0.15, + "output": 0.15 } - }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + } + } + }, + "zhipuai-coding-plan": { + "id": "zhipuai-coding-plan", + "env": [ + "ZHIPU_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://open.bigmodel.cn/api/coding/paas/v4", + "name": "Zhipu AI Coding Plan", + "doc": "https://docs.bigmodel.cn/cn/coding-plan/overview", + "models": { + "glm-5.1": { + "id": "glm-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, @@ -323052,8 +193980,8 @@ }, "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "release_date": "2026-03-27", + "last_updated": "2026-03-27", "modalities": { "input": [ "text" @@ -323062,46 +193990,43 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 200000, + "output": 131072 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.21 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai-gpt-6-sol": { - "id": "openai-gpt-6-sol", - "name": "OpenAI GPT-6 Sol", - "description": "OpenAI model for complex coding and agentic workflows", - "family": "gpt-sol", + "glm-5v-turbo": { + "id": "glm-5v-turbo", + "name": "GLM-5V-Turbo", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] + { + "type": "toggle" } ], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2026-04-20", - "release_date": "2026-09-22", - "last_updated": "2026-09-22", + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-04-01", + "last_updated": "2026-04-01", "modalities": { "input": [ "text", - "image" + "image", + "video", + "pdf" ], "output": [ "text" @@ -323109,58 +194034,39 @@ }, "open_weights": false, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 200000, + "output": 131072 }, "cost": { - "input": 2, - "output": 10, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 15, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 15, - "cache_read": 0.4 - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "anthropic-claude-4.5-sonnet": { - "id": "anthropic-claude-4.5-sonnet", - "name": "Anthropic Claude 4.5 Sonnet", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", - "attachment": true, + "glm-5-turbo": { + "id": "glm-5-turbo", + "name": "GLM-5-Turbo", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-07-31", - "release_date": "2025-09-29", - "last_updated": "2025-09-29", + "release_date": "2026-03-16", + "last_updated": "2026-03-16", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -323169,44 +194075,32 @@ "open_weights": false, "limit": { "context": 200000, - "output": 64000 + "output": 131072 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.6, - "cache_write": 7.5 - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "bge-m3": { - "id": "bge-m3", - "name": "BGE M3", - "description": "Flagship model for demanding analysis, coding, and production agent workflows", - "family": "bge", + "glm-4.5-air": { + "id": "glm-4.5-air", + "name": "GLM-4.5-Air", + "description": "Efficient GLM model for fast reasoning, coding, and agent workflows", + "family": "glm-air", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-01-30", - "last_updated": "2026-04-30", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07-28", + "last_updated": "2025-07-28", "modalities": { "input": [ "text" @@ -323217,29 +194111,38 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 1024 + "context": 131072, + "output": 98304 }, "cost": { - "input": 0.02, - "output": 0 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "ministral-3-8b-instruct-2512": { - "id": "ministral-3-8b-instruct-2512", - "name": "Ministral 3 8B", - "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", - "family": "ministral", + "glm-4.6v": { + "id": "glm-4.6v", + "name": "GLM-4.6V", + "description": "GLM vision model for visual reasoning, documents, and multimodal agents", + "family": "glm", "attachment": true, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, "temperature": true, - "release_date": "2025-12-15", - "last_updated": "2025-12-15", + "knowledge": "2025-04", + "release_date": "2025-12-08", + "last_updated": "2025-12-08", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" @@ -323247,98 +194150,78 @@ }, "open_weights": true, "limit": { - "context": 262144, - "output": 262144 + "context": 128000, + "output": 32768 + }, + "cost": { + "input": 0.3, + "output": 0.9 } }, - "openai-gpt-5.6-sol": { - "id": "openai-gpt-5.6-sol", - "name": "OpenAI GPT-5.6 Sol", - "description": "Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows", - "family": "gpt-sol", - "attachment": true, + "glm-5.2": { + "id": "glm-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ - "none", - "low", - "medium", "high", - "xhigh", "max" ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 1000000, + "output": 131072 }, "cost": { - "input": 4, - "output": 20, - "cache_read": 0.4, - "tiers": [ - { - "input": 8, - "output": 30, - "cache_read": 0.8, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 8, - "output": 30, - "cache_read": 0.8 - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "glm-4.7": { + "id": "glm-4.7", + "name": "GLM-4.7", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "medium", - "high", - "xhigh" - ] + "type": "toggle" } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-07", - "last_updated": "2026-04-07", + "knowledge": "2025-04", + "release_date": "2025-12-22", + "last_updated": "2025-12-22", "modalities": { "input": [ "text" @@ -323349,74 +194232,88 @@ }, "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 204800, + "output": 131072 }, "cost": { - "input": 1.3, - "output": 4.3, - "cache_read": 0.26 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "openai-gpt-5.4-nano": { - "id": "openai-gpt-5.4-nano", - "name": "OpenAI GPT-5.4 Nano", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-nano", + } + } + }, + "chutes": { + "id": "chutes", + "env": [ + "CHUTES_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://llm.chutes.ai/v1", + "name": "Chutes", + "doc": "https://llm.chutes.ai/v1/models", + "models": { + "moonshotai/Kimi-K2.6-TEE": { + "id": "moonshotai/Kimi-K2.6-TEE", + "name": "Kimi K2.6 TEE", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2025-12", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 65535 }, "cost": { - "input": 0.2, - "output": 1.25, - "cache_read": 0.02 + "input": 0.66, + "output": 3.5, + "cache_read": 0.33 } }, - "bge-reranker-v2-m3": { - "id": "bge-reranker-v2-m3", - "name": "BGE Reranker v2 M3", - "description": "Reranking model for improving retrieval quality in search and recommendation systems", - "family": "bge", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-03-12", - "last_updated": "2026-04-30", + "moonshotai/Kimi-K2.5-TEE": { + "id": "moonshotai/Kimi-K2.5-TEE", + "name": "Kimi K2.5 TEE", + "description": "Earlier Kimi frontier model for long-context agents, coding, and multimodal work", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2024-10", + "release_date": "2026-01", + "last_updated": "2026-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" @@ -323424,42 +194321,32 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 1 + "context": 262144, + "output": 65535 }, "cost": { - "input": 0.01, - "output": 0 + "input": 0.44, + "output": 2, + "cache_read": 0.22 } }, - "nemotron-nano-12b-v2-vl": { - "id": "nemotron-nano-12b-v2-vl", - "name": "Nemotron-nano 12b v2-vl", - "description": "Nemotron multimodal model for visual reasoning and agentic AI workflows", - "family": "nemotron", + "google/gemma-4-31B-turbo-TEE": { + "id": "google/gemma-4-31B-turbo-TEE", + "name": "gemma 4 31B turbo TEE", + "description": "Largest Gemma 4 instruction model for open, self-hosted chat and reasoning", + "family": "gemma", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-10-28", - "last_updated": "2025-10-28", + "release_date": "2026-04-02", + "last_updated": "2026-04-02", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -323467,83 +194354,61 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 131072, + "output": 65536 }, "cost": { - "input": 0.2, - "output": 0.6 + "input": 0.12, + "output": 0.37, + "cache_read": 0.06 } }, - "openai-gpt-5.4-mini": { - "id": "openai-gpt-5.4-mini", - "name": "OpenAI GPT-5.4 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", - "attachment": true, + "Qwen/Qwen3-32B-TEE": { + "id": "Qwen/Qwen3-32B-TEE", + "name": "Qwen3 32B TEE", + "description": "Dense open Qwen model for self-hosted chat, reasoning, and coding", + "family": "qwen", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-17", - "last_updated": "2026-03-17", + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-04", + "last_updated": "2025-04", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 40960, + "output": 40960 }, "cost": { - "input": 0.75, - "output": 4.5, - "cache_read": 0.075 + "input": 0.104, + "output": 0.416, + "cache_read": 0.052 } }, - "openai-gpt-5-mini": { - "id": "openai-gpt-5-mini", - "name": "OpenAI GPT-5 Mini", - "description": "Compact GPT model for low-latency assistance and high-volume workloads", - "family": "gpt-mini", + "Qwen/Qwen3.6-27B-TEE": { + "id": "Qwen/Qwen3.6-27B-TEE", + "name": "Qwen3.6 27B TEE", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "minimal", - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, "structured_output": true, - "temperature": false, - "knowledge": "2024-05-30", - "release_date": "2025-08-07", - "last_updated": "2025-08-07", + "temperature": true, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", @@ -323553,40 +194418,69 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, - "output": 128000 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.25, + "input": 0.3, "output": 2, - "cache_read": 0.025 + "cache_read": 0.15 } }, - "anthropic-claude-opus-4.6": { - "id": "anthropic-claude-opus-4.6", - "name": "Anthropic Claude Opus 4.6", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", + "Qwen/Qwen3-235B-A22B-Thinking-2507-TEE": { + "id": "Qwen/Qwen3-235B-A22B-Thinking-2507-TEE", + "name": "Qwen3 235B A22B Thinking 2507 TEE", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", + "attachment": false, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-04", + "release_date": "2025-07", + "last_updated": "2026-06-21", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 262144 + }, + "cost": { + "input": 0.2989, + "output": 1.1957, + "cache_read": 0.14945 + } + }, + "Qwen/Qwen3.5-397B-A17B-TEE": { + "id": "Qwen/Qwen3.5-397B-A17B-TEE", + "name": "Qwen3.5 397B A17B TEE", + "description": "Large open Qwen multimodal MoE for visual agents and long technical tasks", + "family": "qwen", "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-05-31", - "release_date": "2026-02-05", - "last_updated": "2026-03-13", + "release_date": "2026-02-15", + "last_updated": "2026-02-15", "modalities": { "input": [ "text", @@ -323596,47 +194490,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 + "context": 262144, + "output": 65536 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25, - "tiers": [ - { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 10, - "output": 37.5, - "cache_read": 1, - "cache_write": 12.5 - } + "input": 0.45, + "output": 3, + "cache_read": 0.225 } }, - "e5-large-v2": { - "id": "e5-large-v2", - "name": "E5 Large v2", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + "unsloth/Mistral-Nemo-Instruct-2407-TEE": { + "id": "unsloth/Mistral-Nemo-Instruct-2407-TEE", + "name": "Mistral Nemo Instruct 2407 TEE", + "description": "Efficient Mistral-NVIDIA open model for multilingual chat and local deployment", + "family": "mistral-nemo", "attachment": false, "reasoning": false, "tool_call": false, - "temperature": false, - "release_date": "2023-05-19", - "last_updated": "2026-04-30", + "temperature": true, + "knowledge": "2024-07", + "release_date": "2024-07-01", + "last_updated": "2024-07-01", "modalities": { "input": [ "text" @@ -323647,85 +194523,66 @@ }, "open_weights": true, "limit": { - "context": 512, - "output": 1024 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.02, - "output": 0 + "input": 0.0245, + "output": 0.0978, + "cache_read": 0.01225 } }, - "anthropic-claude-opus-4.5": { - "id": "anthropic-claude-opus-4.5", - "name": "Anthropic Claude Opus 4.5", - "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", - "family": "claude-opus", - "attachment": true, + "zai-org/GLM-5-TEE": { + "id": "zai-org/GLM-5-TEE", + "name": "GLM 5 TEE", + "description": "General GLM flagship for coding, analysis, and tool-heavy engineering workflows", + "family": "glm", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-11-24", - "last_updated": "2025-11-24", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 200000, - "output": 8192 + "context": 202752, + "output": 65535 }, "cost": { - "input": 5, - "output": 25, - "cache_read": 0.5, - "cache_write": 6.25 + "input": 0.95, + "output": 2.55, + "cache_read": 0.475 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "Deepseek V4 Pro", - "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", - "family": "deepseek-thinking", + "zai-org/GLM-5.1-TEE": { + "id": "zai-org/GLM-5.1-TEE", + "name": "GLM 5.1 TEE", + "description": "Strong GLM coding model for agentic engineering, terminals, and repository generation", + "family": "glm", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, "interleaved": { "field": "reasoning_content" }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "release_date": "2026-04-07", + "last_updated": "2026-04-07", "modalities": { "input": [ "text" @@ -323736,32 +194593,34 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 384000 + "context": 202752, + "output": 65535 }, "cost": { - "input": 1.74, - "output": 3.48, - "cache_read": 0.348 + "input": 0.98, + "output": 3.08, + "cache_read": 0.49 } }, - "gemma-4-31B-it": { - "id": "gemma-4-31B-it", - "name": "Gemma 4", - "description": "Open Gemma instruction model for efficient chat and self-hosted deployments", - "family": "gemma", - "attachment": true, - "reasoning": false, + "zai-org/GLM-5.2-TEE": { + "id": "zai-org/GLM-5.2-TEE", + "name": "GLM 5.2 TEE", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", + "attachment": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-22", - "last_updated": "2026-04-30", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" @@ -323769,100 +194628,66 @@ }, "open_weights": true, "limit": { - "context": 256000, - "output": 8192 + "context": 1048576, + "output": 65535 }, "cost": { - "input": 0.18, - "output": 0.5, - "cache_read": 0.036 + "input": 1.4, + "output": 4.4, + "cache_read": 0.7 } }, - "openai-gpt-5.6-terra": { - "id": "openai-gpt-5.6-terra", - "name": "OpenAI GPT-5.6 Terra", - "description": "Balanced GPT-5.6 model for capable, cost-efficient everyday work", - "family": "gpt-terra", - "attachment": true, + "deepseek-ai/DeepSeek-V3.2-TEE": { + "id": "deepseek-ai/DeepSeek-V3.2-TEE", + "name": "DeepSeek V3.2 TEE", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, - "temperature": false, - "knowledge": "2026-02-16", - "release_date": "2026-07-09", - "last_updated": "2026-07-09", + "temperature": true, + "release_date": "2025-12", + "last_updated": "2026-06-21", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1050000, - "input": 922000, - "output": 128000 + "context": 131072, + "output": 65536 }, "cost": { - "input": 2, - "output": 12, - "cache_read": 0.2, - "tiers": [ - { - "input": 4, - "output": 18, - "cache_read": 0.4, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 4, - "output": 18, - "cache_read": 0.4 - } + "input": 1, + "output": 1, + "cache_read": 0.5 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "MiniMaxAI/MiniMax-M2.5-TEE": { + "id": "MiniMaxAI/MiniMax-M2.5-TEE", + "name": "MiniMax M2.5 TEE", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -323873,26 +194698,39 @@ }, "open_weights": true, "limit": { - "context": 1048576, - "output": 128000 + "context": 196608, + "output": 65536 }, "cost": { - "input": 1.4, - "output": 4.4, - "cache_read": 0.26 + "input": 0.15, + "output": 1.2, + "cache_read": 0.075 } - }, - "gte-large-en-v1.5": { - "id": "gte-large-en-v1.5", - "name": "GTE Large (v1.5)", - "description": "Embedding model for semantic search, retrieval, clustering, and ranking pipelines", - "family": "text-embedding", + } + } + }, + "minimax-cn-coding-plan": { + "id": "minimax-cn-coding-plan", + "env": [ + "MINIMAX_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://api.minimaxi.com/anthropic/v1", + "name": "MiniMax Token Plan (minimaxi.com)", + "doc": "https://platform.minimaxi.com/docs/token-plan/intro", + "models": { + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-03-27", - "last_updated": "2026-04-16", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -323903,81 +194741,58 @@ }, "open_weights": true, "limit": { - "context": 8192, - "output": 1024 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.09, + "input": 0, "output": 0 } }, - "anthropic-claude-fable-5": { - "id": "anthropic-claude-fable-5", - "name": "Anthropic Claude Fable 5", - "description": "Claude model for creative writing, analysis, and controlled agent workflows", - "family": "claude-fable", - "attachment": true, + "MiniMax-M2.5-highspeed": { + "id": "MiniMax-M2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high", - "xhigh", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "temperature": false, - "release_date": "2026-06-09", - "last_updated": "2026-06-12", + "temperature": true, + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 204800, + "output": 131072 }, "cost": { - "input": 10, - "output": 50, - "cache_read": 1, - "cache_write": 12.5 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai-gpt-oss-20b": { - "id": "openai-gpt-oss-20b", - "name": "OpenAI GPT-oss-20b", - "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", - "family": "gpt-oss", + "MiniMax-M2.7-highspeed": { + "id": "MiniMax-M2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-06", - "release_date": "2025-08-05", - "last_updated": "2026-04-16", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -323988,86 +194803,58 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 4096 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.05, - "output": 0.45 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "openai-gpt-5.4": { - "id": "openai-gpt-5.4", - "name": "OpenAI GPT-5.4", - "description": "Frontier GPT model for professional reasoning, coding, and multimodal work", - "family": "gpt", - "attachment": true, + "MiniMax-M2": { + "id": "MiniMax-M2", + "name": "MiniMax-M2", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", + "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "none", - "low", - "medium", - "high", - "xhigh" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "structured_output": true, - "temperature": false, - "knowledge": "2025-08-31", - "release_date": "2026-03-05", - "last_updated": "2026-03-05", + "temperature": true, + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 400000, + "context": 196608, "output": 128000 }, "cost": { - "input": 2.5, - "output": 15, - "cache_read": 0.25, - "tiers": [ - { - "input": 5, - "output": 22.5, - "cache_read": 0.5, - "tier": { - "type": "context", - "size": 272000 - } - } - ], - "context_over_200k": { - "input": 5, - "output": 22.5, - "cache_read": 0.5 - } + "input": 0, + "output": 0 } }, - "mistral-nemo-instruct-2407": { - "id": "mistral-nemo-instruct-2407", - "name": "Mistral Nemo Instruct", - "description": "Legacy model retained for compatibility with older integrations", - "family": "mistral", + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2024-07-18", - "last_updated": "2024-07-18", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -324078,239 +194865,275 @@ }, "open_weights": true, "limit": { - "context": 128000, - "output": 16384 + "context": 204800, + "output": 131072 }, - "status": "deprecated", "cost": { - "input": 0.3, - "output": 0.3 + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "anthropic-claude-sonnet-4": { - "id": "anthropic-claude-sonnet-4", - "name": "Claude Sonnet 4", - "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", - "family": "claude-sonnet", + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal coding model for long-context reasoning and agent tasks", + "family": "minimax", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "effort", - "values": [ - "low", - "medium", - "high" - ] + "type": "toggle" } ], "tool_call": true, "temperature": true, - "knowledge": "2025-03-31", - "release_date": "2025-05-22", - "last_updated": "2025-05-22", + "release_date": "2026-06-01", + "last_updated": "2026-06-25", "modalities": { "input": [ "text", "image", - "pdf" + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 64000 + "output": 128000 }, "cost": { - "input": 3, - "output": 15, - "cache_read": 0.3, - "cache_write": 3.75, - "tiers": [ - { - "input": 6, - "output": 22.5, - "cache_read": 0.3, - "cache_write": 3.75, - "tier": { - "type": "context", - "size": 200000 - } - } - ], - "context_over_200k": { - "input": 6, - "output": 22.5, - "cache_read": 0.3, - "cache_write": 3.75 - } + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } }, - "fal-ai/fast-sdxl": { - "id": "fal-ai/fast-sdxl", - "name": "Fast SDXL", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "stable-diffusion", + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2023-07-26", - "last_updated": "2026-04-16", + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "temperature": true, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0, + "output": 0, + "cache_read": 0, + "cache_write": 0 } - }, - "fal-ai/elevenlabs/tts/multilingual-v2": { - "id": "fal-ai/elevenlabs/tts/multilingual-v2", - "name": "ElevenLabs Multilingual TTS v2", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "family": "elevenlabs", + } + } + }, + "deepseek": { + "id": "deepseek", + "env": [ + "DEEPSEEK_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.deepseek.com", + "name": "DeepSeek", + "doc": "https://api-docs.deepseek.com/quick_start/pricing", + "models": { + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", + "family": "deepseek-flash", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2023-08-22", - "last_updated": "2026-04-16", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } }, - "fal-ai/flux/schnell": { - "id": "fal-ai/flux/schnell", - "name": "FLUX.1 [schnell]", - "description": "Image model for prompt-driven generation, editing, and visual design workflows", - "family": "flux", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Open MoE flagship with million-token context for coding and long agent runs", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2024-08-01", - "last_updated": "2026-04-16", + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": true, + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-04-24", "modalities": { "input": [ "text" ], "output": [ - "image" + "text" ] }, "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.435, + "output": 0.87, + "cache_read": 0.003625 } }, - "fal-ai/stable-audio-25/text-to-audio": { - "id": "fal-ai/stable-audio-25/text-to-audio", - "name": "Stable Audio 2.5 (Text-to-Audio)", - "description": "Speech generation model for controllable voice, narration, and audio delivery", - "attachment": false, - "reasoning": false, - "tool_call": false, - "temperature": false, - "release_date": "2025-10-08", - "last_updated": "2026-04-16", + "deepseek-reasoner": { + "id": "deepseek-reasoner", + "name": "DeepSeek Reasoner", + "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", + "family": "deepseek-thinking", + "attachment": true, + "reasoning": true, + "reasoning_options": [], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ "text" ], "output": [ - "audio" + "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 0, - "output": 0 + "context": 1000000, + "output": 384000 + }, + "cost": { + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } - } - } - }, - "aixy": { - "id": "aixy", - "env": [ - "AIXY_API_KEY" - ], - "npm": "@ai-sdk/openai-compatible", - "api": "https://api.aixy-gateway.com/v1", - "name": "Aixy", - "doc": "https://docs.aixy-gateway.com/integrations/overview", - "models": { - "openai/gpt-4.1-mini": { - "id": "openai/gpt-4.1-mini", - "name": "GPT-4.1 mini", - "description": "Affordable GPT-4.1 lane for fast coding help and structured extraction", - "family": "gpt-mini", + }, + "deepseek-chat": { + "id": "deepseek-chat", + "name": "DeepSeek Chat", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": true, "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-04-14", - "last_updated": "2025-04-14", + "knowledge": "2025-09", + "release_date": "2025-12-01", + "last_updated": "2026-02-28", "modalities": { "input": [ - "text", - "image", - "pdf" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1047576, - "output": 32768 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.4, - "output": 1.6, - "cache_read": 0.1 + "input": 0.14, + "output": 0.28, + "cache_read": 0.0028 } } } }, - "alibaba-cn": { - "id": "alibaba-cn", + "wafer.ai": { + "id": "wafer.ai", "env": [ - "DASHSCOPE_API_KEY" + "WAFER_API_KEY" ], "npm": "@ai-sdk/openai-compatible", - "api": "https://dashscope.aliyuncs.com/compatible-mode/v1", - "name": "Alibaba (China)", - "doc": "https://www.alibabacloud.com/help/en/model-studio/models", + "api": "https://pass.wafer.ai/v1", + "name": "Wafer", + "doc": "https://docs.wafer.ai/wafer-pass", "models": { - "qwen-flash": { - "id": "qwen-flash", - "name": "Qwen Flash", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", + "deepseek-v4-flash": { + "id": "deepseek-v4-flash", + "name": "DeepSeek V4 Flash", + "description": "Fast DeepSeek model for efficient chat, coding help, and agent loops", + "family": "deepseek-flash", "attachment": false, "reasoning": true, "reasoning_options": [ @@ -324318,15 +195141,22 @@ "type": "toggle" }, { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "high", + "max" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-05-30", "modalities": { "input": [ "text" @@ -324335,19 +195165,21 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 1000000, - "output": 32768 + "output": 384000 }, "cost": { - "input": 0.022, - "output": 0.216 + "input": 0.14, + "output": 0.28, + "cache_read": 0.01, + "cache_write": 0 } }, - "qwen3.5-flash": { - "id": "qwen3.5-flash", - "name": "Qwen3.5 Flash", + "Qwen3.6-35B-A3B": { + "id": "Qwen3.6-35B-A3B", + "name": "Qwen3.6-35B-A3B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", "attachment": true, @@ -324355,21 +195187,14 @@ "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2026-02-23", - "last_updated": "2026-09-22", + "release_date": "2026-05-11", + "last_updated": "2026-05-30", "modalities": { "input": [ "text", @@ -324380,49 +195205,36 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 256000, + "input": 229376, "output": 65536 }, "cost": { - "input": 0.029, - "output": 0.287, - "reasoning": 0.287, - "tiers": [ - { - "input": 0.115, - "output": 1.147, - "reasoning": 1.147, - "tier": { - "type": "context", - "size": 128000 - } - }, - { - "input": 0.172, - "output": 1.72, - "reasoning": 1.72, - "tier": { - "type": "context", - "size": 256000 - } - } - ] + "input": 0.15, + "output": 1, + "cache_read": 0.02, + "cache_write": 0 } }, - "qwen2-5-coder-32b-instruct": { - "id": "qwen2-5-coder-32b-instruct", - "name": "Qwen2.5-Coder 32B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "qwen3.7-max": { + "id": "qwen3.7-max", + "name": "Qwen3.7-Max", + "description": "Flagship model for demanding analysis, coding, and production agent workflows", + "family": "qwen3.7-max", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11", - "last_updated": "2024-11", + "release_date": "2026-05-21", + "last_updated": "2026-05-30", "modalities": { "input": [ "text" @@ -324431,27 +195243,46 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 256000, + "output": 65536 }, "cost": { - "input": 0.287, - "output": 0.861 + "input": 5, + "output": 15, + "cache_read": 0.5, + "cache_write": 0 } }, - "deepseek-v3-2-exp": { - "id": "deepseek-v3-2-exp", - "name": "DeepSeek V3.2 Exp", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "deepseek-v4-pro": { + "id": "deepseek-v4-pro", + "name": "DeepSeek V4 Pro", + "description": "Flagship DeepSeek model for coding, reasoning, and agentic work", + "family": "deepseek-thinking", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "high", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2025-05", + "release_date": "2026-04-24", + "last_updated": "2026-05-30", "modalities": { "input": [ "text" @@ -324460,88 +195291,129 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 1000000, + "output": 384000 }, "cost": { - "input": 0.287, - "output": 0.431 + "input": 1.74, + "output": 3.48, + "cache_read": 0.02, + "cache_write": 0 } }, - "deepseek-r1-distill-qwen-14b": { - "id": "deepseek-r1-distill-qwen-14b", - "name": "DeepSeek R1 Distill Qwen 14B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "Qwen3.5-397B-A17B": { + "id": "Qwen3.5-397B-A17B", + "name": "Qwen3.5-397B-A17B", + "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2025-04", + "release_date": "2026-02-16", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 16384 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.144, - "output": 0.431 + "input": 0.43, + "output": 2.6, + "cache_read": 0.04, + "cache_write": 0 } }, - "qwen-math-plus": { - "id": "qwen-math-plus", - "name": "Qwen Math Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "Kimi-K2.6": { + "id": "Kimi-K2.6", + "name": "Kimi-K2.6", + "description": "Kimi multimodal agent model for visual understanding, coding, and planning", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-08-16", - "last_updated": "2024-09-19", + "knowledge": "2025-01", + "release_date": "2026-05-13", + "last_updated": "2026-06-01", "modalities": { "input": [ - "text" + "text", + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4096, - "output": 3072 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.574, - "output": 1.721 + "input": 0.68, + "output": 3.15, + "cache_read": 0.07, + "cache_write": 0 } }, - "qwen-deep-research": { - "id": "qwen-deep-research", - "name": "Qwen Deep Research", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "GLM-5.2": { + "id": "GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01", - "last_updated": "2024-01", + "release_date": "2026-06-13", + "last_updated": "2026-06-22", "modalities": { "input": [ "text" @@ -324550,36 +195422,36 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 32768 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 7.742, - "output": 23.367 + "input": 1.2, + "output": 4.1, + "cache_read": 0.2, + "cache_write": 0 } }, - "qwen3.7-max": { - "id": "qwen3.7-max", - "name": "Qwen3.7 Max", - "description": "Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks", - "family": "qwen", + "GLM-5.1": { + "id": "GLM-5.1", + "name": "GLM-5.1", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2026-05-21", - "last_updated": "2026-05-21", + "knowledge": "2025-04", + "release_date": "2026-04-07", + "last_updated": "2026-06-01", "modalities": { "input": [ "text" @@ -324588,30 +195460,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 202752, + "output": 131072 }, "cost": { - "input": 2.5, - "output": 7.5, - "cache_read": 0.5, - "cache_write": 3.125 + "input": 1, + "output": 3.2, + "cache_read": 0.1, + "cache_write": 0 } - }, - "qwen2-5-32b-instruct": { - "id": "qwen2-5-32b-instruct", - "name": "Qwen2.5 32B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + } + } + }, + "minimax": { + "id": "minimax", + "env": [ + "MINIMAX_API_KEY" + ], + "npm": "@ai-sdk/anthropic", + "api": "https://api.minimax.io/anthropic/v1", + "name": "MiniMax (minimax.io)", + "doc": "https://platform.minimax.io/docs/guides/quickstart", + "models": { + "MiniMax-M2.1": { + "id": "MiniMax-M2.1", + "name": "MiniMax-M2.1", + "description": "Earlier MiniMax agent model for practical coding and productivity tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "release_date": "2025-12-23", + "last_updated": "2025-12-23", "modalities": { "input": [ "text" @@ -324622,27 +195506,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.287, - "output": 0.861 + "input": 0.3, + "output": 1.2 } }, - "qwq-plus": { - "id": "qwq-plus", - "name": "QwQ Plus", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "MiniMax-M2.5-highspeed": { + "id": "MiniMax-M2.5-highspeed", + "name": "MiniMax-M2.5-highspeed", + "description": "High-speed MiniMax model for low-latency coding and agent workflows", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-03-05", - "last_updated": "2025-03-05", + "release_date": "2026-02-13", + "last_updated": "2026-02-13", "modalities": { "input": [ "text" @@ -324651,29 +195534,30 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.23, - "output": 0.574 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "qwq-32b": { - "id": "qwq-32b", - "name": "QwQ 32B", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", + "MiniMax-M2.7-highspeed": { + "id": "MiniMax-M2.7-highspeed", + "name": "MiniMax-M2.7-highspeed", + "description": "Low-latency M2.7 variant for interactive coding plans and agent loops", + "family": "minimax", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-12", - "last_updated": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -324684,30 +195568,31 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.287, - "output": 0.861 + "input": 0.6, + "output": 2.4, + "cache_read": 0.06, + "cache_write": 0.375 } }, - "qwen2-5-vl-72b-instruct": { - "id": "qwen2-5-vl-72b-instruct", - "name": "Qwen2.5-VL 72B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "MiniMax-M2": { + "id": "MiniMax-M2", + "name": "MiniMax-M2", + "description": "Efficient open MiniMax model built for coding agents and tool-heavy workflows", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "release_date": "2025-10-27", + "last_updated": "2025-10-27", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -324715,25 +195600,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 196608, + "output": 128000 }, "cost": { - "input": 2.294, - "output": 6.881 + "input": 0.3, + "output": 1.2 } }, - "deepseek-v3-1": { - "id": "deepseek-v3-1", - "name": "DeepSeek V3.1", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "MiniMax-M2.5": { + "id": "MiniMax-M2.5", + "name": "MiniMax-M2.5", + "description": "Prior MiniMax coding model for agent workflows, office edits, and automation", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-02-12", + "last_updated": "2026-02-12", "modalities": { "input": [ "text" @@ -324742,69 +195628,83 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 65536 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.574, - "output": 1.721 + "input": 0.3, + "output": 1.2, + "cache_read": 0.03, + "cache_write": 0.375 } }, - "qwen3-vl-plus": { - "id": "qwen3-vl-plus", - "name": "Qwen3-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "MiniMax-M3": { + "id": "MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09-23", - "last_updated": "2025-09-23", + "release_date": "2026-06-01", + "last_updated": "2026-06-25", "modalities": { "input": [ "text", - "image" + "image", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 262144, - "output": 32768 + "context": 1000000, + "output": 128000 }, "cost": { - "input": 0.143353, - "output": 1.433525, - "reasoning": 4.300576 + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "tiers": [ + { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12, + "tier": { + "type": "context", + "size": 512000 + } + } + ], + "context_over_200k": { + "input": 0.6, + "output": 2.4, + "cache_read": 0.12 + } } }, - "qwen-plus-character": { - "id": "qwen-plus-character", - "name": "Qwen Plus Character", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "MiniMax-M2.7": { + "id": "MiniMax-M2.7", + "name": "MiniMax-M2.7", + "description": "Open MiniMax flagship for coding agents, office automation, and complex environments", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01", - "last_updated": "2024-01", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" @@ -324813,31 +195713,90 @@ "text" ] }, + "open_weights": true, + "limit": { + "context": 204800, + "output": 131072 + }, + "cost": { + "input": 0.3, + "output": 1.2, + "cache_read": 0.06, + "cache_write": 0.375 + } + } + } + }, + "github-copilot": { + "id": "github-copilot", + "env": [ + "GITHUB_TOKEN" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.githubcopilot.com", + "name": "GitHub Copilot", + "doc": "https://docs.github.com/en/copilot", + "models": { + "claude-sonnet-4.5": { + "id": "claude-sonnet-4.5", + "name": "Claude Sonnet 4.5 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "budget_tokens", + "min": 1024, + "max": 32000 + } + ], + "tool_call": true, + "temperature": true, + "knowledge": "2025-07-31", + "release_date": "2025-09-29", + "last_updated": "2025-09-29", + "modalities": { + "input": [ + "text", + "image", + "pdf" + ], + "output": [ + "text" + ] + }, "open_weights": false, "limit": { - "context": 32768, - "output": 4096 + "context": 200000, + "input": 168000, + "output": 32000 }, "cost": { - "input": 0.115, - "output": 0.287 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "qwen-max": { - "id": "qwen-max", - "name": "Qwen Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, - "reasoning": false, + "claude-sonnet-4": { + "id": "claude-sonnet-4", + "name": "Claude Sonnet 4 (latest)", + "description": "Balanced Claude model for coding, analysis, agent workflows, and cost control", + "family": "claude-sonnet", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-03", - "last_updated": "2025-01-25", + "knowledge": "2025-03-31", + "release_date": "2025-05-22", + "last_updated": "2025-05-22", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -324845,87 +195804,100 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 216000, + "input": 128000, + "output": 16000 }, "cost": { - "input": 0.345, - "output": 1.377 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "qwen3-next-80b-a3b-thinking": { - "id": "qwen3-next-80b-a3b-thinking", - "name": "Qwen3-Next 80B-A3B (Thinking)", - "description": "Qwen reasoning model for deliberate problem solving, math, and coding", - "family": "qwen", - "attachment": false, + "gemini-2.5-pro": { + "id": "gemini-2.5-pro", + "name": "Gemini 2.5 Pro", + "description": "Google's proven reasoning model for coding, math, and multimodal analysis", + "family": "gemini-pro", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "budget_tokens", + "min": 128, + "max": 32768 } ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "knowledge": "2025-01", + "release_date": "2025-06-17", + "last_updated": "2025-06-17", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 + "context": 128000, + "input": 128000, + "output": 64000 }, "cost": { - "input": 0.144, - "output": 1.434 + "input": 1.25, + "output": 10, + "cache_read": 0.125, + "tiers": [ + { + "input": 2.5, + "output": 15, + "cache_read": 0.25, + "tier": { + "type": "context", + "size": 200000 + } + } + ], + "context_over_200k": { + "input": 2.5, + "output": 15, + "cache_read": 0.25 + } } }, - "qwen3.8-max": { - "id": "qwen3.8-max", - "name": "Qwen3.8 Max", - "description": "2.4-trillion-parameter MoE flagship for coding, professional work, multimodal understanding, and long-horizon agentic workflows", - "family": "qwen", + "claude-haiku-4.5": { + "id": "claude-haiku-4.5", + "name": "Claude Haiku 4.5 (latest)", + "description": "Fast Claude lane for lightweight agents, office tasks, and responsive chat", + "family": "claude-haiku", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "low", - "medium", - "xhigh" - ] - }, { "type": "budget_tokens", - "min": 0, - "max": 262144 + "min": 1024, + "max": 32000 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-03", - "last_updated": "2026-08-03", + "knowledge": "2025-02-28", + "release_date": "2025-10-15", + "last_updated": "2025-10-15", "modalities": { "input": [ "text", "image", - "video", "pdf" ], "output": [ @@ -324934,76 +195906,89 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 200000, + "input": 136000, + "output": 64000 }, "cost": { - "input": 1.77744, - "output": 5.33231, - "cache_read": 0.22218, - "cache_write": 2.22179 + "input": 1, + "output": 5, + "cache_read": 0.1, + "cache_write": 1.25 } }, - "kimi-k3": { - "id": "kimi-k3", - "name": "Kimi K3", - "description": "Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work", - "family": "kimi-k3", + "gemini-3.5-flash": { + "id": "gemini-3.5-flash", + "name": "Gemini 3.5 Flash", + "description": "Fast Gemini model balancing multimodal reasoning, tool use, and cost", + "family": "gemini-flash", "attachment": true, "reasoning": true, "reasoning_options": [ { "type": "effort", "values": [ + "minimal", "low", - "high", - "max" + "medium", + "high" ] + }, + { + "type": "budget_tokens", + "min": 256, + "max": 24000 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": false, - "release_date": "2026-07-16", - "last_updated": "2026-07-16", + "temperature": true, + "knowledge": "2025-01", + "release_date": "2026-05-19", + "last_updated": "2026-05-19", "modalities": { "input": [ "text", - "image" + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 1048576 + "context": 200000, + "input": 128000, + "output": 64000 }, "cost": { - "input": 2.827, - "output": 14.133, - "cache_read": 0.283 + "input": 1.5, + "output": 9, + "cache_read": 0.15, + "input_audio": 1.5 } }, - "qwen2-5-math-7b-instruct": { - "id": "qwen2-5-math-7b-instruct", - "name": "Qwen2.5-Math 7B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "kimi-k2.7-code": { + "id": "kimi-k2.7-code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -325011,44 +195996,45 @@ }, "open_weights": true, "limit": { - "context": 4096, - "output": 3072 + "context": 256000, + "input": 224000, + "output": 32000 }, "cost": { - "input": 0.144, - "output": 0.287 + "input": 0.95, + "output": 4, + "cache_read": 0.19 } }, - "qwen3.5-plus": { - "id": "qwen3.5-plus", - "name": "Qwen3.5 Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "claude-sonnet-5": { + "id": "claude-sonnet-5", + "name": "Claude Sonnet 5", + "description": "Everyday Claude agent model for coding, planning, browsing, and general work", + "family": "claude-sonnet", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-09-22", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-30", + "last_updated": "2026-06-30", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -325057,90 +196043,79 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 65536 + "output": 128000 }, "cost": { - "input": 0.115, - "output": 0.688, - "reasoning": 0.688, - "tiers": [ - { - "input": 0.287, - "output": 1.72, - "reasoning": 1.72, - "tier": { - "type": "context", - "size": 128000 - } - }, - { - "input": 0.573, - "output": 3.44, - "reasoning": 3.44, - "tier": { - "type": "context", - "size": 256000 - } - } - ] + "input": 2, + "output": 10, + "cache_read": 0.2, + "cache_write": 2.5 } }, - "qwen2-5-72b-instruct": { - "id": "qwen2-5-72b-instruct", - "name": "Qwen2.5 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gpt-5.4-nano": { + "id": "gpt-5.4-nano", + "name": "GPT-5.4 nano", + "description": "Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation", + "family": "gpt-nano", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.574, - "output": 1.721 + "input": 0.2, + "output": 1.25, + "cache_read": 0.02 } }, - "glm-5": { - "id": "glm-5", - "name": "GLM-5", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", - "attachment": false, + "claude-opus-4.7": { + "id": "claude-opus-4.7", + "name": "Claude Opus 4.7", + "description": "Stronger Opus tier for advanced software work and high-stakes reasoning", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 32768 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "temperature": true, - "release_date": "2026-02-11", - "last_updated": "2026-02-11", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-04-16", + "last_updated": "2026-04-16", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -325148,88 +196123,98 @@ }, "open_weights": false, "limit": { - "context": 202752, - "output": 16384 + "context": 200000, + "input": 168000, + "output": 32000 }, - "cost": { - "input": 0.573, - "output": 2.58, - "tiers": [ - { - "input": 0.86, - "output": 3.154, - "tier": { - "type": "context", - "size": 32000 + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } } } - ] + } + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "deepseek-v4.1-flash": { - "id": "deepseek-v4.1-flash", - "name": "DeepSeek V4.1 Flash", - "description": "DeepSeek V4.1 Flash model for reasoning and agentic coding", - "family": "deepseek-flash", - "attachment": true, + "mai-code-1-flash-picker": { + "id": "mai-code-1-flash-picker", + "name": "MAI-Code-1-Flash", + "description": "Microsoft coding model built for fast, efficient assistance in everyday developer workflows", + "family": "mai", + "attachment": false, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "effort", "values": [ "low", - "high", - "max" + "medium", + "high" ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-09-10", - "last_updated": "2026-09-10", + "knowledge": "2025-12", + "release_date": "2026-06-02", + "last_updated": "2026-06-08", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 384000 + "context": 256000, + "input": 128000, + "output": 128000 }, "cost": { - "input": 0.29754, - "output": 1.19015, - "cache_read": 0.01488 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "deepseek-r1-distill-llama-8b": { - "id": "deepseek-r1-distill-llama-8b", - "name": "DeepSeek R1 Distill Llama 8B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "gpt-5.2": { + "id": "gpt-5.2", + "name": "GPT-5.2", + "description": "Reliable GPT generation for broad coding, writing, and tool-assisted product work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -325237,108 +196222,91 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0, - "output": 0 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "qwen3-32b": { - "id": "qwen3-32b", - "name": "Qwen3 32B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "gpt-5.3-codex": { + "id": "gpt-5.3-codex", + "name": "GPT-5.3 Codex", + "description": "Coding-optimized GPT model for repository edits, reviews, and agentic software work", + "family": "gpt-codex", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 38912 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 131072, - "output": 16384 - }, - "cost": { - "input": 0.287, - "output": 1.147, - "reasoning": 2.868 - } - }, - "qwen2-5-coder-7b-instruct": { - "id": "qwen2-5-coder-7b-instruct", - "name": "Qwen2.5-Coder 7B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11", - "last_updated": "2024-11", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-02-05", + "last_updated": "2026-02-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.144, - "output": 0.287 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "qwen-plus": { - "id": "qwen-plus", - "name": "Qwen Plus", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "claude-opus-4.8": { + "id": "claude-opus-4.8", + "name": "Claude Opus 4.8", + "description": "Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents", + "family": "claude-opus", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-09-11", + "temperature": false, + "knowledge": "2026-01", + "release_date": "2026-05-28", + "last_updated": "2026-05-28", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -325346,43 +196314,66 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 32768 + "context": 200000, + "input": 168000, + "output": 64000 + }, + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } + } + } + } }, "cost": { - "input": 0.115, - "output": 0.287, - "reasoning": 1.147, - "cache_read": 0.012, - "cache_write": 0.144 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen3.7-flash": { - "id": "qwen3.7-flash", - "name": "Qwen3.7 Flash", - "description": "Lightweight multimodal Qwen model for high-throughput text, image, and video tasks", - "family": "qwen", + "claude-fable-5": { + "id": "claude-fable-5", + "name": "Claude Fable 5", + "description": "Claude model for creative writing, analysis, and controlled agent workflows", + "family": "claude-fable", "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] } ], "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2026-07-15", - "last_updated": "2026-07-15", + "temperature": false, + "knowledge": "2026-01-31", + "release_date": "2026-06-09", + "last_updated": "2026-06-09", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" @@ -325391,111 +196382,87 @@ "open_weights": false, "limit": { "context": 1000000, - "input": 991000, - "output": 65536 + "output": 128000 }, "cost": { - "input": 0.02962, - "output": 0.1185, - "cache_read": 0.002962, - "cache_write": 0.03703, - "tiers": [ - { - "input": 0.08887, - "output": 0.35549, - "cache_read": 0.008887, - "cache_write": 0.11109, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.17774, - "output": 0.71098, - "cache_read": 0.017774, - "cache_write": 0.22218, - "tier": { - "type": "context", - "size": 256000 - } - } - ] + "input": 10, + "output": 50, + "cache_read": 1, + "cache_write": 12.5 } }, - "kimi-k2.6": { - "id": "kimi-k2.6", - "name": "Moonshot Kimi K2.6", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", + "claude-opus-4.5": { + "id": "claude-opus-4.5", + "name": "Claude Opus 4.5 (latest)", + "description": "Flagship Claude model for deep reasoning, coding, and long-horizon agents", + "family": "claude-opus", "attachment": true, "reasoning": true, "reasoning_options": [ - { - "type": "toggle" - }, { "type": "budget_tokens", - "max": 81920 + "min": 1024, + "max": 32000 } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-04-21", - "last_updated": "2026-04-21", + "knowledge": "2025-05", + "release_date": "2025-11-24", + "last_updated": "2025-11-24", "modalities": { "input": [ "text", "image", - "video" + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 200000, + "input": 168000, + "output": 32000 }, "cost": { - "input": 0.929, - "output": 3.858 + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen3-max": { - "id": "qwen3-max", - "name": "Qwen3 Max", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "gpt-5.4": { + "id": "gpt-5.4", + "name": "GPT-5.4", + "description": "Agent-ready GPT for coding and computer-use workflows at a lower cost", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-01-23", - "last_updated": "2026-09-22", + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-05", + "last_updated": "2026-03-05", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -325503,84 +196470,96 @@ }, "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.359, - "output": 1.434, - "reasoning": 1.434, + "input": 2.5, + "output": 15, + "cache_read": 0.25, "tiers": [ { - "input": 0.574, - "output": 2.294, - "reasoning": 2.294, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.004, - "output": 4.014, - "reasoning": 4.014, + "input": 5, + "output": 22.5, + "cache_read": 0.5, "tier": { "type": "context", - "size": 128000 + "size": 272000 } } - ] + ], + "context_over_200k": { + "input": 5, + "output": 22.5, + "cache_read": 0.5 + } } }, - "qwen-omni-turbo-realtime": { - "id": "qwen-omni-turbo-realtime", - "name": "Qwen-Omni Turbo Realtime", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gpt-5.4-mini": { + "id": "gpt-5.4-mini", + "name": "GPT-5.4 mini", + "description": "Strong small GPT for coding subagents, quick tool use, and high-volume work", + "family": "gpt-mini", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] + } + ], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-05-08", - "last_updated": "2025-05-08", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2026-03-17", + "last_updated": "2026-03-17", "modalities": { "input": [ "text", - "image", - "audio" + "image" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 2048 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.23, - "output": 0.918, - "input_audio": 3.584, - "output_audio": 7.168 + "input": 0.75, + "output": 4.5, + "cache_read": 0.075 } }, - "deepseek-v3": { - "id": "deepseek-v3", - "name": "DeepSeek V3", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "gpt-4.1": { + "id": "gpt-4.1", + "name": "GPT-4.1", + "description": "Long-lived GPT workhorse for coding, instruction following, and production apps", + "family": "gpt", + "attachment": true, "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2024-12-01", - "last_updated": "2024-12-01", + "knowledge": "2024-04", + "release_date": "2025-04-14", + "last_updated": "2025-04-14", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -325588,29 +196567,51 @@ }, "open_weights": false, "limit": { - "context": 65536, - "output": 8192 + "context": 128000, + "input": 128000, + "output": 16384 }, "cost": { - "input": 0.287, - "output": 1.147 + "input": 2, + "output": 8, + "cache_read": 0.5 } }, - "deepseek-r1-distill-llama-70b": { - "id": "deepseek-r1-distill-llama-70b", - "name": "DeepSeek R1 Distill Llama 70B", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", - "attachment": false, + "gemini-3.1-pro-preview": { + "id": "gemini-3.1-pro-preview", + "name": "Gemini 3.1 Pro Preview", + "description": "Reasoning-first Gemini preview for agentic coding and complex problem solving", + "family": "gemini-pro", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 256, + "max": 32000 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2025-01", + "release_date": "2026-02-19", + "last_updated": "2026-02-19", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" @@ -325618,86 +196619,65 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 - }, - "cost": { - "input": 0.287, - "output": 0.861 - } - }, - "qwen3-coder-480b-a35b-instruct": { - "id": "qwen3-coder-480b-a35b-instruct", - "name": "Qwen3-Coder 480B-A35B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", - "modalities": { - "input": [ - "text" - ], - "output": [ - "text" - ] - }, - "open_weights": true, - "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "input": 136000, + "output": 64000 }, "cost": { - "input": 0.861, - "output": 3.441, + "input": 2, + "output": 12, + "cache_read": 0.2, "tiers": [ { - "input": 1.291, - "output": 5.161, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 2.151, - "output": 8.602, + "input": 4, + "output": 18, + "cache_read": 0.4, "tier": { "type": "context", - "size": 128000 + "size": 200000 } } - ] + ], + "context_over_200k": { + "input": 4, + "output": 18, + "cache_read": 0.4 + } } }, - "qwen-turbo": { - "id": "qwen-turbo", - "name": "Qwen Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, + "claude-sonnet-4.6": { + "id": "claude-sonnet-4.6", + "name": "Claude Sonnet 4.6", + "description": "Claude workhorse for coding agents, careful analysis, and production cost control", + "family": "claude-sonnet", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] }, { "type": "budget_tokens", - "max": 38912 + "min": 1024, + "max": 32000 } ], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-11-01", - "last_updated": "2025-07-15", + "knowledge": "2025-08-31", + "release_date": "2026-02-17", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -325705,30 +196685,44 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 16384 + "context": 200000, + "input": 168000, + "output": 32000 }, "cost": { - "input": 0.044, - "output": 0.087, - "reasoning": 0.431 + "input": 3, + "output": 15, + "cache_read": 0.3, + "cache_write": 3.75 } }, - "deepseek-r1-distill-qwen-7b": { - "id": "deepseek-r1-distill-qwen-7b", - "name": "DeepSeek R1 Distill Qwen 7B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "gpt-5-mini": { + "id": "gpt-5-mini", + "name": "GPT-5 Mini", + "description": "Small GPT-5 for responsive agents, coding help, and everyday automation", + "family": "gpt-mini", + "attachment": true, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "structured_output": true, + "temperature": false, + "knowledge": "2024-05-30", + "release_date": "2025-08-07", + "last_updated": "2025-08-07", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -325736,107 +196730,154 @@ }, "open_weights": false, "limit": { - "context": 32768, - "output": 16384 + "context": 264000, + "input": 128000, + "output": 64000 }, "cost": { - "input": 0.072, - "output": 0.144 + "input": 0.25, + "output": 2, + "cache_read": 0.025 } }, - "qwen2-5-math-72b-instruct": { - "id": "qwen2-5-math-72b-instruct", - "name": "Qwen2.5-Math 72B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gemini-3-flash-preview": { + "id": "gemini-3-flash-preview", + "name": "Gemini 3 Flash Preview", + "description": "New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs", + "family": "gemini-flash", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + }, + { + "type": "budget_tokens", + "min": 256, + "max": 32000 + } + ], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "knowledge": "2025-01", + "release_date": "2025-12-17", + "last_updated": "2025-12-17", "modalities": { "input": [ - "text" + "text", + "image", + "video", + "audio", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 4096, - "output": 3072 + "context": 128000, + "input": 128000, + "output": 64000 }, "cost": { - "input": 0.574, - "output": 1.721 + "input": 0.5, + "output": 3, + "cache_read": 0.05, + "input_audio": 1 } }, - "qwen3-coder-30b-a3b-instruct": { - "id": "qwen3-coder-30b-a3b-instruct", - "name": "Qwen3-Coder 30B-A3B Instruct", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", - "attachment": false, - "reasoning": false, + "claude-opus-4.6": { + "id": "claude-opus-4.6", + "name": "Claude Opus 4.6", + "description": "High-end Claude for difficult coding, planning, and slower expert reasoning", + "family": "claude-opus", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "max" + ] + } + ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "knowledge": "2025-05-31", + "release_date": "2026-02-05", + "last_updated": "2026-03-13", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 65536 + "context": 200000, + "input": 168000, + "output": 32000 }, - "cost": { - "input": 0.216, - "output": 0.861, - "tiers": [ - { - "input": 0.323, - "output": 1.291, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 0.538, - "output": 2.151, - "tier": { - "type": "context", - "size": 128000 + "experimental": { + "modes": { + "fast": { + "cost": { + "input": 30, + "output": 150, + "cache_read": 3, + "cache_write": 37.5 + }, + "provider": { + "body": { + "speed": "fast" + }, + "headers": { + "anthropic-beta": "fast-mode-2026-02-01" + } } } - ] + } + }, + "cost": { + "input": 5, + "output": 25, + "cache_read": 0.5, + "cache_write": 6.25 } }, - "qwen-doc-turbo": { - "id": "qwen-doc-turbo", - "name": "Qwen Doc Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", - "attachment": false, - "reasoning": false, + "gpt-5.2-codex": { + "id": "gpt-5.2-codex", + "name": "GPT-5.2 Codex", + "description": "Code-specialist GPT for repository edits, reviews, and long-running software agents", + "family": "gpt-codex", + "attachment": true, + "reasoning": true, + "reasoning_options": [], "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01", - "last_updated": "2024-01", + "structured_output": true, + "temperature": false, + "knowledge": "2025-08-31", + "release_date": "2025-12-11", + "last_updated": "2025-12-11", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" @@ -325844,65 +196885,108 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.087, - "output": 0.144 + "input": 1.75, + "output": 14, + "cache_read": 0.175 } }, - "kimi-k2-thinking": { - "id": "kimi-k2-thinking", - "name": "Moonshot Kimi K2 Thinking", - "description": "Kimi reasoning model for long-horizon research, planning, and tool use", - "family": "kimi-thinking", - "attachment": false, + "gpt-5.5": { + "id": "gpt-5.5", + "name": "GPT-5.5", + "description": "Default frontier GPT for coding, computer use, research, and knowledge work", + "family": "gpt", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "budget_tokens" + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high", + "xhigh" + ] } ], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, "structured_output": true, - "temperature": true, - "release_date": "2025-11-06", - "last_updated": "2025-11-06", + "temperature": false, + "knowledge": "2025-12-01", + "release_date": "2026-04-23", + "last_updated": "2026-04-23", "modalities": { "input": [ - "text" + "text", + "image", + "pdf" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 16384 + "context": 400000, + "input": 272000, + "output": 128000 }, "cost": { - "input": 0.574, - "output": 2.294 + "input": 5, + "output": 30, + "cache_read": 0.5, + "tiers": [ + { + "input": 10, + "output": 45, + "cache_read": 1, + "tier": { + "type": "context", + "size": 272000 + } + } + ], + "context_over_200k": { + "input": 10, + "output": 45, + "cache_read": 1 + } } - }, - "qvq-max": { - "id": "qvq-max", - "name": "QVQ Max", - "description": "Multimodal reasoning model for visual analysis, planning, and tool use", - "family": "qvq", - "attachment": false, + } + } + }, + "clarifai": { + "id": "clarifai", + "env": [ + "CLARIFAI_PAT" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.clarifai.com/v2/ext/openai/v1", + "name": "Clarifai", + "doc": "https://docs.clarifai.com/compute/inference/", + "models": { + "moonshotai/chat-completion/models/Kimi-K2_6": { + "id": "moonshotai/chat-completion/models/Kimi-K2_6", + "name": "Kimi K2.6", + "description": "Multimodal Kimi workhorse for agent loops, coding tasks, and visual context", + "family": "kimi-k2", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-03-25", - "last_updated": "2025-03-25", + "knowledge": "2025-01", + "release_date": "2026-04-21", + "last_updated": "2026-04-21", "modalities": { "input": [ "text", @@ -325912,28 +196996,28 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 262144, + "output": 262144 }, "cost": { - "input": 1.147, - "output": 4.588 + "input": 0.95, + "output": 4 } }, - "qwen2-5-7b-instruct": { - "id": "qwen2-5-7b-instruct", - "name": "Qwen2.5 7B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput": { + "id": "minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput", + "name": "MiniMax-M2.5 High Throughput", + "description": "MiniMax model for chat, coding, office work, and agentic tasks", + "family": "minimax", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "release_date": "2026-02-12", + "last_updated": "2026-02-25", "modalities": { "input": [ "text" @@ -325944,107 +197028,74 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 204800, + "output": 131072 }, "cost": { - "input": 0.072, - "output": 0.144 + "input": 0.3, + "output": 1.2 } }, - "qwen3-omni-flash": { - "id": "qwen3-omni-flash", - "name": "Qwen3-Omni Flash", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", + "openai/chat-completion/models/gpt-oss-120b-high-throughput": { + "id": "openai/chat-completion/models/gpt-oss-120b-high-throughput", + "name": "GPT OSS 120B High Throughput", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", - "modalities": { - "input": [ - "text", - "image", - "audio", - "video" - ], - "output": [ - "text", - "audio" - ] - }, - "open_weights": false, - "limit": { - "context": 65536, - "output": 16384 - }, - "cost": { - "input": 0.058, - "output": 0.23, - "input_audio": 3.584, - "output_audio": 7.168 - } - }, - "qwen-vl-max": { - "id": "qwen-vl-max", - "name": "Qwen-VL Max", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, - "reasoning": false, - "tool_call": true, - "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-04-08", - "last_updated": "2025-08-13", + "release_date": "2025-08-05", + "last_updated": "2026-02-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { "context": 131072, - "output": 8192 + "output": 16384 }, "cost": { - "input": 0.23, - "output": 0.574 + "input": 0.09, + "output": 0.36 } }, - "qwen3-235b-a22b": { - "id": "qwen3-235b-a22b", - "name": "Qwen3 235B-A22B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "openai/chat-completion/models/gpt-oss-20b": { + "id": "openai/chat-completion/models/gpt-oss-20b", + "name": "GPT OSS 20B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 38912 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-08-05", + "last_updated": "2025-12-12", "modalities": { "input": [ "text" @@ -326059,24 +197110,23 @@ "output": 16384 }, "cost": { - "input": 0.287, - "output": 1.147, - "reasoning": 2.868 + "input": 0.045, + "output": 0.18 } }, - "qwen3-vl-30b-a3b": { - "id": "qwen3-vl-30b-a3b", - "name": "Qwen3-VL 30B-A3B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": false, + "mistralai/completion/models/Ministral-3-14B-Reasoning-2512": { + "id": "mistralai/completion/models/Ministral-3-14B-Reasoning-2512", + "name": "Ministral 3 14B Reasoning 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "knowledge": "2025-12", + "release_date": "2025-12-01", + "last_updated": "2025-12-12", "modalities": { "input": [ "text", @@ -326088,45 +197138,30 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0.108, - "output": 0.431, - "reasoning": 1.076 + "input": 2.5, + "output": 1.7 } }, - "kimi-k2.5": { - "id": "kimi-k2.5", - "name": "Moonshot Kimi K2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": false, + "mistralai/completion/models/Ministral-3-3B-Reasoning-2512": { + "id": "mistralai/completion/models/Ministral-3-3B-Reasoning-2512", + "name": "Ministral 3 3B Reasoning 2512", + "description": "Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads", + "family": "ministral", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": false, "temperature": true, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "release_date": "2025-12", + "last_updated": "2026-02-25", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -326135,28 +197170,28 @@ "open_weights": true, "limit": { "context": 262144, - "output": 32768 + "output": 262144 }, "cost": { - "input": 0.574, - "output": 2.411 + "input": 1.039, + "output": 0.54825 } }, - "qwen3-next-80b-a3b-instruct": { - "id": "qwen3-next-80b-a3b-instruct", - "name": "Qwen3-Next 80B-A3B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", - "attachment": false, + "deepseek-ai/deepseek-ocr/models/DeepSeek-OCR": { + "id": "deepseek-ai/deepseek-ocr/models/DeepSeek-OCR", + "name": "DeepSeek OCR", + "description": "OCR model for extracting structured text from documents and screenshots", + "family": "deepseek", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-09", - "last_updated": "2025-09", + "release_date": "2025-10-20", + "last_updated": "2026-02-25", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -326164,57 +197199,57 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 32768 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.144, - "output": 0.574 + "input": 0.2, + "output": 0.7 } }, - "qwen-vl-plus": { - "id": "qwen-vl-plus", - "name": "Qwen-VL Plus", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", + "qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507": { + "id": "qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507", + "name": "Qwen3 30B A3B Thinking 2507", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", "family": "qwen", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01-25", - "last_updated": "2025-08-15", + "release_date": "2025-07-31", + "last_updated": "2026-02-25", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 262144, + "output": 131072 }, "cost": { - "input": 0.115, - "output": 0.287 + "input": 0.36, + "output": 1.3 } }, - "deepseek-r1-distill-qwen-1-5b": { - "id": "deepseek-r1-distill-qwen-1-5b", - "name": "DeepSeek R1 Distill Qwen 1.5B", + "qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507": { + "id": "qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507", + "name": "Qwen3 30B A3B Instruct 2507", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, + "structured_output": true, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2025-07-30", + "last_updated": "2026-02-25", "modalities": { "input": [ "text" @@ -326223,70 +197258,59 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 16384 + "context": 262144, + "output": 262144 }, "cost": { - "input": 0, - "output": 0 + "input": 0.3, + "output": 0.5 } }, - "qwen3.6-flash": { - "id": "qwen3.6-flash", - "name": "Qwen3.6 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen3.6", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 131072 - } - ], + "qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct": { + "id": "qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct", + "name": "Qwen3 Coder 30B A3B Instruct", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, "tool_call": true, - "structured_output": true, "temperature": true, - "release_date": "2026-04-27", - "last_updated": "2026-04-27", + "knowledge": "2025-04", + "release_date": "2025-07-31", + "last_updated": "2026-02-12", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 262144, "output": 65536 }, "cost": { - "input": 0.1875, - "output": 1.125, - "cache_write": 0.234375 + "input": 0.11458, + "output": 0.74812 } }, - "qwen3-coder-flash": { - "id": "qwen3-coder-flash", - "name": "Qwen3 Coder Flash", - "description": "Qwen coding model for software agents, repository edits, and code reasoning", - "family": "qwen", + "arcee_ai/AFM/models/trinity-mini": { + "id": "arcee_ai/AFM/models/trinity-mini", + "name": "Trinity Mini", + "description": "Efficient model for low-latency assistance, extraction, and routine automation", + "family": "trinity-mini", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [], "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-28", - "last_updated": "2025-07-28", + "knowledge": "2024-10", + "release_date": "2025-12", + "last_updated": "2026-02-25", "modalities": { "input": [ "text" @@ -326295,63 +197319,83 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, - "output": 65536 + "context": 131072, + "output": 131072 }, "cost": { - "input": 0.144, - "output": 0.574 + "input": 0.045, + "output": 0.15 } }, - "qwen2-5-omni-7b": { - "id": "qwen2-5-omni-7b", - "name": "Qwen2.5-Omni 7B", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", - "attachment": false, + "clarifai/main/models/mm-poly-8b": { + "id": "clarifai/main/models/mm-poly-8b", + "name": "MM Poly 8B", + "description": "Multimodal model for analyzing text, images, documents, and rich media", + "family": "mm-poly", + "attachment": true, "reasoning": false, - "tool_call": true, + "tool_call": false, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-12", - "last_updated": "2024-12", + "release_date": "2025-06", + "last_updated": "2026-02-25", "modalities": { "input": [ "text", "image", - "audio", "video" ], "output": [ - "text", - "audio" + "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { "context": 32768, - "output": 2048 + "output": 4096 }, "cost": { - "input": 0.087, - "output": 0.345, - "input_audio": 5.448 + "input": 0.658, + "output": 1.11 } - }, - "qwen-mt-plus": { - "id": "qwen-mt-plus", - "name": "Qwen-MT Plus", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "family": "qwen", + } + } + }, + "the-grid-ai": { + "id": "the-grid-ai", + "env": [ + "THEGRIDAI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.thegrid.ai/v1", + "name": "The Grid AI", + "doc": "https://thegrid.ai/docs", + "models": { + "agent-prime": { + "id": "agent-prime", + "name": "Agent Prime", + "description": "Reliable models for dependable agentic applications, multi-step tool use, and reasoning workflows. Any model that meets the contract spec can serve your request.", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01", - "last_updated": "2025-01", + "release_date": "2026-05-04", + "last_updated": "2026-07-06", "modalities": { "input": [ "text" @@ -326362,29 +197406,45 @@ }, "open_weights": false, "limit": { - "context": 16384, - "output": 8192 + "context": 196608, + "input": 120000, + "output": 30000 }, - "cost": { - "input": 0.259, - "output": 0.775 - } + "status": "beta" }, - "tongyi-intent-detect-v3": { - "id": "tongyi-intent-detect-v3", - "name": "Tongyi Intent Detect V3", - "description": "General-purpose chat model for instruction following, writing, and analysis", - "family": "yi", + "agent-max": { + "id": "agent-max", + "name": "Agent Max", + "description": "Frontier models for autonomous research, deep multi-step tool chains, and complex long-horizon tasks. Any model that meets the contract spec can serve your request.", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-01", - "last_updated": "2024-01", + "release_date": "2026-05-04", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" @@ -326392,59 +197452,75 @@ }, "open_weights": false, "limit": { - "context": 8192, - "output": 1024 + "context": 1000000, + "input": 922000, + "output": 128000 }, - "cost": { - "input": 0.058, - "output": 0.144 - } + "status": "beta" }, - "qwen3-vl-235b-a22b": { - "id": "qwen3-vl-235b-a22b", - "name": "Qwen3-VL 235B-A22B", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "text-standard": { + "id": "text-standard", + "name": "Text Standard", + "description": "Price-optimized models with low-latency, high-throughput and shorter maximum outputs. Any model that meets the contract spec can serve your request.", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2026-02-26", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 32768 - }, - "cost": { - "input": 0.286705, - "output": 1.14682, - "reasoning": 2.867051 + "context": 128000, + "input": 120000, + "output": 16000 } }, - "qwen3-coder-plus": { - "id": "qwen3-coder-plus", - "name": "Qwen3 Coder Plus", - "description": "Hosted Qwen coder for software agents, repo edits, and long-context code", - "family": "qwen", + "code-prime": { + "id": "code-prime", + "name": "Code Prime", + "description": "Reliable models for everyday software tasks, code completion, review, and standard debugging. Any model that meets the contract spec can serve your request.", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2025-07-23", - "last_updated": "2026-09-11", + "release_date": "2026-05-04", + "last_updated": "2026-07-06", "modalities": { "input": [ "text" @@ -326453,54 +197529,38 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1048576, - "output": 65536 + "context": 196608, + "input": 120000, + "output": 30000 }, - "cost": { - "input": 0.574, - "output": 2.296, - "tiers": [ - { - "input": 0.861, - "output": 3.444, - "tier": { - "type": "context", - "size": 32000 - } - }, - { - "input": 1.435, - "output": 5.74, - "tier": { - "type": "context", - "size": 128000 - } - }, - { - "input": 2.87, - "output": 28.7, - "tier": { - "type": "context", - "size": 256000 - } - } - ] - } + "status": "beta" }, - "moonshot-kimi-k2-instruct": { - "id": "moonshot-kimi-k2-instruct", - "name": "Moonshot Kimi K2 Instruct", - "description": "Kimi model for long-context chat, coding, and agentic reasoning", - "family": "kimi-k2", + "text-prime": { + "id": "text-prime", + "name": "Text Prime", + "description": "Reliable models for everyday text generation, editing, and analysis across diverse workflows. Any model that meets the contract spec can serve your request.", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, - "structured_output": false, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "release_date": "2026-02-26", + "last_updated": "2026-07-06", "modalities": { "input": [ "text" @@ -326509,61 +197569,83 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 - }, - "cost": { - "input": 0.574, - "output": 2.294 + "context": 196608, + "input": 120000, + "output": 30000 } }, - "MiniMax-M2.5": { - "id": "MiniMax-M2.5", - "name": "MiniMax-M2.5", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", + "code-max": { + "id": "code-max", + "name": "Code Max", + "description": "Frontier models for complex research, architectural decisions, debugging, and multi-file development. Any model that meets the contract spec can serve your request.", "attachment": false, "reasoning": true, - "reasoning_options": [], + "reasoning_options": [ + { + "type": "toggle" + }, + { + "type": "effort", + "values": [ + "low", + "medium", + "high", + "xhigh", + "max" + ] + } + ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, + "structured_output": true, "temperature": true, - "release_date": "2026-02-12", - "last_updated": "2026-02-12", + "release_date": "2026-05-04", + "last_updated": "2026-07-06", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 1000000, + "input": 922000, + "output": 128000 }, - "cost": { - "input": 0.3, - "output": 1.2 - } + "status": "beta" }, - "qwen2-5-14b-instruct": { - "id": "qwen2-5-14b-instruct", - "name": "Qwen2.5 14B Instruct", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "agent-standard": { + "id": "agent-standard", + "name": "Agent Standard", + "description": "Price-optimized models for fast tool calls, simple agent loops, high-throughput automation, and orchestration. Any model that meets the contract spec can serve your request.", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "release_date": "2026-05-04", + "last_updated": "2026-07-06", "modalities": { "input": [ "text" @@ -326572,22 +197654,19 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 131072, - "output": 8192 + "context": 128000, + "input": 120000, + "output": 16000 }, - "cost": { - "input": 0.144, - "output": 0.431 - } + "status": "beta" }, - "qwen3.8-flash": { - "id": "qwen3.8-flash", - "name": "Qwen3.8 Flash", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", - "attachment": true, + "text-max": { + "id": "text-max", + "name": "Text Max", + "description": "Frontier models for deep reasoning, long context, and complex workflows. Any model that meets the contract spec can serve your request.", + "attachment": false, "reasoning": true, "reasoning_options": [ { @@ -326598,12 +197677,10 @@ "values": [ "low", "medium", - "xhigh" + "high", + "xhigh", + "max" ] - }, - { - "type": "budget_tokens", - "max": 262144 } ], "tool_call": true, @@ -326611,13 +197688,13 @@ "field": "reasoning_content" }, "structured_output": true, - "release_date": "2026-08-26", - "last_updated": "2026-08-26", + "temperature": true, + "release_date": "2026-03-24", + "last_updated": "2026-07-06", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -326626,27 +197703,88 @@ "open_weights": false, "limit": { "context": 1000000, - "output": 131072 - }, - "cost": { - "input": 0.11875, - "output": 0.40073, - "cache_read": 0.01187, - "cache_write": 0.14844 + "input": 922000, + "output": 128000 } }, - "qwen2-5-vl-7b-instruct": { - "id": "qwen2-5-vl-7b-instruct", - "name": "Qwen2.5-VL 7B Instruct", - "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", - "family": "qwen", + "code-standard": { + "id": "code-standard", + "name": "Code Standard", + "description": "Price-optimized models for rapid autocomplete, linting, high-frequency suggestions, and batch edits. Any model that meets the contract spec can serve your request.", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09", - "last_updated": "2024-09", + "release_date": "2026-05-04", + "last_updated": "2026-07-06", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": false, + "limit": { + "context": 128000, + "input": 120000, + "output": 16000 + }, + "status": "beta" + } + } + }, + "synthetic": { + "id": "synthetic", + "env": [ + "SYNTHETIC_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.synthetic.new/openai/v1", + "name": "Synthetic", + "doc": "https://synthetic.new/pricing", + "models": { + "hf:moonshotai/Kimi-K2.7-Code": { + "id": "hf:moonshotai/Kimi-K2.7-Code", + "name": "Kimi K2.7 Code", + "description": "Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking", + "family": "kimi-k2", + "attachment": true, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, + "temperature": false, + "knowledge": "2025-01", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ "text", @@ -326658,26 +197796,35 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 262144, + "output": 65536 }, "cost": { - "input": 0.287, - "output": 0.717 + "input": 0.95, + "output": 4, + "cache_read": 0.95 } }, - "qwen-math-turbo": { - "id": "qwen-math-turbo", - "name": "Qwen Math Turbo", - "description": "Efficient Qwen model for fast chat, extraction, and high-volume workloads", - "family": "qwen", + "hf:zai-org/GLM-4.7-Flash": { + "id": "hf:zai-org/GLM-4.7-Flash", + "name": "GLM-4.7-Flash", + "description": "Budget GLM lane for fast coding help, routing, and everyday automation", + "family": "glm-flash", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-09-19", - "last_updated": "2024-09-19", + "knowledge": "2025-04", + "release_date": "2026-01-19", + "last_updated": "2026-01-19", "modalities": { "input": [ "text" @@ -326686,28 +197833,42 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 4096, - "output": 3072 + "context": 196608, + "output": 65536 }, "cost": { - "input": 0.287, - "output": 0.861 + "input": 0.1, + "output": 0.5, + "cache_read": 0.1 } }, - "qwen-long": { - "id": "qwen-long", - "name": "Qwen Long", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "hf:zai-org/GLM-5.2": { + "id": "hf:zai-org/GLM-5.2", + "name": "GLM-5.2", + "description": "Open flagship GLM for long-horizon coding agents and million-token context work", + "family": "glm", "attachment": false, - "reasoning": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "none", + "high", + "xhigh" + ] + } + ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01-25", - "last_updated": "2025-01-25", + "release_date": "2026-06-13", + "last_updated": "2026-06-13", "modalities": { "input": [ "text" @@ -326716,68 +197877,83 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 10000000, - "output": 8192 + "context": 524288, + "output": 65536 }, "cost": { - "input": 0.072, - "output": 0.287 + "input": 1.4, + "output": 4.4, + "cache_read": 1.4 } }, - "qwen3.6-max-preview": { - "id": "qwen3.6-max-preview", - "name": "Qwen3.6 Max Preview", - "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", - "family": "qwen", - "attachment": false, + "hf:MiniMaxAI/MiniMax-M3": { + "id": "hf:MiniMaxAI/MiniMax-M3", + "name": "MiniMax-M3", + "description": "MiniMax multimodal model for long-context coding, perception, and agent planning", + "family": "minimax", + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 131072 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "structured_output": true, "temperature": true, - "release_date": "2026-04-20", - "last_updated": "2026-04-21", + "release_date": "2026-06-12", + "last_updated": "2026-06-12", "modalities": { "input": [ - "text" + "text", + "image" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 245800, + "context": 524288, "output": 65536 }, "cost": { - "input": 1.32, - "output": 7.9, - "cache_read": 0.132 + "input": 0.6, + "output": 1.2, + "cache_read": 0.6 } }, - "qwen-mt-turbo": { - "id": "qwen-mt-turbo", - "name": "Qwen-MT Turbo", - "description": "Translation model for multilingual conversion, localization, and cross-language workflows", - "family": "qwen", + "hf:openai/gpt-oss-120b": { + "id": "hf:openai/gpt-oss-120b", + "name": "GPT OSS 120B", + "description": "Open-weight GPT model for self-hosted reasoning and instruction-following workloads", + "family": "gpt-oss", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] + } + ], + "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01", - "last_updated": "2025-01", + "release_date": "2025-08-05", + "last_updated": "2025-08-05", "modalities": { "input": [ "text" @@ -326786,42 +197962,47 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 16384, - "output": 8192 + "context": 131072, + "output": 32768 }, "cost": { - "input": 0.101, - "output": 0.28 + "input": 0.1, + "output": 0.1, + "cache_read": 0.1 } }, - "qwen3.5-397b-a17b": { - "id": "qwen3.5-397b-a17b", - "name": "Qwen3.5 397B-A17B", + "hf:Qwen/Qwen3.6-27B": { + "id": "hf:Qwen/Qwen3.6-27B", + "name": "Qwen3.6 27B", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, + "attachment": true, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 + "type": "effort", + "values": [ + "none", + "low", + "medium", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "structured_output": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-02-16", - "last_updated": "2026-02-16", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -326833,43 +198014,78 @@ "output": 65536 }, "cost": { - "input": 0.172, - "output": 1.032, - "reasoning": 1.032, - "tiers": [ - { - "input": 0.43, - "output": 2.58, - "reasoning": 2.58, - "tier": { - "type": "context", - "size": 128000 - } - } - ] + "input": 0.45, + "output": 3.6, + "cache_read": 0.45 } }, - "qwen3-8b": { - "id": "qwen3-8b", - "name": "Qwen3 8B", - "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", - "family": "qwen", + "hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4": { + "id": "hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4", + "name": "Nemotron 3 Super 120B A12B", + "description": "Nemotron middle tier for collaborative agents and high-volume reasoning workloads", + "family": "nemotron", "attachment": false, "reasoning": true, "reasoning_options": [ { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 38912 + "type": "effort", + "values": [ + "low", + "medium", + "high" + ] } ], "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, + "temperature": true, + "release_date": "2026-03-11", + "last_updated": "2026-03-11", + "modalities": { + "input": [ + "text" + ], + "output": [ + "text" + ] + }, + "open_weights": true, + "limit": { + "context": 262144, + "output": 65536 + }, + "cost": { + "input": 0.3, + "output": 1, + "cache_read": 0.3 + } + } + } + }, + "iflowcn": { + "id": "iflowcn", + "env": [ + "IFLOW_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://apis.iflow.cn/v1", + "name": "iFlow", + "doc": "https://platform.iflow.cn/en/docs", + "models": { + "qwen3-coder-plus": { + "id": "qwen3-coder-plus", + "name": "Qwen3-Coder-Plus", + "description": "Qwen coding model for software agents, repository edits, and code reasoning", + "family": "qwen", + "attachment": false, + "reasoning": false, + "tool_call": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ "text" @@ -326880,31 +198096,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.072, - "output": 0.287, - "reasoning": 0.717 + "input": 0, + "output": 0 } }, - "glm-5.2": { - "id": "glm-5.2", - "name": "GLM-5.2", - "description": "Open flagship GLM for long-horizon coding agents and million-token context work", - "family": "glm", + "deepseek-v3": { + "id": "deepseek-v3", + "name": "DeepSeek-V3", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", "attachment": false, - "reasoning": true, - "reasoning_options": [], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-06-13", - "last_updated": "2026-06-13", + "knowledge": "2024-10", + "release_date": "2024-12-26", + "last_updated": "2024-12-26", "modalities": { "input": [ "text" @@ -326915,76 +198126,56 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 128000 + "context": 128000, + "output": 32000 }, "cost": { - "input": 1.1, - "output": 3.851, - "cache_read": 0.275, - "cache_write": 0 + "input": 0, + "output": 0 } }, - "qwen-omni-turbo": { - "id": "qwen-omni-turbo", - "name": "Qwen-Omni Turbo", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", - "family": "qwen", + "kimi-k2": { + "id": "kimi-k2", + "name": "Kimi-K2", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-01-19", - "last_updated": "2025-03-26", + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ - "text", - "image", - "audio", - "video" + "text" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 32768, - "output": 2048 + "context": 128000, + "output": 64000 }, "cost": { - "input": 0.058, - "output": 0.23, - "input_audio": 3.584, - "output_audio": 7.168 + "input": 0, + "output": 0 } }, - "glm-5.1": { - "id": "glm-5.1", - "name": "GLM-5.1", - "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", - "family": "glm", + "qwen3-32b": { + "id": "qwen3-32b", + "name": "Qwen3-32B", + "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", + "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 131072 - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-04-14", - "last_updated": "2026-04-14", + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ "text" @@ -326995,75 +198186,59 @@ }, "open_weights": true, "limit": { - "context": 202752, - "output": 128000 + "context": 128000, + "output": 32000 }, "cost": { - "input": 0.825, - "output": 3.301, - "cache_read": 0.17, - "tiers": [ - { - "input": 1.1, - "output": 3.851, - "tier": { - "type": "context", - "size": 32000 - } - } - ] + "input": 0, + "output": 0 } }, - "qwen3-omni-flash-realtime": { - "id": "qwen3-omni-flash-realtime", - "name": "Qwen3-Omni Flash Realtime", - "description": "Qwen omni model for text, vision, audio, and multimodal agent tasks", + "qwen3-max-preview": { + "id": "qwen3-max-preview", + "name": "Qwen3-Max-Preview", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", "family": "qwen", "attachment": false, "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2024-04", - "release_date": "2025-09-15", - "last_updated": "2025-09-15", + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image", - "audio" + "text" ], "output": [ - "text", - "audio" + "text" ] }, "open_weights": false, "limit": { - "context": 65536, - "output": 16384 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.23, - "output": 0.918, - "input_audio": 3.584, - "output_audio": 7.168 + "input": 0, + "output": 0 } }, - "qwen3-asr-flash": { - "id": "qwen3-asr-flash", - "name": "Qwen3-ASR Flash", - "description": "Speech transcription model for accurate audio-to-text and captioning workflows", + "qwen3-max": { + "id": "qwen3-max", + "name": "Qwen3-Max", + "description": "Flagship Qwen model for complex reasoning, coding, and agentic workflows", "family": "qwen", "attachment": false, "reasoning": false, - "tool_call": false, - "temperature": false, - "knowledge": "2024-04", - "release_date": "2025-09-08", - "last_updated": "2025-09-08", + "tool_call": true, + "temperature": true, + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "audio" + "text" ], "output": [ "text" @@ -327071,17 +198246,17 @@ }, "open_weights": false, "limit": { - "context": 53248, - "output": 4096 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.032, - "output": 0.032 + "input": 0, + "output": 0 } }, - "deepseek-r1-distill-qwen-32b": { - "id": "deepseek-r1-distill-qwen-32b", - "name": "DeepSeek R1 Distill Qwen 32B", + "qwen3-235b": { + "id": "qwen3-235b", + "name": "Qwen3-235B-A22B", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, @@ -327089,8 +198264,9 @@ "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2024-12-01", "modalities": { "input": [ "text" @@ -327099,28 +198275,29 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 32768, - "output": 16384 + "context": 128000, + "output": 32000 }, "cost": { - "input": 0.287, - "output": 0.861 + "input": 0, + "output": 0 } }, - "deepseek-r1": { - "id": "deepseek-r1", - "name": "DeepSeek R1", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "glm-4.6": { + "id": "glm-4.6", + "name": "GLM-4.6", + "description": "Flagship GLM model for hybrid reasoning, coding, and agentic engineering", + "family": "glm", "attachment": false, "reasoning": true, "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-01-01", - "last_updated": "2025-01-01", + "knowledge": "2024-10", + "release_date": "2024-12-01", + "last_updated": "2025-11-13", "modalities": { "input": [ "text" @@ -327131,42 +198308,27 @@ }, "open_weights": false, "limit": { - "context": 131072, - "output": 16384 + "context": 200000, + "output": 128000 }, "cost": { - "input": 0.574, - "output": 2.294 + "input": 0, + "output": 0 } }, - "deepseek-v4-pro": { - "id": "deepseek-v4-pro", - "name": "DeepSeek V4 Pro", - "description": "Open MoE flagship with million-token context for coding and long agent runs", - "family": "deepseek-thinking", + "qwen3-235b-a22b-thinking-2507": { + "id": "qwen3-235b-a22b-thinking-2507", + "name": "Qwen3-235B-A22B-Thinking", + "description": "Qwen reasoning model for deliberate problem solving, math, and coding", + "family": "qwen", "attachment": false, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2025-04", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ "text" @@ -327177,18 +198339,17 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.435, - "output": 0.87, - "cache_read": 0.003625 + "input": 0, + "output": 0 } }, - "deepseek-r1-0528": { - "id": "deepseek-r1-0528", - "name": "DeepSeek R1 0528", + "deepseek-r1": { + "id": "deepseek-r1", + "name": "DeepSeek-R1", "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", "family": "deepseek-thinking", "attachment": false, @@ -327196,8 +198357,9 @@ "reasoning_options": [], "tool_call": true, "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-05-28", + "knowledge": "2024-12", + "release_date": "2025-01-20", + "last_updated": "2025-01-20", "modalities": { "input": [ "text" @@ -327206,42 +198368,32 @@ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 131072, - "output": 16384 + "context": 128000, + "output": 32000 }, "cost": { - "input": 0.574, - "output": 2.294 + "input": 0, + "output": 0 } }, - "qwen3.6-plus": { - "id": "qwen3.6-plus", - "name": "Qwen3.6 Plus", + "qwen3-vl-plus": { + "id": "qwen3-vl-plus", + "name": "Qwen3-VL-Plus", "description": "Qwen vision-language model for visual reasoning, documents, and agent tasks", "family": "qwen", - "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 81920 - } - ], + "attachment": true, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-04-02", - "last_updated": "2026-04-02", + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ "text", - "image", - "video" + "image" ], "output": [ "text" @@ -327249,55 +198401,26 @@ }, "open_weights": false, "limit": { - "context": 1000000, - "output": 65536 + "context": 256000, + "output": 32000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 256000 - } - } - ], - "context_over_200k": { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5 - } + "input": 0, + "output": 0 } }, - "qwen3-14b": { - "id": "qwen3-14b", - "name": "Qwen3 14B", + "qwen3-235b-a22b-instruct": { + "id": "qwen3-235b-a22b-instruct", + "name": "Qwen3-235B-A22B-Instruct", "description": "Qwen instruction model for multilingual chat, reasoning, and tool use", "family": "qwen", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 38912 - } - ], + "reasoning": false, "tool_call": true, "temperature": true, "knowledge": "2025-04", - "release_date": "2025-04", - "last_updated": "2025-04", + "release_date": "2025-07-01", + "last_updated": "2025-07-01", "modalities": { "input": [ "text" @@ -327308,40 +198431,26 @@ }, "open_weights": true, "limit": { - "context": 131072, - "output": 8192 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.144, - "output": 0.574, - "reasoning": 1.434 + "input": 0, + "output": 0 } }, - "glm-5.3": { - "id": "glm-5.3", - "name": "GLM-5.3", - "description": "Flagship GLM model for long-horizon coding, agents, and complex project delivery", - "family": "glm", + "kimi-k2-0905": { + "id": "kimi-k2-0905", + "name": "Kimi-K2-0905", + "description": "Kimi model for long-context chat, coding, and agentic reasoning", + "family": "kimi-k2", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "effort", - "values": [ - "low", - "high", - "max" - ] - } - ], + "reasoning": false, "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, "temperature": true, - "release_date": "2026-08-14", - "last_updated": "2026-08-14", + "knowledge": "2024-12", + "release_date": "2025-09-05", + "last_updated": "2025-09-05", "modalities": { "input": [ "text" @@ -327350,101 +198459,106 @@ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 1000000, - "output": 131072 + "context": 256000, + "output": 64000 }, "cost": { - "input": 1.1, - "output": 3.851, - "cache_read": 0.275, - "cache_write": 0 + "input": 0, + "output": 0 } }, - "qwen3.7-plus": { - "id": "qwen3.7-plus", - "name": "Qwen3.7 Plus", - "description": "Multimodal Qwen workhorse for long-context agents, visual inputs, and coding", - "family": "qwen", - "attachment": true, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - }, - { - "type": "budget_tokens", - "max": 262144 - } - ], + "deepseek-v3.2": { + "id": "deepseek-v3.2", + "name": "DeepSeek-V3.2-Exp", + "description": "DeepSeek chat model for instruction following, coding, and analysis", + "family": "deepseek", + "attachment": false, + "reasoning": false, "tool_call": true, "temperature": true, - "knowledge": "2025-04", - "release_date": "2026-06-02", - "last_updated": "2026-06-02", + "knowledge": "2024-12", + "release_date": "2025-01-01", + "last_updated": "2025-01-01", "modalities": { "input": [ - "text", - "image", - "video" + "text" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 1000000, + "context": 128000, "output": 64000 }, "cost": { - "input": 0.5, - "output": 3, - "cache_read": 0.05, - "cache_write": 0.625, - "tiers": [ - { - "input": 2, - "output": 6, - "cache_read": 0.2, - "cache_write": 2.5, - "tier": { - "type": "context", - "size": 128000 - } - } + "input": 0, + "output": 0 + } + } + } + }, + "xiaomi-token-plan-sgp": { + "id": "xiaomi-token-plan-sgp", + "env": [ + "XIAOMI_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://token-plan-sgp.xiaomimimo.com/v1", + "name": "Xiaomi Token Plan (Singapore)", + "doc": "https://platform.xiaomimimo.com/#/docs", + "models": { + "mimo-v2.5-tts": { + "id": "mimo-v2.5-tts", + "name": "MiMo-V2.5-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", + "attachment": false, + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", + "modalities": { + "input": [ + "text" + ], + "output": [ + "audio" ] + }, + "open_weights": true, + "limit": { + "context": 8192, + "output": 8192 + }, + "cost": { + "input": 0, + "output": 0 } }, - "deepseek-v4-flash": { - "id": "deepseek-v4-flash", - "name": "DeepSeek V4 Flash", - "description": "Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work", - "family": "deepseek-flash", + "mimo-v2.5-pro": { + "id": "mimo-v2.5-pro", + "name": "MiMo-V2.5-Pro", + "description": "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution", + "family": "mimo", "attachment": false, "reasoning": true, "reasoning_options": [ { "type": "toggle" - }, - { - "type": "effort", - "values": [ - "high", - "max" - ] } ], "tool_call": true, "interleaved": { "field": "reasoning_content" }, - "structured_output": true, "temperature": true, - "knowledge": "2025-05", - "release_date": "2026-04-24", - "last_updated": "2026-04-24", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" @@ -327455,31 +198569,38 @@ }, "open_weights": true, "limit": { - "context": 1000000, - "output": 384000 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.14, - "output": 0.28, - "cache_read": 0.0028 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "qwen-vl-ocr": { - "id": "qwen-vl-ocr", - "name": "Qwen-VL OCR", - "description": "OCR model for extracting structured text from documents and screenshots", - "family": "qwen", + "mimo-v2-pro": { + "id": "mimo-v2-pro", + "name": "MiMo-V2-Pro", + "description": "Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks", + "family": "mimo", "attachment": false, - "reasoning": false, - "tool_call": false, + "reasoning": true, + "reasoning_options": [ + { + "type": "toggle" + } + ], + "tool_call": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "knowledge": "2024-04", - "release_date": "2024-10-28", - "last_updated": "2026-09-11", + "knowledge": "2024-12", + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ - "text", - "image" + "text" ], "output": [ "text" @@ -327487,50 +198608,50 @@ }, "open_weights": false, "limit": { - "context": 34096, - "output": 4096 + "context": 1048576, + "output": 131072 }, + "status": "deprecated", "cost": { - "input": 0.043, - "output": 0.072 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "siliconflow/deepseek-v3-0324": { - "id": "siliconflow/deepseek-v3-0324", - "name": "siliconflow/deepseek-v3-0324", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "mimo-v2-tts": { + "id": "mimo-v2-tts", + "name": "MiMo-V2-TTS", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", "attachment": false, "reasoning": false, - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2024-12-26", - "last_updated": "2025-11-25", + "tool_call": false, + "release_date": "2026-03-18", + "last_updated": "2026-03-18", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163840, - "output": 163840 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.25, - "output": 1 + "input": 0, + "output": 0 } }, - "siliconflow/deepseek-v3.1-terminus": { - "id": "siliconflow/deepseek-v3.1-terminus", - "name": "siliconflow/deepseek-v3.1-terminus", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", - "attachment": false, + "mimo-v2.5": { + "id": "mimo-v2.5", + "name": "MiMo-V2.5", + "description": "Open MiMo model for multimodal coding agents and long-context automation", + "family": "mimo", + "attachment": true, "reasoning": true, "reasoning_options": [ { @@ -327538,166 +198659,167 @@ } ], "tool_call": true, - "structured_output": true, + "interleaved": { + "field": "reasoning_content" + }, "temperature": true, - "release_date": "2025-09-29", - "last_updated": "2025-11-25", + "knowledge": "2024-12", + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163840, - "output": 65536 + "context": 1048576, + "output": 131072 }, "cost": { - "input": 0.27, - "output": 1 + "input": 0, + "output": 0, + "cache_read": 0 } }, - "siliconflow/deepseek-r1-0528": { - "id": "siliconflow/deepseek-r1-0528", - "name": "siliconflow/deepseek-r1-0528", - "description": "DeepSeek reasoning model for multi-step analysis, math, coding, and tools", - "family": "deepseek-thinking", + "mimo-v2.5-tts-voicedesign": { + "id": "mimo-v2.5-tts-voicedesign", + "name": "MiMo-V2.5-TTS-VoiceDesign", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", "attachment": false, - "reasoning": true, - "reasoning_options": [], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-05-28", - "last_updated": "2025-11-25", + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163840, - "output": 32768 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.5, - "output": 2.18 + "input": 0, + "output": 0 } }, - "siliconflow/deepseek-v3.2": { - "id": "siliconflow/deepseek-v3.2", - "name": "siliconflow/deepseek-v3.2", - "description": "DeepSeek chat model for instruction following, coding, and analysis", - "family": "deepseek", + "mimo-v2.5-tts-voiceclone": { + "id": "mimo-v2.5-tts-voiceclone", + "name": "MiMo-V2.5-TTS-VoiceClone", + "description": "Speech generation model for controllable voice, narration, and audio delivery", + "family": "mimo", "attachment": false, - "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], - "tool_call": true, - "structured_output": true, - "temperature": true, - "release_date": "2025-12-03", - "last_updated": "2025-12-03", + "reasoning": false, + "tool_call": false, + "release_date": "2026-04-22", + "last_updated": "2026-04-22", "modalities": { "input": [ "text" ], "output": [ - "text" + "audio" ] }, - "open_weights": false, + "open_weights": true, "limit": { - "context": 163840, - "output": 65536 + "context": 8192, + "output": 8192 }, "cost": { - "input": 0.27, - "output": 0.42 + "input": 0, + "output": 0 } - }, - "kimi/kimi-k2.5": { - "id": "kimi/kimi-k2.5", - "name": "kimi/kimi-k2.5", - "description": "Kimi multimodal agent model for visual understanding, coding, and planning", - "family": "kimi-k2", - "attachment": false, + } + } + }, + "claudinio": { + "id": "claudinio", + "env": [ + "CLAUDINIO_API_KEY" + ], + "npm": "@ai-sdk/openai-compatible", + "api": "https://api.claudin.io/v1", + "name": "Claudinio", + "doc": "https://claudin.io", + "models": { + "claudinio": { + "id": "claudinio", + "name": "Claudinio", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, - "reasoning_options": [ - { - "type": "toggle" - } - ], + "reasoning_options": [], "tool_call": true, - "interleaved": { - "field": "reasoning_content" - }, - "structured_output": true, - "temperature": false, - "knowledge": "2025-01", - "release_date": "2026-01-27", - "last_updated": "2026-01-27", + "knowledge": "2026-05", + "release_date": "2026-05-12", + "last_updated": "2026-06-02", "modalities": { "input": [ "text", "image", + "audio", "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 262144, - "output": 262144 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.6, - "output": 3, - "cache_read": 0.1 + "input": 0.5, + "output": 2, + "cache_read": 0.15 } }, - "MiniMax/MiniMax-M2.7": { - "id": "MiniMax/MiniMax-M2.7", - "name": "MiniMax-M2.7", - "description": "MiniMax model for chat, coding, office work, and agentic tasks", - "family": "minimax", - "attachment": false, + "claudius": { + "id": "claudius", + "name": "Claudius", + "description": "Multimodal reasoning model for visual analysis, planning, and tool use", + "attachment": true, "reasoning": true, "reasoning_options": [], "tool_call": true, - "temperature": true, - "release_date": "2026-03-18", - "last_updated": "2026-03-18", + "knowledge": "2026-05", + "release_date": "2026-05-12", + "last_updated": "2026-05-12", "modalities": { "input": [ - "text" + "text", + "image", + "audio", + "video" ], "output": [ "text" ] }, - "open_weights": true, + "open_weights": false, "limit": { - "context": 204800, - "output": 131072 + "context": 256000, + "output": 64000 }, "cost": { - "input": 0.3, - "output": 1.2, - "cache_read": 0.06, - "cache_write": 0.375 + "input": 3, + "output": 8, + "cache_read": 0.9 } } } diff --git a/packages/agentlayer-core/test/models.test.ts b/packages/agentlayer-core/test/models.test.ts index 811d530..16c9f7f 100644 --- a/packages/agentlayer-core/test/models.test.ts +++ b/packages/agentlayer-core/test/models.test.ts @@ -57,8 +57,8 @@ describe('ModelProvider.getModelLimits', () => { // reported as `undefined`, so Azure AI Foundry sessions recorded tokens and no // dollars at all — silently, and unrecoverably, since cost is frozen at ingest. expect(provider.getModelPricing(`${CUSTOM_RESPONSES_PROVIDER}/gpt-5.6-sol`)).toMatchObject({ - input: 4, - output: 20, + input: 5, + output: 30, }) // The public Responses API, not the private Codex one, so it keeps the public window. @@ -70,9 +70,9 @@ describe('ModelProvider.getModelLimits', () => { expect(limits?.context).toBe(1_050_000) expect(limits?.output).toBe(128_000) - expect(provider.getModelPricing('openai/gpt-5.6-sol')).toMatchObject({ input: 4, output: 20 }) - expect(provider.getModelPricing('openai/gpt-5.6-terra')).toMatchObject({ input: 2, output: 12 }) - expect(provider.getModelPricing('openai/gpt-5.6-luna')).toMatchObject({ input: 0.2, output: 1.2 }) + expect(provider.getModelPricing('openai/gpt-5.6-sol')).toMatchObject({ input: 5, output: 30 }) + expect(provider.getModelPricing('openai/gpt-5.6-terra')).toMatchObject({ input: 2.5, output: 15 }) + expect(provider.getModelPricing('openai/gpt-5.6-luna')).toMatchObject({ input: 1, output: 6 }) }) test('openai/gpt-6-astra keeps the public OpenAI API context window and pricing', () => {